SECTION 22 Haematological disorders

22.1 Introduction to haematology 5169 Chris Hatton
22.1 Introduction to haematology 5169 Chris Hatton
ESSENTIALS
Haematology is the study of the composition, function, and diseases of the
blood. The approach to a patient suspected of having a haematological
disorder begins with taking a history (particularly noting fatigue, weight
loss, fever, and history of bleeding) and performing a clinical examination
(looking for signs of anaemia, infection, bleeding, and signs of cellular
infiltration causing splenomegaly and/​or lymphadenopathy). Key inves-
tigations include a full blood count, a blood film, and (in selected cases)
examination of the bone marrow. Further diagnostic tests now routinely
performed on blood and marrow samples include immunophenotyping
and cytogenetic and molecular analysis. Mutational signatures may be
diagnostically useful and potentially define treatment, keeping haema-
tology in the vanguard of advances in modern medicine.
Introduction
Haematology has always been in the vanguard of advances towards truly
modern medicine. The first disease defined at a molecular level was
haematological (sickle cell disease); the first molecularly targeted treat-
ment was designed for a haematological disorder (imatinib in chronic
myeloid leukaemia); and the revolution of immunological treatments,
whether in the form of allogeneic bone marrow transplantation, fo-
cused therapies (such as rituximab) or cellular therapies (e.g. chimeric
antigen receptor T-cells), was also begun from within haematology.
This remains the case today, and the discipline is now advancing
at an almost bewildering pace. However, despite the huge advances
made in diagnostic techniques and imaging, the starting point for
evaluating a patient relies on basic clinical skills.
In this introduction, we outline the scope of haematology as a dis-
cipline, give an overview of the nature and function of blood cells,
and provide a system for the newcomer to haematology to consider
the likelihood of haematological disease in his or her patient.
The scope of haematology
Haematology is the study of the composition, function, and diseases
of the blood. Its diversity as a specialism therefore immediately
reflects the complexity of its subject. At its simplest, blood is div-
ided into the plasma component (water, electrolytes, clotting factors,
and fibrinogen—​with serum being the same substance without the
clotting factors) and the cellular component, comprising red cells,
platelets, granulocytes, and lymphocytes. Each has its specific and
irreplaceable role in the normal function of blood, which impacts in
turn the function of every tissue in the body. Not only do diseases of
the blood influence every downstream organ, systemic diseases will
also manifest in the blood. An appreciation of normal blood counts
and appearances is therefore central to many fields of medicine.
Diseases and the blood
Even a cell as apparently simple as the red blood cell, anucleate and
devoid of intracellular organelles in its mature form, can manifest
a variety of disorders. Inherited defects in the synthesis of globin
genes, needed for the transport of oxygen to the peripheral tissues,
constitute the commonest genetic diseases in the world. A host of
additional genetic defects in glycolytic enzymes also impact on the
survival of the red cell and the ability of the marrow to maintain a
normal haemoglobin level. Meanwhile, the iron deficiency resulting
from chronic occult blood loss may be the only clue to the presence
of a malignant colonic tumour, and the failure to absorb vitamin
B12 in pernicious anaemia may highlight the possibility of a range
of additional autoimmune disorders. Erythropoietin, the key hor-
mone controlling red cell production, is synthesized principally
in peritubular interstitial fibroblasts in the juxtamedullary region
of the renal cortex and renal disease may therefore result in either
insufficient or excess marrow stimulation. Thus the finding of an-
aemia, a low haemoglobin level, may point to either haematological
disorders, or may reflect primary disease elsewhere. The careful in-
vestigation of the cause of anaemia (see Chapter 22.6.2) is an im-
portant part of general medical practice.
Granulocytes (neutrophils, eosinophils, and basophils, so termed
to reflect the staining characteristics of the granules that are crit-
ical for their function) may also reflect both primary haematological
disease and reactive conditions. The granules of neutrophils contain
myeloperoxidase, needed in the cellular response to bacterial infec-
tion, and a high neutrophil count (neutrophilia) is commonly seen
22.1
Introduction to haematology
Chris Hatton
SECTION 22  Haematological disorders
5170
in the context of infection or inflammation. The specific function
of eosinophils in combating multicellular parasites means that a re-
active eosinophilia may also be seen in infection with these organ-
isms, as well as in allergic reactions. The uncontrolled proliferation
of granulocytes of all kinds is seen in the myeloproliferative disorder
chronic myeloid leukaemia, one of the first haematological malig-
nancies to be defined at the molecular level.
Neutropenia, by contrast, describes an inadequate number of
circulating neutrophils, and may reflect primary marrow dysfunc-
tion (e.g. aplastic anaemia), the result of myelotoxic chemotherapy
administration, or immune attack. The key role of neutrophils in
maintaining the integrity of mucosal surfaces is highlighted by the
increased risk of Gram-​negative infection in severe neutropenia,
and the rapidly progressive sepsis that accompanies it is one of
haematology’s most urgent medical emergencies.
Along with red cells and granulocytes, platelets form the last of the
‘myeloid’ components of the blood. Their role in primary haemo-
stasis (again effected in part by the presence of cell-​specific granules)
is highlighted in Chapter 22.7.3; they are the target of some of the
most widely prescribed agents used in medical practice (aspirin and
other antiplatelet agents are discussed in more detail in the Section
16 on cardiovascular disease).
Lymphocytes divide into B cells, T cells, and NK cells. Each has
its distinct role in the immune process, from the production of
antibodies to cell-​mediated immunity and the development of
antitumour action (e.g. through perforins secreted in the gran-
ules of cytotoxic T cells). As well as a reactive lymphocytosis or
lymphopenia seen in response to viral infection, malignant trans-
formation of lymphoid cells may result in a circulating excess of
clonal lymphocytes or lymphoid precursor cells, or in the develop-
ment of lymphadenopathy. Perhaps the most protean of haemato-
logical malignancies, lymphomas can affect any organ in the body.
A discussion of the nature and treatment of these varied disorders is
given in Chapters 22.4.3 and 22.4.4.
Disorders of haemostasis, whether hereditary or acquired, may
reflect a lack of key components of the clotting cascade, platelet
lack, or platelet dysfunction. The modulation of the haemostatic
machinery for therapeutic purposes also highlights the increasing
awareness of overefficient haemostasis—​for example, in the her-
editary thrombophilias. These are discussed in more detail in
Chapters 22.7.4 and 22.7.5.
How does blood develop?
Haematopoiesis is the term used to describe the cellular pro-
cesses that produce blood cells. The sheer magnitude of the process
is apparent in the observation that the bone marrow produces
2.5 × 1011 red cells, 1 × 1011 platelets, and 1 × 1010 white cells per
day. Beginning in the yolk sac in utero, the process of haemato-
poiesis switches to the spleen and the liver in the developing fetus
before moving to the bone marrow. With increasing age, haemato-
poiesis becomes confined to the axial skeleton, with very little red
(haematopoietically active) marrow present in the long bones of the
limbs in adults.
The bone marrow contains an as yet undetermined number
of pluripotent stem cells that are capable of both continuous self-​
renewal and differentiation. More mature cells lose the capacity for
self-​renewal as they differentiate into fully functional mature blood
cells or form the structural cellular matrix of the bone marrow
stroma. As they divide, progenitors have a progressively restrictive
lineage potential manifested by their use of specific transcription
factors. A  number of cytokines such as erythropoietin, granulo-
cyte colony-​stimulating factor (G-​CSF), granulocyte–​macrophage
colony-​stimulating factor (GM-​CSF), and thrombopoietin induce
proliferation of specific lineages; these, together with complex cel-
lular interactions, lead to development of mature blood cells in the
marrow and subsequent release into the blood. Although a detailed
treatment of haematopoiesis is given in Chapter 22.2.1, it is clinic-
ally useful to consider two different populations arising from this
process: these differentiate into the two main lineages of myeloid
and lymphoid cells. As described previously, myeloid maturation,
or myelopoiesis, produces red cells (erythrocytes), granulocytes,
and platelets; while lymphopoiesis describes the development of
lymphoid cells into mature B cells, T cells, and NK cells.
The identification of haematopoietic stem cells (HSCs) that
traffic from the bone marrow to the blood and back provided a
major step forward in haematological practice. The subsequent rec-
ognition that it was possible to harvest these HSCs from humans,
and that after reinfusion they could re-​establish normal haemato-
poiesis, has led to the development of the flourishing practice of
HSC transplantation. Although very few HSCs are present in the
peripheral blood, it is possible to increase their numbers in blood
using the growth factors G-​CSF or GM-​CSF, or by blocking the
molecule that anchors stem cells in the marrow matrix, CXCR4.
In clinical practice, autologous or allogeneic stem cell infusion can
be used to reconstitute haematopoiesis after appropriate chemo-
therapy. The subject of HSC transplantation is covered in detail
in Chapter 22.8.2. Bone marrow transplantation has had a huge
impact, enabling long-​term remissions to be achieved in patients
suffering from aggressive haematological malignancies and bone
marrow failure syndromes.
Initial approach to the patient
The approach to a patient suspected of having a haematological
disorder begins with taking a history and performing a clinical
examination.
Potential features of the history may include fatigue (perhaps re-
flecting anaemia or underlying malignancy), weight loss, and fever
(again suggestive of the hypercatabolic picture of malignancy).
A careful bleeding history, including responses to previous haemo-
static challenges such as surgery and dental work, will be useful
in delineating a possible bleeding diathesis. A general impression
of the patient’s overall health status—​perhaps via recording his/​
her Eastern Cooperative Oncology Group (ECOG) performance
score—​is important in assessing tolerance for treatment and in cat-
egorizing patients entering clinical trials.
Examination of the patient with a suspected blood disorder should
concentrate on looking for signs of anaemia, infection, bleeding, and
signs of cellular infiltration causing splenomegaly and lymphaden-
opathy. Pallor is a frequent finding in patients with anaemia though
normal pigment differences in the skin make this an unreliable sign.
Pallor of the mucous membranes or palmar creases may be more
useful. Jaundice, commonly seen in liver disease, is also a prominent
22.1  Introduction to haematology
5171
sign in patients with premature red cell destruction (haemolysis),
and is readily detected in the sclerae. Signs of a bleeding tendency
should be sought in the skin, mucous membranes, and the retina.
Haemorrhage into the skin characteristically produces petechiae
and ecchymoses. Petechial haemorrhages are small (1–​2 mm), often
seen in areas with high venous pressure such as around the ankles,
and are a common finding in patients with severe thrombocytopenia
(platelet count <20 × 109/​litre). Ecchymoses (commonly known as
bruises) are larger subcutaneous haemorrhages, a frequent finding
after trauma but also occurring spontaneously in patients with a low
platelet count or a functional platelet defect.
Lymphadenopathy is a common finding in patients with lympho­
proliferative disorders and is sometimes present in patients with mye-
loid disease. Enlargement of a lymph node becomes significant when
greater than 1 cm and nodes of this size should be biopsied when pre-
sent for longer than 6 weeks without obvious cause. Splenomegaly
may also be seen as a result of infiltration of the white pulp by a
lymphoma or leukaemia, or more rarely in storage disorders such as
Gaucher’s disease; expansion of the red pulp in chronic haemolysis
may also cause splenomegaly. Rarely, the spleen may enlarge as a result
of extramedullary haematopoiesis in conditions where there is bone
marrow failure, such as myelofibrosis. In this situation, the spleen
takes over the function of the bone marrow in producing blood.
Investigation of a suspected blood disorder
Laboratory investigation for malignant haematological disorders is
covered in detail in Chapter 22.2.2. In the investigation of a ­patient
with a suspected haematological disorder, a careful inspection of the
blood film is essential. Morphological abnormalities of red cells or
white cells can be diagnostic. For example, the finding of immature
‘blast’ cells on a peripheral blood film is suggestive of acute leukaemia
or marrow stress due to severe sepsis. The presence of immature red
cell and white cell precursors (leucoerythroblastic anaemia) on the
peripheral blood film is often indicative of bone marrow infiltra-
tion by malignancy or marrow stress due to severe sepsis. There is
a huge range of morphological changes affecting all blood cell lin-
eages; the haematologist becomes accustomed to identifying those
changes that are diagnostically significant. If necessary, inspection
of the peripheral blood film is followed by bone marrow aspiration
and biopsy, enabling the haematologist to assess the maturation of
­precursor cells and to look for infiltration by malignancy.
The close link between immunology and haematology is em-
phasized by the importance of immunological analyses for the
diagnosis of lymphoid disease. The finding of a paraprotein may
suggest a mature B-​cell malignancy or plasma cell clone (myeloma).
Further diagnostic tests are now routinely performed on blood
and marrow samples obtained from the patient. Abnormal popu-
lations of cells identified morphologically (e.g. blasts in acute leu-
kaemia, or lymphoid populations in lymphoproliferative diseases)
can be immunophenotyped using flow cytometry carried out on
blood, bone marrow aspirates, and if indicated on cerebrospinal
fluid, pleural fluid, or ascitic fluid. Histopathology of bone marrow,
lymph nodes, and other affected organs which have been biopsied
provide additional morphological and immunohistochemical diag-
nostic information. Furthermore, cytogenetic and molecular ana-
lysis may provide evidence of clonality, define specific diagnostic
translocations and chromosomal rearrangements, and identify the
presence or absence of specific mutations. Increasingly multigene
sequencing panels are being used to identify mutational signa-
tures that may be diagnostically useful and potentially define treat-
ment. Large haematology laboratories offer this complete range of
diagnostic services, integrating the results into a single diagnostic
report.
These data add to those achieved through increasingly sophisti-
cated imaging techniques such as positron emission tomography
and magnetic resonance imaging, which give functional as well as
anatomical data to define haematological diseases.
Commentary
In subsequent chapters, the detailed biology of normal and malig-
nant HSCs and their progeny will be described. A single chapter is
devoted to the laboratory analysis of malignant blood cells with a
strong emphasis on determining the lineage of the malignant clone.
The importance of the myeloid/​lymphoid split is emphasized, and
it will become obvious to the reader that the treatment of different
leukaemias and lymphomas depends on the cell of origin of the
malignant clone.
It will be evident that much of haematology either relates to defi-
ciency or dysfunction of blood cells or the noncellular components
of blood. In addition, many treatments given for haematological dis-
ease have the necessary side effect of depleting normal blood com-
ponents. Therefore one of the major challenges facing haematology
is the need to develop safe, readily available, and cost-​effective
blood component replacements, either from blood donors or via
biotechnology. Blood transfusion services have developed to an
extraordinary degree since their inception at the beginning of the
last century, but still depend exclusively on community altruism.
Stem cell technology particularly holds promise for the future of this
aspect of our specialty.
Few specialties occupy so wide a range as haematology, taking in
both malignant and nonmalignant disease, chronic and acute care,
clinical and laboratory practice, and aligning basic physicianly skills
with the most up-​to-​date genetic and molecular advances. Keeping
so disparate a specialty together, and maintaining skills in all areas,
is increasingly difficult for clinical haematologists and will only be-
come more so as our understanding of the pathological basis of dis-
ease continues to expand. We hope this section of the textbook will
inspire future haematologists to meet this challenge.

22.2 Haematopoiesis 5172 22.2.1 Cellular and molec
22.2 Haematopoiesis 5172 22.2.1 Cellular and molecular basis of haematopoiesis 5172 Paresh Vyas and N. Asger Jakobsen
CONTENTS
22.2.1	 Cellular and molecular basis of haematopoiesis  5172
Paresh Vyas and N. Asger Jakobsen
22.2.2	 Diagnostic techniques in the assessment of
haematological malignancies  5181
Wendy N. Erber
22.2.1  Cellular and molecular basis
of haematopoiesis
Paresh Vyas and N. Asger Jakobsen
ESSENTIALS
Haematopoiesis involves a regulated set of developmental stages by
which haematopoietic stem cells (HSCs) produce haematopoietic
progenitor cells which in turn differentiate into more mature haem-
atopoietic lineages. These then provide all the key functions of the
haematopoietic system.
Development
Haematopoiesis occurs in distinct waves during development.
Definitive HSCs first develop within the embryo in specialized re-
gions of the dorsal aorta and umbilical arteries and then seed the
fetal liver and bone marrow. HSC characteristics differ based on their
site of development and age of the organism.
Haematopoietic stem cells
At the single-​cell level, these have the ability to reconstitute and
maintain a functional haematopoietic system over extended periods
of time in vivo. They (1) have a self-​renewing capacity during the life of
an organism, or even after transplantation; (2) are multipotent, with
the ability to make all types of blood cells; and (3) are relatively qui-
escent, with the ability to serve as a deep reserve of cells to replenish
short-​lived, rapidly proliferating progenitors. In vivo transplantation
models are currently the only reliable assays of HSC activity.
Haematopoietic progenitor cells
These are unable to maintain long-​term haematopoiesis in vivo
due to limited or absent capacity for self-​renewal. Their rapid pro-
liferation and cytokine responsiveness enables increased blood cell
production under conditions of stress. Lineage commitment means
limited cell type production.
The haematopoietic stem cell niche
An anatomically and functionally defined regulatory environment
for stem cells modulates self-​renewal, differentiation, and prolif-
erative activity of stem cells, thereby regulating stem cell number.
Niche function is important in maintaining haematopoietic integrity
and niche dysfunction may contribute to haematopoietic disease.
Niches for HSCs are dynamic, changing during development and
with physiological stress. HSCs naturally traffic into and out of the
niche, a feature that can be exploited for stem cell transplantation or
harvesting, respectively.
Bone marrow transplantation
Haematopoietic reconstitution during bone marrow transplant-
ation is mediated by a succession of cells at various stages of
development. More mature cells contribute to repopulation im-
mediately following transplantation. With time, cells at progres-
sively earlier stages of development are involved, with the final
stable repopulation being provided by long-​lived, multipotent
HSCs. Long-​term haematopoiesis is sustained by a relatively small
number of HSCs.
Haematopoiesis through development
and into adult life
Adult humans produce approximately 300 to 1000 billion blood
cells per day (Table 22.2.1.1). The vast majority of these are myeloid
cells—​platelets, erythroid cells, and granulocytes. Haematopoiesis
is the process by which blood cells are made throughout develop-
ment (embryonic life) and adult life by transiting through a hier-
archy of haematopoietic stem and progenitor cells (HSPCs). This
chapter will summarize the cellular and molecular basis of haem-
atopoiesis as a prelude to a deeper understanding of benign and
malignant blood diseases.
22.2
Haematopoiesis
22.2.1  Cellular and molecular basis of haematopoiesis
5173
Primitive haematopoiesis
Haematopoiesis occurs in waves and at multiple discrete anatomical
sites that change through development (Fig. 22.2.1.1).
In humans, like other vertebrates, the initial wave of haemato-
poiesis occurs in the extraembryonic yolk sac blood islands from
weeks 3 to 6 of gestation (Fig. 22.2.1.2). The yolk sac primarily pro-
duces primitive erythroid cells (termed primitive erythropoiesis)
expressing embryonic globins that deliver oxygen to tissues in the
rapidly growing embryo. Primitive haematopoiesis also produces
myeloid and lymphoid cells (macrophages and natural killer cells).
Interestingly, the developmental potential of embryonic haem-
atopoiesis closely resembles haematopoietic cells derived from
human embryonic stem cells (hESCs) and induced pluripotent
stem cells (iPSCs).
Embryonic definitive haematopoiesis
Primitive haematopoiesis is transient and is replaced by definitive
haematopoiesis that sustains blood production throughout devel-
opment and postnatal life. Embryonic definitive haematopoietic ac-
tivity is detected at 4 to 5 weeks of gestation, in a region around the
ventral wall of the dorsal aorta (Fig. 22.2.1.3). In early development,
blood cells arise in close connection with the vascular structures
(both in the yolk sac and the dorsal aorta) giving rise to the notion that
there may either be a common precursor cell population that gives
rise to both blood and blood vessel cells called a haemangioblast,
or that haematopoiesis arises directly from specialized ‘haemogenic’
endothelial cells such as those lining the ventral aspect of the dorsal
aorta. In mice and other animals, studies have shown that defini-
tive haematopoietic stem cells (HSCs) with serially transplantable
activity and long-​term engraftment capacity are found in the dorsal
aorta. It is still a matter of debate whether HSCs arise from the em-
bryo proper (the dorsal aorta) or by colonization from the yolk sac.
A large transient pool of HSCs has been identified in the placenta of
mice around the time of aorta–​gonad–​mesonephros HSC develop-
ment. It remains to be determined whether an equivalent population
of HSCs exists in the developing human placenta.
Fetal liver haematopoiesis
HSCs are then detected in the developing fetal liver, spleen, and
thymus from 6 to 22 weeks in humans, where they expand and dif-
ferentiate into committed progenitor cells. Recent evidence sug-
gests that the critical niche supporting HSCs is the fetal vasculature.
Expansion and differentiation of HSCs allows for development of
definitive red cells, myeloid cells, and lymphoid cells (T cells that
YOLK SAC
AGM
Weeks 2.5–4 of gestation
Weeks 4–10 of gestation
Week 5 of gestation until term
Mainly postnatal
LIVER
SPLEEN
BONE
MARROW
Fig. 22.2.1.1  Changing anatomical locations of haematopoiesis through development. Haematopoiesis is initially detected in the
extraembryonic yolk sac, then in the embryo, in the aorta–​gonad–​mesonephros (AGM), the adjacent umbilical arteries and vitelline
vessels, and the placenta. It then shifts to the fetal liver and finally to the bone marrow.
Table 22.2.1.1  Cellular components of blood
Red cells
White cells
Platelets
Number/​litre of blood
4–​6 × 1012
4–​11 × 109
c.60% are myeloid cells
150–​400 × 109
Total cell number (blood volume 5 litres)
20–​30 × 1012
20–​55 × 109
750–​2000 × 109
Cell lifespan
120 days
Myeloid cells: 1–​5 days
Lymphoid cells: weeks to years
2–​4 days
Daily myeloid cell production
1.7–​2.5 × 1011
12–​30 × 109
3.5–​10 × 1011
SECTION 22  Haematological disorders
5174
develop in the thymus and B cells in the marrow). It is unclear
whether HSCs from the dorsal aorta migrate to colonize fetal liver
and other embryonic sites or whether they arise de novo at these
other sites.
Adult bone marrow haematopoiesis
HSPCs seed the bone marrow (BM) during fetal life but only make
a small proportion of blood until late in gestation. It is not clear
whether HSCs must first reside in the fetal liver before seeding the
BM. However, at the time of birth, haematopoiesis switches from
fetal liver to BM as the liver is repurposed to serve its roles in adult
life. The mechanisms resulting in this remarkable switch in anatom-
ical location are unclear, but it is very likely that the changing blood
supply to the liver plays a critical role, wherein oxygenated blood from
the umbilical cord is replaced by relatively deoxygenated blood
from the portal vein that is rich in metabolites ingested from the gut.
The marrow then sustains lifelong blood production, where HSPCs
reside in both endosteal and vascular niches of the marrow cavity.
In early life, haematopoiesis occurs in the diaphysis and epiphysis of
the axial skeleton and the long bones. With age, the marrow cavity
becomes fatty and haematopoiesis is progressively restricted to the
axial skeleton.
Given that HSPCs isolated from different developmental time
points and anatomical locations (e.g. fetal liver, placental cord blood,
BM, hESCs, and iPSCs) have been shown to have different func-
tional potentials, this may have implications regarding the choice of
stem cell sources for human transplantation therapies.
Haematopoiesis: a hierarchical
differentiation cascade
HSPC populations give rise to all haematopoietic cells through
a cascade of differentiation. Stem and progenitor are both highly
heterogeneous populations. They are defined by two key proper-
ties: variable ability to self-​renew and a variable potential to differ-
entiate into one or more mature blood cell types. Residing at the top
of this hierarchy are HSCs. They likely number tens to a few hun-
dreds of thousands of cells in adult life, giving rise to hundreds of
millions of hierarchically organized, highly heterogeneous progenitor
cells, which in turn differentiate into precursor cells that eventually
mature into effector cells (Fig. 22.2.1.4).
The current view of the haematopoietic cellular hierarchy will
change over the coming years as we functionally and molecularly
interrogate HSPCs at a single-​cell level. Our understanding of the
hierarchical relationships between populations, the plasticity of
commitment, and the nature of the functional potential of popula-
tions will increase as we better purify HSPC populations.
Haematopoietic stem cells
HSCs have extensive self-​renewal capability and the potential to
­differentiate into all blood cell types. The remarkable ability of HSCs, at
the single-​cell level, to reconstitute and maintain a functional haem-
atopoietic system over extended periods of time in vivo demonstrates
these key properties. Self-​renewal allows HSCs to be transplanted
between individuals, and the surviving HSCs engraft, proliferate,
and differentiate for the life of the recipient. HSCs can be serially
Fig. 22.2.1.3  Transverse section through the human fetal dorsal
aorta at embryonic day 32 showing haematopoietic CD34+ cells
clusters (arrowheads) associated with the ventral wall. CD34+
cells (haematopoietic and endothelial cells) are stained brown.
Scale = 100 microns.
From Tavian M, Hallais MF, Péault B (1999). Emergence of intraembryonic
hematopoietic precursors in the pre-​liver human embryo. Development, 126,
793–​803.
(a)
(b)
Fig. 22.2.1.2  Yolk sac blood islands in a human fetus. (a) Transverse
section of a three-​somite human embryo (21 days) at the truncal level
stained with anti-​CD34 antibody. Paired dorsal aorta (da) ventrolateral
to the neural tube (nt) and above the yolk sac (ys) and blood
islands (bi). Scale = 200 microns. (b) Higher magnification of a solid
haemangioblastic mesodermal cluster of CD34-​expressing cells in a
blood island of the yolk sac (brown). Scale = 25 microns.
From Tavian M, Hallais MF, Péault B (1999). Emergence of intraembryonic
hematopoietic precursors in the pre-​liver human embryo. Development, 126,
793–​803.
22.2.1  Cellular and molecular basis of haematopoiesis
5175
transplanted for many generations between recipients. When HSCs
are recruited into active haematopoiesis, they exit the G0 phase of
the cell cycle, and their daughter cells may either be a replicate of the
parent cells (self-​renewal) or enter into a differentiation programme.
This distinctive, asymmetric division process is the basis for long-​
term preservation of stem cells while enabling continued produc-
tion of mature cells. The daughter cells that undergo differentiation
proceed through a series of maturational cell divisions, culminating
in the generation of progenitor cells.
Most of our detailed understanding of HSCs has come from
murine studies though some of the findings have also been con-
firmed in humans. Contrary to previous assumptions, HSCs
are a heterogeneous cell population through development and
adult life. In the mouse, fetal HSCs show extensive prolifer-
ation and tend towards greater lymphoid output. During adult
life, there is a hierarchy of HSCs. The most primitive HSCs are
rare, representing approximately 1 in 104 to 106 BM cells. These
long-​term stem cells are deeply quiescent and may replicate
only once a year, which protects their genome from replicative
mutational damage. These give rise to HSCs that cycle slightly
more often. Adult HSCs are also heterogeneous with respect
to their lineage output. Some are biased towards producing
megakaryocyte–​erythroid cells, others more myeloid cells, and
finally yet others more lymphoid cells. Yet all HSCs, if required,
can produce all lineages. With age, HSC numbers increase but
they become less functional and those that are biased to pro-
duce lymphocyte lineages diminish relative to myeloid-​biased
HSCs. This potentially contributes to the relative lymphoid de-
ficiency and increase in myeloid diseases (myeloproliferative
disorders, myelodysplasia, and acute myeloid leukaemia) seen
in the elderly.
Progenitor cells
Progenitor cells are also hierarchically arranged. As they differen-
tiate from stem cells and through the progenitor hierarchy they
progressively lose self-​renewal and become restricted in their dif-
ferentiation potential such that multipotent progenitors give rise
to oligopotent and finally monopotent progenitors. Progenitors
are highly proliferative and very cytokine responsive. New studies
of murine haematopoiesis suggest that, in contrast to the trans-
plantation setting, the source of most blood cells produced daily
during normal steady-​state haematopoiesis is maintained by the
continuing expansion of thousands of haematopoietic progenitor
cells, each with a minimal contribution to mature progeny. Single-​
cell transplant studies in mice have also revealed a bypass pathway
that produces long-​term repopulating myeloid progenitors. This
pathway may be operative under stress, as progenitor populations
most readily respond to external stimuli in order to up-​ and down-​
modulate production of specific blood cell types. Progenitors differ-
entiate into lineage-​restricted precursor cells and eventually mature
effector cells of the haematopoietic system. These mature lineages
include erythroid cells for oxygen transport, myeloid and lymphoid
cells that provide immune defence, and megakaryocytes and plate-
lets essential for haemostasis.
HSC
haemopoietic stem cells
MPP
multipotential progenitor/short-term HSC
lymphoid primed multipotential progenitor
myelolymphoid progenitor
common myeloid progenitor
granulocyte myeloid progenitor
megakaryocyte erythroid progenitor
megakaryocyte progenitor
erythroid progenitor
Quiescent
Proliferate
in response
to stimuli
Lymphopoiesis
static cell numbers in steady state
T-cell
B-cells
PC Dendritic Cells
Natural Killer Cells
Neutrophils
Monocytes
Monocyte DC
Myeloid DC
Mks/
Platelets
Erythroid
Cells
ProIiferation
Self Renewal
Quiescence
HSC
MPP
LMPP
CMP
GMP
MLP
MEP
MkP
BFU-E
Myelopoiesis
10 billion cells/day
LMPP
MLP
CMP
GMP
MEP
MkP
BFU-E
Fig. 22.2.1.4  One current model of human adult haematopoiesis. A self-​renewing HSC gives rise to a
multipotent progenitor (MPP) that gives rise to a lymphoid-​primed MPP (LMPP), both of which may have
short-​term self-​renewal capacity. The hierarchy following the LMPP further bifurcates into a multilymphoid
progenitor (MLP) and a granulocyte macrophage precursor (GMP). The common myeloid progenitor (CMP)
is likely to be a heterogeneous population that gives rise to the megakaryocyte/​erythroid progenitor (MEP)
and the GMP.
SECTION 22  Haematological disorders
5176
Phenotypic characterization and
isolation of HSPCs
Attempts to purify stem cell populations have used a combination
of approaches based on physical and biological properties and cell
surface marker expression. Early work on murine BM revealed that
transplantable HSCs co-​purified with lymphocytes and led to the
idea that HSCs are morphologically indistinguishable from lympho-
cytes. Density gradient separation, such as Ficoll and Percoll gra-
dients, are commonly used as a pre-​enrichment step in stem cell
purification protocols. Progenitor cells cycle actively, whereas HSCs
are relatively quiescent. This difference has been exploited in techniques
for HSC enrichment in mouse and human systems. For example,
treatment of mice with the antimetabolite agent fluorouracil mark-
edly reduces progenitor cells, while relatively sparing populations
enriched in HSC activity.
More recently, considerable progress has been made in prospect-
ively isolating HSPCs using flow cytometry and cell surface markers.
However, it is difficult to compare different immunophenotyping
strategies with respect to quantifying the purity of the HSC popula-
tion. Numerous factors, such as the source of HSC (umbilical cord vs
marrow), route of transplant (intravenous vs intrafemoral), the type
of immunodeficient mouse used for xenografting, and pretransplant
manipulation of the cells can all affect the quantification of the purity
of HSC populations.
Figure 22.2.1.5 schematically illustrates how HSCs are isolated
and tested for function. Similar approaches are used to isolate pro-
genitor cells. Haematopoietic tissues are isolated, cells disassociated,
and then labelled with panels of fluorescently conjugated antibodies.
Cell populations can then be analysed and separated on a fluores-
cence activated cell sorter. Early studies to isolate human HSCs found
that approximately 1% of human BM cells express CD34. Isolation
of CD34+ cells enriches for HSPCs and haematopoietic engraftment
YS
BM
disassociate
cells
SSC
FSC
CD34
CD38
CD34
CD38
NOD–SCID mouse
1
2
long-term culture initiator cells (LTC-IC)
cobblestone area forming assay (CAFC)
colony assay
liquid culture assay
in vivo engraftment assay
in vitro clonogenic assay
CD34
CD38
FL
AGM
Fig. 22.2.1.5  Isolation of haematopoietic stem cells. HSCs can be isolated from different
sources. Cells are initially disassociated and stained with multiple antibodies. They are then
separated using a fluorescent activated cell sorter. Here mononuclear live cells are separated in
gate 1. These live cells are then analysed for CD34 and CD38 expression. The live CD34+CD38−
cell population is enriched for stem cell potential. Further purification can be undertaken on the
basis of additional cell surface markers such as CD45RA, CD90, and CD49f and efflux of dyes,
for example, rhodamine. To test functionality of isolated (sorted) cells, cells can be tested in
in vivo assays (transplanted into immunodeficient mice) and in vitro in long-​term culture
(long-​term culture-​initiating cell culture assay and cobblestone-​area forming assay), clonogenic
colony assays, and liquid culture assays.
22.2.1  Cellular and molecular basis of haematopoiesis
5177
when transplanted into irradiated nonhuman primates. Similarly,
human CD34+ selected cells contain HSPCs capable of fully recon-
stituting the ​haematopoietic system in humans after myeloablative
chemotherapy and radiation therapy. Although a CD34-​expressing
population contains long-​term repopulating HSCs, there is some
evidence of an upstream deeply quiescent CD34− population that
gives rise to the CD34+ HSCs. Approximately 5 to 25% of CD34+
cells also express low to moderate levels of CD90. CD90 expression
by human haematopoietic cells decreases with differentiation, and
most lineage-​restricted progenitors are CD34+CD90−/​low cells.
Additional studies demonstrate that human HSCs do not express
mature cell lineage markers (Lin−), CD45RA or CD38. Isolation of
Lin− CD34+CD38−CD45RA−CD90+ cells provides a relatively
easy method to sort for putative HSCs. Sorting for the integrin
CD49f further enriches for HSCs. Others have combined some of
these markers with the ability of HSC to efflux dyes (e.g. the mito-
chondrial dye rhodamine-​123). HSCs, but not progenitor cells, ex-
press high levels of the verapamil-​sensitive multidrug-​resistance
membrane efflux pump (P-​glycoprotein), which confers resistance
to multiple chemotherapeutic agents. This pump also excludes cer-
tain fluorescent dyes, such as rhodamine 123 or Hoechst 33342. By
using these dyes in combination with flow cytometry, it has been
possible to identify a population of haematopoietic cells with low
dye retention, so-​called side population (SP) cells. Although this
population is markedly enriched for HSCs, SP cells still represent
a heterogeneous mix and are not equivalent to pure HSCs. While
the SP phenotype has been useful in characterizing murine HSCs,
this characteristic has not translated as easily into the human system.
Though HSPC populations as currently defined are still impure
these methods have allowed isolation of cell populations of defined
functionality and are useful to identify genes and signalling pathways
that mediate human HSPC differentiation. Isolation of distinct HSPC
populations is also beginning to permit careful dissection of the hier-
archical relationships between different blood cell populations: an
essential step in describing the cellular basis of normal haematopoi-
esis. In turn, this is critical when trying to understand (1) the normal
cellular compartments where genetic and epigenetic changes initially
occur in haematological diseases (i.e. the disease initiating cell popu-
lations); (2)  the cell compartments where subsequent mutations/​
epigenetic change is acquired during disease evolution and how this
changes the haemopoietic hierarchy; and (3)  the cell populations
that propagate haemopoietic disease. Advances in cell sorting, gen-
etic analysis, and single-​cell biology are making analysis of HSPCs
increasingly precise. This progress will certainly provide additional
insight into HSC biology and heterogeneity in the near future.
In routine clinical transplantation practice, isolation of HSCs to
high purity is not necessary. Safe transplantation is routinely possible
with both unfractionated mononuclear cells and CD34+ purified
grafts (usually by magnetic beads that enrich CD34+ cells to c.60–​
85% purity). Indeed, transplantation of highly purified HSCs may
delay engraftment, as initial engraftment in patients is most likely
from progenitors rather than HSCs. Nevertheless, there may be merit
in transplanting purified HSPCs in some clinical situations—​for ex-
ample, to reduce transplantation of contaminating tumour cells in
autologous transplantation. In this regard, there are encouraging clin-
ical studies where transplantation of CD34+Thy1(CD90)+ cells has
been employed. However, this requires a flow cytometry-​based iso-
lation procedure that is difficult to implement on a widespread basis.
Pluripotent stem cells and haematopoiesis
Mouse embryonic stem cells (mESCs) were first isolated in 1981.
mESCs have proven invaluable for studies of basic mammalian
developmental biology, including haematopoietic development.
Unlike adult stem cells (such as HSCs), ESCs are able to undergo
self-​renewal indefinitely in culture, yet maintain the ability to form
all somatic cell lineages (including haematopoietic cells). Studies
with mESCs have been crucial to identifying genes that regulate
haematopoietic development, through gene deletion and/​or ma-
nipulation. Notably, attempts to derive HSCs capable of long-​term
multilineage engraftment have largely only been successful when
mESCs have been genetically manipulated by overexpression of spe-
cific transcription factors, for example, HoxB4.
The first description of human ESCs was in 1998. Like mESCs,
hESCs can be maintained indefinitely as a self-​renewing popu-
lation in culture, yet maintain the ability to form all somatic cell
populations. hESCs have also been used to investigate human
haematopoiesis and have generated considerable interest because
of their potential to produce large numbers of human cells and
tissues suitable for studying disease mechanisms, transplantation,
or transfusion medicine. For example, there has been consider-
able interest in using hESCs to produce red blood cells or platelets
as an adjunct to the supply from blood donation. Additionally,
the potential to produce HSCs from hESCs is of great interest.
To date, however, although most mature blood cell populations
have been produced from hESCs, it has not been possible to iso-
late HSCs to any reasonable extent. Even genetic manipulation
and overexpression of transcription factors, such as HoxB4, has
not been similarly effective in the human system. Considerable ef-
forts to identify strategies to improve development of HSCs from
hESCs are ongoing.
Another important cell type is the induced pluripotent stem cell.
iPSCs can be derived from various somatic cell populations, typic-
ally by expression of a limited number of ‘reprogramming genes’
that are able to convert the somatic cell population into cells that
function like embryonic stem cells. These studies were first per-
formed in mouse cells in 2006 and then from human cells in 2007.
Like their ESC counterparts, iPSCs have been used to derive diverse
haematopoietic cell lineages. Again, to date, transplantable HSCs
have not been derived from iPSCs. This field will continue to mature,
and there is considerable interest in deriving iPSCs from individuals
with genetic deficiencies to model genetic disease. Using iPSCs, gene
correction strategies or other means to overcome the genetic defect
can be analysed. This may lead to effective therapies based on using
iPSCs as a screening resource and would not require direct trans-
plantation of iPSC derived cells. Additionally, future developments
may allow for derivation of iPSCs from individuals with haemato-
logical or other diseases and use of these cells to produce essentially
autologous replacement cell populations.
Haematopoietic niche
In the adult, haematopoietic cell differentiation from HSPCs is regu-
lated by signals provided by the BM microenvironment. The spe-
cific cellular constituents of the microenvironment that influence
blood cell development are still being elucidated. They include
SECTION 22  Haematological disorders
5178
mesenchymal cells, endothelial and neural cells, haematopoietic
cells, and extracellular matrix. The heterogeneous mesenchymal
stromal cell (MSC) population plays a significant role in the haem-
atopoietic niche. Some MSCs are part of the continuum of cells that
produce bone and some are perivascular without a clear role in skel-
etal biology. Both of these cell types influence haematopoiesis. For
example, mature osteoblasts are important in stem cell mobiliza-
tion, nestin-​positive mesenchymal stem cells are important for HSC
persistence, and adipocytes have been implicated as negative regu-
lators of HSC number. In human studies, cord blood co-​cultured
with MSCs undergo a median 30-​fold expansion of CD34+ cells
resulting in significantly improved engraftment. In mouse, leptin
receptor-​expressing (LepR+) cells represent the major proportion
of MSCs. LepR+ cells appear to be the main source of new osteo-
blasts and adipocytes in adult bone marrow and form bony ossicles
supportive of haematopoiesis in vivo. LepR+ MSCs are the major
source of the cytokines stem cell factor (SCF) and chemokine (CXC
motif) ligand 12 (CXCL12) in mouse bone marrow. Conditional
deletion of the stem cell factor gene (Scf) in LepR+ cells leads to
depletion of quiescent HSCs and conditional deletion of the gene
encoding CXCL12 (Cxcl12 also called Sdf1) in LepR+ cells leads to
HSC mobilization.
Other cell types, such as neural cells of the sympathetic nervous
system and nonmyelinated Schwann cells, also support HSCs. The
sympathetic nervous system mediates circadian modulation in
the number of HSCs moving from BM to bloodstream on a daily
basis. Mature haematopoietic cells are also thought to influence
HSC function in the BM. Specifically, macrophages help regulate
HSC mobilization into blood and T cells are thought to influence
HSC engraftment and provide relative protection from immune
attack. Megakaryocytes have been shown to be important for
maintaining HSC quiescence, with various megakaryocyte-​secreted
factors, including CXCL4, transforming growth factor-​β1 (TGFβ1),
and thrombopoietin, implicated in this role. Therefore, a complex
admixture of cells participates in what is designated as the stem cell
niche (Fig. 22.2.1.6).
sinusoidal
endothelial
stem cell niche
HSC
HSC
self-
renewal
stem cell endothelial niche — spindle-shaped osteoblast cells
non-stem cell niche
environmental
asymmetry
MPP
differentiation
divisional
asymmetry
– Myh CXCL4
– osteopontin
niche players
N-cadherin
HSC
SNO cells
TIE2
ANG-1
CXCL12
CXCR4
2 Ca sensory receptor
3 SDF1 CXCL 12 receptor CXCR4
bone
4 Notch/Jagged
5 BMP
6 ICAM-1
7 N-cadherin
1 Osteopontin
(a)
(b)
Fig. 22.2.1.6  (a) The bone marrow niche, which, in part, consists of sinusoidal endothelial cells,
helps control haematopoietic stem cell (HSC) fate. HSCs can be quiescent (G0 of cell cycle) or can
enter the cycle to divide symmetrically or asymmetrically (divisional asymmetry) to self-​renew
and/​or to produce more differentiated cells such as multipotential progenitors (MPPs). HSCs can
also migrate into and out of the niche (environmental asymmetry). The components of the niche
are shown below. (b) This shows an HSC anchored into the niche via binding of: (i) cell surface
receptor TIE-​2/​TEK binding to its ligand angiopoietin-​1 (ANG-​1) on sinusoidal endothelial cells
(SNO cells) and (ii) the CXC-​chemokine ligand 12 on SNO cells binding to its receptor CXCR4.
22.2.1  Cellular and molecular basis of haematopoiesis
5179
The niche serves several functions important for haematopoi-
esis. The first is the regulation of stem cell self-​renewal, a process
that requires expression of molecules such as SCF and members
of the WNT family. The second is control of stem cell number. The
third is the coordinated regulation of proliferation and differenti-
ation of HSCs. When the niche is perturbed in mice, it can lead to
myeloproliferative or myelodysplastic phenotypes. The fourth is
cell localization, a process that is important in the context of either
harvesting stem cells by mobilization into the blood or delivery of
transplanted HSCs to enable engraftment.
Thus, the HSC niche is a critical aspect of the regulated pro-
duction of blood cells throughout life. Unravelling how stem
cells enter and leave the niche will lead to improved methods to
mobilize stem cells for clinical harvest (see ‘HSPC circulation,
homing, and mobilization’). Ongoing efforts to improve stem cell
engraftment into the niche and to discern how the niche contrib-
utes to disease may contribute to future manipulation of the niche
for clinical benefit.
Regulation of haematopoietic differentiation
A complex network of transcription factor and growth factor
signalling pathways regulates HSPC self-​renewal, lineage com-
mitment, and differentiation. Transcription factors (TFs) that are
expressed either exclusively in blood cells, or have restricted tissue-​
specific patterns of expression, play important roles in regulating
blood production. Furthermore, acquired driver mutations of these
TFs are pathogenic in haematological diseases such as lymphoma
and leukaemia. The importance of these TFs is also underscored by
the conserved role they play in haematopoiesis through evolution.
Over the last two decades, this attribute has allowed the function of
these TFs to be extensively investigated in animal models. In these
models, genes encoding critical TFs have been deleted, modified,
overexpressed, and misexpressed. A summary of the site of action of
some of these TFs is shown in Fig. 22.2.1.7.
A thorough description of the function of these proteins is not
possible here. Key points that arise from these studies are as follows:
RUNX1
SCL
LMO2
ETV6
GATA2
GATA2
GATA2
GATA1
ZFPM1
KLF1
SCL
MYB
BCL11A
STAT5
GATA2
GATA1
ZFPM1
ETS Factors
RUNX1
SCL
MYB
GATA2
GATA1
STAT5
GATA 1
CEBPE
STAT3
GFI1
PU.1
CEPBA
CEBPE
TCF3
IKAROS
E2A
PU.1
EBF
PAX5
BLIMP1
PU.1
CEPBA
CEBPE
IRF 8
EGR1/2
E2A
IZKF
NOTCH
GATA3
RUNX1
TCF-1
BCL11B
HEB
MYB
B-cells
Monocytes
Neutrophils
Eosinophils
Mast cells
Megakaryocytes
Red cells
HSC
in development
HSC
in adult life
T-cells
Fig. 22.2.1.7  A schematic representation of the key haematopoietic-​specific transcription factors (in
boxes) required for specification and/​or maturation of different haemopoietic lineages. Thus, for example,
the transcription factors GATA2, RUNX1, SCL, LMO2, and ETV6 are all required for specification of HSCs in
development. GATA2 is required for maintenance of HSCs in adult life. Transcription factors required for each of
eight mature blood lineages are shown.
SECTION 22  Haematological disorders
5180
•	TFs are divided into families that have similar proteins domains.
•	They often bind DNA and interact with other proteins (other TFs
and proteins that control transcription) via specific domains.
•	TFs work in combinations to both activate and repress the expres-
sion of a large number of genes.
•	TFs are required at discrete stages of haematopoiesis and any one
TF often functions at multiple stages within one lineage and can
function in more than one lineage.
•	Ultimately, TFs work in complicated networks that can be mod-
elled much like semiconductor/​computing networks. TFs work
in negative feedback loops, feed-​forward loops, and cross-​
antagonistic loops to mention just three such types of interaction.
•	The function of TFs helps that regulate the cell’s potential to make
blood cells of different lineages, proliferate, undergo apoptosis and
self-​renew.
More specifically, the TFs SCL/​TAL1 and LMO2 are required
to specify HSC from mesoderm. The TFs RUNX1 (AML1), TEL1,
MLL, and GATA2 are required to maintain stem cells once they have
been specified. In myelopoiesis, the TFs PU.1, the C/​EBP family
(C/​EBPα and C/​EBPε), GFI-​1, EGR-​1, and NAB2 all promote the
granulocyte-​macrophage lineage programmes. GATA2 is required
in stem/​early progenitor cells but is also required for mast cell dif-
ferentiation and in the early phases of megakaryocyte–​erythroid
lineage maturation. Working with GATA2 to promote erythropoi-
esis and megakaryopoiesis are GATA1, FOG1, SCL, EKLF, p45NF-​
E2, and FlI-​1. In early lymphopoiesis, the TF Ikaros is required.
In B-​lymphopoiesis, the TFs E2A (and its family members), EBF,
and PAX5 are required and finally the TF BLIMP1 is necessary for
plasma cell formation. In T-​cell maturation, notch signalling acti-
vates the TF CSL, which works with the TFs GATA3, T-​BET, NFATc,
and FOXP3. Of note, the TFs SCL/​TAL1, MLL, RUNX1, LMO2,
PU.1, C/​EBPa, PAX5, E2A, and GATA1 are all implicated in the
pathogenesis of human leukaemia.
In addition to TFs, transcription is also controlled at the level of
accessibility of DNA that is packaged in chromatin in the cell nu-
cleus. A large number of proteins regulate accessibility by reversibly
altering DNA methylation and modifications of histone that package
DNA. These alterations in packaging alter gene expression without
changing the DNA sequence and are known as epigenetic changes.
Protein expression is also regulated by alternative RNA splicing that
alters the structure of the translated protein. Genes encoding epi-
genetic regulators and RNA splicing factors are recurrently mutated
in blood cancers, underscoring their importance in haematopoietic
differentiation. Some of the epigenetic regulators mutated in blood
cancers are shown in Fig. 22.2.1.8.
HSPC circulation, homing, and mobilization
HSPCs migrate from one site of blood cell production, enter the circu-
lation, home, and enter other supportive sites. This certainly is the case
in development before the existence of bones in the fetus. HSPCs even-
tually transit from fetal liver to nascent BM to establish haematopoi-
esis via the bloodstream. Even once haematopoiesis ­becomes restricted
to the bone marrow, HSPCs traffic into and out of the BM regularly.
Experiments using parabiotic mice, in which the circulations of two
separate mice are joined surgically, have indicated that murine HSCs
exit the BM and transit through the peripheral blood system at sur-
prisingly high flux rates (estimated to be c.104–​105 long-​term repopu-
lating HSCs per day in a mouse). Other mouse studies have shown
that macrophages and osteoblasts are critical for granulocyte colony-​
stimulating factor (G-​CSF)-​mediated effects on HSPC trafficking
through regulation of the SDF-​1–​CXCR4 axis (see later in this section).
This ability of HSPCs to move from the BM to the peripheral
bloodstream is exploited for collection of stem cells for clinical
haematopoietic cell transplantation. The rate, timing, and destin-
ation of the HSPCs that circulate from the BM to the periphery
involve chemokines and their receptors, especially CXCL12 (or
SDF-​1) and its receptor CXCR4; integrins, particularly very late
antigen-​4 (VLA-​4, also termed integrin α4β1); selectins, such as P-​
selectin glycoprotein ligand-​1 (PSGL1, also termed CD162); and the
calcium-​sensing receptor and the intracellular-​signalling molecules,
Gαs and Rac1/​Rac2. CXCL12/​SDF-​1 is produced by several marrow
stromal cell types. These include immature osteoblasts located in
the endosteal region adjacent to stem cell niches and the endothe-
lium. The SDF-​1 receptor, CXCR4, is expressed on a wide variety
of haematopoietic cells, and is important in stem cell self-​renewal
IDH1/2
TET2
DNMT3A
EZH2
Trithorax proteins
MLL
HATs
HDACs
ASXL1
Polycomb repressive
complex 2 (PRC2)
hydroxymethylcytosine
methylcytosine
H3K27me3
H3K4me3
H3/H4 Ac
PRMT5
H2A/H4me
+
+
Fig. 22.2.1.8  Schematic representation of DNA (orange) wrapped
around histone octamers (blue discs). DNA can be reversibly methylated
on cytosine (green hexagon) by the DNA methyltransferases (DNMTs).
Demethylation of DNA occurs via intermediates that include
hydroxymethylcytosine (blue octagon). Demethylation requires the
TET2, IDH1, and IDH2 proteins. Histones have peptide tails that protrude
away from the histone octamer (shown as brown lines). Amino acids
in these tails can be post-​translationally modified. Examples of these
modifications include (1) acetylation (Ac) at histones H3 and H4 (shown
by orange ball), by histone acetyltransferases (HATs). Acetyl marks can
be removed by histone deacetylases (HDACs). (2) Methylation on
lysine (H3K27—​purple ball and H3K4—​blue ball) or arginine (H2A and
H4 purple ball). The H3K27 methylation is mediated, in part, by the
polycomb repressive complex 2 (PRC 2). H3K4 methylation is mediated
by the trithorax protein family, which includes MLL as a member. One of
the arginine methyltransferases is PRMT5. Epigenetic changes associated
with gene activation include histone acetylation, H3K4 trimethylation,
and DNA demethylation. Proteins mediating these changes are shown
in a blue font. Conversely, epigenetic changes associated with gene
repression include H3K27 trimethylation, arginine methylation, and DNA
methylation. Proteins mediating these changes are shown in a red font.

22.2.2 Diagnostic techniques in the assessment of
22.2.2 Diagnostic techniques in the assessment of haematological malignancies 5181 Wendy N. Erber
22.2.2  Diagnostic techniques in the assessment of haematological malignancies
5181
and retention of maturing haematopoietic cells within the marrow
as shown by targeting of the gene in mice. Mobilization of stem cells
from the marrow to the peripheral blood by G-​CSF is thought to be
secondary to reduction of SDF-​1 transcripts and proteolytic cleavage
of the protein. Finally, high cell surface expression of CD47 is also
important in allowing circulating HSPCs to evade phagocytosis.
Instead of needing to collect BM as a source of HSPC, ‘mo-
bilized’ peripheral blood HSPCs can be collected by apheresis.
Different preparative regimens and growth factors (e.g. G-​CSF)
can be used to increase the number of HSPCs collected, with enu-
meration of CD34+ cells used as a surrogate marker of HSPCs.
Within the CD34+ population, most cells are progenitors and the
frequency of true HSCs is exceedingly rare. Plerixafor, a drug that
blocks the chemokine receptor CXCR4 from binding to SDF-​1, is
now used clinically when needed to increase the numbers of cir-
culating HSPCs (given with G-​CSF). Ways to improve homing to
the marrow and/​or engraftment are also being studied, particu-
larly with the use of umbilical cord blood transplants. These in-
clude ex vivo fucosylation of cells to enhance selectin interactions,
inhibition of the extracellular peptidase (CD26), which cleaves
CXCL12/​SDF-​1, or pretreatment of donor cells with modified
prostaglandin E2 to expand HSCs and to improve homing to the
marrow. Thus studies of stem cell–​niche interactions may ultim-
ately impact clinical medicine, reducing the numbers of stem cells
needed for transplantation through more efficient mobilization,
homing, and engraftment.
FURTHER READING
Cellular basis of haematopoiesis
Bonig H, Papyannopoulou T (2012). Mobilization of hematopoietic
stem/​progenitor cells:  general principles and molecular mechan-
isms. Methods Mol Biol, 904, 1–​14.
Orkin SH, Zon LI (2008). Hematopoiesis: an evolving paradigm of
stem cell biology. Cell, 132, 631–​44.
Park D, Sykes DB, Scadden DT (2012). The hematopoietic stem cell
niche. Front Biosci, 17, 30–​9.
Ontogeny of haematopoiesis
Dzierzak E, Speck NA (2008). Of lineage and legacy: the development
of mammalian hematopoietic stem cells. Nat Immunol, 9, 129–​36.
Stem cell characteristics
Baum CM, et al. (1992). Isolation of a candidate human hematopoietic
stem-​cell population. Proc Natl Acad Sci U S A, 89, 2804–​8.
Copley MR, Beer PA, Eaves CJ (2012). Hematopoietic stem cell hetero-
geneity takes center stage. Cell Stem Cell, 10, 690–​7.
Görgens A, et  al. (2013). Revision of the human hematopoietic
tree: granulocyte subtypes derive from distinct hematopoietic lin-
eages. Cell Rep, 3, 1539–​52.
Notta F, et al. (2011). Isolation of single human hematopoietic stem
cells capable of long-​term multilineage engraftment. Science, 333,
218–​21.
Stem cell sources, ex vivo expansion of HSCs, pluripotent
stem cells, and haematopoiesis
Anasetti C, et  al. (2012). Peripheral-​blood stem cells versus bone
marrow from unrelated donors. N Engl J Med, 364, 1487–​96.
Boitano AE, et al. (2010). Aryl hydrocarbon receptor antagonists pro-
mote the expansion of human hematopoietic stem cells. Science,
329, 1345–​8.
Brunstein CG, et al. (2010). Allogeneic hematopoietic cell transplant-
ation for hematologic malignancy:  relative risks and benefits of
double umbilical cord blood. Blood, 116, 4693–​9.
Kaufman DS (2009). Toward clinical therapies using hematopoietic
cells derived from human pluripotent stem cells. Blood, 114, 3513–​23.
Rocha V, et al. (2004); Acute Leukemia Working Party of European
Blood and Marrow Transplant Group; Eurocord-​Netcord Registry.
Transplants of umbilical-​cord blood or bone marrow from unrelated
donors in adults with acute leukemia. N Engl J Med, 351, 2276–​85.
Shultz LD, Ishikawa F, Greiner DL (2007). Humanized mice in transla-
tional biomedical research. Nat Rev Immunol, 7, 118–​30.
Stadtfeld M, Hochedlinger K (2010). Induced pluripotency: history,
mechanisms, and applications. Genes Dev, 24, 2239–​63.
Yu J, et al. (2007). Induced pluripotent stem cell lines derived from
human somatic cells. Science, 318, 1917–​20.
22.2.2  Diagnostic techniques in
the assessment of haematological
malignancies
Wendy N. Erber
ESSENTIALS
The diagnosis of haematological malignancies requires an under-
standing of the diseases and the uses and limitations of the range of
available investigations. The relative importance of different investi-
gations varies by disease entity.
The blood count is one of the most widely used tests in all of medi-
cine and often the first indication of an underlying haematological
malignancy. Some blood count features are ‘diagnostic’ and others
may give an indication of a bone marrow defect.
Morphological assessment of a stained blood film adds value to
an abnormal blood count. It may identify abnormal morphology of
red cells, leucocytes, or platelets which may be specific and diag-
nostic (e.g. lymphoma cells), or give clues suggesting a diagnosis (e.g.
red cell rouleaux formation in plasma cell dyscrasias).
The bone marrow aspirate (liquid sample) gives cytological ­detail,
whereas the trephine biopsy provides information about marrow
cellularity, architecture, cellular distribution, and extent of fibrosis.
Immunophenotyping detects cellular antigens in clinical samples
and is essential in the diagnosis and classification of haematological
malignancies. It is also used for disease staging and monitoring, to
detect surrogate markers of genetic aberrations, identify potential
immunotherapeutic targets, and to aid prognostic prediction.
Cytogenetics assesses the number and structure of whole
chromosomes (e.g. the presence of chromosomal translocations)
and chromosomal regions in neoplastic cells and is performed to
diagnose and classify some haematological malignancies.
SECTION 22  Haematological disorders
5182
Molecular genetic methods facilitate the detection of mutations,
rearrangements, or translocations in genes. Applications in malig-
nant haematology include confirming clonality, detecting disease-​
associated genotypes, determining prognosis, disease monitoring
following therapy, predicting imminent clinical relapse, and identifying
patients who are likely (or not) to respond to new targeted inhibitor
therapies.
Introduction
The diagnosis of haematological malignancies is complex and
evolving rapidly with many test types available. Optimal test util-
ization requires an understanding of the diseases and the uses (and
limitations) of the range of available investigations. Morphology,
cell phenotyping, cytogenetics, and molecular genetics all have roles
with their relative importance varying by disease entity. To use the
tests appropriately requires an understanding of the principles of
each test type and how they supplement traditional analyses such
as blood count and morphological assessment of blood and bone
marrow. This chapter provides a guide to the most frequently per-
formed tests in the diagnostic assessment of haematological ma-
lignancies as well as those that are currently under development.
It covers the blood count and blood film, bone marrow examin-
ation, immunophenotyping, cytogenetics, and molecular genetics
(including ‘next-​generation’ sequencing (NGS)). The reader is re-
ferred to other chapters in this textbook that will discuss their appli-
cation to the assessment haematological malignancies.
The blood count
Until the early 1960s the blood count was a manual, laborious test
that required centrifugation, spectrophotometry, and cell counting
using etched grids. Through modern technology and computing
we now have high-​throughput automated analysers which use flow
technology, electrical impedance, optical light scatter, cytochemistry,
and/​or fluorescence to measure and count blood cells. These sophis-
ticated blood count machines generate large amounts of quantitative
numerical data including red cell indices, haemoglobin concentra-
tions, reticulocyte counts, leucocyte counts with differentials, and
platelet counts and indices. They also produce graphical data and
‘flag’ cells with abnormal features. Due to the low cost, simplicity,
and easy access, the blood count is one of the most widely used tests
in all of medicine. As it is commonly used as a ‘screening test’, this
is often the first indication of an underlying haematological malig-
nancy. Some blood count features are ‘diagnostic’. Other results may
give an indication of a bone marrow defect. These ‘indicative’ blood
count features could be in one or more of red cells (anaemia or poly-
cythaemia), leucocytes (-​cytopenias or -​cytoses; abnormal ‘flags’),
or platelets.
Anaemia is a common presenting abnormality for many haem-
atological malignancies as a result of reduced erythropoiesis. The
red cells are generally normochromic and normocytic. Elevated
mean cell volume (MCV) is seen with myelodysplasia and com-
monly with plasma cell myeloma. Dimorphic red cells (elevated red
cell distribution width) are seen in some cases of myelodysplasia
such as those with ring sideroblasts. Polycythaemia vera, charac-
terized by increased red cell production, presents with an elevated
haemoglobin concentration and haematocrit; the red cells may be
normocytic or hypochromic and microcytic (low MCV) if there is
iron depletion.
Abnormal leucocyte number and morphology are also common
at diagnosis of a haematological malignancy. Neutrophilia, although
most commonly secondary to infection, inflammation, haemor-
rhage, or drugs, can also be seen in neoplastic disorders, particu-
larly those of myeloid origin. A mild reactive neutrophilia can be
seen as part of the acute phase response (e.g. in Hodgkin lymphoma)
whereas a neutrophilia of 50–​100 × 109/​litre (‘leukaemoid’ re-
action) may be in response to a nonhaematopoietic malignancy
(e.g. carcinoma of lung, mesothelioma). Neutropenia may be iso-
lated, occur in conjunction with anaemia or thrombocytopenia, or
be part of a pancytopenia. Possible causes include failure or sup-
pression of granulopoiesis due to bone marrow failure, fibrosis
or infiltration, drug therapy, and toxins. The blood count may
highlight ‘left shift’ due to the presence of immature or abnormal
granulocytic progenitors in the circulation; this can occur in
dysgranulopoiesis (e.g. myelodysplastic syndrome) or in response
to peripheral consumption or destruction of mature neutrophils or
their precursors (e.g. immune mechanisms) such as in lymphoid
neoplasms or hypersplenism. Monocytosis (>1 × 109/​litre) is a
defining feature of chronic myelomonocytic leukaemia and juvenile
myelomonocytic leukaemia and can be seen in chronic myeloid leu-
kaemia. Monocytopenia is rare but is seen in hairy cell leukaemia.
Eosinophilia is most commonly secondary to reactions to aller-
gens, parasites, or drugs. Primary eosinophil disorders (i.e. chronic
eosinophilic leukaemia and myeloid or lymphoid disorders with
abnormalities of PDGFRA, PDGFRB, or FGFR1) are rare but eo-
sinophilia may be a ‘bystander’ feature in other haematological ma-
lignancies (e.g. Hodgkin lymphoma). Peripheral blood basophilia
is exceedingly rare, and, when present, should raise suspicion of a
myeloproliferative neoplasm, in particular chronic myeloid leu-
kaemia. Thrombocytopenia, common at presentation of a haemato-
logical malignancy, is usually due to reduced megakaryopoiesis and
platelet morphology is normal. Thrombocytosis can be seen with
myeloproliferative neoplasms, although is more commonly seen in
response to infection or inflammation.
Pancytopenia, a reduction in all blood cells, suggests failure of
normal haematopoiesis (inherited or acquired), bone marrow in-
filtration, or reduced survival of all blood cells as a consequence
of drugs or infections. The diagnostic possibilities that arise from
an abnormal blood count must be interpreted in the context of the
patient’s age, clinical scenario, and associated findings (e.g. physical
examination, biochemistry).
Blood film
Morphological assessment of a stained blood film adds value to
an abnormal blood count as it may provide an explanation for the
quantitative and qualitative (i.e. ‘flags’) abnormalities. The film may
identify abnormal morphology of red cells, leucocytes, or plate-
lets which may be specific and diagnostic (e.g. lymphoma cells).
Alternatively, they may be features that are associated with, but not
diagnostic, of a clinical entity, that is, ‘the company the cells keep’.
22.2.2  Diagnostic techniques in the assessment of haematological malignancies
5183
These accompanying abnormalities (e.g. red cell rouleaux formation
in plasma cell dyscrasias) may help generate a provisional diagnosis.
Some of the blood film abnormalities that may indicate a malignant
bone marrow disorders are described in the following paragraphs.
Abnormal leucocytes in the blood may be diagnostic of a haem-
atological malignancy. Abnormal lymphoid cells on a blood film
are most commonly reactive (e.g. secondary to viral infection) but
on rare occasions may be neoplastic. Some malignant cells have
characteristic morphology (e.g. hairy cell leukaemia, follicular
lymphoma—​see also Chapter 22.4.3). However, this is not always
the case and it can be challenging on morphology alone to distin-
guish between reactive and neoplastic cells. Correlation with clin-
ical history and serology is commonly required and, in unresolved
situations, flow cytometric immunophenotyping may be indi-
cated (see ‘Flow cytometric immunophenotyping’). Most reactive
lymphocytoses are of T-cells whereas neoplastic proliferations
are more commonly of a B-​cell lineage and have restricted kappa/​
lambda light-​chain expression.
The presence of circulating abnormal (‘dysplastic’) neutrophils
(e.g. abnormal size, granularity, nuclear segmentation, or chro-
matin condensation) indicates dysgranulopoiesis within the bone
marrow. The presence of isolated dysplastic promyelocytes in the
absence of other neutrophil precursors is highly suggestive of acute
promyelocytic leukaemia, which also commonly presents with pan-
cytopenia. Blast cells are not normally present in the blood. Their
presence may indicate recovery from bone marrow failure, severe
sepsis, cytokine administration, or underlying bone marrow path-
ology. They may also indicate acute leukaemia, myelodysplastic
syndromes, myelodysplastic/​myeloproliferative neoplasms, chronic
myeloid leukaemia, primary myelofibrosis, or bone marrow in-
filtration by a nonhaematopoietic malignancy. More than 20%
blast cells in the blood (or bone marrow) defines acute leukaemia.
Plasma cells do not commonly appear in the circulation and their
presence indicates a florid B-​cell response to infection or a B-​cell
neoplasm with plasmacytoid differentiation (i.e. plasma cell mye-
loma, plasma cell leukaemia, or lymphoplasmacytic lymphoma/​
Waldenström macroglobulinaemia). Plasma cells may be accom-
panied by cytopenias, high red cell MCV, leucoerythroblastic film,
rouleaux formation, and background protein staining of the film.
Additional biochemical investigations (e.g. serum protein analysis)
and bone marrow examination may be required.
Although red cells and platelets are rarely directly involved in
the malignancy per se, there are commonly ‘accompanying’ abnor-
malities in these lineages. Red blood cell morphological abnormal-
ities include rouleaux (plasma cell neoplasms), spherocytes and red
cell agglutination (autoimmune conditions with a mature B-​cell
neoplasm), teardrop poikilocytes (myelofibrosis or metastatic in-
filtration of the marrow), and hyposplenic features (splenic infiltra-
tion). Circulating erythroblasts (nucleated red cells) with abnormal
morphology imply bone marrow dyserythropoiesis suggestive of
myelodysplastic syndrome or acute myeloid leukaemia. Abnormal
platelet morphology occurs with bone marrow dysmegakaryopoiesis
such as in the myeloproliferative neoplasms and myelodysplastic
syndromes. A  leucoerythroblastic blood film (i.e. presence of
erythroid and leucocyte precursors in the blood) is seen with bone
marrow infiltration (i.e. haematological malignancy, metastatic in-
filtrate, or marrow fibrosis), severe sepsis, cytokine administration,
and prolonged hypoxia. There may be accompanying other features
(as previously mentioned) that may shed light on the diagnosis.
From this it can be seen that blood film review is crucial: the findings
may be diagnostic, and, if not, they guide the next step in the inves-
tigation process. Is a bone marrow examination required to reach a
final diagnosis or can flow cytometric immunophenotyping be used
to determine the diagnosis?
Bone marrow examination
Examination of the bone marrow may be required to determine
the cause of unexplained blood count or film abnormalities or to
confirm a diagnosis suspected from the blood count and film.
Indications for a bone marrow examination have been developed
by the International Council for Standardization in Haematology
and are summarized in Box 22.2.2.1. Bone marrow aspirate and
trephine biopsy specimens are generally both taken and these pro-
vide complementary information. The aspirate (liquid sample) gives
cytological detail, whereas the trephine biopsy provides information
about the marrow cellularity, architecture, cellular distribution, and
extent of fibrosis. For some disorders, the aspirate may provide suffi-
cient information without the need for a trephine biopsy (e.g. acute
leukaemia). For others, the trephine biopsy is the prime diagnostic
material (e.g. lymphoma staging and myelofibrosis). In addition to
morphology, the marrow sample can be used for ancillary biological
tests required to reach a diagnosis and World Health Organization
(WHO) classification (i.e. immunophenotyping and molecular gen-
etics). It is beyond the scope of this chapter to describe the bone
marrow morphological findings in haematological malignancies as
this will be addressed in accompanying chapters.
Immunophenotyping
Immunophenotyping is the method by which antibodies are used
to detect cellular antigens in clinical samples and is essential in the
diagnosis and classification of haematological malignancies (as per
WHO criteria). It is also used for disease staging and monitoring,
to detect surrogate markers of genetic aberrations, identify poten-
tial immunotherapeutic targets, and to aid prognostic prediction
Box 22.2.2.1  Indications for bone marrow examination in the
diagnosis and assessment of haematological malignancies
•	 Investigation of unexplained cytopenia/​s or pancytopenia
•	 Investigation of unexplained blood film morphological abnormalities
•	 Investigation of a paraproteinaemia
•	 To confirm a diagnosis of a haematological malignancy made on
peripheral blood
•	 To classify a haematological malignancy
•	 To determine the extent of bone marrow involvement by a haemato-
logical malignancy
•	 Bone marrow staging of lymphoma
•	 To obtain prognostic information based on the pattern of marrow
infiltration
•	 To obtain specimens for ancillary studies in the investigation of haem-
atological malignancies
•	 Post-​therapy and post-​transplant assessment of haematological
malignancies
SECTION 22  Haematological disorders
5184
(Table 22.2.2.1). Immunophenotyping can be performed on single
cells in solution by flow cytometry or on sections of bone marrow
trephine biopsies by immunohistochemistry. Immunophenotyping
can be used to establish the lineage and stage of differentiation of
cells and provide a surrogate of clonality. It can make or confirm
a diagnosis based on ‘classical’ disease-​associated phenotypic pro-
files and classify according to WHO definitions. The technology
and antibody panels used to achieve this vary by sample type (fresh
or fixed), the suspected neoplasm, and the information required to
best characterize the cells of interest.
Flow cytometric immunophenotyping
Flow cytometric immunophenotyping is the technique of choice
for the assessment of cells in blood or aspirated bone marrow. It
requires only a small sample, performs high-​speed analysis of large
numbers of cells, and allows many cellular parameters to be as-
sessed simultaneously. It assesses individual cells in suspension for
the presence (or absence) of specific antigens. The sample is incu-
bated with preselected antibodies, each of which has an attached
fluorophore. Following exposure to a laser beam, the cells with
bound antibody (and fluorophore) emit light at a specific wave-
length which is captured by detectors. This signal is captured and
indicates the presence of the relevant antigen. Since morphology
cannot be assessed, cells of interest are ‘gated’, i.e., electronically
selected based on predefined criteria. This can be on light scatter
properties (related to cell size and internal complexity) and/​or
fluorescence (i.e. antigen expression, such as CD45). Both surface
membrane and intracellular antigens (cytoplasmic and nuclear)
can be assessed (Fig. 22.2.2.1). Flow cytometers are commonly
fitted with three or more lasers and there are many fluorophores
available for use. Hence, with this combination, it is possible to
assess eight or more antigens simultaneously in one cell. Flow
cytometry can therefore provide high diagnostic precision and
sensitivity. It can identify disease-​associated phenotypes and be
used for low-​level disease monitoring (i.e. ability to detect one cell
with a specific phenotype in 10 000 cells).
Imaging flow cytometry is a new technological development
which further refines flow cytometry but is yet to find its place
in diagnostic practice. In addition to generating standard flow
cytometric data, these instruments capture high-​resolution im-
ages of each cell using digital cameras. This enables the cells being
studied to be directly visualized, thereby overcoming the major
limitation of ‘standard’ flow cytometry. The cell imagery compo-
nent opens possibilities for further study of neoplastic cells, such
as detecting colocalized cellular molecules, ‘spot’ counting (e.g.
intracellular molecules), integrating phenotype and fluorescent in
situ hybridization (FISH), and studying biological process (e.g. cell
cycle, mitosis, or apoptosis).
Immunohistochemistry
Immunohistochemistry (also known as immunocytochemistry)
is another immunophenotyping method but where the testing is
performed on sections of tissue, in this case bone marrow tre-
phines or other haematological biopsies. It is performed using
antibodies (usually monoclonal) and antigen–​antibody binding
is detected with an enzyme (i.e. horseradish peroxidase or alka-
line phosphatase) and a chromogenic substrate. Cells of interest
are identified by their morphology and location by standard light
microscopy. The presence (or absence) of a chromogenic colour
reaction shows whether the antigen in question is expressed.
Both cell membrane and intracellular antigens can be detected.
Fig. 22.2.2.1  Example of flow cytometry of a case of acute myeloid leukaemia. The gated leukaemic (blast) cells on CD45 and side scatter (purple)
are then shown to express CD33 and myeloperoxidase. The cells are HLA-​DR negative and few express CD34 antigen.
Table 22.2.2.1  Clinical applications of immunophenotyping
Diagnosis and
classification
Determine cell lineage
Determine stage of cell differentiation
Classical disease-​associated phenotypes
Aberrant antigen expression
Clonality assessment
Undifferentiated neoplasms
Prognostic
prediction
Antigen expression and prognostic stratification
Staging
Extent of disease
Rare event analysis
Surrogate phenotype–​genotype correlation
Integrated phenotype and genotype
Therapeutic
applications
Detection of potential immunotherapeutic targets
Minimal/measurable residual disease assessment
Early detection of disease relapse
Bone marrow regeneration following therapy
22.2.2  Diagnostic techniques in the assessment of haematological malignancies
5185
Immunohistochemistry is widely used to refine and classify the
diagnosis of haematological malignancies. Other applications
include lymphoma staging, detecting antigens associated with
disease prognosis and potential immunotherapeutic targets, and
disease monitoring.
Cytogenetics
Cytogenetics is performed to diagnose and classify a number of
haematological malignancies according to WHO criteria. It as-
sesses the number and structure of whole chromosomes (e.g. the
presence of chromosomal translocations) and chromosomal re-
gions in neoplastic cells. The ‘gold standard’ tool for basic genetic
diagnosis of haematological malignancies is karyotyping. This de-
pends on the presence of dividing cells in the sample and is there-
fore generally performed on aspirated bone marrow. Although the
resolution of karyotyping is limited, and generally only 20 cells
are studied, it provides a global analysis of the entire genome.
Some karyotypic abnormalities provide a definitive diagnosis (e.g.
the Philadelphia chromosome arising from t(9;22)(q34;q11) in
chronic myeloid leukaemia). Other karyotypic changes have prog-
nostic significance. In childhood lymphoblastic leukaemia, for ex-
ample, near-haploidy (<30 chromosomes) is associated with a poor
prognosis whereas hyperdiploidy (51–​65 chromosomes) carries a
good prognosis. Karyotyping is used at diagnosis but since only
few cells are analysed it lacks sensitivity when only small numbers
of abnormal cells are present (e.g. monitoring for residual disease
following therapy).
In recent years, a range of FISH and high-​resolution array-​
based techniques have become integrated into the broader field
of cytogenetics. These identify chromosomal regions and can
be performed on nondividing cells (including smears and tissue
sections). This is particularly useful for haematological malignan-
cies with a low proliferative index, when chromosome morphology
is poor or when full chromosomal analysis is unsuccessful. FISH
is based on a single-​stranded DNA probe annealing to its comple-
mentary sequence in a target genome; it thereby detects and local-
izes specific DNA sequences in cell metaphases or interphase cells.
Fluorescent probes (i.e. whole chromosome paints or locus-​specific
probes) are used and these bind to those parts of the chromo-
some with which they show a high degree of sequence homology.
Binding of the probe to chromosomes is visualized by fluorescence
microscopy. In some clinical scenarios it may be necessary to carry
out FISH on specific cell types. To achieve this, phenotyping and
genotyping can be integrated in a single analysis. This method is
called ‘FICTION’ (Fluorescence Immunophenotyping and inter-
phase Cytogenetics as a Tool for the Investigation of Neoplasms)
and allows specific genetic abnormalities (e.g. gene transloca-
tions, numerical abnormalities) to be assessed in cells highlighted
(or identified) by their phenotype. Although complex, FICTION
increases the accuracy over standard FISH since the genetic ab-
erration is assessed only in the cell population of interest. Array-​
based comparative genomic hybridization and single nucleotide
polymorphism arrays are two other genetic methods that can be
applied to haematological malignancies. These have facilitated the
identification of novel chromosomal abnormalities but have not
yet been integrated into clinical practice.
Molecular genetics
A major change that has occurred over the past 10 years has been
the integration of molecular genetic methods into the diagnostic
workup of haematological malignancies. We now have a vast array
of techniques in our diagnostic armamentarium to facilitate the de-
tection of mutations, rearrangements, or translocations in genes.
Specific chromosomal changes and perturbed genes can be detected
down to single base changes in individual genes, moving diagnostics
from morphology to mutation. Applications in malignant haema-
tology include confirming clonality, detecting disease-​associated
genotypes, determining prognosis, disease monitoring following
therapy, and predicting imminent clinical relapse. Molecular gen-
etic tests are also increasingly being used to identify patients who
are likely to respond to new targeted inhibitor therapies. Some of the
techniques that are used in or are applicable to haematology practice
are described.
Polymerase chain reaction
Polymerase chain reaction (PCR), by which DNA is amplified to
generate thousands of copies of the same sequence, is one of the most
widely used techniques. A series of 25 to 40 amplification cycles are
repeated at different temperatures. Each cycle commences with ini-
tial denaturation of the complementary DNA strands at high tem-
perature. The temperature is then lowered to allow annealing of the
added test primers of known sequence to the single-​stranded sample
DNA template. Addition of DNA polymerase to the template–​primer
hybrid results in DNA synthesis. Through repeated cycling, multiple
copies of the same DNA sequence are generated. The amplified DNA
products can be visualized by one of a number of methods (e.g. gel
electrophoresis or fragment analysis) or further analysed by a var-
iety of other downstream techniques. PCR is rapid, inexpensive, and
a simple means of producing relatively large numbers of copies of
DNA molecules derived from all haematological sample types (i.e.
blood and bone marrow) even when the DNA is of relatively poor
quality (e.g. extracted from paraffin-​embedded material or air-​dried
smears scraped from glass slides). JAK2 V617F mutation detection
for the investigation of possible myeloproliferative neoplasms is an
example of a commonly performed diagnostic PCR test.
Multiplex PCR
Multiplex PCR uses multiple different primer sets such that a
number of target regions can be assessed in a single PCR reac-
tion. The amplified DNA regions (amplicons) that are generated
in the cycling reaction are specific for each of the different DNA
sequences. One common application is the assessment of clonality
in lymphoid cell proliferations. B-​cell clonality can be identified by
immunoglobulin (IG) heavy-​ and light-​chain gene rearrangement
and T-​cell clonality by rearrangements of the T-​cell receptor (TCR)
gene. During differentiation of B-​ and T-​progenitor cells, DNA re-
arrangements of the IG and TCR genes result in massive diversity
of genotypically different cells. This process of gene rearrangement
of the V, D, and J domains can be utilized for the detection of clonal
lymphoid cells since all the neoplastic cells will have undergone the
same IG or TCR gene rearrangement. In a polyclonal population,
all the lymphoid cells will have undergone different rearrange-
ments. Multiplex PCR assays that use multiple primers have been
SECTION 22  Haematological disorders
5186
developed for the most widely used V, D, and J domains (such as
the BIOMED-​2 PCR protocols). These protocols are used to assess
IG and TCR gene rearrangements in the investigation of lymphoid
proliferations in fresh (blood and bone marrow) and formalin-​fixed
paraffin-​embedded tissue.
Nested PCR
Nested PCR is performed by two consecutive PCR reactions using
two different primer pairs, both covering the region of interest. As
a result, nested PCR yields high analytical specificity and sensitivity
(as low as one cell in a million). Due to this exquisite sensitivity,
nested PCR is used for residual disease monitoring applications fol-
lowing therapy.
Quantitative real-​time PCR
Quantitative real-​time PCR (RQ-​PCR) is used to quantify gene ex-
pression. RQ-​PCR follows the general principles of PCR, but the
amplified DNA is detected in ‘real time’ as the reaction progresses
cycle by cycle. It gives a precise measurement of the amount of a spe-
cific DNA or RNA in a sample. Quantitation is usually performed
by comparing the expression with that of a control (‘normalized’)
gene. The method is highly sensitive (one cell in 1000–​1 000 000)
and is used for residual disease monitoring. It is particularly useful
to monitor changes in level of gene expression during and following
therapy (e.g. BCR-​ABL1 in chronic myeloid leukaemia) and as an
‘early warning system’ to predict disease relapse (e.g. PML-​RARA in
acute promyelocytic leukaemia).
Reverse transcription PCR
Reverse transcription PCR (RT-​PCR) uses mRNA instead of DNA
as the starting point and can be used to identify the expression of
a gene or the sequence of an RNA transcript. It is particularly ap-
plicable to haematological malignancies for mutation detection and
for the detection of chimeric mRNA resulting from chromosomal
translocations. An example is quantitation of BCR-​ABL1 transcripts
in chronic myeloid leukaemia.
Gene expression analysis
Gene expression analysis, or microarray-​based gene expression pro-
filing, allows the simultaneous assessment of the expression of many
thousands of genes, and, potentially, every gene within a cell. This
has some benefits in assessing haematological malignancies but is
not routinely used due to the high test cost. Gene expression analysis
has largely been superseded by newer technologies which are more
rapid, cost-​effective, and simpler to perform.
Sequencing
Sequencing is the ultimate molecular genetic test. The Human
Genome Project, completed over a decade ago, determined the com-
plete sequence of base pairs making up human DNA. One of the out-
comes has been the ability to identify genetic variants or mutations in
malignancies. DNA sequencing methods have changed substantially
over time. The ‘first-​generation’ sequencing, or Sanger sequencing,
is based on the incorporation of fluorescently labelled A, C, G, and
T nucleotides during DNA synthesis using DNA polymerase. This
remains the ‘gold standard’ method for diagnostic applications
but has many limitations. These include the time to perform, low
throughput, being labour intensive, and having low sensitivity (the
limit of detection of a genetic variant is one cell in five).
Massively parallel DNA sequencing
Massively parallel DNA sequencing, commonly known as NGS,
was introduced in the early 2000s and is a radical change from
‘standard’ Sanger sequencing. NGS records the sequence of DNA
while the strand is being synthesized (so-​called sequencing by syn-
thesis). All DNA fragments in the starting material are sequenced
simultaneously thereby generating vast amounts of output and
rapidly. NGS can be used to sequence entire human genomes or
large fractions, such as the exome. It is now finding its place in
diagnostic practice. It has potential for large-​scale introduction
and to replace other molecular testing approaches. At present,
most clinical NGS is ‘targeted’, that is, only a selected number of
genes (tens to hundreds) are sequenced. Targeted sequencing is
the most likely format that will be incorporated into practice. The
principle of NGS is that large numbers (millions or billions) of
genome fragments that have been sheared into 100 to 400 base-​
pair lengths are sequenced in parallel. The fragments are ampli-
fied and fluorescence emission or hydrogen ion release from an
incorporated nucleotide determines the genomic sequence. Each
fragment of DNA is sequenced multiple times: the term ‘coverage’
or ‘read depth’ is a reflection of the number of times a specific re-
gion of a gene has been sequenced (these and other NGS terms
are shown in Table 22.2.2.2). To ensure accuracy, most technolo-
gies aim for greater than 30-​fold coverage at greater than 90%
of bases for whole genome NGS, 100-fold for exome and 1000-
fold for targeted NGS. The sequencing reads are ‘mapped’ to a
reference genome and variants (i.e. putative mutations) called
(Fig. 22.2.2.2). A ‘variant call’ is a conclusion that there is a nu-
cleotide difference from the reference at a given position in the
genome. DNA sequencing variants (i.e. mutations, insertions, de-
letions, translocations, or copy number variations) may indicate
the presence of a mutation or a genetic association.
NGS methods are cheap, high throughput, and rapid (with a run
time of hours to a few days depending on the technology used).
There are a number of commercial platforms available, which differ
in their methods for preparing the template for sequencing, the
sequencing method, detection (i.e. imaging), and data analysis.
All generate huge amounts of data and data interpretation is com-
plex requiring bioinformatics expertise. Errors can occur in the
technical performance, mapping, and variant calling. The whole
human genome, or targeted regions, can be analysed and produce
sequences at a subgenomic level in a clinically useful time frame.
Massively parallel sequencing is now beginning to be used clinically
to characterize individual patient tumours and to select therapies
based on the identified mutations (Box 22.2.2.2). NGS can be used
to perform whole-​genome, whole-​exome, or targeted sequencing.
Whole-​genome sequencing
Whole-​genome sequencing determines the complete DNA sequence
(coding and noncoding regions) of cells in a sample. Genetic vari-
ants are detected by comparing somatic variants between the neo-
plastic population and the germline (normal cells from the same
individual; Fig. 22.2.2.2). As it is comprehensive, it is an attractive
approach, but it generates huge amounts of data making analysis and
22.2.2  Diagnostic techniques in the assessment of haematological malignancies
5187
interpretation complex. The first report of NGS of a haematological
malignancy determined the full sequence and mutation profile of a
case of cytogenetically normal acute myeloid leukaemia. This was a
major breakthrough and illustrative of the enormous potential for
clinical application in the assessment of malignancies. Future devel-
opments and cost reductions may lead to it moving from research
applications to becoming a clinical tool in guiding management.
Whole-​exome sequencing
Whole-​exome sequencing is where only the protein-​coding regions
of a gene are sequenced. There are 180 000 exons, constituting 1% of
the human genome. Mutations in these coding regions of the DNA
are likely to of greater significance than in the noncoding regions.
The sequencing technology used for exons is the same as for other
NGS approaches (i.e. whole genome and targeted). Whole-​exome
sequencing is quicker and cheaper than whole-​genome sequencing.
It is generally used as a discovery tool and also has not yet found a
place in routine assessment of haematological malignancies.
Targeted sequencing
Targeted sequencing determines the sequence of predefined
selected genes and has an analytical sensitivity of one cell in 50
to 100. The genes assessed are determined by the disease type,
the application (i.e. diagnostic, classification, and therapeutic
options), and published data. Targeted analysis is usually per-
formed to screen for somatic mutation ‘hotspots’ of genes or re-
current gene fusions known to occur in the malignant process.
Due to the targeted approach, it cannot detect mutations out-
side the genes being analysed. Several technologies are available
that selectively enrich for relevant genes/​regions (target enrich-
ment) before NGS is performed. Targeted sequencing is quicker,
cheaper, more reliable, and simpler to interpret than whole-​
genome and whole-​exome methods. Targeted sequencing paves
the way for panels of genes to be assessed simultaneously; this
may obviate the need for individual PCR assays which are cur-
rently performed when assessing haematological malignancies. It
also offers the advantage of being able to add additional ‘targets’
to the panel as new discoveries are made. Put simply, targeted
NGS offers the capability for a holistic ‘multiplex’ genomic plat-
form that could replace the current sequential testing methods
which are laborious, expensive, and time-​consuming. A targeted
amplicon sequencing panel for myeloid neoplasms, for example,
would include the most relevant ‘key’ gene targets known to be
mutated in these neoplasms (e.g. NPM1, JAK2, RUNX1, IDH1/​2,
and SF3B1). One comprehensive test could be performed which
would include all relevant genes required for the diagnosis, clas-
sification, and ongoing management. Similar panels could be es-
tablished for lymphoid neoplasms.
Table 22.2.2.2  Glossary of terms used in genomics and sequencing
Adapters
Short DNA oligonucleotides that contain the primer sites used by the sequencer to generate the sequencing read
Amplicons
Amplified regions of DNA
Amplification
A selective increase in the number of copies of a gene
Copy number variation (CNV)
The number of copies of a gene varies from one individual to another which may result from insertions, deletions,
duplications, and complex variants
Deletion
Mutation resulting in the loss of part of a chromosome or a gene
Duplication
A type of mutation resulting in the production of one or more copies of a chromosomal region or gene
Exon
Coding region of a gene
FICTION
Fluorescence Immunophenotyping and interphase Cytogenetics as a Tool for the Investigation of Neoplasms
FISH
Fluorescent in situ hybridization
Genome
The complete set of genetic instructions within a cell
Indel
Insertion or deletion of bases of DNA that cause a shift in a reading frame. A combination of insertion and deletion
Insertion
A type of mutation resulting in the addition of genetic material
Intron
Noncoding portion of a gene
Library
A collection of DNA or complementary DNA (cDNA) fragments prepared for sequencing
Library preparation
Method to prepare DNA or RNA for next-​generation sequencing
Massively parallel sequencing
Sequencing of many DNA templates simultaneously
Mutation
A change in the structure of a gene that is different from the reference leading to a variant
Read depth
Number of times a nucleotide is read
Single nucleotide polymorphism
A single base difference when the same DNA sequence is compared between individuals
Targeted sequencing
Sequencing of a selected subset of interest of a genome
Translocations
Rearrangement of parts between nonhomologous translocations
Variants
Putative mutations
Variants of uncertain significance
An alteration in the sequence of a gene the significance of which is unclear
Whole-​exome sequencing
Sequencing of coding regions of the genome
Whole-​genome sequencing
Sequencing of the entire genome, both coding and noncoding regions
SECTION 22  Haematological disorders
5188
These new sequencing technologies offer the prospect of cheaper,
faster approaches to the genomic analysis of haematological malig-
nancies. However, there are obstacles to be overcome before testing
can become mainstream. Clinical laboratory standards, consensus
guidelines, and integrated reporting methods need to be devel-
oped. These are essential to ensure quality of testing and accurate
relevant reporting especially as the aim is to detect clinically rele-
vant genomic alterations of diagnostic, prognostic, or therapeutic
significance. NGS panels and clear diagnostic algorithms could
revolutionize diagnostic testing and lead to clinical applications and
personalized pharmacogenomics.
Future developments
The field of diagnostic testing in haematological malignancies has
come a long way since the first descriptions of leukaemia in the mid-
nineteenth century. Diagnostic assessment now requires the inte-
gration of a number of testing modalities ranging from ‘basic’ tests
(the blood count) to manual skill (morphology) and advanced ‘high-​
technology’ approaches (immunophenotyping and genomics). All
deliver information about the biology of the neoplastic cell population
from ‘morphology to mutation’. It is the integration of these data that
leads to an accurate diagnosis and optimized management of a haem-
atological malignancy. Current costs may preclude the full profile of
tests being performed on all cases in all settings. As we go forward,
we will develop rational diagnostic algorithms and only apply those
diagnostic techniques that are required for modern clinical practice.
FURTHER READING
Bacher U, Schnittger S, Haferlach T (2010). Molecular genetics in
acute myeloid leukemia. Curr Opin Oncol, 22, 646–​55.
Bacher U, et al. (2018). Challenges in the introduction of next-
generation sequencing (NGS) for diagnostics of myeloid malignan-
cies into clinical routine use. Blood Cancer Journal,  8, 113.
Calvo KR, McCoy CS, Stetler-​Stevenson M (2011). Flow cytometry
immunophenotyping of hematolymphoid neoplasia. Methods Mol
Biol, 699, 295–​316.
Craig FE, Foon KA (2008). Flow cytometric immunophenotyping for
hematologic malignancies. Blood, 111, 3941–​67.
Erber WN (ed) (2010). Diagnostic techniques in hematological malig-
nancies. Cambridge University Press, Cambridge.
Heel K, et al. (2013). Developments in the immunophenotypic analysis
of haematological malignancies. Blood Rev, 27, 193–​207.
Huret JL, et al. (2013). Atlas of genetics and cytogenetics in oncology and
haematology in 2013. Nucleic Acids Res, 41 Database issue, D920–​4.
Johnsen JM, Nickerson DA, Reiner AP (2013). Massively parallel
sequencing: the new frontier of hematologic genomics. Blood, 122,
3268–​75.
Kuo FC (2019). Next generation sequencing in hematolymphoid neo-
plasia.  Seminars in Hematology, 56, 2–6.
Lee SH, et al. (2008). ICSH guidelines for the standardization of bone
marrow specimens and reports. Int J Lab Hematol, 30, 349–​64.
Ley T, et al. (2008). DNA sequencing a cytogenetically normal acute
myeloid leukaemia genome. Nature, 456, 66–​72.
Liu H, et al. (2007). A practical strategy for the routine use of BIOMED-​
2 PCR assays for detection of B-​ and T-​cell clonality in diagnostic
haematopathology. Br J Haematol, 138, 31–​43.
Mullighan CG (2013). Genome sequencing of lymphoid malignancies.
Blood, 122, 3899–​907.
Swerdlow SH, et al. (eds) (2008). WHO classification of tumours of he-
matopoietic and lymphoid tissues. IARC, Lyon.
Ulahannan D, et al. (2013). Technical and implementation issues in
using next-​generation sequencing of cancers in clinical practice. Br
J Cancer, 109, 827–​35.
Yates LR, Campbell PJ (2012). Evolution of the cancer genome. Nat Rev
Gen, 13, 795–​806.
Tumour
Constitutional
Whole-genome sequencing
Whole-exome sequencing
Targeted sequencing
Alignment
Mapping to reference genome
Single nucleotide variants
lndels
Translocations
Copy number variants
Annotation
Filtering
Prioritization of variants
Detect potential actionable targets
Therapeutic decisions
Sample type
Sequencing
Assembly
Variant detection
Bioinformatics and analysis
Fig. 22.2.2.2  Diagrammatic representation of next-​generation
sequencing of a malignancy.
Box 22.2.2.2  Applications of next-​generation sequencing in the
assessment of haematological malignancies
•	 Classification of disease
•	 Detection of mutations in clinically actionable genes (‘personalized
pharmacogenomics’)
•	 Trial recruitment to molecularly targeted therapies
•	 Disease monitoring based on mutation profile
•	 Early detection of relapse
•	 To determine resistance mechanisms
•	 Research and biological discovery

22.3 Myeloid disease 5189 22.3.1 Granulocytes in h
22.3 Myeloid disease 5189 22.3.1 Granulocytes in health and disease 5189 Joseph Sinning and Nancy Berliner
CONTENTS
22.3.1	 Granulocytes in health and disease  5189
Joseph Sinning and Nancy Berliner
22.3.2	 Myelodysplastic syndromes  5197
Charlotte K. Brierley and David P. Steensma
22.3.3	 Acute myeloid leukaemia  5205
Nigel Russell and Alan Burnett
22.3.4	 Chronic myeloid leukaemia  5213
Mhairi Copland and Tessa L. Holyoake
22.3.5	 The polycythaemias  5227
Daniel Aruch and Ronald Hoffman
22.3.6	 Thrombocytosis and essential thrombocythaemia  5239
Daniel Aruch and Ronald Hoffman
22.3.7	 Primary myelofibrosis  5247
Evan M. Braunstein and Jerry L. Spivak
22.3.8	 Eosinophilia  5254
Peter F. Weller
22.3.9	 Histiocytosis  5259
Chris Hatton
22.3.1  Granulocytes in health
and disease
Joseph Sinning and Nancy Berliner
ESSENTIALS
White cells (leucocytes) mediate inflammatory and immune responses
and are key to the defence of the host against microbial pathogens.
Subpopulations of leucocytes include (1) granulocytes—​neutrophils,
eosinophils (see Chapter 22.3.1 and 22.3.8), and basophils; (2) mono-
cytes; and (3) lymphocytes (see Chapter 22.4.1).
Neutrophils and their disorders
Neutrophils comprise half the peripheral circulating leucocytes and
are characterized by (1)  heterogeneous primary and secondary
granules—​with contents including a variety of degradative enzymes,
and (2) a segmented nucleus. Maturation from the haematopoietic
stem cell occurs in the bone marrow and takes 10 to 14 days. How
long neutrophils circulate in the intravascular space is controversial;
it was originally thought they circulated for only 3 to 6 hours, but
current studies suggest that they may be in the blood for as long as
24 hours before migrating through the vascular endothelium into the
extravascular space, where they may survive for 1 to 3 days.
Neutrophilia—​defined as an increase in the circulating neutrophil
count to greater than 7.5 × 106/​µl, usually occurs as an acquired re-
active response to underlying disease. Causes include (1) infection,
particularly bacterial—​the commonest cause of an elevated leuco-
cyte count; (2)  drugs (e.g. steroids); (3)  malignancies—​including
myeloproliferative disorders and nonhaematological cancers; and
(much less commonly) (4) hereditary conditions—​including heredi-
tary neutrophilia, leucocyte adhesion deficiency, and chronic idio-
pathic neutrophilia.
Neutropenia—​defined as a reduction in the absolute neutrophil
count to less than 1.5 × 106/​µl, is of particular importance because,
when severe (<0.5 × 106/​µl), it markedly increases the risk of life-​
threatening infection. Causes include (1) drugs and toxins. Mechanisms
of drug-​induced neutropenia include (a) direct marrow suppression,
(b) immune destruction with antibody-​ or complement-​mediated
damage of myeloid precursors, and (c) peripheral destruction of neu-
trophils; common offending drugs that cause dose-​dependent neu-
tropenia include cancer chemotherapeutic agents, phenothiazines,
anticonvulsants, and ganciclovir; (2) postinfectious—​particularly after
viral infections; (3)  nutritional deficiencies (e.g. vitamins B12, folic
acid); (4)  autoimmune—​usually attributable in adults to disorders
such as systemic lupus erythematosus (SLE) and rheumatoid arthritis;
(5) large granular lymphocytosis; and (6) congenital—​including se-
vere congenital neutropenia and cyclic neutropenia.
Disorders of neutrophil function include (1)  chronic granu-
lomatous disease—​a heterogeneous group of rare disorders (most
X-​linked) characterized by defective production of superoxide by
neutrophils, monocytes, and eosinophils; patients usually present
in childhood with severe infections, often with catalase-​negative
pathogens; (2) leucocyte adhesion deficiency; (3) myeloperoxidase
deficiency; and (4) Chediak–​Higashi syndrome.
Monocytes and their disorders
Monocytes share a common myeloid precursor with granulocytes,
present antigens to T cells, produce several important cytokines
with immunomodulatory and inflammatory functions, and are the
22.3
Myeloid disease
SECTION 22  Haematological disorders
5190
precursors to resident tissue macrophages. They are especially im-
portant in defence against intracellular pathogens.
Causes of monocytosis (>0.9 × 106/​µl) include (1) chronic in-
fection (e.g. tuberculosis, endocarditis), (2) autoimmune diseases
(e.g. SLE), and (3)  malignancy (e.g. primary malignancies of the
marrow or marrow infiltration with solid tumours).
Basophils and their disorders
Basophils are nonphagocytic granulocytes that function in
immediate-​type hypersensitivity. Basophilia (> 0.2 × 106/​µl) is seen
in myeloproliferative disorders, hypersensitivity reactions, and with
some viral infections.
Introduction
Leucocytes perform a critical role in the host defence against patho-
gens. They mediate inflammation and modulate the immune re-
sponse. Leucocytes can be divided into granulocytes (neutrophils,
eosinophils, and basophils; Fig. 22.3.1.1), monocytes, and lympho-
cytes. This chapter will focus on the role of granulocytes and mono-
cytes in the normal host response and pathological manifestations
of abnormalities of their number and/​or function. Lymphocytes are
discussed elsewhere.
Neutrophils
Morphology
Under normal conditions, neutrophils make up over one-​half
of the leucocytes in the peripheral blood. The morphological
hallmarks of these cells include heterogeneous granules and a
multilobated or segmented nucleus. The two predominant types
of granules in the neutrophil’s cytoplasm are the azurophilic
(or primary) granules and the specific (or secondary) granules.
Azurophilic granules arise at the promyelocytic stage of differen-
tiation. They contain myeloperoxidase, proteases, acid hydrolases,
and microbicidal proteins. Specific granules and their content
proteins are synthesized at the myelocytic stage of differentiation.
Their contents include lactoferrin, lysozyme, vitamin B12-​binding
protein, gelatinase, and neutrophil collagenase. The specific gran-
ules are not a uniform population, and their variable content is
determined mainly by the timing of their formation. Those formed
early in the myelocyte stage contain abundant lactoferrin, while
those formed later are enriched for gelatinase, and are often re-
ferred to as ‘tertiary’ granules or gelatinase granules. The specific
granule membrane contains the cytochrome b-​558 component of
the respiratory burst oxidase, as well as chemotactic and opsonic
receptors, which are transferred to the plasma membrane upon ac-
tivation of the neutrophil. Finally, the neutrophil cytoplasm also
contains secretory vesicles that are endocytic vesicles containing
primarily plasma proteins, and are the most rapidly mobilized
fraction of cytoplasmic granules in the neutrophil. The membrane
of secretory vesicles is rich in receptors and cytochrome b, and the
vesicles contribute these proteins to the plasma membrane upon
neutrophil activation.
Common variants of neutrophil morphology include the Pelger–​
Huet anomaly, hypersegmentation of the nucleus, Dohle bodies,
and toxic granulations. The Pelger–​Huet anomaly is a dominantly
inherited defect in nuclear segmentation that results in a dumb-​
bell-​ or rod-​shaped nucleus. Neutrophils with nuclei similar to
this (‘pseudo-​Pelger–​Huet anomaly’) may be seen in acquired
myelodysplastic syndromes. Hypersegmented nuclei (containing
five or more segments) are characteristic of megaloblastic haemato-
poiesis due to folic acid or vitamin B12 deficiency. Dohle bodies are
large basophilic inclusions that may be seen in sepsis, pregnancy,
and following cytotoxic chemotherapy. Toxic granulations are ab-
normally staining primary granules that arise when neutrophils
are released prematurely from the marrow, as in severe bacterial
infections.
Maturation
There are three cellular compartments that contain myeloid cells: the
marrow, the intravascular compartment, and the extravascular space.
Maturation from the haematopoietic stem cell occurs in the bone
marrow and takes from 10 to 14 days. The marrow compartment can
be subdivided into the mitotic compartment and the postmitotic
and storage compartment. In the marrow mitotic compartment,
neutrophils arise through serial division of myeloid precursors. The
mitotic compartment contains myeloid cells with the ability to rep-
licate:  myeloblasts, promyelocytes, and myelocytes. The marrow
postmitotic and storage compartment contains myeloid elem-
ents that have lost the ability to divide, including metamyelocytes,
bands, and segmented neutrophils. Neutrophils are released from
the storage pool into the intravascular space, where they remain for
4 to 24 h. Within this space, approximately one-​half of the neutro-
phils circulate freely in the peripheral blood while the other half re-
main ‘marginated’ along the vascular endothelium. The marginated
and circulating cells are in dynamic equilibrium with one another.
Neutrophils then migrate through the vascular endothelium into the
extravascular space, where they survive for 1 to 3 days. At any given
time, approximately 90% of neutrophils are in the marrow compart-
ment and 2 to 3% are in the intravascular space, with the remainder
in the extravascular space.
Neutrophilia
Neutrophilia is defined as an elevation of the circulating neutrophil
count (>7.5 × 106/​µl). Although it may reflect a primary haemato-
logical process, it usually occurs as a secondary manifestation of an
underlying disease process or drug. The causes of an elevated neu-
trophil count are summarized in Box 22.3.1.1.
(a)
(b)
(c)
Fig. 22.3.1.1  Peripheral blood granulocytes: (a) polymorphonuclear
leucocyte (neutrophil), (b) eosinophil, (c) basophil.
22.3.1  Granulocytes in health and disease
5191
Hereditary neutrophilias
Hereditary neutrophilia
This is a dominantly inherited syndrome manifested by leucocyt-
osis, splenomegaly, and widened diploë of the skull. Laboratory
evaluation reveals a white blood count of 20 000 to 70 000/​µl with
a neutrophilic predominance, and an elevated leucocyte alkaline
phosphatase. Its clinical course is benign.
Chronic idiopathic neutrophilia
This is a sporadically occurring condition that manifests as a white
blood count of 11 000 to 40 000/​µl with a neutrophilic predomin-
ance. Patients are otherwise well and have been followed for up to
20 years without the development of significant pathology.
Leucocyte adhesion deficiency
This is a rare inherited disorder characterized by recurrent life-​
threatening bacterial and fungal infections, cutaneous abscesses, gingi-
vitis, or periodontal infections. Expression of the CD11b/​CD18 integrin
is deficient, resulting in the inability of neutrophils to migrate to sites of
infection (see ‘Disorders of neutrophil function’ for further discussion).
Acquired neutrophilias
Infection
The most common cause of an elevated leucocyte count is infec-
tion. Acute infection often causes a modest rise in the white blood
count, which may be accompanied by an increase in circulating
immature precursors (‘left shift’). This occurs more commonly
with bacterial infection but can also occur with viral processes.
Along with a left shift, morphological changes in the neutrophil
may be seen with bacterial infection, including toxic granulation,
Dohle bodies, and cytoplasmic vacuoles. Neutrophilia resolves
with treatment or resolution of the infectious process. In chronic
inflammation, marrow granulocyte production is stimulated, re-
sulting in moderate neutrophilia, sometimes with monocytosis.
Chronic infections such as osteomyelitis, empyema, and tubercu-
losis can also give rise to a leukaemoid reaction with white blood
counts markedly elevated (>50 000/​µl), usually associated with a
marked left shift.
Drugs
Drugs can cause leucocytosis by several different mechanisms.
Steroids increase the release of mature neutrophils from the marrow
and should not cause a left shift. β-​Agonists acutely raise the neutro-
phil count by inducing the demargination of neutrophils adherent
to the vascular endothelium, and may result in a neutrophil count
twice that of baseline. Acute stress also results in demargination
of neutrophils, which is probably mediated by adrenergic stimula-
tion. Stresses that can cause this include exercise, surgery, seizure,
and myocardial infarction. The cytokines granulocyte colony-​
stimulating factor (G-​CSF) and granulocyte–​macrophage colony-​
stimulating factor (GM-​CSF) stimulate marrow production of
neutrophils and can cause dramatic elevations in the white blood
count. The majority of white cells formed are neutrophils and a
left shift is often seen. The use of these cytokines therefore requires
careful monitoring.
Primary haematological conditions
In other situations, neutrophilia may reflect a primary haem-
atological condition. Marrow hyperstimulation in the setting of
autoimmune haemolytic anaemia, immune thrombocytopenia, or
recovery following chemotherapy or toxic insult to the marrow
may result in a reactive leucocytosis. In autoimmune haemo-
lytic anaemia and immune thrombocytopenia, neutrophilia may
reflect disease activity, but steroid therapy or splenectomy may
contribute. Splenectomy or hyposplenic states (e.g. sickle cell dis-
ease) may also result in modest neutrophilia at baseline with more
marked neutrophilia at times of stress or infection, reflective of
the loss of the spleen as a site of margination and sequestration
of leucocytes.
Myeloproliferative disorders
Neutrophilia is a common feature of the myeloproliferative dis-
orders chronic myeloid leukaemia, polycythaemia vera, and
myelofibrosis as well as familial myeloproliferative disorders.
Elevated eosinophil and basophil counts are also often seen in these
disorders. Leucocyte alkaline phosphatase may be low or undetect-
able in chronic myeloid leukaemia. The myeloproliferative dis-
orders are discussed in further detail elsewhere.
Nonhaematological malignancies
Various nonhaematological malignancies including lung and breast
tumours may also cause neutrophilia. Tumours may secrete colony-​
stimulating factors or may cause a leukaemoid reaction. Tumour
Box 22.3.1.1  Differential diagnosis of neutrophilia
Primary haematological disease
•	 Chronic idiopathic neutrophilia
•	 Hereditary neutrophilia
•	 Leucocyte adhesion deficiency
•	 Myeloproliferative disorders:
—​	 Chronic myeloid leukaemia
—​	 Polycythaemia vera
—​	 Myelofibrosis
Secondary to other disease processes or drugs
•	 Infection:
—​	 Acute
—​	 Chronic
•	 Acute stress:
—​	 Exercise
—​	 Surgery
—​	 Seizure
—​	 Myocardial infarction
•	 Drugs:
—​	 Steroids
—​	 Lithium
—​	 β-​Agonists
—​	 Cytokines (G-​CSF, GM-​CSF)
•	 Chronic inflammation
•	 Marrow infiltration
•	 Marrow hyperstimulation:
—​	 Chronic haemolysis
—​	 Immune thrombocytopenia
—​	 Recovery from marrow suppression
•	 Postsplenectomy/​hyposplenism
•	 Nonhaematological neoplasms
SECTION 22  Haematological disorders
5192
metastatic to the bone marrow may cause leucoerythroblastic
changes, characterized by fragmented erythrocytes, teardrops, and
nucleated red cells, as well as leucocytosis with a left shift.
Evaluation of neutrophilia
The evaluation of neutrophilia should take account of the fact that
leucocytosis is usually reactive, and that primary haematological
aetiologies are relatively rare. The abnormal laboratory value
should be verified to rule out laboratory error or a transient un-
explained leucocytosis that resolves spontaneously. A careful his-
tory and physical examination are essential to evaluate for potential
infectious processes, and to obtain a history of medication use.
Examination of the bone marrow is usually not necessary for the
evaluation of neutrophilia, but examination of a peripheral smear
may be very helpful. Evidence of leucoerythroblastic changes war-
rants examination of the bone marrow to rule out infiltration of
the marrow. If a bone marrow aspirate and biopsy are performed,
evaluation should also include culture of the marrow for fungus or
mycobacteria.
Features that raise the question of myeloproliferative disease in-
clude concomitant elevation of platelets and haematocrit, basophilia
and/​or eosinophilia, and splenomegaly. In that setting, evalu-
ation should include cytogenetics or fluorescent in situ hybridiza-
tion examination for BCR-​ABL1 (diagnostic of chronic myeloid
leukaemia in this setting), and assay for mutations in JAK2 (for
diagnosis of polycythemia vera and other non-​bcr-​abl-​positive
myeloproliferative syndromes). Evaluation for myeloproliferative
disease is discussed in detail elsewhere.
Neutropenia
Neutropenia is defined as an absolute neutrophil count (ANC)
of less than 1.5 × 106/​µl. In some populations, such as Africans
and Yemeni Jews, normal ANCs are lower, with a lower limit of
normal of 1.2 × 106/​µl. Neutropenia may pose a risk of serious
bacterial infection, and this risk is directly related to the degree
of neutropenia. In mild neutropenia (ANC 1000–​1500 × 106/​
µl) the risk of life-​threatening infection is not increased, and in
moderate neutropenia (ANC 500–​1000 × 106/​µl) the risk of se-
vere infection is only mildly elevated. Severe neutropenia (ANC
<500 × 106/​µl) markedly increases the risk of life-​threatening in-
fection. The duration and timecourse of neutropenia may also be
important, as the acute onset of severe neutropenia is associated
with a higher risk of serious infection than is chronic neutropenia
of similar severity. Neutropenia in the setting of marrow failure is
more threatening than neutropenia with an intact marrow, as the
marrow reserve pool may afford protection. Fever of new onset in
the setting of severe neutropenia is a medical emergency requiring
immediate evaluation and treatment. Common causes of infection
in these patients include Gram-​negative enteric pathogens such as
Escherichia coli, pseudomonas, and Klebsiella pneumoniae, as well
as Staphylococcus aureus. The causes of neutropenia are summar-
ized in Box 22.3.1.2.
Congenital neutropenia
Severe congenital neutropenia
Severe congenital neutropenia (SCN), originally characterized by
Rolf Kostmann as an autosomal recessive disorder (Kostmann’s
syndrome), is characterized by severe persistent neutropenia, and
the early onset of frequent, life-​threatening infections. Bone marrow
aspirate reveals a maturation arrest at the promyelocyte stage. This
syndrome was originally described as an autosomal recessive dis-
order, but recent evidence suggests that SCN is a heterogeneous dis-
order with autosomal dominant, autosomal recessive, X-​linked, and
sporadic forms. Autosomal dominant SCN has been linked to muta-
tions in the gene encoding neutrophil elastase (ELANE), a primary
granule protein gene expressed at high levels at the promyelocyte
stage of differentiation. Current evidence suggests that the impact of
the mutations is not related to the enzymatic function of elastase, but
rather reflects the failure of the protein to fold properly. This induces
the ‘unfolded protein response’, a protective response to cellular
stress that leads to decreased protein synthesis, degradation of un-
folded proteins in the endoplasmic reticulum, and increased apop-
tosis. Autosomal recessive SCN (Kostmann’s syndrome) is caused by
mutations in HAX-​1, a mitochondrial protein that is important for
stabilizing the inner mitochondrial membrane. Homozygous loss of
HAX-​1 leads to loss of mitochondrial membrane potential, and also
Box 22.3.1.2  Differential diagnosis of neutropenia
Decreased production of neutrophils
•	 Constitutional neutropenia
•	 Congenital neutropenias:
—​	 Severe congenital neutropenia (including Kostmann’s syndrome)
—​	 Shwachman–​Diamond–​Oski syndrome
—​	 Chediak–​Higashi syndrome
—​	 Reticular dysgenesis
—​	 Dyskeratosis congenita
•	 Cyclic neutropenia
•	 Postinfectious
•	 Nutritional deficiency:
—​	 Vitamin B12
—​	 Folic acid
—​	 Copper
•	 Anorexia nervosa
•	 Drug or toxin induced
•	 Primary marrow failure:
—​	 Aplastic anaemia
—​	 Myelodysplastic syndromes
—​	 Acute leukaemia
—​	 Paroxysmal nocturnal haemoglobinuria
—​	 Pure white-​cell aplasia
—​	 Shwachman–​Diamond–​Oski syndrome
—​	 Chediak–​Higashi syndrome
—​	 Reticular dysgenesis
—​	 Dyskeratosis congenita
—​	 Large granular lymphocytosis
Increased peripheral destruction of neutrophils
•	 Overwhelming infection
•	 Immune destruction:
—​	 Drug related
—​	 Collagen vascular disease associated
—​	 Large granular lymphocytosis
—​	 Felty’s syndrome
—​	 Isoimmune
•	 Hypersplenism/​sequestration
22.3.1  Granulocytes in health and disease
5193
leads to apoptosis. Other rare cases of SCN are linked to mutations
in G6PC3, the Wiskott–​Aldrich protein (WASp), and the transcrip-
tion factor Gfi-​1.
Most patients with SCN respond to G-​CSF with increases in
their ANC and decreased incidence of infection. Haematopoietic
stem cell transplantation is another viable treatment option. With
the prolongation of life offered by G-​CSF therapy, it has become
apparent that patients with SCN have an increased incidence of
myelodysplastic syndrome (MDS) and acute myeloblastic leu-
kaemia (AML). These malignancies often develop in association
with an acquired mutation in the G-​CSF receptor. A relationship
has been speculated to exist between G-​CSF therapy and the de-
velopment of these mutations in the G-​CSF receptor, but this
connection remains unproven, as has the pathogenetic role of the
mutations in G-​CSF receptor and the subsequent development of
MDS/​AML.
Cyclic neutropenia (cyclic haematopoiesis)
This is a rare, dominantly inherited, marrow disorder character-
ized by cyclic fluctuations in neutrophil counts approximately
every 21 days and lasting 3 to 7 days. Along with the neutropenia,
cyclic drops in the reticulocyte and monocyte counts are also
observed. Episodes of neutropenia may be severe, often with an
ANC less than 200 × 106/​µl, and may be accompanied by fevers,
pharyngitis, stomatitis, and other bacterial infections. Cyclic
neutropenia has also been linked to mutations in the neutrophil
elastase gene. Why some mutations give rise to cyclic haemato-
poiesis and others to SCN is still a matter of speculation. This
has led to the hypothesis that the severity of the phenotype is re-
lated to the degree of abnormal protein folding and induction of
the unfolded protein response associated with different ELANE
mutations. Cyclic neutropenia can be treated safely and effect-
ively with G-​CSF. Unlike Kostmann’s syndrome, cyclic haemato-
poiesis is not associated with an increased incidence of AML
and MDS.
Acquired neutropenias
Postinfectious neutropenia
This is commonly seen following viral infections. It usually occurs
several days after the onset of infection and may last several weeks.
Varicella zoster, measles, Epstein–​Barr, cytomegalovirus, influenza
A and B, and hepatitis A and B are some of the viruses most com-
monly associated with postinfectious neutropenia. The neutropenia
resolves spontaneously. Transient neutropenia may also be seen with
parvovirus infection. Neutropenia occurs commonly in patients
with HIV. The causes are multifactorial and may be related directly
to the viral infection, to opportunistic infections or associated con-
ditions, or to the treatment of the virus or its complications.
Several bacterial infections can cause neutropenia, including
rickettsial infections, typhoid fever, brucellosis, and tularaemia.
Bacterial sepsis of any cause can also result in acute neutropenia.
This occurs both as a result of marrow suppression and increased
destruction of neutrophils. Acute neutropenia in bacterial infections
suggest that egress to tissue exceeds the capacity of the marrow re-
serve pool. The neutropenia may be severe and it portends a poor
prognosis. Fungal infections, such as disseminated histoplasmosis,
and mycobacterial diseases may also cause neutropenia.
Nutritional deficiencies
Nutritional deficiencies of vitamins B12 and folic acid result in meg-
aloblastic haematopoiesis with ineffective myelopoiesis. Deficiency
of copper is a rare nutritional cause of neutropenia seen in the set-
ting of severe malnutrition or long-​term parenteral alimentation.
Mild neutropenia may also be seen with anorexia nervosa.
Drugs and toxins
Numerous drugs and toxins are known to cause neutropenia.
Mechanisms of drug-​induced neutropenia include (1)  direct
marrow suppression, (2)  immune destruction with antibody-​ or
complement-​mediated damage of myeloid precursors, and (3) per-
ipheral destruction of neutrophils. In most cases, direct marrow
suppression is dose dependent. Common offending drugs that cause
dose-​dependent neutropenia include cancer chemotherapeutic
agents, phenothiazines, anticonvulsants, and ganciclovir. Alcohol
can also cause neutropenia by marrow suppression. If a drug is
suspected of causing dose-​dependent neutropenia, resolution
will occur promptly upon discontinuation of the offending agent.
However, if it is not possible to stop the drug and the neutropenia
is not severe, the drug may be continued with careful monitoring.
Neutropenia is often related to the dose and duration of therapy. In
contrast, those drugs that cause immune neutropenia usually cause
profound agranulocytosis, resulting from both intramedullary
destruction of myeloid precursors and peripheral destruction of
mature neutrophils. Such drugs include antithyroid medications,
sulphonamides, and semisynthetic penicillins. Examination of the
bone marrow shows a maturation arrest of the myeloid lineage, re-
flecting immune destruction of myeloid precursors. The offending
agent must be stopped and rechallenge should not be attempted.
Recovery of the neutrophil count can be accelerated by the admin-
istration of G-​CSF.
Autoimmune neutropenia
Primary autoimmune neutropenia is a disease of childhood, with an
average age of onset of 6 to 12 months. Patients present with mod-
erate to severe neutropenia that spontaneously remits within 2 years
in 95% of patients. Treatment with prophylactic antibiotics prevents
most serious complications, and G-​CSF therapy is recommended
only in the setting of severe or recurrent infections.
Secondary autoimmune neutropenia is seen primarily in adults,
and may occur in association with collagen vascular disorders such
as systemic lupus erythematosus and rheumatoid arthritis, as well as
with immune thrombocytopenia and autoimmune haemolytic an-
aemia. Destruction may be mediated by IgG or IgM antibodies. The
neutropenia may be severe but the degree of neutropenia frequently
does not correlate as well with the risk of infection as in other con-
ditions. The marrow typically is hypercellular with a late myeloid
maturation arrest. Treatment is indicated in the setting of severe,
recurrent infections.
Treatment options include intravenous immunoglobulin, splenec-
tomy, and other therapies directed at the underlying collagen vascular
disorder. In Felty’s syndrome, neutropenia accompanies rheumatoid
arthritis and splenomegaly and neutropenia probably reflects both
immune destruction and splenic sequestration. Granulopoiesis is
inhibited by either antibodies or T-cells. This can lead to severe and
recurrent infections. It may be managed with G-​CSF. Splenectomy
SECTION 22  Haematological disorders
5194
relieves the neutropenia in the majority of cases. However, given its
close association with large granular lymphocytosis (see following
section), treatment with low-​dose methotrexate or cyclophospha-
mide is the chosen approach in many patients.
Large granular lymphocytosis
Large granular lymphocytosis occurs in an older population, and
is frequently seen in association with rheumatological diseases
such as rheumatoid arthritis. Due to the association with systemic
inflammatory disease, large granular lymphocytosis was origin-
ally hypothesized to be a polyclonal abnormal immune response.
However, gene rearrangement studies have confirmed that large
granular lymphocytosis is frequently a clonal disease representing a
form of T-​cell lymphoma. There are two distinct subtypes, with cells
expressing either an unusual Tγ phenotype (CD3+CD8+CD56–​)
or a natural killer (NK) phenotype (CD56+). When seen in associ-
ation with rheumatoid arthritis, the disease has significant overlap
with Felty’s syndrome. Both large granular lymphocytosis and Felty’s
syndrome are associated with a very high frequency (80–​90%) of
HLA D4, and investigators now believe that these diseases represent
a spectrum of a single disease. Neutropenia related to large granular
lymphocytosis is associated with a myeloid maturation arrest in the
marrow, consistent with immune-​mediated neutrophil destruction.
Surprisingly, however, the neutrophil count will often respond to
G-​CSF. The neutropenia responds well to low-​dose methotrexate or
cyclophosphamide in 50% of patients, and other immunosuppres-
sive agents also have activity in restoring neutrophil counts. The
course of lymphoma in large granular lymphocytosis varies from in-
dolent to rapidly progressive.
Other causes
Aplastic anaemia reflects a primary failure of haematopoi-
esis with neutropenia, anaemia, and thrombocytopenia. In the
myelodysplastic syndromes and acute leukaemias, the marrow does
not produce adequate numbers of neutrophils.
Isoimmune neutropenia occurs in 1 in 500 babies born alive. It is
caused by placental transfer of maternal IgG directed against fetal
neutrophils, and it presents in the first days of life.
Hypersplenism usually causes mild or moderate neutropenia
along with anaemia and thrombocytopenia. Normal myeloid
maturation is seen in the marrow. The neutropenia is rarely
severe.
Evaluation of neutropenia
In contrast to the evaluation of neutrophilia, most patients with
confirmed neutropenia require bone marrow examination. A com-
prehensive history and physical examination may identify the oc-
casional patient with mild neutropenia and no other evidence of
disease that may warrant close observation only. However, recur-
rent infections, including oral and mucosal infections, abnormal-
ities observed in a peripheral blood smear, or severe neutropenia
increase the likelihood of significant marrow pathology and marrow
aspiration and biopsy is indicated. If neutropenia is accompanied
by anaemia or thrombocytopenia, marrow examination is required
to rule out aplasia, leukaemia, myelodysplasia, or other primary
marrow malignancy. A  marrow that shows hyperplastic myeloid
precursors and a maturation arrest supports a diagnosis of periph-
eral neutrophil destruction and/​or immune neutropenia, which
should lead to a search for an underlying collagen vascular disorder
or drug-​induced neutropenia.
Management of neutropenia
Fever of new onset in the setting of severe neutropenia (ANC <500
× 106/​µl) is a medical emergency. A careful history and physical
examination should be performed in a timely fashion. Due to the
lack of neutrophils, sites of infection may be difficult to find as sig-
nificant inflammation or tissue infiltration by neutrophils may not
occur. Blood and bodily fluids should be cultured. Empirical broad-​
spectrum antibiotics should be initiated without delay. In patients
with fever in the setting of neutropenia that is expected to resolve
(usually neutropenia induced by chemotherapy or drug reaction),
antibiotics should be continued until the neutrophil count re-
covers to over 500/​µl. In patients with chronic neutropenia that is
expected to persist indefinitely, antibiotics should be continued for
several days past the resolution of fever. If fever persists for more
than 1 week despite antibiotic therapy, empirical antifungal therapy
should be given. Granulocyte transfusion should be considered in
culture-​positive Gram-​negative sepsis not responsive to antibiotics
in the setting of continued neutropenia.
Granulocyte colony-​stimulating factor
G-​CSF (filgrastim) is a haematopoietic growth factor that has effects
primarily on the neutrophilic myeloid lineage. G-​CSF reduces the
time of maturation of committed neutrophil precursors, prolongs
the lifespan of mature neutrophils, and primes them for enhanced
function of the respiratory burst, phagocytosis, and chemotaxis.
Clinically, G-​CSF is used in the treatment and prevention of neu-
tropenia. When used in conjunction with myelosuppressive
chemotherapy, G-​CSF has been shown to reduce the severity of neu-
tropenia, shorten the duration of neutropenia, reduce the risk of
developing neutropenic fever, and reduce the length of stay in hos-
pital. G-​CSF has also been utilized successfully in the treatment of
severe neutropenia secondary to congenital disorders such as cyclic
neutropenia and SCN, and may be useful in the treatment of auto-
immune neutropenia as seen in Felty’s syndrome and systemic lupus
erythematosus. The neutropenia of marrow failure states, such as the
myelodysplastic syndromes, may respond to G-​CSF.
Neutropenia secondary to the treatment of HIV infection can
also be controlled with G-​CSF. The other major use of G-​CSF is in
the mobilization of haematopoietic progenitor cells from the bone
marrow to the peripheral blood. While in the peripheral blood,
these cells can be collected by cytopheresis for use in haematopoi-
etic cell transplantation.
Disorders of neutrophil function
Chronic granulomatous disease
Chronic granulomatous disease is a heterogeneous group of rare
disorders characterized by defective production of superoxide (O2–​)
by neutrophils, monocytes, and eosinophils. The majority of cases
are inherited in an X-​linked fashion, but autosomal recessive in-
heritance also occurs. The genetic lesions causing chronic granu-
lomatous disease have been characterized, and involve mutations
in any of four genes encoding the proteins of the respiratory burst
oxidase. These include the 91-​kDa (X-​linked) and 22-​kDa (auto-
somal) components of the membrane cytochrome b-​558 complex,
and the 47-​ and 67-​kDa soluble components (autosomal) of the
22.3.1  Granulocytes in health and disease
5195
oxidase complex. Patients usually present in childhood with se-
vere infections, often with catalase-​negative pathogens. The most
common infection in patients with chronic granulomatous disease
is pneumonia, with S. aureus, Burkholderia cepacia, aspergillus, and
enteric Gram-​negative bacteria often implicated. Other common
infections in chronic granulomatous disease include lymphaden-
itis, cutaneous infections, hepatic abscesses, and osteomyelitis.
Noninfectious inflammatory conditions, such as aphthous ulcer-
ation of the oral mucosa are common, as are chronic mucosal in-
flammation, perirectal abscesses or fissures, and granulomas of the
gastrointestinal and genitourinary tract. The diagnosis of chronic
granulomatous disease should be considered in an individual with a
history of multiple severe bacterial and fungal infections or a family
history of the disorder. The diagnosis is established by confirming
abnormal neutrophil oxidative metabolism with tests such as the
nitroblue tetrazolium slide test or measurements of superoxide or
peroxide production. The management of chronic granulomatous
disease is based on aggressive prophylaxis and prompt treatment of
infection. Prophylactic trimethoprim–​sulphamethoxazole or peni-
cillin can significantly decrease the number of bacterial infections
in patients with chronic granulomatous disease. Potentially serious
infections require the prompt initiation of parenteral antibiotics.
Surgical interventions including drainage of abscesses and resec-
tion of infected tissue are an important adjunct to antimicrobial
chemotherapy. Prophylaxis with recombinant human interferon-​γ
was shown in a phase III trial to decrease substantially the number
of serious infections in patients with chronic granulomatous dis-
ease, although oxidase activity was unaffected. Chronic granuloma-
tous disease has also been a target of early gene therapy trials.
Leucocyte adhesion deficiency
Leucocyte adhesion deficiency is an inherited disorder of neutro-
phil function. Two types of leucocyte adhesion deficiency have been
characterized. Type 1 deficiency is a rare autosomal recessive dis-
order resulting from mutations in CD18, the gene encoding the
β-​chain of leucocyte function antigen-​1 (LFA-​1, CD11a/​CD18),
Mac-​1 (CD11b/​CD18, CR3, the receptor for the opsonin C3Bi), and
gp150,95 (CD11c/​CD18). Deficient expression of these three in-
tegrin complexes on the neutrophil cell surface results in decreased
neutrophil adhesion to the endothelium, impaired chemotaxis, and
defective C3Bi-​mediated pathogen ingestion, degranulation, and
respiratory burst activation. Patients with leucocyte adhesion defi-
ciency typically present in early childhood with recurrent pyogenic
infections of the skin, respiratory and digestive tracts, and mucosal
membranes. A history of delayed umbilical cord separation is also
often noted. Common pathogens in patients with type 1 leucocyte
adhesion deficiency include S. aureus and Gram-​negative enterics.
Foci of infection notably lack neutrophil infiltration. A mild leuco-
cytosis persists due to impaired margination. The diagnosis is con-
firmed by flow cytometric measurement of neutrophil CD11b/​
CD18 expression. The treatment of type 1 leucocyte adhesion defi-
ciency includes aggressive use of parenteral antibiotics for pyogenic
infections. Prophylactic trimethoprim–​sulphamethoxazole may
benefit some patients. Patients with a severe phenotype often die in
the first 2 years of life, but patients with mild disease may survive to
early adulthood. Type 2 leucocyte adhesion deficiency is caused by
a deficiency of sialyl–​Lewis X moieties on neutrophil selectins. In
addition to neutrophil function abnormalities, this extremely rare
syndrome also is characterized by mental retardation, short stature,
and the rare Bombay erythrocyte phenotype.
Myeloperoxidase deficiency
Myeloperoxidase deficiency is a relatively common, autosomal
recessively inherited disorder of neutrophil function. Complete
deficiency occurs in 1 in 2000 individuals and partial deficiency
occurs twice as frequently. Myeloperoxidase catalyses the pro-
duction of hypochlorous acid, which is an antimicrobial agent.
Myeloperoxidase deficiency is often of no clinical consequence
because other host defence mechanisms can adequately com-
pensate for the defective myeloperoxidase; however, when
myeloperoxidase deficiency coexists with another defect in host
defence, such as diabetes mellitus, disseminated candidal or fungal
infections may occur. The diagnosis of myeloperoxidase deficiency
is made by histochemical staining of neutrophils and monocytes.
Therapy consists of aggressive treatment of fungal infections as
well as careful control of glucose levels in patients with diabetes.
An acquired form of myeloperoxidase deficiency occurs in some
myeloid leukaemias.
Chediak–​Higashi syndrome
Chediak–​Higashi syndrome (OMIM 214500) is a rare disorder of
neutrophil function. Neutrophils and monocytes contain giant pri-
mary granules and demonstrate impaired degranulation and fu-
sion with phagosomes. Chemotaxis is also defective. Neutropenia
results from defective granulopoiesis. Chediak–​Higashi syndrome
is inherited in an autosomal recessive manner. The gene respon-
sible, LYST, has been cloned, and is homologous to a murine lyso-
somal trafficking protein. Chediak–​Higashi syndrome manifests
in childhood or infancy with infections of the skin, lungs, and mu-
cous membranes. S. aureus, Gram-​negative enterics, candida, and
aspergillus species are responsible for most infections in this syn-
drome. Nonhaematological manifestations of Chediak–​Higashi
syndrome include partial oculocutaneous albinism, progressive
peripheral and cranial neuropathies, and in some cases, mental dis-
ability. The majority of patients will develop an accelerated phase
of the syndrome, manifested by lymphohistiocytic proliferation in
the liver, spleen, bone marrow, and lymphatics. The diagnosis of
Chediak–​Higashi syndrome is made by the demonstration of giant
peroxidase-​containing granules in peripheral blood or bone marrow
myeloid cells, outside of the setting of myelogenous leukaemia.
Chediak–​Higashi syndrome is treated in the early or stable phase
with prophylactic antibiotics and aggressive parenteral antibiotics
for infections. Ascorbic acid may also be of benefit. The accelerated
phase is treated with vinca alkaloids and glucocorticoids, but often
responds poorly to these measures. Allogeneic haematopoietic cell
transplantation from HLA-​compatible donors is the only potentially
curative therapy for Chediak–​Higashi syndrome.
Specific granule deficiency
An extremely rare disorder, neutrophil-​specific granule deficiency
is characterized by absent or empty neutrophil-​specific granules.
Specific granule deficiency is manifested clinically as recurrent skin
and pulmonary infections resulting from the absence of antimicro-
bial neutrophil granule proteins such as lactoferrin and defensins. An
inability to upregulate the expression of integrins stored on the spe-
cific granule membrane may also be responsible for the impairment
SECTION 22  Haematological disorders
5196
of host defence. The diagnosis of specific granule deficiency is made
by microscopic examination of neutrophils. With appropriate anti-
biotic prophylaxis and aggressive treatment of infections, patients
usually live to adulthood. A truncation mutation in the transcription
factor C/​EBPε has been demonstrated to be responsible for some,
but not all, cases of specific granule deficiency.
Monocytes
Monocytes are large circulating cells with a nonsegmented nucleus
and cytoplasmic granules. They function as phagocytes both in
antimicrobial defence and in clearing cellular debris. Their gran-
ules are essentially identical to neutrophil azurophilic granules,
and contain acid hydrolases and myeloperoxidase. Monocytes are
also capable of producing reactive oxygen and nitrogen compounds
with microbicidal activity. Monocytes play a critical role in the im-
mune response as they present antigens in the context of MHC to
T-cells. They also produce a variety of immunomodulatory cyto-
kines including interleukin (IL)-​1 and -​6, tumour necrosis factor-​α,
and interferon-​β.
Monocytes arise from bone marrow stem cells. They share a
common myeloid precursor with granulocytes. The differentiation
to the monocyte is modulated by several cytokines, most import-
antly monocyte CSF and granulocyte–​monocyte CSF. The majority
of monocytes are marginated to the vascular endothelium. Upon
stimulation, they migrate to the tissue where they develop into
macrophages. In the tissue they kill bacteria, mycobacteria, fungi,
and protozoa. They are especially important in defence against intra-
cellular pathogens. Specialized resident tissue macrophages include
the Langerhans’ cells of the skin, dendritic cells of lymph nodes,
Kupffer’s cells of the liver, and alveolar macrophages.
Monocytosis is defined as a monocyte count of greater than 0.9 ×
106/​µl. Disorders causing monocytosis are heterogeneous. Recovery
of the marrow following chemotherapy or agranulocytosis is her-
alded by monocytosis prior to the return of neutrophils. Monocytosis
is also seen in syndromes such as cyclic neutropenia, SCN, and idio-
pathic neutropenia.
The most common causes of monocytosis include chronic infection,
inflammation, or tumour, as well as some primary haematological
disorders (Box 22.3.1.3). Chronic infections leading to monocytosis
include subacute bacterial endocarditis and mycobacterial diseases.
Monocytosis is typically moderate and resolves with treatment of the in-
fection. Autoimmune processes such as systemic lupus erythematosus,
rheumatoid arthritis, and vasculitis also cause moderate monocytosis.
Monocytosis may arise from primary malignancies of the marrow or in
the setting of marrow infiltration with solid tumours.
Primary marrow disorders causing monocytosis include acute
monocytic leukaemia, chronic myeloid leukaemia and other
myeloproliferative disorders, and chronic myelomonocytic
leukaemia, which has features of both myelodysplastic and
myeloproliferative disorders. Juvenile chronic myeloid leukaemia
is a rare disorder occurring in children less than 4 years of age.
Lymphadenopathy and splenomegaly are also prominent features.
Monocytopenia in isolation is uncommon. Monocytopenia is
sometimes seen following steroid administration, endotoxaemia,
or in marrow failure syndromes such as aplastic anaemia. It is also
characteristic of hairy cell leukaemia.
Eosinophils
Morphology
Eosinophils have a bilobate nucleus and contain characteristic ellip-
tical granules that stain with eosin. There are three types of eosinophil
granules. Primary granules are round in shape. Secondary granules
are abundant and contain crystalloid material, and account for the
eosinophil’s staining properties. The third type of granule is small and
contains lysosomal enzymes. Granules contain high concentrations of
eosinophil major basic protein, histaminase, eosinophil cationic pro-
tein, hydrolases, and peroxidase. Eosinophils are capable of phagocytic
function but more commonly release their granule contents to the en-
vironment. Eosinophils are also capable of producing reactive oxygen
species, and produce prostaglandins, thromboxane A2, and leukotriene
C4. Eosinophils play a prominent role in defence against helminths
and parasites. They arise in the marrow from a common myeloid pre-
cursor, and their production is dependent on GM-​CSF, IL-​3, and IL-​5.
Disorders associated with eosinophilia are discussed elsewhere (see
Chapter 22.3.8); causes of eosinophilia are listed in Box 22.3.1.4.
Basophils
Basophils are rare circulating cells, accounting for less than 0.1% of
white blood cells. They are nonphagocytic granulocytes. Their large
heterogeneous granules account for their purple–​black staining. Their
granules contain histamine, heparin, tryptase, chemotactic factors
Box 22.3.1.3  Causes of monocytosis
Inflammatory diseases
•	 Infectious diseases:
—​	 Tuberculosis
—​	 Syphilis
—​	 Subacute bacterial endocarditis
—​	 Fungal infections
—​	 Kala-​azar
—​	 Brucellosis
•	 Autoimmune processes:
—​	 Systemic lupus erythematosus
—​	 Rheumatoid arthritis
—​	 Polyarteritis
—​	 Inflammatory bowel disease
—​	 Sarcoidosis
Malignancy
•	 Acute myeloid leukaemia
•	 Chronic myeloid leukaemia
•	 Chronic myelomonocytic leukaemia
•	 Juvenile chronic myeloid leukaemia
•	 Hodgkin disease
•	 Non-​Hodgkin lymphoma
•	 Histiocytoses
•	 Solid tumours
Miscellaneous
•	 Chronic neutropenia
•	 Post-splenectomy
•	 Marrow recovery

22.3.2 Myelodysplastic syndromes 5197 Charlotte K.
22.3.2 Myelodysplastic syndromes 5197 Charlotte K. Brierley and David P. Steensma
22.3.2  Myelodysplastic syndromes
5197
for neutrophils and eosinophils, leukotrienes, prostaglandins, and
platelet-​activating factor. They arise in the marrow from the same
myeloid precursor as eosinophils. Basophils function in immediate-​
type hypersensitivity. They are structurally similar to mast cells but
the exact relationship between these cell types is not clear. Basophilia
(> 0.2 × 106/​µl) is seen in myeloproliferative disorders such as chronic
myeloid leukaemia and polycythaemia vera, hypersensitivity reac-
tions, and with some viral infections including varicella and influenza.
Mast cell leukaemia is a rare disorder with a poor prognosis.
FURTHER READING
Andersohn F, Konzen C, Garbe E (2007). Systematic review: agran-
ulocytosis induced by nonchemotherapy drugs. Ann Intern Med,
146, 657–​65.
Chiriaco M, et  al. (2015). Chronic granulomatous disease:  clinical,
molecular and therapeutic aspects. Pediatr Allergy Immunol, 27,
242–​53.
Dinauer MC (2016). Primary immune deficiencies with defects in
neutrophil function. Hematology Am Soc Hematol Educ Program,
2016(1), 43–50.
Dinauer MC (2019). Inflammatory consequences of inherited dis-
orders affecting neutrophil function. Blood, 133(20), 2130–9.
Horwitz MS, et al. (2013). ELANE mutations in cyclic and severe con-
genital neutropenia: genetics and pathophysiology. Hematol Oncol
Clin North Am, 27, 19–​41.
Kiehl M, et al. (2019). Management of sepsis in neutropenic cancer
patients: 2018 guidelines from the Infectious Diseases Working
Party (AGIHO) and Intensive Care Working Party (iCHOP) of the
German Society of Hematology and Medical Oncology (DGHO).
Ann Hematol, 98(5), 1051–69.
Klion A (2018). Hypereosinophilic syndrome: approach to treatment
in the era of precision medicine. Hematology Am Soc Hematol Educ
Program, 2018(1), 326–31.
Lane A, Berliner N (2013). Nonmalignant disorders of leukocytes. ACP
Medicine. http://​what-​when-​how.com/​acp-​medicine/​nonmalignant-​
disorders-​of-​leukocytes-​part-​1/​ and http://​what-​when-​how.com/​
acp-​medicine/​nonmalignant-​disorders-​of-​leukocytes-​part-​2/​.
Schram AM, Berliner N (2018). Nonmalignant disorders of Leukocytes.
Scientific American Medicine. Decker Intellectual Properties, Inc.
DOI 10.2310/7900.1010.
22.3.2  Myelodysplastic syndromes
Charlotte K. Brierley and David P. Steensma
ESSENTIALS
The myelodysplastic syndromes (MDS) are marrow failure syn-
dromes characterized by cytopenias, blood cell dysmorphology, ac-
quired clonal cytogenetic and molecular genetic mutations, and a
risk of development of acute myeloid leukaemia. MDS may evolve
in patients previously treated with cytotoxic chemotherapy or radio-
therapy for a solid tumour, but most commonly arise de novo in pa-
tients over 60 years old.
Clinical features, diagnosis, and classification
Most patients present with features of chronic anaemia or manifest-
ations related to thrombocytopenia or infection. The diagnosis may
be suggested by the presence of normocytic or macrocytic anaemia,
with the peripheral blood smear showing dysplastic changes in red
blood cells or neutrophils. Bone marrow aspirate and biopsy permits
detailed cytogenetic study, which is critical for diagnostic classifi-
cation and prognosis. Increasingly, molecular genetic assays (next-​
generation sequencing panels) aid in diagnosis and prognosis.
The World Health Organization (WHO) classification recognizes
various MDS subtypes based on the morphological appearance of
the peripheral blood and bone marrow. These include MDS with ex-
cess blasts, MDS with deletion of the long arm of chromosome 5,
and MDS with ring sideroblasts.
Treatment and prognosis
Treatment is symptomatic in most cases. The only potentially curative
treatment is allogeneic bone marrow transplantation, which is often
precluded due to patients’ advanced age or comorbidity. Higher-​risk
patients may experience a survival benefit from treatment with the
DNA hypomethylating agent azacitidine, and decitabine delays dis-
ease progression to acute myeloid leukaemia. Some patients with
lower-​risk disease may show a response to immunosuppression
with antithymocyte globulin and ciclosporin. Patients with isolated
chromosome 5q deletions and lower-​risk disease may respond dra-
matically to lenalidomide, an immunomodulatory drug.
Prognosis varies widely (median survival <1 year to >10 years)
according to particular subtype and disease features such as pro-
portion of marrow blasts, karyotyping results, and the severity
of cytopenias. Patients with excess blasts, a complex karyotype,
and mutations of TP53 have the poorest prognosis. Most patients
who do not die of unrelated conditions die as a result of either
bleeding or infection, but in some, transformation to leukaemia
proves fatal.
Introduction
The myelodysplastic syndromes (MDS) are clonal haematopoi-
etic neoplasms characterized by bone marrow dysfunction with
peripheral blood cytopenias, dysplastic cell morphology, acquired
(somatic) cytogenetic and molecular genetic aberrations, and a vari-
able risk of progression to acute myeloid leukaemia (AML).
Box 22.3.1.4  Causes of eosinophilia
•	 Allergies
•	 Atopy
•	 Inflammation:
—​	 Collagen vascular diseases (rheumatoid arthritis, polyarteritis
nodosa, eosinophilic fasciitis)
•	 Infection:
—​	 Helminths
—​	 Parasites
•	 Neoplasms:
—​	 Hodgkin lymphoma and non-​Hodgkin lymphoma
—​	 Chronic myeloid leukaemia
—​	 Eosinophilic leukaemia
•	 Job’s syndrome
•	 Idiopathic hypereosinophilic syndromes
•	 Addison’s disease
SECTION 22  Haematological disorders
5198
MDS are clinically and genetically heterogeneous. These syn-
dromes result from malignant transformation and clonal expansion
of a mutated multipotent myeloid progenitor or stem cell. Disease-​
associated manifestations and trajectory vary significantly, from an
indolent condition with a single cytopenia (most commonly an-
aemia) that is stable for years, to fulminant bone marrow failure and
rapid progression to AML within months.
The 2016 World Health Organization (WHO) classification
system delineates several MDS subtypes, based on the percentage
of bone marrow blasts, the number of dysplastic lineages, and
characteristic chromosomal and molecular genetic abnormalities
(Table 22.3.2.1). MDS are classed as having progressed to AML
once the bone marrow or peripheral blood blast count reaches a
threshold of 20%. Although these syndromes were formerly known
as ‘preleukaemia’, only a minority of patients (c.30%) with MDS
progress to AML.
The advent of next-​generation sequencing technology has led
to advances in our understanding of the disease, but treatment
options remain limited and clinical outcomes unsatisfactory. This
chapter sets out to describe our current understanding of the
pathophysiology, epidemiology, clinical features, and therapeutic
options for this heterogeneous group of bone marrow failure
syndromes.
Aetiology
MDS are typically acquired disorders caused by de novo somatic
mutations in a haematopoietic progenitor or stem cell. Common
initiating mutations include those seen in TET2 or DMT3A, genes
which encode factors involved in epigenetic regulation, and SF3B1,
which encodes a component of the spliceosome. Subsequent muta-
tions such as in NRAS or TP53 genes may lead to clonal evolution
and diversity, and functional changes including increased cell prolif-
eration and genomic instability. The primary risk factor for MDS is
age, secondary to the increasing cumulative burden of somatic mu-
tations in ageing stem cells.
The incidence of MDS is mildly increased in petroleum and agri-
cultural workers, possibly due to accelerated mutation rates sec-
ondary to exposure to polycyclic aromatic hydrocarbons. Exposure
to ionizing radiation from the atomic bomb explosions at Hiroshima
and Nagasaki led to continued increased MDS risk among the ex-
posed population for more than 50 years post event. Therapy-​related
MDS (t-​MDS) comprises 5 to 10% of MDS, occurring in recipients
of radiation or chemotherapy as treatment for other disorders such
as solid tumours. Cytotoxic chemotherapies, particularly alkylating
agents and topoisomerase inhibitors, and ionizing radiation predis-
pose to t-​MDS both by selecting for expansion of pre-​existing TP53
mutant clones and by causing DNA damage and loss of chromo-
somal integrity in haematopoietic stem cells. t-​MDS carries an ad-
verse prognosis and occurs an average of 5 to 10 years post exposure
to alkylators or radiation, and sooner (1–​3 years) after topoisom-
erase inhibitors.
Rare familial cases of MDS, accounting for less than 2% of MDS
cases, have enabled identification of germline mutations in genes
such as GATA2, DDX41, SRP72, RUNX1, and others. Some of these
genes are also recognized as recurrent somatic mutations in de
novo/​t-​MDS. Other inherited syndromes conferring an increased
risk of MDS include Down syndrome, Fanconi anaemia, and
telomeropathies such as dyskeratosis congenita.
Epidemiology
MDS are estimated to be one of the most common haematological
malignancies with a reported age-​adjusted incidence of 4.9 to 5.3 per
100 000, though this may be an underestimate.
Ascertaining the exact epidemiology of MDS is difficult. Isolated
cytopenias, especially mild anaemia, are often underinvestigated
in the elderly. In addition, MDS diagnosis and classification are
complex, and cancer registry reporting of MDS cases has been
inconsistent.
The median age of MDS diagnosis in the United States of America
and Western Europe is approximately 70 years. Diagnosis in those
younger than 50 years is rare (unless exposed to the risk factors out-
lined previously) and diagnosis in children is extremely rare, with
an estimated annual incidence of one per million. However, there
is a worldwide variation in age at diagnosis, with a lower average
age documented in patients in Asia and Eastern Europe. The exact
cause of this is unknown, but may be attributable to differing envir-
onmental exposures or genetic background.
For reasons that remain unclear but may be due to either occupa-
tional exposure patterns or a protective effect of two copies of certain
X-​encoded genes, the incidence of MDS is approximately 1.5 times
higher in men than women. The exception to this is the MDS disease
subtype characterized by isolated loss of the long arm of chromo-
some 5, known as del(5q) or 5q− syndrome, which is more common
in women.
Pathogenesis/​pathology
MDS are clonal disorders, characterized by a complex interaction
between sequential acquired mutations in haematopoietic stem
cells, dysfunction of immune surveillance, and permissive alter-
ations in the bone marrow microenvironment.
The MDS cell of origin acquires sequential genetic and epigenetic
abnormalities, rendering it increasingly abnormal, initially unable
to differentiate normally and sustain normal haematopoiesis, and
ultimately unable to differentiate at all beyond the blast stage and
frankly malignant. MDS cells undergo expansion and proliferation
in the bone marrow to the detriment of healthy cells, and may ac-
quire additional mutations enabling the development of subclones
that may differ in behaviour from the ancestral clone. The subse-
quent domination of the bone marrow by one or several mutated
progenitor cells is known as clonal haematopoiesis. Clonally re-
stricted haematopoiesis dominated by mutant cells leads to mor-
phological cell dysplasia and cytopenias; the risk of progression to
AML is a result of the genomic instability of such clones. Eight dif-
ferent MDS subtypes are recognized by the WHO and these are out-
lined in Table 22.3.2.1.
Recent analyses have identified that haematopoietic clones
bearing mutations in DNMT3A, TET2, or other MDS-​associated
genes occur in up to 10% of older adults, often in the absence of other
features of a haematological disorder. Known as ‘clonal haematopoi-
esis of indeterminate potential’ (CHIP), this disease precursor state
22.3.2  Myelodysplastic syndromes
5199
confers a risk of progression to MDS of about 0.5 to 1% per year and
a higher all-​cause mortality. CHIP is also a risk factor for cardiovas-
cular events due to a proinflammatory interaction between clonally
derived monocytes/​macrophages and the vascular endothelium.
In turn, some patients experience persistent cytopenias in the
presence of unremarkable marrow morphology, normal cytogen-
etics, and no MDS-​associated genetic mutations. This condition
is termed ‘idiopathic cytopenia of unknown significance’ (ICUS).
Some patients with ICUS will ultimately be discovered to have an-
other haematological neoplasm or a nonhaematological disorder.
Patients who have idiopathic cytopenias and also have a clonal mu-
tation that may be contributing to the marrow failure, but who lack
other features of MDS such as extensive dysplasia or increased blast
cells, are said to have ‘clonal cytopenia of unknown significance’
(CCUS). CCUS has a natural history akin to lower-​risk MDS, and
individuals with CCUS are at higher risk for disease progression
than patients with ICUS.
Clonal haematopoiesis dominates even in lower-​risk MDS
without excess blast cells in the marrow, and clones identified
in secondary AML can be tracked back to the preceding MDS
stage. Recent studies combining immunophenotyping and deep
sequencing have identified a rare multipotent haematopoietic pro-
genitor cell of origin in lower-​risk MDS and provide evidence that
MDS occur secondary to mutations sustained in a stem cell with
­intrinsic self-​renewal ­capacity, as opposed to a differentiated cell
that acquires self-​renewal ability.
Three different types of genetic anomalies have been identified
in MDS: chromosomal abnormalities, aberrancies in epigenomic
pattern, and single gene mutations.
Chromosomal abnormalities
Chromosomal abnormalities in MDS are of prognostic relevance.
Gains or losses of chromosomal material are detectable on meta-
phase karyotyping in greater than 50% of de novo MDS and greater
Table 22.3.2.1  World Health Organization classification of the adult myelodysplastic syndromes (2016)
MDS subtype
Blast proportion
Dysplasia
Additional notes
Peripheral
blood (%)
Bone
marrow (%)
MDS with single
lineage dysplasia
(MDS-​SLD)
<1
<5
Present in >10% cells in a
single lineage
•	Cases with 2 cytopenias may be included here, but marrow
dysplasia must be limited to 1 lineage
•	Ring sideroblasts represent <15% of erythroid precursors, or if
SF3B1 mutation is present, ring sideroblasts represent <5% of
erythroid precursors
MDS with ring
sideroblasts
(MDS-​RS)
0
<5
Dysplastic erythroid lineage
(>10% of erythroid precursors)
•	Anaemia (normocytic/​macrocytic)
•	Ring sideroblasts comprise ≥15% of erythroid precursors, or if
SF3B1 mutation is present, ring sideroblasts represent ≥5% of
erythroid precursors
•	If multilineage dysplasia is present, it is MDS-​RS-​MLD
MDS with
multilineage
dysplasia
(MDS-​MLD)
<1
<5
In 2+ lineages (>10% of cells in
each lineage affected)
•	1+ cytopenias
•	No Auer rods
•	May or may not have ≥15% ring sideroblasts; if ring sideroblasts are
present, this may be denoted as MDS-​RS-​MLD
MDS with excess
blasts-​1 (MDS-​EB1)
<5
5–​9
In 1+ lineages (>10% of cells in
each lineage affected)
•	1+ cytopenias
•	No Auer rods
MDS with excess
blasts-​2 (MDS-​EB2)
5–​19
10–​19
In 1+ lineages (>10% of cells in
each lineage affected)
•	1+ cytopenias
•	Auer rods present in context of blast count <20%: MDS-​EB2
MDS with isolated
deletion of
chromosome 5q
(5q− syndrome)
<1
<5
High numbers of
megakaryocytes, many small
and with hypolobated/​
nonlobated nuclei
Dysplasia in other lineages rare
•	Anaemia (often macrocytic) with or without other cytopenias/​
thrombocytosis
•	No Auer rods
•	Interstitial or terminal deletion of the long arm of chromosome 5,
either alone or with 1 other clonal cytogenetic alteration
•	Some MDS patients with del5q may better fit other categories
(e.g. if 8% marrow blasts are present, MDS-​EB1 is the diagnosis)
MDS, unclassifiable
(MDS-​U)
£1
<5
Unequivocal, but dysplasia
present in <10% of cells of
1+ lineages
•	Can progress to a specific MDS
•	Can include cases otherwise classified as MDS-​SLD or MDS-​MLD
but with 1% blasts in peripheral blood
•	Can include cases with an MDS-​associated chromosome
abnormality (other than loss of the Y chromosome or trisomy 8
or deletion of chromosome 20, which are not specific enough for
MDS) but without dysplasia
Therapy-​related MDS
or AML
(t-​MDS/​AML)
Any
Any
Variable
•	MDS resulting from prior therapy with DNA damaging
chemotherapy or irradiation, usually associated with abnormalities
of chromosome 5 or 7 or TP53 gene mutation in the case of
alkylating agents
•	WHO does not distinguish t-​MDS from t-​AML since aetiology
similar and prognosis very poor for both
Note: this table does not include MDS subtypes with proliferative features (e.g. chronic myelomonocytic leukaemia), myeloid neoplasms with germ line predisposition (e.g. MDS arising
as a consequence of germline mutations in GATA2, DDX41, SRP72, RUNX1, or a congenital syndrome such as a telomere disorder), or the provisional and highly heterogeneous entity of
MDS-​refractory cytopenias in childhood.
SECTION 22  Haematological disorders
5200
than 80% of t-​MDS. Smaller chromosomal amplifications/​trans-
locations may be detectable by fluorescence in situ hybridization
(FISH) or other techniques. High-​resolution techniques (e.g. com-
parative genomic hybridization), have led to the detection of subtle
chromosomal deletions in approximately 90% of patients.
Deletion of the long arm of chromosome 5, del(5q), is the most
common recurrent karyotypic abnormality, documented in ap-
proximately 15% of MDS patients. As noted above, a subset of pa-
tients, predominantly females, with del(5q) have the ‘5q− syndrome’
characterized by dyserythropoietic anaemia, micromegakaryocytes,
a low risk of AML transformation, and a high clinical response rate
to lenalidomide therapy. The size of the 5q deletion is variable; the
two commonly deleted regions are 5q31.1 and 5q32 to 5q33.3, and
many patients with del(5q) MDS lose both. Loss of only the distal
region is associated with a more favourable course. As most patients
with del(5q) retain a normal 5q arm, haploinsufficiency is enough
to cause the phenotype, and a number of genes on 5q− have been
identified as relevant.
Haploinsufficiency of RPS14, a ribosomal subunit gene at
5q31.2, leads to p53 activation in erythroid progenitors and
dyserythropoiesis. Loss of a copy of CSNK1A1, located at 5q32
and encoding casein kinase 1, engenders lenalidomide-​sensitivity.
Del(5q) may also occur together with other abnormalities, in
which case it bears a more adverse prognosis with poor response
to lenalidomide. Del(5q) and TP53 mutations co-​occur more
often than should be expected by chance, suggesting pathogenic
cooperation.
Loss of one entire copy of chromosome 7 (monosomy 7) occurs
in 5% of MDS patients. Monosomy 7 occurs most frequently after
alkylating agent exposure and is a poor prognostic marker. The
causative genetic loss remains unclear, although a number of recur-
rently mutated genes lie on 7q.
Trisomy 8, a large-​scale amplification of chromosome 8, is present
in about 5% of MDS patients and is nonspecific to MDS.
A complex karyotype, defined as three or more chromosomal
abnormalities, is common in t-​MDS and associated with TP53 mu-
tations in more than 50% of cases. The term ‘monosomal’ karyo-
type describes loss of two or more entire chromosomes or deletion
of one chromosome in association with another structural cyto-
genetic abnormality. Both complex and monosomal karyotypes
most commonly affect chromosomes 5 or 7 and carry an adverse
prognosis.
Aberrancies in epigenomic pattern
Epigenetic changes are heritable alterations in chromatin struc-
ture that affect gene expression, largely via DNA methylation, or
by histone acetylation, methylation, or other modification. The
underlying DNA sequence remains unaltered. MDS patients display
aberrant methylation when compared to healthy controls—​both
hypermethylation in promoters of tumour suppressors and global
hypomethylation elsewhere. How methylation patterns link to
pathogenesis remains unclear. Hypomethylating agents (HMAs; e.g.
the DNA methyltransferase inhibitors decitabine and azacitidine)
have demonstrated clinical responses and survival benefits in the
treatment of MDS. Reactivating tumour suppressor genes silenced
by hypermethylation comprises a potential mechanism of action
for HMAs, but this is yet to be proven, and no single gene or gene
methylation pattern consistently correlates with response.
Single gene (‘point’) mutations
The most common genetic abnormalities in MDS are single gene
mutations and over 50 different recurrent somatic mutations have
been identified, accounting for the clinical heterogeneity of the
disease. Over 90% of patients with MDS carry at least one clonal
somatic mutation. No single mutation dominates, and only a few
mutations occur in more than 20% of cases. Table 22.3.2.2 describes
key recurrent gene mutations identified in MDS. The vast combina-
torial genetic heterogeneity and apparent cooperativity of some mu-
tations hints at the complexity of attributing pathogenesis to single
gene alterations.
Differentiating between the rarer ‘driver’ mutations of pathogenic
consequence and so-​called passenger mutations is a key challenge
in MDS. The identification of a ‘driver’ mutation implies that the
mutation is recurrent in MDS, has a plausible function contributing
to pathogenesis, and, ideally, that the effect of its disruption can be
recapitulated in an in vitro or in vivo model.
Recent insight to the biological organization of proteins encoded
by recurrently mutated genes into cellular pathways and functional
pathways has increased understanding as to how the MDS clone
evolves. Pathways affected include the RNA splicing machinery,
DNA methylation and histone modification regulators, signal trans-
duction apparatus, and haematopoietic growth factors. There is
emerging evidence of the association of individual genetic mutations
with prognosis and of a typical sequence of acquisition. Certain mu-
tations can inform clinical care as they are associated with specific
clinical features, yet personalizing therapy to an individual’s genetic
mutations is not yet a reality except in rare cases such as IDH1/​2 or
BRAF mutations.
Clinical features
The clinical features of MDS are largely a consequence of inef-
fective haematopoiesis. Most patients experience symptomatic
Table 22.3.2.2  Most frequent recurrent gene mutations in MDS
Class of gene
Gene
Frequency (%)
Prognosis
Spliceosome
components
SF3B1
SRSF2
U2AF1
ZRSR2
20–​30
10–​15
5–​12
1–​4
Favourable
Adverse
Neutral or adverse
Adverse
Epigenetic
modifiers
TET2
DNMT3A
ASXL1
EZH2
IDH1/​2
20–​30
8–​13
10–​20
5–​10
<5
Neutral
Unclear
Adverse
Adverse
Unclear
Transcription
factors
RUNX1
GATA2
10–​15
Rare
Adverse
Adverse
Genome stability
TP53
PPM1D
10–​12
<5
Adverse
Adverse
Tyrosine kinase
signalling
JAK2
NRAS
KRAS
<5
5–​10
<5
Unfavourable
Adverse
Adverse
Cohesin complex
STAG2
RAD21
5–​10
<5
Unclear
Unclear
GPCR complex
GNAS
Rare
Unclear
GPCR, G protein-​coupled receptor.
22.3.2  Myelodysplastic syndromes
5201
cytopenias, although MDS can also be diagnosed as a result of
an incidental finding on a full blood count. Key features are
exertional breathlessness due to anaemia, infection due to neu-
tropenia and neutrophil dysfunction, and bleeding and easy
bruising due to thrombocytopenia and platelet dysfunction.
Fatigue is common and does not appear to correlate with the
degree of anaemia, and instead may relate to cytokine release by
clonal cells.
Approximately 15% of MDS patients experience paraneoplastic
features. These range from skin manifestations such as Sweet
syndrome (characterized by neutrophilic dermatosis associated
with painful plaques, fever and arthralgia) to rheumatological
symptoms, such as diffuse arthralgias or inflammatory arthritis.
A  finding of splenomegaly or hepatomegaly should raise sus-
picion of another diagnosis or of an overlap syndrome with a
myeloproliferative neoplasm.
MDS has a variable natural history. About 25 to 30% of patients
progress to AML, which is usually fatal. More than 50% of MDS pa-
tients ultimately succumb to complications of cytopenia. Infection
is the most common cause of nonleukaemic death, followed by
bleeding. The susceptibility to infection is not merely due to pro-
longed periods of neutropenia, but also relates to functional neu-
trophil defects resulting in an impaired inflammatory response.
Similarly, patients with a normal platelet count may experience
spontaneous bleeding due to ineffective platelet function. As pa-
tients are often elderly, a significant proportion of MDS patients die
of unrelated causes.
Differential diagnosis
Not all haematopoietic dysplasia is MDS and in the absence of
clonal markers or excess blasts, MDS becomes a diagnosis of exclu-
sion. A number of nonclonal disorders may exhibit similar morpho-
logical changes and these need to be considered and actively ruled
out in making the diagnosis.
Vitamin and micronutrient deficiencies, specifically vitamin B12,
folate, copper, and iron, may lead to dysplastic marrow appear-
ances. Vitamin B12 and folate deficiency both cause macrocytosis
and the bone marrow may demonstrate megaloblastoid changes
and a preponderance of immature cells such that the appear-
ance can be similar to early AML. Copper deficiency, most
often observed post-gastrectomy or in the context of zinc sup-
plement intake, can lead to severe anaemia and the formation
of ring sideroblasts on bone marrow. Ring sideroblasts are
erythroid progenitors with iron-​laden mitochondria and can
feature in all MDS subtypes. If at least 15% of erythroid pre-
cursors are ring sideroblasts in the absence of other marrow
abnormalities, or at least 5% are ring sideroblasts and a somatic
mutation in SF3B1 (encoding a component of the spliceosome)
is present, this is diagnostic of the WHO category ‘MDS with
ring sideroblasts’ (MDS-​RS). Finding an SF3B1 mutation helps
exclude reactive causes of sideroblasts or late presentations of
congenital sideroblastic anaemias.
HIV infection may confer trilineage dysplasia, particularly in the
erythroid series, and an HIV test should be considered if chromo-
some abnormalities and increased blasts are not present. Alcohol
excess may also induce cytopenias accompanied by sideroblastic
changes or a megaloblastic appearance.
A careful history of exposure to drugs that may cause cytopenias
and marrow changes mimicking MDS needs to be sought when
investigating dysplasia. The list of potential culprits is long, and
includes methotrexate, azathioprine, valproate, ganciclovir, and
mycophenolate mofetil.
MDS may also be confused with other clonal disorders.
Autoimmune-​induced bone marrow failure secondary to clonal
T-cells (e.g. in the chronic lymphoid neoplasm, T-​cell large granular
lymphocyte leukaemia) may have very similar bone marrow
changes. Aplastic anaemia, an oligoclonal disorder usually asso-
ciated with a normal morphology and karyotype, can be difficult
to differentiate from hypoplastic MDS. Finally, there are overlap
syndromes of MDS/​myeloproliferative neoplasm such as chronic
myelomonocytic leukaemia which are categorized separately from
MDS and have a unique mutational spectrum.
Laboratory features
MDS are primarily laboratory diagnoses (see Box 22.3.2.1 for diag-
nostic criteria). Determining the MDS subtype by WHO criteria is
integral to the diagnostic work-​up (Table 22.3.2.1). Relevant inves-
tigations include a full blood count, peripheral blood film, and bone
marrow aspirate and trephine. Metaphase cytogenetic analysis is es-
sential and FISH is useful if karyotyping fails. Increasingly, targeted
sequence analysis contributes valuable diagnostic and prognostic
information.
Full blood count
More than 85% of MDS patients demonstrate anaemia at presenta-
tion, which is usually macrocytic but can be normocytic or in rare
cases microcytic. About 50% of patients are neutropenic and ap-
proximately 25% are thrombocytopenic at diagnosis; the frequency
of these other cytopenias increases with time.
Peripheral blood film
The peripheral blood film is often suggestive of the diagnosis. Red cell
abnormalities on the film vary widely. Red cells may be of unequal
size (anisocytosis), unequal haemoglobinization (anisochromia),
abnormally shaped (poikilocytosis), nucleated, or demonstrate
Box 22.3.2.1  Diagnostic criteria for MDS
Presence of a persistent and otherwise unexplained cytopenia:
1	 Haemoglobin less than or equal to 110 g/​litre
2	 Absolute neutrophil count less than or equal to 1.5 × 109/​litre
3	 Platelet count less than or equal to 100 × 109/​litre
Plus one of the following:
1	 5–​19% marrow blasts
2	 Dysplasia in at least 10% of cells in erythroid, myeloid, or megakaryocyte
lineages
3	 Evidence of a characteristic MDS-​associated cytogenetic abnormality
4	 Exclusion of alternative diagnosis accounting for dysplastic cell
morphology
SECTION 22  Haematological disorders
5202
inclusions of aggregated ribosomes (appearing as small blue dots at
the periphery, known as basophilic stippling). Reticulocyte counts
are disproportionately low. Red cell function may also be aberrant,
evidenced by abnormal surface antigen expression and reduced red
cell enzyme activity.
One MDS subtype—​α-​thalassaemia MDS or acquired haemo-
globin H disease—​confers a red cell morphology akin to
α -​thalassaemia and occurs secondary to a somatic mutation in
ATRX, which encodes a chromatin-​remodelling protein. Red cells
in this subtype are microcytic with target cells, poikilocytosis, and
anisocytosis and beta chain tetramers may be evident on crystal
violet or other supravital stain.
In the myeloid lineage, neutrophils demonstrate visible anom-
alies, such as hypogranularity, hypersegmentation, aberrant ring
shapes, or a specific phenotype of condensed chromatin with bi-
lobed nuclei, known as pseudo-​Pelger–​Huët cells (Fig. 22.3.2.1).
(Pelger–​Huët anomaly is a benign congenital condition of neutro-
phils.) A left shift to immature myeloid cells predominates, and cir-
culating early myeloid cells and blasts can occur.
Platelets may also demonstrate dysplastic features such as large
size or hypogranularity.
Bone marrow
In normal bone marrow, the ratio of myeloid to erythroid progen-
itors is 2:1 to 4:1. In MDS, a profound skew to the erythroid lineage
may occur, often as far as 1:2. A range of characteristic abnormal
erythroid progenitors have been documented. Megaloblastoid fea-
tures with multinucleated red cell precursors and dyssynchronous
nuclear/​cytoplasmic maturation may be seen (Fig. 22.3.2.1).
Chromatin fragments can appear in the cytoplasm as part of the
destruction of the nucleus during cell death, known as nuclear
karyorrhexis. Perls’ stain or the Prussian Blue reaction is required to
identify ring sideroblasts (Fig. 22.3.2.1).
Myeloid progenitors may be hypo-​ or hypergranulated, hyper­
segmented, or feature hypolobated nuclei. Increased numbers of
blasts may be evident and can serve as a marker of increased risk
of AML.
Micromegakaryocytes or hyper-​ or hypolobated megakaryocytes
are seen (Fig. 22.3.2.1), and megakaryocytes may be distributed
anomalously within the marrow in MDS.
Management
General considerations
Treatment options in MDS range from conservative management with
supportive care alone to high-​intensity therapy including allogeneic
stem cell transplantation (alloSCT), which is the only curative option.
Any proposed treatment must be tailored to the patient, considering
age, comorbidity, functional status, disease risk, and patient wishes.
MDS is generally refractory to conventional cytotoxic chemotherapy,
likely because the cell of origin is a quiescent stem cell that is resistant
to cytotoxics—​and also because even when the abnormal clone is
cytoreduced, blood counts may recover poorly since normal haemato-
poietic reserve is limited. The lack of well-​defined targets in MDS means
that there are no approved biologically personalized therapies.
Current management is based on risk scoring algorithms. If disease
risk is classed as low/​intermediate-​1 by the MDS risk scoring system
(International Prognostic Scoring System or IPSS), or very low/​low by the
2012 revised IPSS (IPSS-​R), current algorithms advise close monitoring
for transition of disease and transfusion or haematopoietic growth factor
support, with use of lenalidomide if del(5q) is present or immunosup-
pressive therapy in selected cases, and potential addition of iron chelation
therapy. If disease risk is IPSS intermediate-​2 or higher, or IPSS-​R high or
very high, then life expectancy is below 2 years and immediate disease-​
modifying therapy is advised, including alloSCT if the patient is young
enough and fit. Treatment of IPSS-​R intermediate category can be similar
to either lower-​ or higher-​risk subgroups, depending on specific disease
features. Participation in clinical trials should be offered where possible.
Supportive care
Good supportive care is invaluable in the management of MDS.
Prompt management of febrile episodes with broad-​spectrum anti-
biotics may be life-​saving, especially in the context of neutropenia.
Red cell transfusion with a transfusion trigger of 70 to 80 g/​litre can
minimize symptoms of anaemia. Platelet transfusions should be used ju-
diciously due to the risk of alloimmunization. A randomized controlled
trial (RCT) demonstrated that a prophylactic strategy with trigger of 10
× 109/​litre leads to fewer significant haemorrhages than a therapeutic
strategy of transfusing only in context of bleeding. Antifibrinolytics can
be helpful where recurrent mucosal bleeding is a key feature.
(b)
(a)
Fig. 22.3.2.1  (a) Peripheral blood smear in MDS showing several dysplastic neutrophils,
including a pseudo-​Pelger–​Huët neutrophil with a bilobed nucleus. The chromatin in the
neutrophils is clumped, and the red cells show a range of sizes and appearances. (b) A bone
marrow aspirate with a small, monolobed megakaryocyte, typical in MDS.
22.3.2  Myelodysplastic syndromes
5203
Iron chelation
The use of iron chelators such oral deferasirox and parenteral
desferrioxamine in MDS remains highly controversial. Patients with
MDS are often heavily transfused and retrospective studies indicate
that transfusion dependency and high serum ferritin are markers of
poor outcome. Whether this reflects a more advanced disease state
or complications of iron overload cannot be demonstrated retro-
spectively. Retrospective case cohort studies and meta-​analyses
demonstrating benefit of iron chelators suffer from selection bias. No
definite prognostic benefit has yet been demonstrated prospectively,
although a composite endpoint of potentially iron related events was
reduced with chelation therapy in a randomized trial of deferasirox
versus placebo. Competing risks such as clonal progression and com-
plications of cytopenias dominate the clinical picture in most cases.
However, some patients who undergo chelation will experience
improvement in organ function or haematopoiesis. Current pub-
lished consensus guidelines state that iron chelators can be con-
sidered for patients with lower-​risk disease and multiple transfusions
in whom end-​organ damage is anticipated.
Haematopoietic growth factors
Erythropoiesis-​stimulating agents
Erythropoiesis-​stimulating agents (ESAs) have been extensively in-
vestigated as therapy to reduce transfusion needs in MDS. Results
from over 20 studies demonstrate that 40 to 50% of MDS patients re-
spond to ESAs with modest benefit that lasts a median of 1 to 2 years.
ESAs are most effective in low-​risk disease that is not heavily transfu-
sion dependent and in patients with normal blast counts, low inflam-
matory markers, and low baseline serum erythropoietin levels (<500
U/​litre). In current clinical practice, ESA therapy is initiated when
the haemoglobin falls below 100 g/​litre and continued to a target of
110 to 120 g/​litre with response assessed at 2 to 3 months. For pa-
tients with MDS-associated anaemia in whom ESAs are no longer
effective, luspatercept, an activin receptor ligand trap molecule that
binds to erythropoiesis-inhibitor cytokines of the transforming
growth factor beta superfamily, reduced transfusion needs and in-
creased haemoglobin compared to placebo in a randomized trial.
Granulocyte colony-​stimulating factor and granulocyte–​
monocyte colony-​stimulating factor
While granulocyte colony-​stimulating factor (G-​CSF) and
granulocyte–​monocyte colony-​stimulating factor (GM-​CSF) can
improve the absolute neutrophil count, these myeloid growth fac-
tors do not compensate for neutrophil dysfunction and there is no
evidence that G-​CSF or GM-​CSF improves survival in MDS. In
high-​risk disease it is feared that GM-​CSF may accelerate leukaemic
expansion, though this appears to be rare.
Thrombopoiesis-​stimulating agents
Bleeding is the second most common cause of nonleukaemic death
in MDS, and low platelet count limits the tolerability of disease-​
modifying therapies. Two thrombopoietin receptor activators,
romiplostim and eltrombopag, have demonstrated efficacy in re-
ducing platelet transfusions and bleeding in low-​risk MDS. RCTs
for each agent demonstrated a modest increased risk of disease pro-
gression; the trial with romiplostim was stopped early by the data
monitoring committee, though with 5-​year follow-​up there was no
difference between romiplostim and placebo with respect to AML
progression. Individual patients with platelet  alloimmunization
and severe thrombocytopenia in whom the bleeding risk outweighs
leukaemia progression risk can be considered for thrombopoiesis-​
stimulating agent use on a case-​by-​case basis.
Disease modification
Lenalidomide
RCT evidence supports the use of lenalidomide, a derivative of thalido-
mide, in reversal of anaemia in lower-​risk MDS associated with isolated
del(5q). Patients with del(5q) MDS achieve high response rates with
lenalidomide, with 70% achieving transfusion independence and more
than 30% cytogenetic normalization, lasting a median of 2 years. This is
currently the only example in MDS of a genetic abnormality dictating
treatment choice. Responses are more likely with a higher dose (10 mg
daily vs 5 mg daily) and AML progression risk is not affected.
A novel mechanism of action of lenalidomide has recently been
elucidated. Lenalidomide binds cereblon, a component of an E3 ubi-
quitin ligase complex, which modifies its affinity for ubiquitination
of casein kinase 1α, accelerating proteasomal degradation of the
latter. The gene encoding casein kinase 1α, CSNK1A1, lies on 5q and
haploinsufficiency renders the 5q− cells sensitive to lenalidomide
with a therapeutic window.
In the absence of 5q−, response rates to lenalidomide lie at 25%
and last a median of 8 to 9 months. Side effects of lenalidomide in-
clude diarrhoea, rash, and cytopenias.
Hypomethylating agents
HMAs are the mainstay of treatment for higher-​risk MDS and can
be considered for those with lower-​risk disease who are refractory
to other treatments. The use of HMAs resulted from the recognition
that MDS genomes display highly aberrant methylation patterns.
Two agents, azacitidine (AZA) and decitabine, are azanucleoside ana-
logues that bind to and inhibit DNA methyltransferase irreversibly,
thereby reducing methylation status and changing gene expression.
HMAs are also able to exert direct cytotoxicity via incorporation into
DNA as a false nucleotide similarly to cytarabine, and it remains un-
clear which is the more relevant mechanism of response.
HMA efficacy is underpinned by RCT evidence. AZA is the only
drug with a demonstrated survival benefit in MDS; in one multicentre
RCT, AZA prolonged life by 9 months compared with a supportive
care or conventional chemotherapy control arm. Approximately
40 to 50% of MDS patients achieve a response with AZA, while
15% have complete pathological response. Most respond within 6-​
monthly cycles of therapy, and patients experience better quality of
life and delayed disease progression with AZA when compared to
conventional care. Side effects are mild, most commonly cytopenias
and gastrointestinal upset. TET2 mutation status predicts a slightly
higher likelihood of response to AZA and ASXL1 mutation predicts
a lower likelihood of response, but is not currently part of treatment
choice algorithms since many patients who are TET2 wildtype or
ASXL1 mutant will also respond to AZA.
Decitabine is a structurally similar compound to AZA but is
directly incorporated into DNA whereas most AZA is incorpor-
ated into RNA. At low doses, decitabine acts as a hypomethylating
agent, whereas at higher doses it induces DNA crosslinking and cell
cycle arrest. RCTs have not demonstrated a survival benefit with
decitabine compared to observation, possibly due to suboptimal
SECTION 22  Haematological disorders
5204
dosing/​scheduling in trials. High response rates have been reported
with decitabine treatment of patients with higher-​risk MDS or AML
with TP53 mutation; these responses are not durable.
HMAs are not curative, and once they fail the patient life expect-
ancy is sub 6 months with no useful treatment options beyond sup-
portive care. Numerous agents are being combined with HMA in
the clinical trial setting with early promise, including venetoclax,
immune checkpoint inhibitors, and targeted agents.
Cytotoxic chemotherapy
Intensive induction chemotherapy with regimens similar to those
used for AML is sometimes used in young patients with high-​risk
MDS, but is generally not effective because of clonal resistance and
impaired blood count recovery. Low-​dose cytarabine, low-​dose
melphalan, and other cytotoxics are also used but are of limited
value. Remission rates are less than 20% and there is a theoretical
risk of selecting for a quiescent, malignant clone.
Immunosuppression
Some cases of MDS feature an autoimmune pathophysiology akin to
aplastic anaemia. In such cases, autoreactive T-cells inhibit haemato-
poiesis, and are susceptible to immune suppression with antithymocyte
globulin (ATG) or calcineurin inhibitors such as tacrolimus or ciclosporin.
In retrospective analyses, these immunosuppressive therapy agents have
demonstrated improved overall survival and reduced risk of AML trans-
formation in responding patients, but selection of appropriate MDS pa-
tients for immunosuppressive therapy has proven challenging. In various
series, hypoplastic marrow, trisomy 8 or normal karyotype, younger age
and female sex, HLA DR15, and presence of a paroxysmal nocturnal
haemoglobinuria clone predicted a higher likelihood of response to
immunosuppressive therapy. In the largest study of immunosuppres-
sive therapy to date—​367 patients from 13 centres—​only hypocellular
marrow predicted response, and the response rate to equine ATG plus
ciclosporin was higher than to rabbit ATG or ATG without ciclosporin.
Allogeneic haematopoietic stem cell transplantation
AlloSCT is an option only for those who are young and fit enough to
undergo high-​intensity treatment. Despite recent advances in trans-
plant medicine including reduced intensity conditioning regimens,
improved supportive care, and availability of new stem cell sources,
alloSCT remains a high-​risk procedure conferring a transplant-​
related mortality of 15 to 20%. However, at least one-​third of patients
achieve long-​term disease-​free survival. The decision to transplant is
difficult, as alloSCT imposes an immediate risk of mortality for the
possibility of long-​term survival and there is a lack of randomized,
quality data to guide clinicians. Many practical questions remain,
including defining the optimal conditioning regimen, the timing of
alloSCT, and the role for bridging therapy or reducing disease burden
prior to alloSCT. Mathematical models indicate that alloSCT should
be considered in all fit patients with higher-​risk disease under the
age of 75. Patients with TP53 mutations or Ras pathway mutations
have a higher relapse rate than other genotypes.
Prognosis
The heterogeneous nature of MDS renders prognostication difficult.
In response, a number of scoring systems have been developed to
aid clinical decision-​making and differentiate low-​risk patients with
stable disease from high-​risk patients at risk of fulminant cytopenia
and rapid progression to AML.
The most widely used scoring system for prognostication is the
IPSS, originally published in 1997 and revised in 2012 as the IPSS-​
R. The IPSS-​R is now the standard method to predict risk of death
in MDS with supportive care only. It stratifies patients into five risk
categories by marrow cytogenetics, blast percentage, and number
and degree of cytopenias.
Other adverse markers of prognosis not included in the IPSS-​R
include the presence of comorbidities, high ferritin or lactate de-
hydrogenase levels, aberrant myeloid cell surface marker expres-
sion, and the presence of certain high-​risk molecular abnormalities,
such as mutations in TP53 or EZH2. No scoring system to date in-
cludes all identified prognostic variables to predict individual dis-
ease trajectory.
Areas of uncertainty, controversy, and
future developments
Clinical outcomes for MDS remain disappointing. Recent major ad-
vances in our understanding of MDS disease biology will advance
clinical management in coming years. Clinically defined subtypes
are genetically highly heterogeneous, which to date has posed a
major barrier to development of reliable therapeutic targets. As our
understanding of pathogenic driver mutations in MDS grows, the
chances of identifying treatable molecular targets are increasing.
Yet many more basic questions remain. Genetic profiling is cur-
rently formally integrated into disease classification only with re-
spect to SF3B1 and MDS-​RS. The frequency of complications
secondary to iron overload and the consequent role for chelation
therapy in MDS remains unclear. Few agents in clinical trials today
offer hope of major advances in current treatment, as most focus
on refinement of drug dosing or combine established treatments.
Going forward, prospective recruitment for MDS registry studies
and trials will be critical in engendering a step change in clinical
outcomes.
FURTHER READING
Bejar R, Steensma DP (2014). Recent developments in myelodysplastic
syndromes. Blood, 124, 2793–​803.
Bejar R, et  al. (2011). Clinical effect of point mutations in
myelodysplastic syndromes. N Engl J Med, 364, 2496–​506.
Fenaux P, et  al. (2009). Efficacy of azacitidine compared with that
of conventional care regimens in the treatment of higher-​risk
myelodysplastic syndromes:  a randomised, open-​label, phase III
study. Lancet Oncol, 10, 223–​32.
Fenaux P, et al. (2011). A randomized phase 3 study of lenalidomide
versus placebo in RBC transfusion-​dependent patients with low-​/​
intermediate-​1-​risk myelodysplastic syndromes with del5q. Blood,
118, 3765–​76.
Gattermann N (2008). Overview of guidelines on iron chelation therapy
in patients with myelodysplastic syndromes and transfusional iron
overload. Int J Hematol, 88, 24–​9.
Genovese G, et al. (2014). Clonal hematopoiesis and blood-​cancer risk
inferred from blood DNA sequence. N Engl J Med, 371, 2477–​87.

22.3.3 Acute myeloid leukaemia 5205 Nigel Russell
22.3.3 Acute myeloid leukaemia 5205 Nigel Russell and Alan Burnett
22.3.3  Acute myeloid leukaemia
5205
Greenberg PL, et al. (2012). Revised international prognostic scoring
system for myelodysplastic syndromes. Blood, 120, 2454–​65.
Haferlach T, et al. (2014). Landscape of genetic lesions in 944 patients
with myelodysplastic syndromes. Leukemia, 28, 241–​7.
Jaiswal S, et al. (2014). Age-​related clonal hematopoiesis associated
with adverse outcomes. N Engl J Med, 371, 2488–​98.
Lubbert M, et  al. (2011). Low-​dose decitabine versus best sup-
portive care in elderly patients with intermediate-​ or high-​risk
myelodysplastic syndrome (MDS) ineligible for intensive chemo-
therapy:  final results of the randomized phase III study of the
European Organisation for Research and Treatment of Cancer
Leukemia Group and the German MDS Study Group. J Clin Oncol,
29, 1987–​96.
Platzbecker U (2019). Treatment of MDS. Blood, 133(10), 1096–107.
Silverman LR, et al. (2002). Randomized controlled trial of azacitidine
in patients with the myelodysplastic syndrome: a study of the cancer
and leukemia group B. J Clin Oncol, 20, 2429–​40.
Steensma DP (2018). How I use molecular genetic tests to evaluate
patients who have or may have myelodysplastic syndromes. Blood,
132(16), 1657–63.
Steensma DP, et al. (2015). Clonal hematopoiesis of indeterminate
potential and its distinction from myelodysplastic syndromes.
Blood, 126, 9–​16.
Stone RM (2009). How I  treat patients with myelodysplastic syn-
dromes. Blood, 113, 6296–​303.
22.3.3  Acute myeloid leukaemia
Nigel Russell and Alan Burnett
ESSENTIALS
Acute myeloblastic leukaemia arises in a haematopoietic stem cell as
a result of mutations which promote growth or inhibit apoptosis in
association with mutations that inhibit differentiation. There is usu-
ally no obvious cause, but exposure to chemical and ionizing radi-
ation may be relevant, including previous chemotherapy for solid
tumours. This leukaemia arises particularly in older patients, often
in the context of antecedent haematological disorders such as
myelodysplastic syndromes or myeloproliferative neoplasms.
Clinical features—​these are of marrow failure, with anaemia,
bleeding (petechiae, purpura, from mucous membranes), and
infection. Acute promyelocytic leukaemia should be viewed
as a medical emergency characterized by disseminated intra-
vascular coagulation which requires urgent treatment to avoid
haemorrhagic death. Diagnosis relies on examination of periph-
eral blood and bone marrow for blast cell infiltration, with clas-
sification and prognosis of disease depending on morphology,
immunophenotyping, karyotyping, and definition of particular
molecular mutations.
General approach to treatment—​aside from providing appropriate
supportive care, the first clinical decision to be made is whether to give
conventional intensive chemotherapy aiming for disease eradica-
tion, or to adopt a more palliative approach. Intensive chemotherapy
is the norm up to age 65 to 70 years. Above this age, the biology
of the disease tends to be less favourable and patients may have
significant comorbidities limiting treatment tolerance. A  decision
about embarking on conventional intensive therapy in older patients
should be made only after a careful medical assessment.
Initial chemotherapy—​(1) intensive chemotherapy—​this typically in-
volves the combination of daunorubicin (or other anthracycline-​like
drug) and cytosine arabinoside (cytarabine, ara-​C), which achieves
complete remission in 50 to 80% of cases depending on age. This
is followed by consolidation chemotherapy (a second course of in-
duction treatment and then further ara-​C with or without additional
agents). (2)  Less-​intensive chemotherapy—​has most often com-
prised of hydroxycarbamide (hydroxyurea), low doses of ara-​C, or
a demethylating agent such as azacitidine. (3) Acute promyelocytic
leukaemia is exquisitely sensitive to all-​trans-​retinoic acid (ATRA) or
arsenic trioxide (ATO), each of which can be given concurrently with
chemotherapy. Recent data demonstrate that the combination of
ATRA and ATO alone is highly effective, at least in low-​risk cases.
In patients under 60 years, 75 to 80% will achieve initial remission
and about 45 to 50% will survive. In older patients given intensive
treatment, 50 to 60% will enter remission but only 15 to 20% will sur-
vive 2 years. With nonintensive treatments, remissions are seen in 15
to 25% with the median survival being 6 to 9 months.
Relapsed disease—​more than 50% of patients will ultimately relapse
and their overall outcome is generally very poor. The best curative
option is allogeneic stem cell transplantation if a second complete
remission can be achieved with reinduction chemotherapy.
Prospects for the future
Increasing knowledge of the underlying molecular characteristics of
acute myeloid leukaemia may permit more targeted therapy; how-
ever, most cases have more than one of the many mutations which
can occur and the resulting small patient groups will be a challenge
for therapy development in what is already a relatively rare disease.
Currently available treatments may be improved in some patients
using minimal/​measurable residual disease assessment to determine
risk on a more individualized basis.
Epidemiology and causation
The median age at presentation is 68 to 70 years. Acute myeloid
leukaemia (AML) occurs at all ages, ranging in frequency from
3 per million up to 12 to 15 per million patients in their 70s and
80s (Fig. 22.3.3.1). This age range has implications for treatment.
In most cases there is no obvious cause; however, it is known that
chemical and ionizing radiation exposure can be leukaemogenic.
Chemotherapy for other cancers may be leukaemogenic, and the
development of AML represents a late complication of treatment of
some solid tumours, which is being seen more frequently as chemo-
therapy becomes more successful. In many cases, AML in older pa-
tients evolves from the related disorder myelodysplastic syndrome
or other myeloproliferative disease, and indeed the boundary be-
tween the diseases is sometimes hard to define. The current inter-
national consensus defines AML as over 20% blast cells in the bone
marrow. At this level, marrow function is usually compromised re-
quiring intervention.
SECTION 22  Haematological disorders
5206
At the molecular level, it is believed that AML is an example of
multi-hit pathogenesis, with mutations which promote growth or
inhibit apoptosis arising in association with mutations that inhibit
differentiation. Dysregulation of gene expression by methylation
is also involved. Molecular abnormalities that have such effects are
being increasingly recognized.
Diagnosis
The heterogeneity of morphology which reflects the ability of the
leukaemic blast to achieve some degree of differentiation gave
rise to a commonly accepted morphological classification known
as the French–​American–​British (FAB) classification. With the
increasing availability of high-​quality monoclonal antibodies,
immunophenotypic confirmation gave some objectivity to the diag-
nostic process. Over 30 years ago it was being recognized that various
chromosome abnormalities were present in the blast cells. These
were nonrandom and comprised balanced translocations, deletions,
and trisomies. In about 40% of cases only, a normal karyotype could
be found. Some of these abnormalities corresponded to the morpho-
logical subtype. It took a number of years for the clinical community
to feel that this was useful knowledge. However, it is now recognized
that the karyotype is the strongest predictor of response to therapy,
and is now an essential part of the diagnostic process. Since some of
the cytogenetic breakpoints have been cloned, they have revealed
targets for therapy or sensitive disease monitoring.
Currently a similar awakening is underway with the recognition
that molecular abnormalities are often found. The most common
are mutations of the FMS receptor FLT3, NPM1, DNMT3A, IDH1,
IDH2, CEBPα, RAS, and c-​KIT. These represent in some cases add-
itional strong independent prognostic factors, but may become the
target of a new generation of molecular-​based therapies. There is
increasing interest in the role of noncoding DNA and epigenetic
abnormalities
Prognostic factors
Several characteristics independently predict how a patient with
AML will respond to treatment. Most apply to the prospect of
initial treatment achieving complete disease remission, and to
overall survival. Performance score is mainly useful for predict­
ing response to induction treatment. It should be mentioned
here that these factors have been defined in the setting of large
clinical trials, into which poor performance patients may not be
entered.
The key factors are shown in Box 22.3.3.1. Cytogenetics has
been most widely adopted, and can guide treatment decisions such
as who should be subjected to allogeneic transplant. Grouping
the more favourable abnormalities (t(8;21), inv(16), and t(15;17))
and the poorer group abnormalities (of chromosomes 5 or 7,
3q–​, complex) leaves about 60% of patients as at standard risk. As
can be seen in Fig. 22.3.3.2, this subdivision has a major impact on
survival. One of the reasons that older patients respond less well to
the same chemotherapy is that older patients tend to have a high
proportion of adverse features, in contrast to younger patients.
Some molecular abnormalities may provide prognostic informa-
tion particularly in normal karyotype AML, for example, the asso-
ciation with an NPM1 mutation in the absence of a FLT3 mutation
or biphenotypic mutation of CEBPα are generally regarded as
having a favourable prognosis with chemotherapy, and as such can
avoid stem cell transplant as part of initial treatment. Molecular
findings are continuously influencing prognostic estimates. This
is highly complex because of the coexistence of mutations which
may modulate each other’s impact compared with when they occur
alone. Furthermore, the allelic burden of a mutation can impact on
its prognostic importance.
Treatment of acute myeloid leukaemia
General considerations
Because of the variation of disease and patient biology, treatment
and the assessment of treatment is complex. Treatment outcomes
have improved over the years in children and adults under 60 years
0
25
75
50
100
20
15
10
5
0
age
rate per 100 000
Median age:
65–70 years
Fig. 22.3.3.1  Age-​specific incidence of acute myeloid leukaemia (AML).
Data from Wingo PA, et al. (1995). Cancer statistics, 1995. CA Cancer J Clin, 45, 8.
Box 22.3.3.1  Prognostic factors
•	 Age
•	 Cytogenetics
•	 Presenting WCC
•	 Secondary disease
•	 Performance score
•	 FLT3/​NPM1 mutation status
•	 Expression of resistance phenotype
•	 Marrow response to first chemotherapy course
100
75
50
25
0
0
1
2
3
4
5
years from randomization
68%
44%
18%
2P <0.00001
% still alive
Good
Standard
Poor
Fig. 22.3.3.2  Survival from complete remission (CR) by the Medical
Research Council (MRC) risk group.
22.3.3  Acute myeloid leukaemia
5207
(Fig. 22.3.3.3), but progress in older patients is much less clear.
Chemotherapy has been relatively unchanged over the years in
terms of drugs used, and it is easy to attribute the better outcomes to
improved supportive care, which in turn has allowed treatment to be
given in a more intensive way.
Definition of remission
The standard definition of ‘remission’ is that the bone marrow
should show evidence of trilineage activity with less than 5%
blasts, and peripheral blood counts should have returned to at
least 100 × 109/​litre for platelets and 1.0 × 109/​litre for neutrophils.
Molecular and other markers of residual disease may still indicate
the presence of the leukaemic clone, and it is well established that
failure to deliver further courses of treatment to consolidate the
response will result in rapid regrowth of disease. It is increasingly
likely that the application on minimal residual disease detection
by flow cytometry or polymerase chain reaction (PCR) may refine
these definitions as is the case in other haematological malignan-
cies. The first clinical decision to be made in an individual patient
is whether to undertake conventional intensive chemotherapy
aiming for disease eradication, or to adopt a more palliative ap-
proach. The aim of intensive chemotherapy is to kill off the leu-
kaemic population, which then enables normal haematopoiesis to
re-​establish itself. This will inevitably mean several days of pan-
cytopenia, a high risk of gut toxicity, and extreme vulnerability to
infections of microbiological, fungal, or viral causation.
Chemotherapeutic regimens
The combination of the anthracycline daunorubicin and the nucleo-
side analogue cytosine arabinoside (cytarabine, ara-​C) has been the
mainstay of treatment of AML for nearly 40 years with the intention
of inducing complete remission (CR) and curing the disease. Such
a combination is extremely myelosuppressive and is associated with
a significant risk of infection and severe systemic toxicity. Among
younger patients, there is a risk of death during induction therapy of
almost 10%, usually due to infection or bleeding. Such toxicity may
limit the applicability of such a regimen to the treatment of older
adults who constitute the majority of patients with AML. Patients
who achieve CR will continue with further courses of chemotherapy,
at approximately monthly intervals, to consolidate the remission
and reduce the risk of relapse.
In the United Kingdom, initial induction therapy for younger
people (<60 years of age) comprises 3 days of daunorubicin at doses
of 60 mg/​m2 (usually given on days 1, 3, and 5) with 10 days of
ara-​C at 200 mg/​m2 daily. In the United States of America, this is
usually shortened to a 7-​day continuous infusion of ara-​C, the so-​
called 3+7 approach. Recent studies have shown a dose effect of
daunorubicin, with 90 mg/​m2 being superior to 45 mg/​m2. However
a large recent study showed that 60 mg/​m2 was equivalent to 90 mg/​
m2. Similarly higher doses of ara-​C have been tested in induction
without convincing evidence of improving survival. There are other
apparent differences in treatment preferences between the United
States of America and the United Kingdom. Comparison of large
national trials has not shown convincing evidence that the addition
of a third drug, usually etoposide or thioguanine, improves the out-
come of induction treatment in younger adults since the additional
leukaemia cell kill is potentially offset by increased toxicity. Equally,
attempts to improve outcomes with alternative anthracyclines
(idarubicin or mitoxantrone) have shown no advantage. The United
Kingdom Medical Research Council (MRC) AML12 trial compared
MRC AML Trials: overall survival
age 0–14
MRC AML Trials: overall survival
age 15–59
100
75
73%
63%
63%
54%
46%
24%
3%
50
% still alive
25
0
100
75
53%
43%
35%
0
5
10
15
20
25
years from entry
29%
23%
15%
5%
50
% still alive
25
0
0
5
10
years from entry
(a)
(c)
(b)
15
20
25
MRC AML Trials: overall survival
age 60+
100
75
23%
10%
7%
4%
2%
1%
50
% still alive
25
0
0
5
10
years from entry
15
20
25
2%
Fig. 22.3.3.3  Medical Research Council (MRC) trial results by patient ages. (a) 0 to 14 years (n = 1096); (b) younger adults, 15 to
59 years (n = 7704); (c) older patients, 60+ years (n = 3541).
SECTION 22  Haematological disorders
5208
mitoxantrone with daunorubicin, in combination with either
etoposide or thioguanine, and found no overall difference in long-​
term outcome.
In recent years, it has become feasible to conjugate chemothera­
peutics with antibodies as a means of targeting the antileukaemic
potential without increasing toxicity. The archetypal agent in AML
is gemtuzumab ozogamicin, which is a CD33-​targeted immuno­
conjugate of the anthracycline-​like drug, calicheamicin. CD33 is a
transmembrane protein expressed on the cells of 95% of patients
with AML. On binding antibody, it is rapidly internalized, and free
calicheamicin is released causing genotoxic damage. A recent meta-​
analysis of five large trials in over 3000 patients where gemtuzumab
ozogamicin was combined with induction chemotherapy reported
that this provides a significant survival benefit for patients without
adverse cytogenetic characteristics.
Cytarabine remains one of the most active drugs in the treatment
of AML. Several groups have increased the dose up to 3 g/​m2 in in-
duction with mixed results. At these high doses there is significant
skin, renal, and central nervous system toxicity. However, certain
subsets of AML, particularly the core binding factor leukaemia as-
sociated with t(8;21) and inv(16), are particularly sensitive to ara-​C
and may benefit from this approach. Newer analogues of cytarabine
such as cladribine are now entering randomized trials. The combin-
ation of fludarabine/​ara-​C/​granulocyte-​colony stimulating factor
(G-​CSF) and idarubicin (FLAG-​Ida) is a very effective induction
combination capable of producing high remissions with one in-
duction course, but results in profound myelosuppression, and may
limit the ability to deliver consolidation treatment.
Outcomes of treatment
Between 40 and 80% of patients, depending on age, will achieve CR
with this approach and over 70% will do so after a single course of
induction. Factors influencing the chance of CR include age at diag-
nosis, cytogenetic abnormalities, and expression of multidrug resist-
ance genes (e.g. p-​glycoprotein). For example, 75 to 80% of patients
aged less than 60 years will achieve a CR, compared with 45 to 55%
of older patients given the same treatment schedule.
Consolidation of chemotherapy
Patients who achieve CR will require further consolidation chemo-
therapy. This may take the form of a second course of the initial
induction treatment, followed by two or three further courses of al-
ternative drugs. High doses of ara-​C (1.5 g/​m2 or 3.0 g/​m2) are com-
monly used particularly in patients with favourable and standard
risk disease. There are probably better treatments for consolidation
of adverse risk patients such as FLAG-​Ida, but such patients should
proceed to an allogeneic stem cell transplant. It is not clear how
many courses of treatment is optimal pretransplant or in total if a
transplant is not feasible.
Treatment of older patients
Age is a major factor determining CR rates and long-​term outcome.
Older age is associated with the onset of significant comorbidities
such as hypertension, lung disease, or renal or cardiac impairment,
which limit chemotherapy delivery. Similarly, the biology of leu-
kaemia is more adverse in older people: there is a higher incidence
of unfavourable genetic changes and multidrug resistance, and AML
in this age group more frequently results from secondary disease.
Consequently, older people tend to do less well than younger pa-
tients with AML given the same treatment.
The age of 60 years is frequently taken as an arbitrary threshold
for the use of the term ‘older’, but this is arbitrary. Up to this age
there is little doubt that intensive chemotherapy should be the
norm; however, with increasing age, patients develop comorbidities
and become less generally fit. This makes intensive treatment more
risky particularly in patients older than 70 years. In addition, the
biology of the disease is less favourable, with a higher proportion
of patients with secondary AML, more with leukaemia that has
associated resistance proteins within the leukaemic population,
and a higher proportion with an adverse risk karyotype. A number
of epidemiological studies indicate that at least 40% of older pa-
tients are not treated with intensive chemotherapy and receive
supportive care only with hydroxycarbamide (hydroxyurea) or
some other low-​intensity treatment. There has been much recent
interest in the development of treatments for this patient group.
A major national trial in the United Kingdom was able to demon-
strate that low doses of ara-​C given subcutaneously (20 mg twice
daily for 10 days) repeated at 4-​ to 6-​week intervals is superior to
hydroxycarbamide. Indeed, one in six patients will gain CR with
this approach, although these remissions tend to be less durable
than those seen in younger patients treated with conventional
doses of chemotherapy.
Demethylation treatments, azacitidine or decitabine, are more
widely used and are approved in Europe for this patient group.
These agents do not improve the rate of remission compared with
low-​dose ara-​C, but seem capable of maintaining those who do not
enter remission in a stable haematological condition. Such ran-
domized data as are available do not show a significant difference
in survival between these treatment options. Azacitidine also has
activity in AML, particularly in patients with unfavourable cytogen-
etics or myelodysplasia-​related changes although randomized trials
did not result in a significantly better survival compared with low-​
dose cytarabine.
What these studies show is that overall the outlook for older pa-
tients with AML remains very poor but that the majority of older
patients should be considered for specific chemotherapy as well as
optimal supportive care with prophylactic antibiotics, antifungals,
and blood product support.
Treatment of relapsed AML
Despite 60 to 80% of patients with AML attaining CR with induction
chemotherapy, more than 50% of these will ultimately experience a
relapse of their disease. Reinduction with intensive chemotherapy
remains an option for these patients, but the overall outcome is gen-
erally very poor. Remission rates with reinduction are lower than in
first presentation, and remissions are generally of shorter duration.
The factors which predict outcome, irrespective of what treatment
is used, are patient age, the risk group based on cytogenetics and
other factors at the original presentation, the duration of first remis-
sion, and whether or not a stem cell transplant has already been per-
formed. The only curative option is to proceed to an allogeneic bone
marrow transplantation when in second complete remission (CR2).
No randomized trial has shown a superiority of one reinduction
regimen over another, but the combination of fludarabine (a purine
analogue) with ara-​C and idarubicin with G-​CSF support is one
favoured regimen.
22.3.3  Acute myeloid leukaemia
5209
Targeted therapy
In recent years the molecular heterogeneity of AML has become
better understood and has opened an era of targeted treatments,
some of which have received regulatory approval. None is a ‘cure-
all’ for the particular subset targeted. Gentuzumab ozogamicin
has been mentioned and because its addition to induction chemo-
therapy has reduced the risk of relapse it can improve the survival of
intermediate and favourable, but not adverse risk groups. The first
molecularly targeted treatment has been midostaurin against FLT3
mutated disease which occurs in 30% of younger patients and about
15% of older patients. The survival benefit is significant but still less
than 10% at 5 years when added as part of first line treatment. More
recently a second FLT3 inhibitor (gilteritinib) has been approved
for relapsed disease, and there are other FLT3 ­inhibitors in devel-
opment. More recently mutations of the iso-citrate dehydrogenase
(IDH) genes have been found in 10–15% of cases. There are now
IDH inhibitors of IDH1 (ivosidenib) and IDH2 (enasidenib) which
have been approved for relapsed disease. It has been recognized that
one reason why AML treatment is less successful in older patients is
that the cells more frequently express anti-apoptotic proteins such
as BCL2. Previous efforts to target BCL2 have not been successful,
however venetoclax has caused great interest in other haemato-
logical malignancies and preliminary studies in older patients with
AML in combination with low dose cytaribine or azacitidine or
dectibine have led to its regulatory approval.
A development, which was surprising to some, was that by com-
bining daunorubicin and cytarabine in a fixed 1:5 ratio in a lipo-
some, showed superior survival when compared with daunorubicin
and cytarabine alone. However the benefit was restricted to patients
with secondary AML or adverse cytogenetics. This drug, Vyxeos,
provides a new option for poor risk patients.
These new agents which have become available in the last two
years, provide many opportunities to further refine treatment.
Maintenance chemotherapy and DNA methylation
Traditionally, maintenance therapy using lower doses of chemo-
therapy has had no or only marginal benefit, and has dropped out
of use. There has been interest over the years in immunomodulation
using interleukin-​2 (IL-​2). This agent did not show benefit in ran-
domized trials and was associated with toxicity. However, efforts
were made to find treatments which could be combined with IL-​
2 which could retain the efficacy, but with lower doses of IL-​2 and
thereby less toxicity. Studies with the combination of histamine
dichloride and low-​dose IL-​2 validated this approach, but confirma-
tory studies are lacking. There is emerging interest as to whether
maintenance therapy with a demethylation agent could be effective,
arguably through an immunomodulatory mechanism.
Bone marrow transplantation
Having entered remission, the main challenge is to prevent re-
lapse. There is no doubt that the most effective approach is to ad-
minister myeloablative chemoradiotherapy followed by infusion
of donor haematopoietic stem cells. These have been obtained from
an HLA-​matched sibling donor as bone marrow or mobilized stem
cells collected from the peripheral blood. This approach reduces the
risk of disease relapse from 40 to 50%, to 10 to 15%. This powerful
antileukaemic effect is partly due to the myeloablation and partly due
to the associated ‘graft versus leukaemia’ mediated by donor T lympho-
cytes in the graft. Unfortunately, it has not been possible to fully realize
the antileukaemic potential of allogeneic transplantation due to the
associated risks of graft-​versus-​host disease and immunosuppression
which usually involve a life-​threatening risk of up to 30%. The risks
increase with advancing age and the presence of comorbidities particu-
larly when intensive myeloablative conditioning is used. The develop-
ment of large donor banks has made the finding of a matched unrelated
donor a practical possibility for the majority of patients, and the results
of this approach are now equivalent to those using a matched sibling.
Allografting involving reduced-​intensity conditioning (RIC)
has become well developed. Here the mechanism is primarily im-
munological, mediated by a graft-​versus-​leukaemia effect. Data now
emerging from the National Cancer Research Institute AML 15 and
16 trials suggest that this is also a feasible approach which can im-
prove survival compared to chemotherapy in older patients up to the
age of 70 years who have few comorbidities.
As discussed previously, once patients enter remission, they are
at differing risks of relapse based on their cytogenetic group. It is
unlikely that patients with favourable cytogenetics will benefit from
allogeneic transplant, because the additional reduction in relapse
risk is more than outweighed by the treatment risks. For patients
at intermediate risk, there are mixed opinions, but no convincing
evidence of overall survival benefit in United Kingdom prospective
trials. Most would accept that patients with bad risk disease should
undergo a transplant as soon as the risk is known. Such patients
will have a higher risk of relapse after the transplant, but they have a
very poor outcome with chemotherapy alone. The debate about who
should receive a transplant, and of which type if the option is avail-
able, has been going on for many years. The debate centres on the
60% of younger patients who have intermediate risk. It is generally
accepted that any benefit from a myeloablative transplant is limited
to patients less than 40 years of age. As well as cytogenetics and
the other risk factors mentioned previously, the various molecular
abnormalities can now be taken into account. For example, many
investigators use the presence of a FLT3 mutation as an indication
for transplant because of the high relapse risk. However, the results
from transplantation in these patients is less good than expected,
and furthermore the prognosis of a FLT3 mutated patient can be
affected by the alleleic ratio of the mutation and whether or not its
adverse effect is neutralized by the coexistence of a nucleophosmin
1c (NPM1c) mutation. Studies from the United Kingdom currently
suggest that for patients aged less than 40 years a matched (related
or unrelated) myeloablative transplant is beneficial in patients with
adverse risk and those with intermediate risk who have a FLT3
mutation-​positive/​NPM1c mutation-​negative genotype. For older
patients, a RIC allograft is beneficial for intermediate risk patients
who have a matched sibling donor. There is a lack of evidence for
benefit for a RIC in intermediate risk with an unrelated donor, or
patients with adverse risk. These observations are subject to change
as novel preparative schedules for RIC transplants are developed.
For patients who relapse following chemotherapy and enter
a second remission, the prospects for cure are poor and are
largely dictated by the length of their first remission. It is axio-
matic that second remissions will be shorter than first remissions.
Transplantation is the only treatment that can change this and is
indicated for all patients who relapse, where it offers a 30 to 40%
chance of salvage.
SECTION 22  Haematological disorders
5210
Supportive care in AML
The steady but significant improvements seen in disease survival
rates over the last 30 to 40 years have been facilitated by the de-
velopment of better supportive care strategies which have allowed
the safer intensification of chemotherapy regimens. There is clear
evidence that the 30-​ and 60-​day mortalities have dropped over
recent years.
Following diagnosis of AML, early mortality can result either dir-
ectly from presenting complications of the disease, from the direct
consequences of treatment initiation, or from problems arising
during the 3 to 4 weeks of profound pancytopenia that inevitably
follow remission-​induction chemotherapy. Effective supportive care
during this period requires close coordination between special-
ists from a number of disciplines including haemato-​oncologists,
microbiologists, radiologists, intensivists, specialist nurses, phar-
macists, and dieticians working in facilities dedicated to the care
of this type of patient. Clear written standards should be adhered
to, including local policies for infection prophylaxis and treatment,
national guideline documents, and, where appropriate, clinical trial
protocols.
Supportive care at the initiation of therapy
The presenting clinical features of AML vary according to both the
depth of bone marrow failure and the rate of turnover of the leu-
kaemic clone. Prompt chemotherapeutic intervention is required in
cases with high rates of blast proliferation but, paradoxically, rapid
cell kill may lead to life-​threatening metabolic disturbances.
Hyperleucocytosis
Although a feature of only a minority of cases, a high presenting
white cell count (WCC) is a well-​established poor prognostic factor
in AML. Patients with hyperleucocytosis (WCC >100 × 109/​litre)
are at a threefold greater risk of early mortality (15%) than those
with lower counts. Hyperleucocytosis predisposes to hyperviscosity
and leucostasis. ‘Sludging’ in the microvasculature, particularly of
the lungs and brain, clinically manifests most frequently as hyp-
oxia and central nervous system dysfunction and carries significant
risks of both thrombotic and haemorrhagic sequelae. The effects of
hyperviscosity may be partially offset at presentation by the pres-
ence of concurrent anaemia:  red cell transfusion should thus be
delayed, unless absolutely unavoidable, until the WCC has been re-
duced. Oral hydroxycarbamide may be of practical value in reducing
the WCC prior to the commencement of formal induction chemo-
therapy. Leucapheresis is generally safe and, although evidence is
lacking, may be considered in patients presenting with symptomatic
hyperleucocytosis. However, leucapheresis may fatally exacerbate
the presenting coagulopathy of APL and should be avoided in this
setting.
Tumour lysis syndrome and metabolic complications
Acute tumour lysis syndrome describes a collection of metabolic
abnormalities including hyperuricaemia, hyperphosphataemia,
hypocalcaemia, and hyperkalaemia that result from the release of
nuclear and cytoplasmic degradation products from malignant cells
and may precipitate acute kidney injury. It is vital that treating phys-
icians are aware of the risks of acute tumour lysis syndrome, par-
ticularly when instituting cytoreductive therapy in AML patients
with hyperleucocytosis or bulky extramedullary disease. Emergency
haemodialysis may be required in the event of acute kidney injury,
rising potassium levels, or recalcitrant hyperphosphataemia.
Standard measures to prevent acute tumour lysis syndrome prior
to the commencement of chemotherapy include use of the xanthine
oxidase inhibitor allopurinol (300 mg daily) coupled with vigorous
intravenous hydration and with meticulous monitoring of fluid
balance and electrolyte levels as induction therapy commences.
Alkalinization of the urine using intravenous bicarbonate has been
used historically to reduce tubular uric acid crystal deposition, but
it remains controversial as it carries the potential for both reducing
tubular xanthine solubility and exacerbating calcium pyrophosphate
deposition in organs including the heart. The recombinant urate oxi-
dase enzyme rasburicase is able to rapidly reverse hyperuricaemia
by promoting the breakdown of uric acid into allantoin. It is now
the treatment of choice in patients with hyperleucocytosis at pres-
entation, renal failure, or early evidence of evolving acute tumour
lysis syndrome. Rasburicase also avoids any need for urinary
alkalinization.
Hypokalaemia is also frequently encountered in AML patients,
both at presentation (due to high serum lysozyme levels particu-
larly in monocytic subtypes M4 and M5) or later as a consequence
of prolonged diarrhoea or the renal tubular effects of amphotericin.
Vigorous intravenous electrolyte supplementation is frequently
required.
Other supportive measures prior to starting cytotoxic therapy
Secure central venous access is usually established through inser-
tion of a tunnelled Hickman line or temporary central line, allowing
safe administration of vesicant drugs, blood products, and intra-
venous antibiotics, as well as facilitating frequent blood-​sampling
procedures. Young men should be counselled regarding potential
loss of fertility and, whenever possible, offered the opportunity to
store sperm. Loss of fertility due to chemotherapy is less common
in women: in vitro preservation of unfertilized ova is not yet under-
taken routinely. There is a high risk of severe emesis with intensive
chemotherapy, and strenuous efforts should be made to prevent
this distressing complication. Serotonin antagonists (ondansetron
or granisetron) are a standard first choice, although combination
therapy is often necessary.
Supportive care during chemotherapy-​induced pancytopenia
Clearance of leukaemic blasts by induction chemotherapy is
achieved at the expense of 3 to 4 weeks of severe pancytopenia, and
similar cytopenic episodes will follow subsequent courses of con-
solidation therapy. During these periods, patients remain at high
risk: prompt access to blood product support and robust procedures
to prevent and manage neutropenic infections are vital.
Blood product support
By the time of initial disease presentation, the ability of most patients
to produce red cells and platelets is severely impaired. Due to the
often rapid onset of anaemia, there may be little time for haemo-
dynamic compensation making many patients symptomatic due
to acute impairment of oxygen-​carrying capacity. In the absence of
hyperleucocytosis, red cells should be transfused promptly.
Following intensive chemotherapy, patients will inevitably be de-
pendent on regular transfusion support until bone marrow recovery.
22.3.3  Acute myeloid leukaemia
5211
Although there is no firm evidence to support a particular red cell
transfusion threshold, many units operate a policy of transfusing
as required to maintain haemoglobin levels in excess of 80 g/​litre.
Patients treated with cytotoxic regimens containing purine ana-
logues (fludarabine or clofarabine) should receive irradiated blood
products to minimize the risk of transfusion-​associated graft-​
versus-​host disease.
In general, one adult therapeutic dose of platelets should be
transfused whenever the platelet count falls to below 10 × 109/​litre.
Platelet survival may be further compromised by sepsis or the use of
concurrent intravenous antibiotics, and in these situations or in the
presence of additional haemostatic abnormalities a higher transfu-
sion threshold of 20 × 109/​litre is usually observed. Antifibrinolytic
agents such as tranexamic acid may be useful for local mucosal
bleeding but are contraindicated in the presence of haematuria due
to the potential for ureteric clot formation.
Infection
The risk of infection in AML is influenced by both the degree and
duration of neutropenia and increases markedly during episodes of
chemotherapy-​induced bone marrow aplasia. Changes to the bac-
terial flora as a consequence of broad-​spectrum antibiotic use, and
poor nutritional status following prolonged periods of hospitaliza-
tion also contribute significantly. The vast majority of AML patients
will become febrile at some point, although only a minority of these
episodes will be accompanied by symptoms or signs of localizing
infection. Sepsis should be suspected in the presence of any sudden
nonspecific clinical deterioration; inflammatory responses may
be muted in the neutropenic setting and may be associated with
hypothermia, declining mental status, myalgia, or increasing leth-
argy. Potential portals of bacterial entry include indwelling lines
and chemotherapy-​induced breaches in the integrity of the bowel
mucosa.
Neutropenic patients should be advised to pay particular atten-
tion to personal hygiene and dental care. Careful hand-​washing and
decontamination before patient contact is mandatory for healthcare
workers. The role of prophylactic antibiotic therapy remains con-
tentious. Data from studies of adults with leukaemia, and stem cell
transplantation, suggest that prophylaxis with fluoroquinolone anti-
biotics reduces the risk of neutropenic sepsis and this approach is
recommended in National Institute for Health and Care Excellence
guidance but with the corollary that rates of antibiotic resistance
and infection patterns within individual institutions should be
monitored.
Patients should be made aware of their susceptibility to infection
and provided with emergency contact details to allow rapid clin-
ical assessment. In the presence of neutropenic sepsis, the prompt
institution of broad-​spectrum antibacterial therapy is potentially
life-​saving. Patterns of infection and pathogen isolation will vary
between hospitals, and clear written guidelines for the emergency
management of patients with febrile neutropenia should be decided
in discussion with local microbiologists. Examples of empirical
antibiotic regimens include monotherapy with a third-​generation
cephalosporin or carbapenem, or combination therapy with a
broad-​spectrum antipseudomonal penicillin and aminoglycoside.
Vancomycin or teicoplanin may be added to broaden Gram-​
positive coverage if there are particular clinical concerns regarding
indwelling line infection. Mandatory investigations include central
and peripheral blood cultures, cultures of urine and stool, and a chest
radiograph. Further modifications to the initial antibiotic regimen
should be based on culture results and regular clinical examination,
although surveys demonstrate that the rate of proven bacteraemia
during episodes of febrile neutropenia has remained between 20 and
25% for many years. Persistent infection or blood culture isolation
of Gram-​negative organisms or candida should prompt indwelling
central line removal.
The risk of invasive fungal infection is high in AML patients re-
ceiving intensive chemotherapy, and its incidence increases with
the severity and duration of neutropenia, often occurring in the
aftermath of bacterial sepsis. Established fungal infections carry a
high mortality. The diagnosis of invasive fungal infection should be
confirmed wherever possible, and there is an increasing move away
from the empirical use of antifungal agents as treatment of fever of
unknown origin. High-​resolution CT scanning of the chest in pa-
tients with persistent pyrexia refractory to antibiotic therapy, and
screening of patients using the sandwich enzyme-​linked immuno-
sorbent assay (ELISA) for Aspergillus galactomannan aid the early
detection of invasive pulmonary aspergillosis, allowing the tar-
geted implementation of antifungal therapy with agents including
liposomal amphotericin B and caspofungin. Azole antifungal agents
are widely prescribed prophylactically during neutropenia in AML.
Posaconazole is the drug of choice in this setting and has activity
against a broader spectrum of fungal organisms than fluconazole
which is inactive against moulds including aspergillus species.
A modest reduction in duration (but not depth) of neutropenia
may be achieved with the use of recombinant growth factors (G-​
CSF or granulocyte–​macrophage colony-​stimulating factor) fol-
lowing induction and consolidation chemotherapy. Large controlled
trials show variable effects on the incidence of severe infection and
no clear overall survival benefit. Routine growth factor use is not
recommended, although there may be cost–​benefit advantages in
terms of reduction in both antibiotic usage and the duration of hos-
pital admissions.
Acute promyelocytic leukaemia
APL is a medical emergency characterized by bleeding and dis-
seminated intravascular coagulation. It usually presents with pan-
cytopenia rather than a raised WCC. The diagnosis is made on
morphology with a typical hypercellular marrow with characteristic
heavily granulated promyelocytes. Immunophenotyping by flow
cytometry may provide supportive data, and PML body staining
shows a characteristic perinuclear speckled pattern on staining for
PML. The diagnosis is confirmed by cytogenetic or molecular studies
for the PML-​RARα rearrangement of the t(15;17) translocation.
Treatment of APL
This disease is exquisitely sensitive to treatment with all-​trans-​
retinoic acid (ATRA), a vitamin A analogue normally present in
blood. The translocation renders cells resistant to physiological con-
centrations of ATRA, but sensitive to pharmacological doses. APL
is particularly sensitive to anthracyclines so ATRA is given con-
currently with idarubicin. Whether other anthracyclines could be
equally effective is not clear. Whether other chemotherapy drugs
are need is a matter of debate, in particular the need for any ara-​C,
SECTION 22  Haematological disorders
5212
and the requirement for maintenance is now in doubt. The disease
is also highly sensitive to arsenic trioxide. High rates of CR and an
excellent long-​term prognosis with 5-​year disease-​free survival rate
of greater than 80% are now expected. However, treatment is associ-
ated with an increased risk of bleeding in these patients, particularly
in the first 3 or 4 weeks of treatment, and intense support of the
coagulation system is required with blood products. Once a patient
reaches 1 month into treatment, the prognosis is greater than 90%
survival, and the risk of relapse is lower with APL than with other
types of AML.
Due to the specific nature of the molecular lesion in APL, moni-
toring of minimal residual disease by PCR allows very sensitive de-
tection of any impending relapse. Restarting treatment at the time
of molecular detection of disease recurrence improves overall out-
come. For relapsed APL the treatment of choice is arsenic trioxide
which induces degradation of PML-​RARα.
In recent years, the aim of most clinical trials has been to de-​
escalate treatment and the most recent randomized data impres-
sively show that the ‘chemo-​free’ combination of ATRA and arsenic
trioxide is highly effective, giving an overall survival of higher than
90%. In these studies, the risk of disease relapse is very low once in
molecular remission, which raises the question of the need for rou-
tine minimal residual disease monitoring in the context of that treat-
ment, although it probably has benefit if the ATRA/​anthracycline
chemotherapy is used.
Supportive care issues specific to treatment
initiation in APL
Although abnormalities of coagulation may contribute to a bleeding
tendency at presentation in any subtype of AML, APL, and more es-
pecially its hypogranular variant (M3v), are particularly associated
with a high risk of early haemorrhagic death due to a combination
of disseminated intravascular coagulation and increased fibrin-
olysis. Eighty per cent of APL patients have clinically significant
coagulopathy: this condition constitutes a genuine haematological
emergency that requires rapid diagnosis and prompt initiation of
therapy. There is now considerable evidence that the early introduc-
tion of ‘differentiation therapy’ with ATRA alongside anthracycline-​
based chemotherapy improves the coagulopathy associated with
APL. The platelet count and coagulation profile should be checked at
least twice daily during the early stages of treatment. By using an ag-
gressive transfusion policy, the platelet count should be maintained
above 50 × 109/​litre, and coagulation times kept within the normal
range using fresh frozen plasma replacement. Cryoprecipitate or fi-
brinogen concentrates should be used to maintain a fibrinogen level
close to 2 g/​litre. The use of heparin is no longer recommended.
Complications of treatment in APL
Retinoic acid syndrome (also known as differentiation syndrome)
is a potentially life-​threatening complication of the use of ATRA or
arsenic trioxide therapy in APL. It is caused by cytokine release from
differentiating APL cells and characterized by a rising WCC with ac-
companying features of fluid retention and capillary leak including
pulmonary infiltrates, pleural and pericardial effusions, peripheral
oedema, hypoxia, and progressive respiratory failure. Standard treat-
ment on first suspicion of retinoic acid syndrome is dexamethasone
10 mg intravenously twice daily, with interruption of ATRA therapy
and provision of respiratory support until all symptoms and signs
have resolved.
Minimal/​measurable residual disease measurement
Techniques for sensitive measurement of residual disease beyond
the sensitivity of the microscope or cytogenetics are developing in
haematological cancer and in some circumstances are approved as
a surrogate endpoint for testing new treatments. There are several
molecular targets for which semi-​quantitative reverse transcription
PCR methods have been developed. However, they are only suitable
for 50 to 60% of patients.
Aberrant phenotypes can also be defined at diagnosis by
immunophenotyping using several antibodies. This is demanding
technology requiring a high level of standardization, but it is ap-
plicable to most patients depending how many antibodies are in
the panel. There is little doubt that tests on marrow which is clearly
in morphological remission will detect residual disease at a level of
1 cell per 1000. This is usually a reliable predictor of imminent re-
lapse thus enabling pre-​emptive treatment to be given. While this
approach is attractive for individualizing patient choices, at present
it may not be ready for routine practice. It is prognostic, probably
adding additional information to what is already known to enable
refinement of current treatments, but the predictive value will re-
quire prospective confirmation.
FURTHER READING
Appelbaum FR, et al. (2006). Age and acute myeloid leukemia. Blood,
107, 3481–​5.
Burnett A, Wetzler M, Löwenberg B (2011). Therapeutic advances in
acute myeloid leukemia. J Clin Oncol, 29, 487–​94.
Burnett AK (2012). Treatment of acute myeloid leukemia:  are we
making progress? Hematology Am Soc Hematol Educ Program,
2012, 1–​6.
Cancer Genome Atlas Research Network (2013). Genomic and
epigenomic landscapes of adult de novo acute myeloid leukemia. N
Engl J Med, 368, 2059–​74.
Dohner H, et al. (2010). Diagnosis and management of acute myeloid
leukemia in adults: recommendations from an international expert
panel, on behalf of the European LeukemiaNet. Blood, 115, 453–​74.
Estey E, Dohner H (2006). Acute myeloid leukaemia. Lancet, 368,
1894–​1907.
Grimwade D, Freeman SD (2014). Defining minimal residual disease
in acute myeloid leukemia: which platforms are ready for ‘Prime
Time’? Blood, 124, 3345–​55.
Grimwade D, et al. (1998). The importance of diagnostic cytogenetics
on outcome in AML: Analysis of 1,612 patients entered into the
MRC AML:10 Trial. Blood, 92, 2322–​3.
Grunwald MR, Levis MJ (2013). FLT3 inhibitors for acute myeloid leu-
kemia: a review of their efficacy and mechanisms of resistance. Int J
Hematol, 97, 683–​94.
Milligan DW, et al. (2006). Guidelines on the management of acute
myeloid leukaemia in adults. Br J Haematol, 135, 450–​74.
Sanz MA, et  al. (2009). Management of acute promyelocytic leu-
kemia:  recommendations from an expert panel on behalf of the
European LeukemiaNet. Blood, 113, 1875–​91.

22.3.4 Chronic myeloid leukaemia 5213 Mhairi Copla
22.3.4 Chronic myeloid leukaemia 5213 Mhairi Copland and Tessa L. Holyoake†
22.3.4  Chronic myeloid leukaemia
5213
22.3.4  Chronic myeloid leukaemia
Mhairi Copland and Tessa L. Holyoake†
ESSENTIALS
Chronic myeloid leukaemia (CML) has a worldwide incidence of 1
to 2 per 100 000 of the population. Most cases are caused by trans-
location of the distal end of chromosome 9 on to chromosome
22 (known as a Philadelphia (Ph) chromosome), which leads to
the creation of a fusion protein expressed from the fusion gene
formed by juxtaposition of parts of the BCR (breakpoint cluster
region) and ABL1 (Abelson 1) genes. The resulting oncoprotein is
a constitutively active tyrosine kinase and appears to operate as
an initiator for the development of the leukaemia. It is not known
why this precise translocation occurs recurrently.
Clinical features, diagnosis, and (historical) prognosis
Clinical features—​many patients are asymptomatic at diagnosis, which
is made following a routine blood test. Others present with signs and
symptoms including fatigue, sweats, fever, weight loss, haemorrhagic
manifestations, and abdominal discomfort (due to splenomegaly).
Diagnosis—​this is typically made by the examination of a periph-
eral blood film (revealing features including increased numbers of
neutrophils, eosinophils, basophils, immature myeloid cells, and, in
some cases, platelets) and the demonstration of the Ph chromosome
by conventional cytogenetics in a bone marrow aspirate or periph-
eral blood sample. Polymerase chain reaction analysis of peripheral
blood confirms the presence of a BCR-​ABL1 transcript and charac-
terizes the BCR-​ABL1 junction.
Prognosis—​before the introduction of tyrosine kinase inhibitors
(TKIs) the condition, having usually been diagnosed in the chronic
phase, then spontaneously progressed after (typically) 3 to 6 years to
myeloid (or less commonly lymphoid) blast transformation, which
had a very poor prognosis. This has changed significantly since the
introduction of TKIs.
Treatment
The original TKI, imatinib, has had a very significant impact on
the first-​line management of patients with CML. It induces durable
complete cytogenetic responses in the majority of patients and
­prolongs overall survival substantially. Although the incidence is
­unchanged, the improvement in survival resulting from TKI treatment
has led to an ever increasing prevalence of this form of leukaemia.
Imatinib, however, does not totally eradicate the leukaemia in most
cases and therapy is usually lifelong. Second-​ and third-​generation
TKIs, including dasatinib, nilotinib, bosutinib, and ponatinib,
show enhanced potency against BCR-​ABL1 activity and are
­licensed within Europe for first-​line (dasatinib, nilotinib, bosutinib)
or second-​line or subsequent (dasatinib, nilotinib, bosutinib,
ponatinib) use in CML. Patients with suboptimal responses to first-​
line treatment can be offered (1) a different second-​line TKI; or
(2) a third-​line TKI, such as ponatinib; or (3) allogeneic stem cell
transplantation—​for patients less than 65 years of age and with a
suitable donor. A second indication to switch TKI is drug intoler-
ance and each agent is associated with a range of similar and
nonoverlapping toxicities, although cardiovascular toxicity appears
to be a particular concern for nilotinib and ponatinib, pleural effu-
sion and pulmonary arterial hypertension for dasatinib, and diar-
rhoea for bosutinib.
Introduction
Patients with chronic myeloid leukaemia (CML) have been well
served by translational research over the past half century. Though
the disease was first described in 1845 and characterized by the
1920s, it was a further 60 years before the unravelling of initiating
molecular events paved the way to define specific targets for treat-
ment. CML is a clonal disease that results from an acquired mo-
lecular change in a pluripotent haematopoietic stem cell. The
leukaemia cells have a consistent cytogenetic abnormality, the
Philadelphia (Ph) chromosome, which carries a BCR-​ABL1 fusion
gene (Fig. 22.3.4.1). This gene encodes a BCR-​ABL1 oncoprotein
with enhanced tyrosine kinase activity, which is generally con-
sidered to be the ‘initiating event’ in the chronic phase of CML,
though there remains some debate as to whether this is indeed the
first molecular event in all cases.
TKIs can inhibit the enzymatic activity of the dysregulated
BCR-​ABL1 tyrosine kinase and have now become the preferred
treatment for all newly diagnosed patients with CML, including
children. TKIs substantially reduce the number of leukaemia cells
in a patient’s body, and comparison with historical data confirms
the notion that they prolong overall survival very substantially,
leading to an increased prevalence of the disease. However, very
deep molecular responses (MR4.5, 4.5-​log reduction in BCR-​ABL1
transcripts from baseline) occur in only a minority of patients, and
allogeneic stem cell transplantation (alloSCT) remains the only
treatment that can reliably produce complete and durable MR4.5
due presumably to eradication of all residual leukaemia stem cells.
The authors and editors gratefully acknowledge the inclusion in this
chapter of material contributed to previous editions of the Oxford
Textbook of Medicine by Tariq I. Mughal and John M. Goldman (who
died on 24 December 2013).
† It is with great regret that we report that Tessa L. Holyoake died on 30
August, 2017.
22q-(Ph)
bcr-abl1
abl
22
bcr
9
9q+
Expresses a fusion
protein with
tyrosine kinase activity
abl-bcr
Fig. 22.3.4.1  A schematic representation of how the t(9;22)
translocation produces the Philadelphia (Ph) chromosome.
SECTION 22  Haematological disorders
5214
More recently, TKI discontinuation trials have been conducted for
patients achieving MR4.5 (and in some cases less deep molecular
response) and consistently demonstrate that around 40% of pa-
tients in deep molecular remission (at least MR4) can safely stop a
TKI without suffering an inevitable relapse. These studies are cur-
rently being extended worldwide with results eagerly awaited. The
second-​generation TKIs, notably dasatinib, nilotinib and bosutinib,
have become firmly established in clinical use for the treatment of
imatinib-​resistant/​refractory CML and Ph-​positive acute lympho-
blastic leukaemia (ALL) and are now used by many specialists for
first-​line treatment also.
Epidemiology
The annual incidence of CML is constant worldwide at about 1 to
2 per 100 000 of the population per annum. In the Western world, it
represents approximately 15% of all adult leukaemias and less than
5% of all childhood leukaemias, although these figures are changing
because of the increasing prevalence of CML resulting from suc-
cessful therapy. In the Western world, the median age of onset is 50
to 60 years, and there is a slight male excess. In contrast, the median
age of onset may be considerably younger in some other countries,
such as India.
Importantly, as we become more successful in treating this rare
malignancy, the annual CML-​related death rates are declining fur-
ther: the current estimate is around 2% and this predicts that the
prevalence will plateau at around 35 times the incidence by 2050
(Fig. 22.3.4.2).
Aetiology
For most patients with CML, possibly for all, there appear to be no
obvious predisposing factors, and the disease arises sporadically.
Epidemiology studies have suggested a marginal increment in the
number of cases of CML following exposure to high doses of irradi-
ation as occurred in survivors of the Hiroshima and Nagasaki atomic
bombs in 1945. A small number of families with a high incidence of
the disease have also been reported, though no specific HLA geno-
types have been identified. One convincing case has been reported
of CML recurring in cells of donor origin following related alloSCT.
Natural history
CML is a remarkably heterogeneous disease. Before the introduc-
tion of TKIs, it typically ran a biphasic or triphasic course. It was
usually diagnosed in the chronic phase, which typically lasted 3 to
6 years; the leukaemia then spontaneously progressed to blast trans-
formation. About 70 to 80% of patients had a myeloid blast trans-
formation, and they usually survived 2 to 6 months; the 20 to 30% of
patients with a lymphoid blast transformation had a slightly better
survival. About half the patients in the chronic phase transformed
directly into blast transformation, and the remainder did so fol-
lowing a period of accelerated phase.
Soon after the introduction of imatinib, it was observed that the
natural history for most patients with CML who received this drug
as initial therapy, particularly for patients who remain in complete
cytogenetic response (CCyR) beyond the fourth year of therapy,
200000
180000
160000
140000
120000
100000
Number of cases
80000
60000
40000
20000
2000
2005
2010
2015
2020
2025
Year
2030
2035
2040
2045
2050
Prevalence
Fig. 22.3.4.2  Estimated prevalence of CML in the United States of America.
Reproduced with permission from Huang X, Cortes J, Kantarjian H (2012). Estimations of the increasing prevalence and plateau
prevalence of chronic myeloid leukemia in the era of tyrosine kinase inhibitor therapy. Cancer, 118, 3123–​7. Copyright © 2012, John
Wiley and Sons.
22.3.4  Chronic myeloid leukaemia
5215
was very greatly improved. The 8-​year follow-​up of a phase III pro-
spective trial, the International Randomized Study of Interferon
plus Cytarabine vs STI571 (IRIS), which compared imatinib to
the previous best nontransplant therapy, interferon-​α (IFN-​α) and
cytarabine, showed that 55% of the original cohort randomized
to receive the imatinib were still taking the drug and the majority
was still in CCyR 8 years after starting treatment (Fig. 22.3.4.3 and
Table 22.3.4.1). Patients presenting in the late chronic phase appear
to fare less well, and those in the advanced phases, particularly the
blast phase, generally do poorly, including those who did initially
respond to imatinib. In patients with lymphoid blast phase CML,
there appear to be no durable responses beyond 6 months.
Clinical features and diagnosis
Current estimates suggest that one-​third to one-​half of patients
with CML are totally asymptomatic at diagnosis, which is made
following a routine blood test. The remainder present with signs and
symptoms often of about 3 months’ duration and related to altered
haemopoiesis, particularly anaemia and platelet dysfunction and
increasing disease burden, resulting in splenomegaly. Most patients
will have leucocytosis due to increased numbers of myeloid cells at
all stages of maturation; basophilia is almost universal, and some pa-
tients have an eosinophilia (Box 22.3.4.1). The anaemia tends to be
mild and normochromic normocytic in nature; some patients have
3.3
7.5
4.8
1.7
0.8
0.3
1.5
2.8
1.8
0.9
0.5
0
0
1
2
3
4
5
6
7
8
With Event %
Event:
Loss of CHR
Loss of MCyR, AP/BC
Death during treatment
AP/BC
Year
6
7
8
5
4
3
2
1
1.4
0
1.3
0.4
Fig. 22.3.4.3  IRIS 8-​year follow-​up. Loss of response or progression events are early (years 1–​4) and decline thereafter.
CHR, complete haematological response; MCyR, major cytogenetic response; AP/BC, accelerated phase/blast crisis.
Source data from Deininger, M, et al. (2009). International Randomized Study of Interferon Vs STI571 (IRIS) 8-Year Follow up:
Sustained Survival and Low Risk for Progression or Events in Patients with Newly Diagnosed Chronic Myeloid Leukemia in Chronic
Phase (CML-CP) Treated with Imatinib, Blood, 114, 1126.
Table 22.3.4.1  The 8-​year follow-​up results of the IRIS trial
Still on first-​line imatinib
304 (55%)
Discontinued imatinib
249 (45%)
Adverse events/​abnormal labs
30 (5.4%)
Suboptimal response
77 (13.9%)
Death
16 (2.9%)
SCT
16 (2.9%)
Withdrawal consent
44 (8.0%)
No reconsent to amendment
19 (3.4%)
Crossed over to IFN + Ara-​Ca
14 (2.5%)
Other reasonsb
  3 (6.0%)
Ara-​C, cytosine arabinoside; IFN, interferon; SCT, stem cell transplantation.
a Due to intolerance (0.7%), lack of minor cytogenetic response at 12 months or
progression (1.8%).
b Includes administrative problems, protocol violation, lost to follow-​up.
Box 22.3.4.1  WHO criteria for accelerated and blast
phases of CML
CML, accelerated phase (AP)
Diagnose if one or more of the following is present:
•	 Blasts 10 to 19% of peripheral blood white cells or bone marrow cells.
•	 Peripheral blood basophils at least 20%.
•	 Persistent thrombocytopenia (<100 × 109/​litre) unrelated to therapy, or
persistent thrombocytosis (>1000 × 109/​litre) unresponsive to therapy.
•	 Increasing spleen size and increasing white blood cell count unre-
sponsive to therapy.
•	 Cytogenetic evidence of clonal evolution (i.e. the appearance of an
additional genetic abnormality that was not present in the initial spe-
cimen at the time of diagnosis of chronic phase CML).
•	 Megakaryocytic proliferation in sizable sheets and clusters, associated
with marked reticulin or collagen fibrosis, and/​or severe granulocytic
dysplasia, should be considered as suggestive of CML-​AP. These findings
have not yet been analysed in large clinical studies, however, so it is not
clear if they are independent criteria for accelerated phase. They often
occur simultaneously with one or more of the other features listed.
CML, blast phase (BP)
Diagnose if one or more of the following is present:
•	 Blasts 20% or more of peripheral blood white cells or bone marrow cells.
•	 Extramedullary blast proliferation.
•	 Large foci or clusters of blasts in bone marrow biopsy.
Source data from Vardiman, JW (2008). Chronic myelogenous leukaemia,
BCR-ABL1 positive. WHO classification of tumours of haematopoietic and
lymphoid tissues, 32–37.
SECTION 22  Haematological disorders
5216
a degree of thrombocytosis. Nearly all patients diagnosed in the ad-
vanced phases of CML are symptomatic. Occasionally patients may
present with extramedullary disease, such as a chloroma.
Classical clinical features include sweats, weight loss, haemor-
rhagic manifestations, such as spontaneous bruising and retinal
haemorrhages, abdominal discomfort due to splenomegaly, fatigue
(often but not always related to anaemia), and fever (Box 22.3.4.2).
The diagnosis is typically made by the examination of a peripheral
blood film and the demonstration of the Ph chromosome by conven-
tional cytogenetics on a bone marrow aspirate sample. Most haema-
tologists also carry out a bone marrow trephine examination; this is
often hypercellular with complete or near complete loss of fat spaces
and a high myeloid to erythroid ratio. The presence of less than 10%
of blast cells is compatible with chronic disease, but a higher per-
centage suggests that the patient may be in accelerated or blast phase
(Fig. 22.3.4.4).
Sometimes the diagnosis is made by demonstrating the presence
of a BCR-​ABL1 gene by fluorescence in situ hybridization on a per-
ipheral blood sample. Modern practice dictates the use of a baseline
real-​time quantitative polymerase chain reaction analysis of periph-
eral blood to confirm the presence of a BCR-​ABL1 gene and charac-
terize the BCR-​ABL1 junction. Such an analysis is particularly useful
in the subsequent monitoring of patients.
Molecular biology
The Ph chromosome is an acquired cytogenetic abnormality pre-
sent in all leukaemic cells of the myeloid lineage and in some B-
cells. It is formed as a result of a reciprocal translocation of DNA
from chromosomes 9 and 22, t(9; 22)(q34;q11) (Fig. 22.3.4.1). The
classical Ph chromosome is easily identified in about 90% of CML
patients. A  further 5% of patients have variant translocations in
which chromosomes 9, 22, and other additional chromosomes are
involved. About 5% of patients with clinical and haematological fea-
tures typical of CML lack the Ph chromosome and are referred to
as having ‘Ph-​negative’ CML. These patients have a BCR-​ABL1 chi-
meric gene and are referred to as Ph-​negative, BCR-​ABL1-​positive
cases. Patients who are BCR-​ABL1 negative are not considered to
have CML but an unclassified form of myeloproliferative disorder.
Some patients acquire additional clonal cytogenetic abnormal-
ities, in particular +8, +Ph, iso17q–​, and +19, as their disease pro-
gresses. The emergence of such clones may herald the onset of blast
transformation.
The various genetic events have now been elucidated, and the
chimeric BCR-​ABL1 gene is believed to play a central role in the
pathogenesis of CML, though the precise mechanism(s) are still
not fully understood. Three distinct breakpoint locations in the
BCR gene in chromosome 22 have been identified (Fig. 22.3.4.5).
The break in the major breakpoint cluster region (M-​bcr) occurs
in the intron between exon e13 and e14 or in the intron between
exon e14 and e15 (toward the telomere). By contrast, the position of
the breakpoint in the ABL1 gene on chromosome 9 is highly vari-
able and may occur at almost any position upstream of exon a2.
The Ph translocation results in the juxtaposition of 5′ sequences
from the BCR gene with 3′ sequences from the ABL1 gene. This
event results in the generation of the chimeric BCR-​ABL1 fusion
gene transcribed as an 8.5-​kbp mRNA. This mRNA encodes a pro-
tein of 210 kDa (p210BCR-​ABL1) that has a greater tyrosine kinase
activity compared with the normal ABL protein. The different
breakpoints in the M-​bcr result in two slightly different chimeric
BCR-​ABL1 genes, resulting in either an e13a2 or an e14a2 tran-
script. The type of BCR-​ABL1 transcript has no important prog-
nostic significance.
Box 22.3.4.2  Clinical features of patients with chronic phase
CML seen at the Hammersmith Hospital, London
•	 Fatigue: 33.5%
•	 Bleeding: 21.3%
•	 Weight loss: 20.0%
•	 Abdominal discomfort (left upper quadrant): 18.6%
•	 Sweats: 14.6%
•	 Bone pain: 7.4%
•	 Splenomegaly: 75.8%
•	 Hepatomegaly: 2.2%
Adapted with permission from Savage D, et al. (1997). Clinical features at
diagnosis in 430 patients with chronic myeloid leukaemia seen at a referral
centre over a 16-​year period. Br J Haematol, 96, 111–​16.
Fig. 22.3.4.4  A peripheral blood film from a patient with CML in
chronic phase.
Alternative fusion genes in CML
BCR-ABL1
mRNAs
e14a2
e13a2
11
a2
e14
e13
e12
1
1
e12 e13 a2
e1
e1
ABL1
BCR
kinase domain
Oncoproteins p210BCR-ABL1
Fig. 22.3.4.5  The various breakpoints identified in the CML.
22.3.4  Chronic myeloid leukaemia
5217
The second breakpoint location in the BCR gene occurs be-
tween exons e1 and e2 in an area designated the minor breakpoint
cluster region (m-​bcr) and forms a smaller BCR-​ABL1 fusion gene.
This is transcribed as an e1a2 mRNA which encodes a p190BCR-​ABL1
oncoprotein. This protein characterizes about two-​thirds of patients
with Ph-​positive ALL. A third breakpoint location is found in pa-
tients with the very rare Ph-​positive chronic neutrophilic leukaemia.
This has been designated as a micro breakpoint cluster region
(µ-​bcr) and results in e19a2 mRNA, which encodes a larger protein
of 230 kDa (p230BCR-​ABL1).
The recognition of several features in the BCR-​ABL1 oncoprotein
that are essential for cellular transformation led to the identifi-
cation of signal transduction pathways activated in BCR-​ABL1-​
positive cells (Fig. 22.3.4.6). Much attention has since focused
on determining the precise role played by the various BCR-​ABL1
downstream proteins in the pathogenesis of CML. A  number of
possible mechanisms of BCR-​ABL1-​mediated malignant transform-
ation have been implicated, which are not necessarily mutually ex-
clusive. These include constitutive activation of mitogenic signalling,
reduced apoptosis, impaired adhesion of cells to the stroma and
extracellular matrix, and proteasome-​mediated degradation of ABL
inhibitory proteins. The deregulation of the ABL tyrosine kinase
facilitates autophosphorylation, resulting in a marked increase of
phosphotyrosine on BCR-​ABL1 itself, which creates binding sites
for the SH2 domains of other proteins. A variety of such substrates,
which can be tyrosine phosphorylated, have now been identified.
Although much is known of the abnormal interactions between the
BCR-​ABL1 oncoprotein and other cytoplasmic molecules, the finer
details of the pathways through which the ‘rogue’ proliferative signal
is mediated, such as the RAS-​MAP kinase, JAK-​STAT, and the PI3
kinase pathways, are incomplete, and the relative contributions to
the leukaemic ‘phenotype’ are still unknown. Moreover, the multiple
signals initiated by the BCR-​ABL1 oncoprotein have both prolif-
erative and antiapoptotic qualities, which are often difficult to sep-
arate. Much remains to be learned about the significance of tyrosine
phosphatases in the transformation process.
It is generally believed that some CML stem cells, at a cytokinetic
level, transit through a ‘quiescent’ or ‘dormant’ (G0) phase. These
quiescent CML cells appear to be able to shift between quiescence
and active cycling, allowing them to proliferate under certain cir-
cumstances. There is also evidence that some Ph-​positive cells are
quiescent and less likely to be eradicated by cycle-​dependent cyto-
toxic drugs, even at high doses, or indeed by any of the currently
available TKIs.
It is likely that the acquisition of a BCR-​ABL1 fusion gene by a
haemopoietic stem cell and the ensuing expansion of the Ph-​positive
clone sets the scene for acquisition and expansion of one or more
Ph-​positive subclones that are genetically more aggressive than the
original Ph-​positive population. The propensity of the Ph-​positive
clone to acquire such additional genetic changes is an example of
‘genomic instability’, but the molecular mechanisms underlying this
instability are poorly defined. Such new events may occur in the
BCR-​ABL1 fusion gene or indeed in other genes in the Ph-​positive
population of cells and presumably underlie the progression to ad-
vanced phases of the disease. The average length of chromosomal
telomeres in the Ph-​positive cells is generally less than that in cor-
responding normal cells, and the enzyme telomerase, which is re-
quired to maintain the length of telomere, is up-​regulated as the
patient’s disease enters the advanced phases. About 25% of patients
with CML in myeloid blast transformation have point mutations or
Haematopoietic stem cell
(a)
(b)
JAK2
JAK2
BCR-ABL
GRB2-SOS
Ras-GTP
Raf-MEK-ERK
GRB2-SOS
Ras-GTP
Raf-MEK-ERK
mTOR
PI3K
mTOR
PI3K
STAT5
STAT5
Transcription of target genes for
differentiation or proliferation
Transcription of target genes for
inhibition of apoptosis or drug resistance
CML cell
Fig. 22.3.4.6  Signal transduction pathways which are potentially important in CML in chronic phase. BCR-ABL1
enables JAK2-independent phosphorylation of downstream pathways.
Reprinted by permission from Springer Nature: Fabbro D (2012). BCR-​ABL signaling: A new STATus in CML. Nat Chem Biol, 8,
228–​9. Copyright © 2012, Springer Nature.
SECTION 22  Haematological disorders
5218
deletions in the p53 gene, and about half of all patients in lymphoid
blast transformation show homozygous deletion in the p16 gene.
There is also evidence supporting the role of the RB (retinoblastoma)
and the MYC genes in disease progression. In the future, whole-​
genome and targeted exome sequencing should largely clarify the
mutational landscape of CML, whether it differs between patients,
and how that then modifies the response to TKIs.
Prognostic factors
Various efforts have been made to establish criteria definable at diag-
nosis, both prognostic (disease related) and predictive (treatment
related), that may help to predict survival for individual patients.
Historically, the most frequently used method was that proposed
by Sokal in 1984, whereby patients can be divided into various risk
categories based on a mathematical formula that takes into account
the patient’s age, blast cell percentage, spleen size, and platelet count
at diagnosis. The Euro or Hasford system is an updated Sokal index,
which includes consideration of basophil and eosinophil num-
bers. Stratifying patients into good-​, intermediate-​, and poor-​risk
categories may assist in the decision-​making process regarding ap-
propriate treatment options. Recent observations, however, suggest
that age per se might not influence the biology of the disease, but ra-
ther increases the probability of treatment-​related adverse effects by
virtue of potential comorbid conditions. In 2011, Hasford and col-
leagues proposed a new prognostic score, European Treatment and
Outcome Study (EUTOS), which requires only an assessment of the
spleen size and percent basophils in blood. This method has since
been validated and found to be predictive for CCyR, progression-​
free survival, and overall survival in an independent large series
of patients treated with first-​line imatinib outside of prospective
studies (Table 22.3.4.2).
More recently, the response to TKIs at a given time point is being
increasingly used to assess prognosis (and response). Several inves-
tigators have identified BCR-​ABL1 transcript numbers at 3 months
following the initiation of treatment as the single most important
prognostic factor (Table 22.3.4.3). This has been included in the
National Comprehensive Cancer Network (NCCN) 2019 treatment
guidelines in the United States of America and the updated 2013
European LeukemiaNet recommendations. More recently, rather
than use an individual time point, the rate of decline or halving time
in the early period after start of TKI has been assessed and may prove
even more useful.
Management
First-​line therapy
Clearly TKIs have had a significant impact on the worldwide
standard practice to treat patients with CML. Until 2000, it was con-
ventional to recommend an alloSCT to all patients younger than
50 years of age with newly diagnosed CML in the chronic phase, pro-
vided they had suitable HLA-​identical sibling or ‘matched’ unrelated
donors. Patients presenting in the advanced phases of CML usually
received combination chemotherapy, often followed by an alloSCT if
a ‘second’ chronic phase could be achieved. The treatment algorithm
for newly diagnosed patients changed dramatically once the impres-
sive success of imatinib in inducing durable CCyR in the majority
of newly diagnosed patients with CML in the chronic phase was
recognized. Worldwide, imatinib, or one of the second-​generation
TKIs (nilotinib, bosutinib or dasatinib), are now the preferred treat-
ment for most, if not all, newly diagnosed patients with CML in the
chronic phase, and are also useful in the management of patients
presenting in the advanced phase Fig. 22.3.4.7.
Importantly, the question whether an adult patient should start
with imatinib (at the ‘standard’ dose), or dasatinib, bosutinib or
nilotinib cannot be resolved at present. We have approaching
20 years of experience with imatinib and the drug’s unprecedented
clinical success is notable; however, despite approaching 15 years of
experience with dasatinib and nilotinib, although we see convincing
evidence of more rapid and deeper molecular responses, this has not
translated into a convincing improvement in overall survival. Even
longer-​term follow-​up with assessment of side effects is needed to
address the issue of preferred first-​line therapy. The current indica-
tions for alloSCT are summarized in Table 22.3.4.4.
Imatinib, a 2-​phenylaminopyrimidine, inhibits the enzymatic ac-
tion of the activated BCR-​ABL1 tyrosine kinase by occupying the
ATP-​binding pocket of the kinase component of the BCR-​ABL1
oncoprotein, thereby blocking the capacity of the enzyme to phos-
phorylate and activate downstream effector molecules that cause
the leukaemic phenotype. It also binds to an adjacent part of the
kinase domain in a manner that holds the ABL activation loop of the
oncoprotein in an inactive configuration.
Imatinib induces ‘cumulative best’ CCyR in 82% of all previously
untreated patients with CML in the chronic phase. About 2% of all
patients in the chronic phase progress to advanced phase disease
each year, which contrasts with estimated annual progression rates
for historical therapies of more than 15% for patients treated with
hydroxycarbamide and about 10% for patients receiving IFN-​α,
­either with or without cytarabine. The event-​free survival was 83%
and the estimated overall survival was 93% (corrected for CML-​
related deaths only; Fig. 22.3.4.8), confirming that imatinib prolongs
overall survival very substantially compared with historical patients
who received IFN-​α or hydroxycarbamide. A substantial proportion
of the patients in CCyR also achieve at least a 3-​log reduction in
BCR-​AB1 transcripts (major molecular response (MMR)), and this
proportion seems to have continued to increase steadily with time
on imatinib; a small minority of patients achieve MR4.5.
The standard starting dose of imatinib is 400 mg/​day, although
several single-​arm studies suggest that higher doses, up to 800 mg/​
day, might give better results with a greater proportion of pa-
tients achieving CCyR and MMR. Such studies also suggest better
Table 22.3.4.2  Prediction of prognosis
Sokal 1984
EURO 1998
EUTOS 2011
Parameters
Age
Age
Spleen
Spleen
Spleen
Blasts
Blasts
Platelets
Platelets
Eosinophils
Basophils
Basophils
Treatment
Chemotherapy
IFN
Imatinib
Endpoint
Survival
Survival
CCyR, survival
22.3.4  Chronic myeloid leukaemia
5219
progression-​free survival and transformation-​free survival but
with potentially more side effects, particularly myelosuppression.
The safety analysis of imatinib is also quite impressive, with very
few potentially serious long-​term side effects. When imatinib is
used at the standard starting dose of 400 mg/​day, side effects in-
clude nausea, headache, various skin reactions, infraorbital oe-
dema, bone pains, and sometimes, generalized fluid retention.
Significant cytopenias and hepatotoxicity occur less commonly
and usually in the first 6 to 12 months of therapy. Very rare cases of
severe or fatal cerebral oedema have been reported, and there have
been some concerns about potential cardiomyopathy, although the
longer-​term (8 years) IRIS study analysis reassures us that this is
not a major problem, except for older patients, who might have
other predisposing cardiac risks and have anaemia. Another con-
cern is the potential teratogenicity of imatinib. One study assessing
outcomes in 125 of 180 study patients exposed to the agent during
pregnancy concluded that about half of the offspring born were
normal; 28% of the study cohort elected to undergo termination
of pregnancy, including three after identification of fetal abnor-
malities. In total, there were 12 infants in whom abnormalities
were identified, including three who had strikingly similar com-
plex malformations. It would therefore appear sensible to avoid
imatinib exposure during pregnancy. However, there appear to
be no risks of fetal malformations of children from men taking
imatinib at the time of conception.
Second-​generation TKIs as potential first-​line therapy
Following the successful treatment of patients with CML in chronic
phase resistant/​refractory or intolerant to imatinib, dasatinib,
nilotinib, and bosutinib entered clinical trials for first-​line therapy
of the newly diagnosed patient. Dasatinib at a dose of 100 mg once
daily was tested in a trial known as Dasatinib versus Imatinib Study
in Treatment-​Naïve CML Patients (DASISION), nilotinib at two
dosages, either 300 or 400 mg twice daily in the Evaluating Nilotinib
Efficacy and Safety in Newly Diagnosed Patients (ENESTnd), and
bosutinib at a dose of 500 mg once daily in the Bosutinib Safety
and Efficacy in Newly Diagnosed CML (BELA) trial. Dasatinib and
nilotinib received regulatory approval for first-​line therapy following
the initial results in 2010. Table 22.3.4.5 depicts the current results
of IRIS, DASISION, ENESTnd, and BELA at 12, 24, and 60 months
(where available). Currently dasatinib is recommended at a dose
of 100 mg once daily, nilotinib at 300 mg twice daily and bosutinib
400 mg once daily (following a second trial called BFORE) in the
first-​line setting.
Overall, the results reported suggest that frontline therapy with
dasatinib, nilotinib (at either dose) or bosutinib, renders higher re-
sponse rates with a comparable toxicity profile compared to imatinib
with up to 60 months of follow-​up. It remains unknown whether
these higher rates of early response will translate into improved
event-​free and/​or overall survival. Thus far, no statistically signifi-
cant differences in survival have been observed. It is of course of
Table 22.3.4.3  Early (3-​month) response, rate of decline and outcome in chronic phase CML
Drug
Response at 3 months
Level
Outcomes
Reference
IM
MR
10% IS
OS, PFS, CCyR, MR4.5
Marin et al. J Clin Oncol, 2012, 30, 232–​6
IM
MR
10% IS
OS
Hanfstein et al. Leukemia, 2012, 9, 2096–​102
NIL second line
MR
10% IS
CCyR, MMR, EFS
Branford et al. J Clin Oncol, 2012, 30, 4323–​9
IM
MR
0.35-​fold of baseline by 3 months
OS
Hanfstein et al. Leukemia, 2014, 10, 1988–​92
EFS, event-​free survival; IM, imatinib; IS, International scale; MR, molecular response; NIL, nilotinib; MMR, major molecular response; MR4.5, undetectable disease in cDNA with
32 000 ABL control transcripts; OS, overall survival; PFS, progression-​free survival.
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0
2
4
6
8
10
12
14
Year after diagnosis
n = 3615
(CMLI, II)
Imatinib, 2002–2011 (CML IV)
5–year survival 90%
8–year survival 88%
IFN or SCT, 1997–2003
(CML IIIA) 5–year survival 71%
IFN or SCT, 1995–2001 (CML III)
5–year survival 63%
IFN ,1986–1994
5–year survival 53%
Busulfan, 1983–1994 5-year survival 38%
Hydroxyurea, 1983–1994 5-yr surv. 44%
Survival probability
16
18
20
22
24
26
Fig. 22.3.4.7  Improvement of survival of CML by therapy 1983–​2011.
Courtesy of Professor Rudiger Hehlmann and German CML Study Group.
SECTION 22  Haematological disorders
5220
Estimated overall survival
at 8 years was 93%, considering
only CML-related deaths, and 85%
for all deaths
100
90
80
70
60
50
40
30
20
10
0
0
12
24
36
Months since randomization
48
60
72
84
96
108
Survival: deaths associated with CML
Overall survival
Fig. 22.3.4.8  Leukaemia-​free survival in patients with CML based on the IRIS trial (an intention
to treat analysis).
Courtesy of Professor Michael Deininger, presented at ASH 2009.
Table 22.3.4.5  Results of clinical trials of imatinib, dasatinib, nilotinib, and bosutinib as initial therapy in CML in chronic phase
Trial
N
Response rates (intention to treat)
12 months
24 months
60 months
CCyR
MMR
MR4.5
CCyR
MMR
MR4.5
CCyR
MMR
MR4.5
IRISa
Imatinib 400 mg od
553
69%
NA
NA
73%
NA
NA
87%
NA
NA
DASISIONb
Dasatinib 100 mg od
259
85%
46%
NA
85%
64%
17%
83%
76%
42%
Imatinib 400 mg od
260
73%
28%
NA
82%
46%
8%
78%
64%
33%
ENESTndc
Nilotinib 300 mg bd
282
65%
55%
11%
87%
71%
26%
NA
77%
54%
Nilotinib 400 mg bd
281
55%
51%
7%
85%
67%
21%
NA
77%
52%
Imatinib 400 mg od
283
22%
27%
1%
77%
44%
10%
NA
60%
31%
BELAd
Bosutinib 500 mg od
250
70%
41%
NA
79%
59%
NA
NA
NA
NA
Imatinib 400 mg od
252
68%
27%
NA
80%
49%
NA
NA
NA
NA
bd, twice daily; CCyR, complete cytogenetic remission; MMR, major molecular response; MR4.5, undetectable disease in cDNA with >32 000 ABL control transcripts; N, number of
patients; NA, not applicable; od, once daily.
a IRIS Trial: Druker BJ, et al. N Engl J Med, 2006, 355, 2408–​17.
b DASISION Trial: Kantarjian HM, et al. Blood, 2012, 119, 1123–​9.
c ENESTnd Trial: Larson RA, et al. Leukemia, 2012, 26, 2197–​203.
d BELA Trail: Brummendorf TH, et al. Br J Haematology, 2015, 168, 69–​81.
Table 22.3.4.4  Potential indications for an alloSCT in CML in 2019
First chronic phase
Third line after failure of and/​or intolerance to two TKIs
T315I mutation and not optimally responding to ponatinib
Accelerated phase
De novo accelerated phase at diagnosis: alloSCT recommended for all patients that do not achieve an optimal response to TKIs
Progression to accelerated phase from chronic phase on TKI: alloSCT recommended once disease control re-​established
Blast crisis
Urgently in all eligible patients once chronic phase is re-​established with TKI or chemotherapy; consider second-​ or third-​generation
TKI post allograft (maintenance)
AlloSCT is not recommended in uncontrolled resistant blast phase CML
22.3.4  Chronic myeloid leukaemia
5221
considerable interest that almost twice the number of patients treated
with dasatinib or nilotinib achieved an MR4.5 compared to imatinib
and therefore might be candidates for treatment discontinuation in
the future. Despite showing superior molecular responses compared
to imatinib in the first line, bosutinib 500 mg daily failed to achieve
its primary endpoint of superior CCyR in the BELA study. However,
a second study comparing imatinib with bosutinib 400 mg daily
(BFORE) has shown superior cytogenetic and molecular responses
compared to imatinib.
The European LeukemiaNet has published recommendations for
optimal response, warning features and treatment failure for chronic
phase CML patients treated with TKIs. These guidelines were last
updated in 2013 (Table 22.3.4.6).
The challenge of how long to continue imatinib in optimally re-
sponding patients remains unresolved at present. For patients who
achieve a durable MR4.5, stopping the drug in the context of a clinical
trial is reasonable. Several clinical trials (STIM, TWISTER, STOP
2G TKI, and EUROSKI) have all reported successful discontinu-
ation of therapy in a subgroup of patients who have responded op-
timally to their TKI and have been in a stable MR4.5 for a prolonged
period. In the French STIM study, in which imatinib was discon-
tinued in CML patients who had undetectable BCR-​ABL transcripts
for at least 2 years, 39% of patients did not develop molecular relapse
and remained with undetectable transcripts (Fig. 22.3.4.9). Similar
results were obtained in the Australian TWISTER study with 47.1%
of patients who had sustained undetectable transcripts obtaining
a stable treatment-​free remission. Interestingly, the majority of re-
lapses in both studies occurred within the first 6 months of stopping
imatinib. These important results raise the possibility that imatinib is
able to eradicate CML in some cases, but not in others. More recent
studies (STOP 2G TKI and EUROSKI) have used loss of MMR as
the trigger for restarting therapy. With this approach, treatment-​free
Table 22.3.4.6  Revised European LeukemiaNet (ELN 2013) criteria for responses in patients with chronic myeloid leukaemia in chronic
phase initially treated with TKIs
Milestone
Response definition and criteria
Optimal
Warning
Failure
Baseline
NA
High risk or ACA/​Ph+, major route
NA
3 months
BCR-​ABL1 ≤10% and/​or Ph+ ≤35%
BCR-​ABL1 >10% and/​or Ph+ 36–​95%
No CHR and/​or Ph+ >95%
6 months
BCR-​ABL1 ≤1% and/​or Ph+ 0 (CCyR)
BCR-​ABL1 1-​10% and/​or Ph+ 1–​35%
BCR-​ABL1 >10% and/​or Ph+ >35%
12 months
BCR-​ABL1 ≤0.1% (MMR)
BCR-​ABL1 >0.1–​1%
BCR-​ABL1 >1% and/​or Ph+ >0
Any time
Stable or improving MMR
ACA/​Ph− (−7 or 7q−)
Loss of CHR, loss of CCyR, confirmed loss of MMR,
mutations, ACA in Ph+ cells
ACA, additional cytogenetic abnormalities; CHR, complete haematological response; NA, not applicable; Ph+, Philadelphia chromosome positive.
Baccarani M, et al. (2013). Blood, 122, 872–​84.
0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
3
6
9
12
15
18
21
Months (m) since discontinuation of imatinib
56 had a recurrence (loss of CMR) within
the first 6 months and one at M7 one at
M19 and at M24
At 18 months
43% (95% confidence interval [CI]: 33–52)
Survival without molecular relapse
24
27
30
33
36
Fig. 22.3.4.9  Preliminary Kaplan–​Meier estimates of sustained complete molecular response (CMR) after
discontinuation of imatinib from the French STIM (Stop Imatinib) study. For 100 patients, the estimated
molecular relapse-​free survival is 45% (95% CI 34–​55%) at 6 months, 43% (33–​53%) at 12 months, 41% (34–​55%)
at 24 months, and 35% (22–​46%) at 30 months.
Adapted from Lancet Oncology, Vol. 11, Mahon FX, et al., Discontinuation of imatinib in patients with chronic myeloid leukaemia
who have maintained complete molecular remission for at least 2 years: the prospective, multicentre Stop Imatinib (STIM) trial,
Pages 1029–​35, Copyright © 2010, with permission from Elsevier.
SECTION 22  Haematological disorders
5222
remission rates are higher, 61% at 12 months and 57% at 24 months
in the STOP 2G TKI study and similar results at 12 months in the
EUROSKI trial. The STIM study identified patients with a low Sokal
risk score, male sex, and longer duration of imatinib treatment as
potential prognostic factors for the maintenance of treatment-​free
remission after discontinuing imatinib, and current research is fo-
cusing on identifying factors which will predict maintenance of
treatment-​free remission. However, unless a TKI cessation clinical
trial is available, at present, the best advice for the responding patient
is to continue the drug indefinitely.
The best initial treatment for children is still uncertain, though
most paediatric haematologists would now advocate the use of
imatinib in the first instance. For those children failing imatinib,
a second-​generation TKI (dasatinib or nilotinib) would usually be
considered for second-​line therapy. AlloSCT would only be con-
sidered for treatment failure/​multiple intolerances and if the child
has a suitable HLA donor available. In addition to the expected side
effects also encountered in adults, imatinib can cause significant
growth retardation in children.
Second-​line therapy
Definitions of treatment ‘failure’ and ‘warning’ to TKIs are shown
in Table 22.3.4.6. Primary resistance or refractoriness to the drug
appears to be very rare and when seen may be related to poor drug
compliance, poor gastrointestinal absorption, cytochrome P450
polymorphisms, interactions with other medications, and abnormal
drug efflux and influx at the cellular level. In a small cohort of patients,
a correlation between the expression of human organic cation trans-
porter type 1 (hOCT-​1) and response has been observed: the higher
the levels of OCT-​1, the better the molecular responses. Clearly, it is
prudent to confirm compliance in all patients in whom resistance is
suspected since clinical outcome is known to be ­adversely affected
when adherence is less than 90% (Fig. 22.3.4.10).
A somewhat larger proportion of patients, about 20% in the
chronic phase, respond initially to imatinib and then lose their
response. This acquired or ‘secondary’ resistance results from a
variety of mechanisms, including amplification of the BCR-​ABL1
fusion gene, relative overexpression of the BCR-​ABL1 protein, and
expansion of subclones with point mutations in the BCR-​ABL1
kinase domain. Such point mutations code for amino acid substitu-
tions that may impede binding of imatinib but do not impair phos-
phorylation of downstream substrates that mediate the leukaemia
signal. The precise position of the mutation appears to dictate the
degree of resistance to imatinib; some mutations are associated
with minor degrees of drug resistance, whereas one notorious mu-
tation, the replacement of threonine by isoleucine at position 315
(T315I), is associated with near-​total unresponsiveness to imatinib,
dasatinib, nilotinib, and bosutinib. The precise clinical significance
and indeed the kinetics of the over 100 currently well-​characterized
mutations remain largely unknown (Fig. 22.3.4.11).
The majority of patients who are resistant/​intolerant to imatinib
should receive dasatinib, nilotinib or bosutinib (Fig. 22.3.4.12). For
those patients demonstrating resistance to first-​line therapy with
nilotinib, dasatinib or bosutinib, an alternative second-​generation
TKI (nilotinib, dasatinib, or bosutinib) should be considered.
Currently, ponatinib should be reserved for third-​line therapy or
those with a demonstrable T315I mutation in which case it should
be considered for any line of therapy.
Dasatinib is a thiazole-​carboxamide structurally unrelated to
imatinib. Furthermore, it binds to the ABL kinase domain regard-
less of the conformation of the activation loop—​whether open or
closed. It also inhibits some of the Src family kinases. Preclinical
studies showed that dasatinib is 300-​fold more potent than imatinib
and is active against 18 of 19 tested imatinib-​resistant kinase do-
main mutant subclones, with the notable exception of the T315I
1.0
(a)
(b)
(c)
P < .001
0.9
0.8
Time Since Start of Imatinib Therapy (months)
Probability of MMR
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
6
12
18
24
30
36
42
48
54
60
66
72
1.0
P < .001
0.9
0.8
Time Since Start of Imatinib Therapy (months)
Probability of 4-Log Reduction
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
6
12
18
24
30
36
42
48
54
60
66
72
1.0
P < .002
0.9
0.8
Time Since Start of Imatinib Therapy (months)
Probability of CMR
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
6
12
18
24
30
36
42
48
54
60
66
72
Adherence > 90% (n = 64)
Adherence ≤ 90% (n = 23)
Adherence > 90% (n = 64)
Adherence ≤ 90% (n = 23)
Adherence > 90% (n = 64)
Adherence ≤ 90% (n = 23)
Fig. 22.3.4.10  Six-​year probability of major molecular response (MMR)
according to measured adherence to treatment.
From Marin D, et al. (2010). Adherence is the critical factor for achieving molecular
responses in patients with chronic myeloid leukemia who achieve complete
cytogenetic responses on imatinib. J Clin Oncol, 28, 2381–​8.
22.3.4  Chronic myeloid leukaemia
5223
1242T
M244V
K247R
L248V
G250E/R
Q252R/H
Y253F/H
E255K/V
M237V
P-loop
SH3
contact
SH2 contact
A-loop
E258D
W261L
L273M
E275K/Q
V299L
The 10 most frequent mutations, accounting for
~70% of resistance cases, are highlighted in red
L298V
I293V
E292V/Q
Y342H
M343T
A344V
A350V
M351T
E355D/G/A
V379I
A380T
F382L
L384M
L387M/F/V
M388L
Y393C
H396P/R/A
A397P
M472I
P480L
F486S
E507G
D276G
T277A
E279K
V280A
V289A/I
F311L/l
T315I
F317L/V/1/C
Y320C
L324Q
F359V/1/C/L
D363Y
L3641
A365V
A366G
L370P
V371A
E373K
S417F/Y
I418S/V
A433T
S438C
E450K/G/A/V
E453G/K/V/Q
E459K/V/G/Q
Fig. 22.3.4.11  A schematic depiction of some of the currently established ABL1 kinase domain mutations.
Courtesy of Dr Simona Soverini.
N
N
N
Imatinib
(Gleevec, STI-571)
Nilotinib
(Tasigna, AMN107)
Bosutinib
(SKI-606)
Dasatinib
(Sprycel, BMS-354825)
Ponatinib
H
N
H
N
H
N
H
N
H
N
H
N
N
N
N
N
N
OH
N
O
O
CI
CI
O
CN
HN
N
N
S
O
CI
O
N
N
N
N
N
O
F
N
N
F
F
N
H
N
O
CF3
N
N
N
N
Fig. 22.3.4.12  BCR-​ABL1 inhibitors: imatinib, nilotinib, dasatinib, bosutinib, and ponatinib.
SECTION 22  Haematological disorders
5224
mutant. Current experience with dasatinib in patients with chronic
phase CML resistant/​refractory to imatinib suggests that about 90%
of the patients have a complete haematological response and 52%
have a CCyR. About 25% of patients with the more advanced phases
of CML and Ph-​positive ALL also achieve reasonable responses.
Haematological toxicities are common, particularly in those with
advanced phases of CML and Ph-​positive ALL. These include
­neutropenia (49%), thrombocytopenia (48%), and anaemia (20%).
Nonhaematological toxicities include diarrhoea, headaches, superfi-
cial oedema, pleural effusions, occasional pericardial effusions, and
pulmonary arterial hypertension (rare, c.1%). Grade 3/​4 side effects
are rare, and grade 3/​4 pleural effusions occur in less than 10% of
patients.
Dasatinib has also been tested in patients with CML in advanced
phases whose disease was resistant to both imatinib and nilotinib;
remarkably, 57% of patients achieved haematological responses.
Among those patients who had a haematological ­response, 32% had
a cytogenetic response, including two ­patients who achieved CCyR.
However, these responses are seldom durable in blast phase or in
Ph-​positive ALL, with the majority of ­patients developing resistance
within 6 months. In advanced phase CML, the dasatinib dose is ei-
ther 140 mg once daily or 70 mg twice daily.
Nilotinib, like imatinib, acts by binding to the closed (inactive)
conformation of the ABL kinase domain, but with a much higher
affinity. Like imatinib, it inhibits the deregulated tyrosine kinase
activity of ABL by occupying the ATP-​binding pocket of the
oncoprotein and blocking the capacity of the enzyme to phos-
phorylate downstream effector molecules. In vitro studies suggest
that nilotinib is about 30-​ to 50-​fold more potent than imatinib.
Nilotinib is also active in 32 of 33 imatinib-​resistant cell lines with
mutant ABL kinases, but like imatinib and dasatinib has no ac-
tivity against the Bcr-​Abl1T315I mutation. Phase II studies in pa-
tients who are resistant or intolerant to imatinib show a CCyR
rate of 45% at 4  years and progression-​free survival of 57%.
Nilotinib is recommended at a dose of 400 mg twice daily second
line. Patients in the advanced phases of CML also respond, but to
a lesser degree. The most common treatment-​related toxicity is
myelosuppression, followed by headaches, pruritus, and rashes.
Overall, 22% of the patients in the chronic phase experienced
thrombocytopenia, with 19% having either grade 3 or 4 severity;
16% had neutropenia and a further 16% had anaemia. Most of the
nonhaematological adverse effects were of a grade 1/​2 severity.
More recently, arterial thrombotic events and onset of diabetes/​
metabolic syndrome have been described with nilotinib therapy
and caution should be taken when commencing nilotinib in pa-
tients with cardiovascular risk factors or pre-​existing diabetes. It
is recommended that all patients should have a cardiovascular
risk assessment, including blood pressure measurement, lipids,
glucose, and HbA1c prior to commencing nilotinib and at least
annually thereafter.
Third-​line therapy
Bosutinib (formerly SKI-​606), an oral dual ABL and SRC inhibitor,
is chemically different from both dasatinib and nilotinib. Following
single-​arm studies in patients with CML in all phases intolerant or
resistant to one or more TKI, it was noted that 53.4% of the study co-
hort achieved a durable major cytogenetic response. Grade 3 to 4 side
effects included diarrhoea, anaphylactic shock, myelosuppression,
fluid retention, hepatotoxicity, and rash. Based on these results, the
drug was approved for the treatment of adult CML patients with
chronic phase or advanced phase disease who were resistant to prior
therapy.
Ponatinib (formerly AP24534) is a third-​generation TKI which
has an interesting chemical structure based on a purine scaffold
and a central triple carbon–​carbon bond with a substructure that
is similar to imatinib. The drug inhibits ABL, SRC, FLT3, and a var-
iety of other kinases. It was developed initially for patients who were
considered to have become resistant to TKIs as a result of a T315I
subclone and subsequently, in a phase II trial, found to have consid-
erable activity in all patients with CML who were resistant/​refractory
or indeed intolerant to prior TKIs, including imatinib, dasatinib,
and/​or nilotinib. Like nilotinib, ponatinib has been associated with
a significant number of arterial thrombotic events, which led to it
being temporarily withdrawn by the Food and Drug Administration
in 2013. In addition, a phase III trial assessing its candidacy as first-​
line therapy, compared to imatinib was abandoned due to concerns
about cardiovascular toxicity.
The indications for alloSCT are shown in Table 22.3.4.4. For pa-
tients who are resistant/​refractory to two or more TKIs, and are
under the age of 65 years, it is probably best to consider an alloSCT,
provided a suitable donor is identified and the patient remains in the
chronic phase of the disease. A risk score has been developed by the
European Society for Blood and Marrow Transplantation (EBMT)
which is helpful in determining those patients that may benefit
from an alloSCT (Fig. 22.3.4.13). It is also timely to note that as our
nontransplant efforts to improve CML patients’ outcomes have im-
proved, there have been some improvements in transplant outcomes
also (Figs. 22.3.4.14 and 22.3.4.15).
Patients who proceed to a transplant should stop the TKI at
least 2 weeks prior to the transplant. Current data also suggest that
prior treatment with any TKI does not increase the probability of
transplant-​related mortality, but clearly our experience with alloSCT
following any TKI is still limited and we should continue to be vigi-
lant. Moreover, patients with kinase domain mutations appear to
fare as well post-transplant as those lacking such mutations.
Advanced phase CML
For those with advanced phase disease, the choice of treatment is de-
pendent on whether this arises de novo or as a result of progression
from chronic phase CML while on TKI therapy. Current European
LeukemiaNet guidelines recommend either imatinib 400 mg twice
daily or dasatinib 70 mg twice daily or 140 mg once daily for treat-
ment of de novo advanced phase CML. AlloSCT is recommended
for all blast phase patients and accelerated phase patients without
an optimal response who have a suitable donor. Additional conven-
tional chemotherapy may be required for disease control prior to
alloSCT. For patients progressing to accelerated or blast phase CML
while on TKI therapy, further therapy should be with any TKI not
previously used and ponatinib for cases with the T315I mutation. All
eligible patients should then proceed to alloSCT; and conventional
chemotherapy is frequently required for disease control in this poor
risk group of patients. None of the currently available agents has
made a major impact in this area, in particular for patients who are
in lymphoid blast crisis.
22.3.4  Chronic myeloid leukaemia
5225
Investigational approaches
Immunotherapy
Following the realization that a molecular remission and ‘cure’
might not be possible with imatinib alone in the majority of patients,
many efforts were directed to exploring the potential of developing
an active specific immunotherapy strategy for patients with CML
by inducing an immune response to a tumour-​specific or tumour-​
associated antigen. The principle involves generating an immune re-
sponse to the unique amino acid sequence of p210 at the fusion point.
Clinical responses to the BCR-​ABL1 peptide vaccination, including
CCyR, have been reported in a small series. In contrast to previous
earlier unsuccessful attempts, administration of granulocyte–​
macrophage colony-​stimulating factor was included as an immune
adjuvant, and patients were only enrolled if they had measurable
residual disease and human leucocyte antigen known to which the
selected fusion peptides were predicted to bind avidly. However,
these results have not been confirmed in other studies, and the con-
tinuing success of TKIs has resulted in limited development of other
approaches. Nonetheless, vaccine development against BCR-​ABL1
and other CML-​specific antigens could become an attractive treat-
ment for patients who have a minimal residual disease status with
TKIs as a potential additional strategy to enable treatment-​free re-
mission. Other targets for vaccine therapy studied include peptides
derived from the Wilms tumour 1 protein, proteinase 3, and elastase,
all of which are overexpressed in CML cells. Furthermore, interferon
is also considered to have an immunological mode of action and
superior responses to the combination of imatinib and pegylated
interferon compared to imatinib alone were demonstrated in the
100
Donor
HLA identical sibling
Unrelated donor
First chronic phase
Accelerated phase
Blast crisis
<20 years
20–40 years
40 Years
All, except:
Male recipient/female donor
<12 months
12 months
1
0
1
0
2
1
0
2
1
0
1
0
Stage
Age
Score 0 or 1 (n = 634)
Score 2 (n = 881)
Score 5–7 (n = 275)
Score 4 (n = 485)
Score 3 (n = 867)
Sex match
Time to
transplantation
Survival
Variable
Categories
Score
75
50
25
0
0
12
Survival probability (%)
24
36
48
60
72
84
Fig. 22.3.4.13  Transplantation risk: EBMT score.
Adapted from The Lancet, Vol. 352, Gratwohl A, et al., Risk assessment for patients with chronic myeloid leukaemia before allogeneic blood or marrow transplantation,
Pages 1087–​92, Copyright © 1988, with permission from Elsevier.
Overall survival among good risk patients (score = 0, 1)
N = 645
2000–2003
1991–1999
1980–1990
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0
24
48
72
96 120 144 168 192 216 240 months
N = 1466
N = 594
Fig. 22.3.4.14  Improvements in survival rates by decades of
transplantation for patients with CML in chronic phase.
The EBMT registry data; courtesy of the EBMT Group.
100
80
60
40
20
0
0
1
2
P<.001
Score = >4 (N = 34)
Score = 2 (N = 35)
Score = 0.1 (N = 18)
Score = 3 (N = 41)
Score = 4 (N = 45)
3
4
5
6
Years post SCT
Probability of survival %
7
8
9
10
Fig. 22.3.4.15  Survival of patients allografted for CML at the
Hammersmith Hospital, London, from January 2000 to December 2010
stratified by EBMT risk score.
Reproduced, with permission, from Pavlu J, et al. (2011). Blood 117(3), 755–​63.
SECTION 22  Haematological disorders
5226
French SPIRIT trial. Unfortunately, many patients find interferon
difficult to tolerate, limiting its utility in this setting.
Investigational drugs
Other specific inhibitors of signal transduction pathways downstream
of BCR-​ABL1 have been tested alone and in combination with TKIs
in vitro. However, in chronic phase in particular, translation of these
effective combinations into successful clinical trials has been dif-
ficult due to concerns about side effects. Recently, early-​phase clin-
ical trials of SMO inhibitors (inhibiting the developmental hedgehog
pathway which is upregulated in many cancers) in combination with
various TKIs in resistant CML closed early due to poor recruitment,
lack of efficacy and an unacceptable side effect profile. The phase II
CHOICES clinical trial which evaluated the autophagy inhibitor
hydroxychloroquine in combination with imatinib in patients with
detectable transcripts has recently closed to recruitment. Asciminib,
(ABL001), an allosteric inhibitor of BCR-​ABL, is currently in a phase
II clinical trials alone or in combination with other TKIs. Omacetaxine
mepesuccinate (formerly homoharringtonine) is a semisynthetic
plant alkaloid (cetaxine) that enhances apoptosis of CML cells, and is
active in combination with imatinib in drug-​resistant/​refractory pa-
tients, including those who harbour the T315I mutation. Other po-
tential agents include rapamycin, an mTOR inhibitor, venetoclax, a
BCL-​2 inhibitor, idasanutlin, an MDM2 inhibitor, tazemetostat, an
EZH2 inhibitor, and ruxolitinib, a Janus kinase inhibitor.
Conclusion
The substantial understanding of the molecular features and patho-
genesis of CML has provided important insights into targeting treat-
ment to specific molecular defects. The successful introduction of
TKIs, commencing with imatinib and now with the addition of
dasatinib and nilotinib, as targeted therapy for CML has made the
approach to management of the newly diagnosed patient fairly com-
plex. Furthermore, a third second-​generation TKI, bosutinib, and a
third-​generation TKI, ponatinib, have significant activity in selected
patients in both chronic and the more advanced phases of the dis-
ease, who are resistant to imatinib, dasatinib, and/​or nilotinib.
For the moment, the various treatment options should be assessed
carefully in terms of the relative risk:benefit ratios, and a manage-
ment strategy should be developed accordingly. Consideration
of long-​term side effects, particularly cardiovascular events is
developing increasing prominence with nilotinib and ponatinib
in particular. With increasing data demonstrating the safety and
efficacy of imatinib in particular, early alloSCT should no longer
be considered in any patient cohort, other than in exceptional cir-
cumstances. Newly diagnosed chronic phase patients should com-
mence either imatinib 400 mg once daily, nilotinib 300 mg twice
daily, dasatinib 100 mg once daily, or bosutinib 400 mg once daily,
with regular review and cytogenetic/​molecular testing to ensure
they are achieving treatment milestones. It is important to consider
that although nilotinib, dasatinib and bosutinib result in superior
cytogenetic and molecular responses, compared to imatinib, there is
no current evidence that they prolong life any more than imatinib.
Patients who are resistant/​refractory to imatinib, dasatinib, or
nilotinib should probably receive bosutinib or ponatinib (including
those with the T315I mutation in particular), and/​or be considered
for an alloSCT provided that a suitable donor is identified. As we
see more patients obtaining more rapid and deeper molecular re-
sponses, it is likely that the criteria for an optimal response and treat-
ment failure will become more stringent with a view to obtaining
a future treatment-​free remission for as many patients as possible.
FURTHER READING
Baccarani M, et al. (2013). European LeukemiaNet recommendations
for the management of chronic myeloid leukemia: 2013. Blood, 122,
872–​84.
Brummendorf TH, et al. (2015). Bosutinib versus imatinib in newly
diagnosed chronic-​phase chronic myeloid leukaemia: results from
the 24-​month follow-​up of the BELA trial. Br J Haematol, 168, 69–​81.
Cortes J, et al. (2015). Cardiovascular and pulmonary adverse events
in patients treated with BCR-​ABL inhibitors: data from the FDA
Adverse Event Reporting System. Am J Hematol, 90, E66–​72.
Cortes JE, et al. (2013). A phase 2 trial of ponatinib in Philadelphia
chromosome-​positive leukemias. N Eng J Med, 369, 1783–​96.
Daley GQ, et al. (1990). Induction of chronic myelogenous leukemia
in mice by the p210 bcr/​abl gene of the Philadelphia chromosome.
Science, 24, 824–​30.
Druker BJ, et al. (2001). Efficacy and safety of a specific inhibitor of
the BCR-​ABL tyrosine kinase in chronic myeloid leukemia. N Engl
J Med, 344, 1031–​7.
Druker BJ, et  al. (2006). Five-​year follow-​up of patients receiving
imatinib for chronic myeloid leukemia. N Engl J Med, 355, 2408–​17.
Fialkow PJ (1981). Evidence for a multistep origin of chronic myeloid
leukemia. Blood, 58, 158–​63.
Gambacorti-​Passerini C, et al. (2014). Bosutinib efficacy and safety in
chronic phase chronic myeloid leukemia after imatinib resistance or in-
tolerance: minimum 24-​month follow-​up. Am J Hematol, 89, 732–​42.
Gratwohl A, et al. (1998). Risk assessment for patients with chronic
myeloid leukaemia before allogeneic blood or marrow transplant-
ation. Lancet, 352, 1078–​92.
Gratwohl A, et al (2016). Long-​term outcome of patients with newly
diagnosed chronic myeloid leukemia: a randomized comparison of
stem cell transplantation with drug treatment. Leukaemia, 30, 562–​9.
Hijiya N, et al. (2015). Chronic myeloid leukemia in children: clin-
ical findings, management, and unanswered questions. Pediatr Clin
North Am, 62, 107–​19.
Holyoake T, et al (1999). Isolation of a highly quiescent subpopulation
of primitive leukemic cells in chronic myeloid leukemia. Blood, 94,
2056–​64.
Huang X, et al. (2012). Estimations of the increasing prevalence and
plateau prevalence of chronic myeloid leukemia in the era of tyro-
sine kinase inhibitor therapy. Cancer, 118, 3123–​7.
Kabarowski JH, et al. (2000). Consequences of BCR-​ABL expression
within the hematopoietic stem cell in chronic myeloid leukemia.
Stem Cells, 18, 399–​408.
Kantarjian H, et al. (2010). Dasatinib versus imatinib in newly diag-
nosed chronic-​phase chronic myeloid leukemia. N Engl J Med, 362,
2260–​70.
Kantarjian H, et al. (2012). Dasatinib or imatinib in newly diagnosed
chronic-​phase chronic myeloid leukemia: 2-​year follow-​up from a
randomized phase 3 trial (DASISION). Blood, 119, 1123–​9.
Krause DS, Van Etten RA (2008). Bedside to bench: interfering with
leukemic stem cells. Nat Med, 14, 494–​5.
Larson RA, et al. (2012). Nilotinib vs imatinib in patients with newly diag-
nosed Philadelphia chromosome-​positive chronic myeloid leukemia
in chronic phase: ENESTnd 3-​year follow-​up. Leukemia, 26, 2197–​203.

22.3.5 The polycythaemias 5227 Daniel Aruch and Ro
22.3.5 The polycythaemias 5227 Daniel Aruch and Ronald Hoffman
22.3.5  The polycythaemias
5227
Latif A-​L, et al. (2013). Allogeneic stem cell transplantation for Chronic
Myeloid Leukemia is safe and effective in high risk patients fol-
lowing second generation tyrosine kinase inhibitors: a single centre
experience. Leuk Res Rep, 2, 47–​50.
Mahon FX, et al. (2010). Discontinuation of imatinib in patients with
chronic myeloid leukaemia who have maintained complete mo-
lecular remission for at least 2 years: the prospective, multicentre
Stop Imatinib (STIM) trial. Lancet Oncol, 11, 1029–​35.
Marin D, et al. (2010). Adherence is the critical factor for achieving
molecular responses in patients with chronic myeloid leukemia who
achieve complete cytogenetic responses on imatinib. J Clin Oncol,
28, 2381–​8.
Marin D, et al. (2011). Assessment of BCR-​ABL1 transcript levels at
3 months is the only requirement for predicting outcome for pa-
tients with chronic myeloid leukemia treated with tyrosine kinase
inhibitors. J Clin Oncol, 30, 232–​8.
Nowell PC, Hungerford DA (1960). A minute chromosome in human
chronic granulocytic leukemia. Science, 132, 1497.
Palani R, et al. (2015). Managing pregnancy in chronic myeloid leu-
kaemia. Ann Hematol, 94 Suppl 2, S167–​76.
Pavlu J, et al. (2011). Three decades of transplantation for chronic mye-
loid leukemia: what have we learned? Blood, 117, 755–​63.
Pinilla-​Ibarz J, et  al. (2000). Vaccination of patients with chronic
myelogenous leukemia with BCR-​ABL oncogene breakpoint fusion
peptides generates specific immune responses. Blood, 95, 1781–​7.
Preudhomme C, et al. (2010). Imatinib plus peginterferon alfa-​2a in
chronic myeloid leukemia. N Eng J Med, 363, 2511–​21.
Rea D (2015). Management of adverse events associated with tyrosine
kinase inhibitors in chronic myeloid leukemia. Ann Hematol, 94
Suppl 2, S149–​58.
Ross DM, et al. (2013). Safety and efficacy of imatinib cessation for
CML patients with stable undetectable minimal residual disease: re-
sults from the TWISTER study. Blood, 122, 515–​22.
Sasaki K, et al. (2016). Conditional survival in patients with chronic
myeloid leukemia in chronic phase in the era of tyrosine kinase in-
hibitors. Cancer, 122, 238–​48.
Shah NP, et al. (2004). Overriding imatinib resistance with a novel ABL
kinase inhibitor. Science, 305, 399–​401.
Soverini S, et al. (2011). BCR-​ABL kinase domain mutation analysis
in chronic myeloid leukemia patients treated with tyrosine kinase
inhibitors:  recommendations from an expert panel on behalf of
European LeukemiaNet. Blood, 118, 1208–​15.
Vardiman, JW (2008). Chronic myelogenous leukaemia, BCR-ABL1
positive. WHO classification of tumours of haematopoietic and
lymphoid tissues, 32–7.
22.3.5  The polycythaemias
Daniel Aruch and Ronald Hoffman
ESSENTIALS
Polycythaemia or erythrocytosis is characterized by an abnormal in-
crease in the numbers of red blood cells, leading to an elevation in
the haemoglobin concentration and haematocrit (>49% in men and
48% in women). The cause may be either (1) primary—​due to an
intrinsic defect of haematopoietic stem cells; or (2) secondary—​due
to extrinsic stimulation of progenitor erythroid cells by circulating
growth factors; and the condition needs to be distinguished from
(3) pseudopolycythaemia—​in which haematocrit is raised because
the plasma volume is decreased.
Normal erythropoiesis
The primary controlling factor for erythropoiesis is the glycoprotein
hormone erythropoietin. This is produced by the kidney in response
to hypoxia, which leads to the accumulation of a transcriptional
factor, hypoxia-​inducible factor-​1 (HIF-​1), the principal regulator of
numerous genes that participate in the hypoxic response. Mutation
in genes encoding for proteins involved in the oxygen-​sensing mech-
anism, in the erythropoietin receptor, or in pathways downstream of
the receptor can all (rarely) lead to polycythaemia.
Secondary polycythaemias
Associated with appropriate erythropoietin secretion—​conditions that
are ultimately the result of tissue hypoxia and subsequent exces-
sive erythropoietin production include (1)  living at high altitude,
(2) chronic lung disease, (3) cyanotic congenital heart disease with
right-​to-​left shunting, (4) carbon monoxide intoxication—​as occurs
in heavy smokers, (5) haemoglobin variants with increased oxygen
affinity, and (6) mutations in genes involved in the oxygen sensing
pathway—​e.g. von Hippel–​Lindau gene (Chuvash polycythaemia)
and prolyl hydroxylases.
Associated with inappropriate erythropoietin secretion—​in the absence
of tissue hypoxia, inappropriate erythropoietin production commonly
originates from the kidney and many renal disorders are associated
with erythrocytosis (e.g. renal artery stenosis, polycystic kidney disease,
and tumours). Tumour-​associated polycythaemia may also result from
cerebellar haemangioblastoma, hepatocellular carcinoma, phaeo-
chromocytoma, and other adrenal tumours.
Primary polycythaemia—​polycythaemia vera
This is a clonal, chronic progressive haematological malignancy
characterized by excessive proliferation of erythroid, myeloid, and
megakaryocytic elements in the bone marrow.
Aetiology—​up to 95% of cases are caused by somatic mutations
in the pluripotent haemopoietic stem cells leading to replacement
of a key valine residue by phenylalanine at position 617 of the JAK2
kinase (V617F), which releases it from autoinhibition. Less common
mutations have been described recently, primarily JAK2 exon 12 and
LNK mutations.
Clinical features—​may be detected on a full blood count in asymp-
tomatic patients, or may present with a wide range of nonspecific
symptoms (notably pruritus). Signs include those directly related to
polycythaemia (e.g. ruddy complexion), also splenomegaly and hep-
atomegaly. Complications of particular note include (1) a thrombotic
tendency—​deep venous thrombosis/​pulmonary embolism, hepatic
or portal venous thrombosis, venous thrombosis in unusual sites, or
transient ischaemic attack/​stroke; (2) other neurological syndromes—​
a wide variety are described; (3) a haemorrhagic tendency—​due to
abnormalities of platelet function; and (4)  gout—​associated with
hyperuricaemia. Myelofibrosis with marrow failure develops in
about half of patients with polycythaemia vera at 20 years.
Diagnosis—​using the World Health Organization 2016 criteria,
the major criteria are (1) haemoglobin concentration greater than
SECTION 22  Haematological disorders
5228
165 g/​litre in men or greater than 160 g/​litre in women, haemato-
crit concentration greater than 49% in men or greater than 48% in
women, or increased red cell mass; (2) bone marrow biopsy showing
hypercellularity for age with trilineage growth (panmyelosis) including
prominent erythroid, granulocytic, and megakaryocytic proliferation
with pleomorphic, mature megakaryocytes (differences in size); and
(3) presence of JAK2 (V617F) or JAK2 exon 12 mutation. Minor cri-
teria are (1) trilineage myeloproliferation in the bone marrow, (2) a
low serum erythropoietin level, and (3) abnormal marrow prolifera-
tive capacity as manifested by the formation of erythroid colonies
in the absence of exogenous erythropoietin. Diagnosis requires all
three major criteria, or the first two major criteria and the minor cri-
terion. Major criterion 2, bone marrow biopsy, may not be required
with sustained absolute erythrocytosis: haemoglobin levels greater
than 185 g/​litre in men (haematocrit 55.5%) or greater than 165 g/​
litre in women (haematocrit 49.5%) if major criterion 3 and the minor
criterion are present.
Treatment—​patients should be strongly advised to stop smoking.
Phlebotomy should be initiated as soon as the diagnosis is es-
tablished to reduce and maintain the haematocrit level at less
than 45% in men and less than 42% in women. Low-​dose aspirin
should be given. Myelosuppressive therapy with hydroxycarbamide
(hydroxyurea) or other agents should be considered in older pa-
tients intolerant of phlebotomies and in those with repeated throm-
botic episodes and/​or high platelet counts. Those intolerant of
hydroxycarbamide may be treated with interferon or ruxolitinib.
The optimal myelosuppressive therapy for polycythaemia vera
(hydroxycarbamide, pegylated interferon-​α, or ruxolitinib) remains
an area of controversy. The resolution of this controversy will require
the completion of carefully performed phase III trials. In lieu of such
trials, it remains the physician’s choice when to initiate therapy and
with which agent. Such therapy should be reserved for patients with
high-​risk disease who have repeated thrombotic events in the face of
strict haematocrit control. Although interferon has been reported to
transiently eliminate cells with the JAK2 V16F mutation, the effects of
this on the incidence of thrombotic episodes and disease evolution
remain the subject of speculation. Haematopoietic stem cell trans-
plantation is a potentially curative option for myelofibrosis but is not
indicated in polycythaemia vera.
Prognosis—​survival is about 18  months in untreated patients
whereas with appropriate management, survival of over 10 years
is now common. Patients previously treated with alkylating agents
and/​or radioactive phosphorous have an increased long-​term risk
of leukaemia.
Rare causes of primary polycythaemia
These include primary familial and congenital polycythaemia—​
caused by germ-​line mutations in the erythropoietin receptor gene
and genes encoding components of the JAK2–​STAT pathway; may be
suggested by family history (autosomal dominant).
Introduction
Erythropoiesis is the process responsible for maintaining a normal
red blood cell mass. This is a tightly regulated process, which
maintains a balance between the production and destruction of
erythrocytes. Polycythaemia or erythrocytosis is a distinct group of
disorders characterized by an abnormal increase in the numbers of
red blood cells, leading to an elevation in the haemoglobin concen-
tration and haematocrit. Absolute polycythaemias (increased red
cell mass) can be attributed to either an intrinsic defect of haem-
atopoietic stem cells (primary) or to the stimulation of progenitor
cells by excessive levels of circulating growth factors (secondary).
A pathophysiological classification of polycythaemia is provided in
Box 22.3.5.1. Patients with absolute polycythaemias should be dis-
tinguished from individuals in whom a minimally elevated haem-
atocrit is not accompanied by a corresponding absolute increase in
the red cell mass (spurious polycythaemia, stress erythrocytosis,
Gaisbock’s syndrome), but rather by a contraction of plasma volume.
Haematocrit levels above 49% in men and 48% in women are ab-
normal and require further evaluation to determine the cause of the
polycythaemia.
Erythropoietin (EPO), a 34.4-​kDa glycoprotein hormone, is the
primary humoral regulator of erythropoiesis. Alterations in its pro-
duction are accompanied by adjustments in the rate of red cell pro-
duction. Production of EPO is controlled by the relative supply of
oxygen to the kidney and can increase by 1000-​fold in states of se-
vere hypoxia. Under normal conditions, EPO production is medi-
ated by decreased oxygen content of haemoglobin within red cells,
termed hypoxaemia, which leads to decreased oxygen delivery to
tissues.
Box 22.3.5.1  Classification of polycythaemia
Relative polycythaemias
Associated with volume loss or contraction:
•	 Gastrointestinal losses: diarrhoea, vomiting, ileostomy
•	 Renal losses: osmotic diuresis, therapeutic diuresis, Addison’s disease,
hypercalcaemia
•	 Insensible losses: profuse sweating, fever
•	 Stress or Gaisbock’s polycythaemia
Absolute polycythaemias
Primary polycythaemias:
•	 Primary familial and congenital polycythaemia
•	 Polycythaemia vera
Secondary polycythaemias associated with appropriate secretion
of EPO:
•	 Smokers’/​hookah polycythaemia
•	 Hypobaric hypoxia (high altitude)
•	 Chronic pulmonary disease
•	 Alveolar hypoventilation (Pickwickian syndrome)
•	 Congenital heart diseases associated with right-​to-​left shunts
•	 High-​affinity haemoglobins
•	 2,3-​DPG deficiency
•	 Methaemoglobinaemias
•	 Chuvash polycythaemia (VHL mutations)
•	 HIF-​1α mutations
•	 Prolyl hydroxylase mutations
Secondary polycythaemias associated with inappropriate secretion
of EPO:
•	 Polycythaemia of renal disease
•	 Tumour-​associated polycythaemia
•	 Endocrine disorders—​phaeochromocytomas, aldosterone-​producing
adenomas, Cushing’s syndrome, iatrogenic androgen administration,
acromegaly
22.3.5  The polycythaemias
5229
Decreased tissue oxygenation (partial pressure of oxygen (Po2)
<60 mmHg) is associated with accumulation of hypoxia-​inducible
factor-​1 (HIF-​1), the major transcriptional factor responsible for
activation of the EPO gene. The HIF transcriptional system is a
master regulator of the hypoxic response controlling a large number
of genes including phosphoglycerate kinase, glucose transporter-​1,
vascular endothelial growth factor, and EPO. HIF-​1α is continu-
ously synthesized irrespective of oxygen availability. It is barely de-
tected in steady-​state cells because of its rapid degradation by the
ubiquitin proteasome pathway. During hypoxic conditions, a rapid
accumulation of HIF-​1α occurs due to a blockade of its degradation.
This pathway permits a rapid response to hypoxia without activating
transcriptional/​translational machinery. Increased HIF-​1α mRNA
and protein levels are induced by hypoxia while protein levels rap-
idly decay during normoxia. The degradation of HIF-​1α requires the
von Hippel–​Lindau (VHL) protein, oxygen, and three different iron
requiring proline hydroxylase (PH) enzymes. The VHL gene is a tu-
mour suppressor gene, which participates in the hypoxia-​sensing
pathway by binding HIF-​1α facilitating its degradation by the pro-
teasome. The PH enzymes, which hydroxylate HIF, are required for
HIF proteolytic degradation by promoting the interaction between
HIF and VHL. As oxygen levels decrease, hydroxylation of HIF de-
creases and HIF-​1α is no longer able to bind VHL and becomes
stabilized. In normal individuals, VHL protein binds to hydroxyl-
ated HIF-​1α and causes its ubiquitination and proteasomal degrad-
ation. Genetic mutations in VHL or PH are frequently capable of
leading to inherited forms of erythrocytosis by causing alterations
in the binding of the VHL to HIF-​1α, leading to its accumulation
and activation of hypoxia-​related genes including EPO and vascular
endothelial growth factor (VEGF). EPO binds to its receptor, and ini-
tiates downstream effects via the Janus tyrosine kinase 2 (JAK2) and
signal transducer and activator of transcription (STAT) intracellular
signalling pathways. Binding of EPO causes dimerization of the EPO
receptor, phosphorylation of intracytoplasmic residues, and activa-
tion of STAT, which shuttles from the cytoplasm into the nucleus
and initiates protein transcription. JAK2–​STAT activation promotes
erythroid cell division, differentiation, proliferation, and preven-
tion of precursor cell apoptosis. Mutations in the EPO receptor gene
leading to its constitutive activation have been observed in some pa-
tients with familial and congenital forms of polycythaemia.
Oxygen transport is a complex process dependent on a number
of variables, including ambient oxygen levels, minute ventilation
rates, lung diffusion capacity, cardiac output, red cell mass, regional
blood flow, tissue capillary density, and haemoglobin–​oxygen af-
finity. Acute changes in tissue oxygen demands or in environmental
oxygen levels are compensated not only by increased EPO produc-
tion but also increased ventilation rates, cardiac output, blood flow
distribution, and haemoglobin–​oxygen affinity (through the modu-
lation of 2,3-​disphosphoglycerate (2,3-​DPG) production). Sustained
hypoxia is required for polycythaemia to occur as a compensatory
mechanism.
Relative polycythaemias
These disorders are characterized by an elevated haemoglobin
or haematocrit level, which occurs as the result of contraction in
plasma volume. The red cell mass remains normal. There are two
major groups of patients with relative forms of polycythaemia. The
first includes patients with more acute conditions associated with
significant degrees of dehydration, with a consequential decrease
in plasma volume, for example, gastrointestinal fluid losses, thera-
peutic diuresis, endocrine disorders such as Addison’s disease, and
hypercalcaemia. In most cases, the consequences of volume contrac-
tion are clinically obvious. The aetiology of the increase in haemato-
crit does not usually present a diagnostic challenge.
The second group is associated with a slight but sustained increase
in the haematocrit. These patients are frequently active, middle-​aged,
mildly hypertensive, obese men subjected to considerable stress who
present with persistent polycythaemia. Characteristically, they ap-
pear plethoric but without any of the other typical features of poly-
cythaemia vera. The cause for the contraction in the plasma volume
is poorly understood, but autonomic dysregulation with changes in
venous capacitance may be responsible. These patients also have a
normal red cell mass.
The usual range of haemoglobin levels in these individuals is
between 160 and 180 g/​litre with haematocrits ranging from 48 to
55%. Most of these patients seek medical evaluation for an unrelated
condition, and are incidentally found to have increased haemo-
globin and haematocrit values. Suitable advice regarding weight re-
duction, control of hypertension, and smoking cessation is usually
provided to these patients. The optimal therapy is unknown but is
generally directed to correcting the patient’s underlying cardiovas-
cular risk factors.
Overfilling blood collection tubes can cause artefact, or pseudo­
polycythaemia, due to inadequate sample collection. Attention to
this error can frequently avoid unnecessary investigation.
Absolute polycythaemia
Absolute polycythaemias may be classified as being primarily due
to autonomous cell growth or to an enhanced response to growth
factors that promotes the proliferation of developing erythroid cells,
or secondary, due to excessive production of EPO in response to
a variety of stimuli. Primary polycythaemia, caused by defects in
haematopoietic stem cells, is accompanied, in general, by low levels
of circulating EPO. Germ-​line mutations of the EPO receptor that
lead to enhanced erythropoiesis cause primary familial congenital
polycythaemias. Polycythaemia vera, the most common primary
polycythaemia, is caused by an acquired defect in haematopoietic
stem cells resulting in an excessive proliferation of myeloid cells.
A mutation in the autoinhibitory domain of JAK2 tyrosine kinase
(V617F) has been associated with polycythaemia vera in up to 95%
of the cases; JAK2 exon 12 mutations make up the majority of the
remaining cases. Most recently LNK mutations have been described.
By contrast, secondary polycythaemia is generally caused by ele-
vated EPO production and is not associated with mutation in the
JAK2 tyrosine kinase. Elevated levels of plasma EPO can accompany
systemic hypoxaemia, certain neoplasms, and disorders that impair
oxygen delivery to tissues (Box 22.3.5.1).
Absolute polycythaemias are accompanied by an increased red
cell mass. Documentation of such an increased red cell mass may
require a blood volume study with direct determination of both the
red cell mass and plasma volume if available. This test is presently
available in select referral centres and the need for its performance
SECTION 22  Haematological disorders
5230
is limited to special situations due to the availability of JAK2 mu-
tation analysis. A haematocrit value greater than 60% in men and
greater than 55% in women, however, is almost always associated
with absolute erythrocytosis. In such cases, it is usually unnecessary
to perform blood volume studies to be assured that the patient has
an absolute polycythaemia. The presence of a JAK2 mutation with
intermediate elevations of haemoglobin or haematocrit elevations is
diagnostic of polycythaemia vera.
Secondary polycythaemias associated with appropriate
EPO secretion
This group of polycythaemias encompasses a number of conditions
that are ultimately the result of tissue hypoxia and subsequent exces-
sive EPO production leading to erythrocytosis. These disorders are
collectively regarded as hypoxic erythrocytoses.
Hypobaric hypoxia
At high altitudes, the barometric pressure and, consequently, the
ambient oxygen tension are reduced, resulting in alveolar and ar-
terial hypoxia. Natives of the Andes (South America) who live above
4200 m have been reported to have haematocrit values 30% higher
than individuals who live at sea level. Acutely, changes in minute
ventilation, heart rate, blood flow, and haemoglobin–​oxygen affinity
occur as an individual reaches a high altitude. Serum EPO is ele-
vated initially, but eventually returns to the normal range in the ab-
sence of extreme hypoxia. This decline will not prevent the increase
in red cell mass, which will be sustained, because early unsustained
elevations of EPO promote expansion of the erythroid progenitor
pool. Only very small quantities of the hormone are subsequently
required to sustain the red cell mass under normal circumstances.
Healthy highlanders develop pulmonary hypertension, right ven-
tricular hypertrophy, and increased amounts of smooth muscle
cells in distal pulmonary arterial branches, which leads to increased
pulmonary vascular resistance and pulmonary artery pressures as
compared to individuals who live at sea level. Due to these adap-
tive changes, healthy highlanders are able to perform physical ac-
tivities similar to or often even more strenuous than those living at
sea level. On the other hand, many individuals living at high alti-
tude do not demonstrate an increased haemoglobin concentration.
A  high-​frequency missense mutation in the EGLN1 gene, which
encodes prolyl hydroxylase domain 2 (PHD2), has been described
in Tibetans. This mutation reduces hypoxia-​induced erythropoiesis
secondary to HIFs that is otherwise the cause of erythrocytosis at
altitude elevations.
Chronic mountain sickness is a pathological loss of adaptation to
high altitude by highlanders that occurs in native or lifelong resi-
dents living above 2500 m. They suffer from headaches, fatigue,
impaired exercise tolerance, cyanosis, clubbing, right heart failure,
and absolute polycythaemia. These symptoms frequently resolve
with descent to lower altitudes. The prevalence of chronic moun-
tain sickness is higher in men than in women and increases with
altitude, ageing, associated lung disease, history of smoking, and air
pollution. The chronic response to high altitudes is probably deter-
mined by poorly defined genetic factors, which likely contribute to
the development of chronic mountain sickness. The major mech-
anism underlying chronic mountain sickness is relative hypoventi-
lation, since healthy highlanders characteristically hyperventilate.
Chronic mountain sickness is a common problem, affecting 6 to 8%
of males in La Paz (Peru), for instance. The definitive treatment of
chronic mountain sickness is descent to lower altitudes or sea level.
Phlebotomy and acetazolamide therapy are recommended for af-
fected individuals who must remain at high altitudes.
Chronic pulmonary disease
Pulmonary diseases are a common cause of secondary polycy-
thaemia. Defects in gas exchange result in hypoxia, with conse-
quent increases in EPO and red cell mass. Not every patient with
hypoxia secondary to respiratory disease develops polycythaemia;
however, the presence of concurrent inflammation or infection may
blunt the marrow response to hypoxia. It is important to be aware
that smoking itself may also contribute significantly to the polycy-
thaemia associated with chronic respiratory disease. Phlebotomy
may be indicated in patients with relatively high haematocrit levels
(55–​60%), given the known deleterious effects of hyperviscosity.
Chronic oxygen therapy in patients with severe chronic obstructive
pulmonary disease has resulted in relief of hypoxia and modest re-
ductions of haematocrit levels.
Smokers’/​hookah polycythaemia
Smoking even in the absence of chronic lung disease is associated
with secondary erythrocytosis. Increasingly recognized as well is the
association with chronic hookah use. These disorders are associated
with elevated carboxyhaemoglobin levels. Smoking cessation is both
diagnostic and therapeutic as the erythrocytosis will resolve.
Alveolar hypoventilation
Hypoventilation may lead to hypoxia and an EPO-​mediated in-
crease in red cell mass. These disorders include the sleep apnoea
syndrome and supine hypoventilation. Up to 25% of patients with
unexplained polycythaemia are subsequently found to have sleep
apnoea. Common symptoms include loud snoring, breathing
pauses, feelings of nonrefreshing sleep, and excessive daytime
sleeping. In these conditions, significant degrees of hypoxia may
occur without evident parenchymal pulmonary disease. Decreases
in blood oxygen content may occur intermittently; consequently,
EPO levels and arterial blood gas values may be normal. Diseases
affecting the central nervous system may also impair respiratory
centre function and trigger hypoventilation. These defects have
been described in association with encephalitis, cerebrovascular
accidents, and drug intoxication (e.g. barbiturates). Impaired skel-
etal muscle function of the chest wall or diaphragm may also suffi-
ciently compromise alveolar ventilation to trigger polycythaemia.
In these cases, correction of hypoxia is warranted. The role of phle-
botomy is unclear, but not unreasonable in patients with significant
elevations in haematocrit and associated cardiovascular or cere-
brovascular disease. The obesity–​hypoventilation syndrome seen
in morbidly obese individuals is characterized by chronic hypoxia
and hypercapnia due to alveolar hypoventilation with a resultant
increase in EPO production, polycythaemia, and cor pulmonale.
Effective treatment includes surgically induced weight loss, nasal
continuous positive airway pressure ventilation, and, occasionally,
the use of respiratory stimulants.
Cardiovascular disease
Cyanotic congenital heart diseases with an associated right-​to-​left
shunt result in oxygen desaturation and an elevation of EPO, causing
22.3.5  The polycythaemias
5231
secondary polycythaemia. After compensatory erythrocytosis in re-
sponse to oxygen desaturation occurs, serum EPO levels may return
to normal levels. An extremely elevated haematocrit may be detri-
mental to optimal oxygen delivery. Some children with congenital
heart disease may develop extreme haematocrit values (≥80%),
which leads to a significant risk of a thrombotic event, especially
during periods of dehydration due to hyperviscosity and sludging
within the microcirculation. The treatment of erythrocytosis in pa-
tients with cyanotic congenital heart disease is controversial and
should be individualized. Phlebotomy may be indicated in some
instances where it has been shown to improve cerebral blood flow
and neurological symptoms and to increase exercise capacity. To
date, there is no consensus defining the precise target haematocrit
values for therapeutic phlebotomy in the management of patients
with these disorders. It is also important to note that serial phle-
botomy will cause iron deficiency and subsequent microcytosis.
Microcytosis may actually worsen the symptoms of hyperviscosity
and a trial of iron replacement may be warranted.
Carbon monoxide intoxication
Chronic carbon monoxide intoxication most commonly occurs as
a consequence of smoking. Elevated haematocrits have been re-
ported in 3% of all smokers. Other less common causes include
work-​related exposures such as those seen in caisson workers, truck
drivers, or tunnel toll-​collectors. Carbon monoxide has a much
higher affinity for haemoglobin than oxygen does, thereby reducing
the amount of oxygen that can be bound and transported by haemo-
globin. It also shifts the oxygen–​haemoglobin dissociation curve to
the left, decreasing the ability of haemoglobin to release oxygen to
peripheral tissues. Furthermore, carbon monoxide impairs normal
compensatory mechanisms: carboxyhaemoglobin is known to de-
crease 2,3-​DPG production by red cells and to reduce the affinity
of haemoglobin for 2,3-​DPG. Polycythaemia due to chronic carbon
monoxide intoxication may be associated with an increased risk of
thromboembolic phenomena. Phlebotomy may be indicated in pa-
tients with very high haematocrit values (>55–​60%).
The decreased oxygen-​carrying capacity associated with carbon
monoxide intoxication is not detected by standard blood gas meas-
urements; therefore, a direct measure of carboxyhaemoglobin levels
is required. Morning carboxyhaemoglobin levels ranging from 4 to
20% have been reported. Individuals with chronic carbon monoxide
poisoning may experience neuropsychiatric and cardiac abnormal-
ities. The treatment is smoking cessation or removal of the patient
from the source of carbon monoxide.
High-​affinity haemoglobins
At least 50 haemoglobin variants exhibit increased avidity for
oxygen. These mutations are inherited in an autosomal dominant
manner. Oxygen transport by haemoglobin occurs as a function of
the oxygen–​haemoglobin affinity curve. This function is represented
by a sigmoid curve and is a reflection of the initial binding of oxygen
by deoxygenated haemoglobin occurring with significant difficulty.
As oxygen molecules are bound to normal haemoglobin, further
binding is facilitated by structural changes that occur in the haemo-
globin molecule. High-​affinity haemoglobin variants arise when
mutations alter key amino acid residues in regions of haemoglobin
that affect these rearrangements, or at the interface between α and
β chains. Another group of mutations induces changes in oxygen
affinity indirectly, by causing structural changes in haemoglobin re-
gions that are critical for the binding of 2,3-​DPG.
Increases in oxygen affinity result in a shift of the oxygen dissoci-
ation curve to the left. Consequently, haemoglobin binds oxygen
more readily and retains more oxygen at lower Po2 levels. This ultim-
ately results in decreased delivery of oxygen to tissues where capil-
lary Po2 is low (35–​45 mmHg). Mild tissue hypoxia then triggers an
increase in the production of EPO with consequent polycythaemia.
Oxygen affinity by variant haemoglobin is usually measured as the
P50o2, which represents the partial oxygen pressure at which 50% of
haemoglobin is saturated with oxygen. This analysis is necessary for
the identification of patients with high-​affinity haemoglobins. High-​
affinity haemoglobins are associated with lower than normal values
of P50o2; values below 17 mmHg are usually diagnostic of such an
abnormal haemoglobin. Haemoglobin electrophoresis or high-
performance liquid chromatography may, on occasion, aid in the
recognition of an abnormal haemoglobin, but many high-​affinity
haemoglobins display normal electrophoretic mobility or retention
times. Conversely, the presence of an abnormal band per se does
not provide information regarding oxygen affinity. A study of family
members is important, but a negative family history does not negate
the diagnosis since there is a high rate of spontaneous mutations.
Most patients with high-​affinity haemoglobins have mild poly-
cythaemia and are asymptomatic since the compensatory polycy-
thaemia results in normal oxygen delivery to tissues. Phlebotomy
therapy of such patients has been reported to be of no value and has
been shown to reduce exercise tolerance.
Methaemoglobinaemia
Hereditary methaemoglobinaemia may be associated with a mild
polycythaemia. Methaemoglobin results from the oxidation of ferrous
ions (Fe2+) to the ferric state (Fe3+). Oxygen does not bind reversibly
to methaemoglobin, resulting in a left shift of the oxygen dissociation
curve, impaired oxygen delivery, and chronic tissue hypoxia.
2,3-​DPG deficiency
This rare familial form of polycythaemia is due to a deficiency of the
enzyme biphosphoglycerate mutase. Deficiency leads to a decrease
in 2,3-​DPG, resulting in the increased affinity of oxygen to haemo-
globin, peripheral tissue hypoxia, and hypoxic erythrocytosis.
This disorder should be suspected in patients with familial polycy-
thaemia with a low P50o2 in the absence of a mutant haemoglobin.
Measurements of 2,3-​DPG in fresh red cell reveals reduced levels.
Chuvash polycythaemia
Chuvash polycythaemia is a recognized form of congenital and fa-
milial polycythaemia endemic to the Chuvash population of the
Russian Federation, which has also been reported to occur in a var-
iety of other racial and ethnic groups. The extreme elevations of
haemoglobin in this autosomal recessive disorder are accompanied
by increased EPO levels. Chuvash polycythaemia is caused by germ-​
line mutations in the von Hippel–​Lindau gene, most common being
the 598 C→T mutation. Stimulated EPO production is caused by the
inability of the mutated von Hippel–​Lindau gene to bind to HIF-​
1α, which escapes degradation and accumulates. These patients
present with isolated erythrocytosis without elevations of white
cell or platelet counts, have low blood pressure, varicose veins, and
vertebral haemangiomas probably due to increased levels of VEGF.
SECTION 22  Haematological disorders
5232
Mortality, often attributable to cerebral infarction or haemorrhage,
is higher than 25% by age 40. The roles of both aspirin and venesec-
tion in the treatment of Chuvash polycythaemia are unclear. There is
no effective therapy for reducing symptoms, complications, or mor-
tality. JAK2 inhibition has been investigated as a possible therapy
and is currently the subject of a clinical trial (NCT01730755).
Prolyl hydroxylase mutations
Mutations in the prolyl hydroxylases are inherited as an autosomal
dominant condition and are associated with failure of hydroxylation
of prolyl residues of HIF-​1α, increased HIF levels, and increased
EPO production leading to the development of erythrocytosis. In
contrast to the patients with Chuvash polycythaemia, patients with
mutations of prolyl hydroxylases do not have the clinical conse-
quences described previously.
Secondary polycythaemias associated with the
inappropriate secretion of EPO
Enhanced EPO levels and secretion occur in the absence of tissue
hypoxia in this group of disorders. The EPO response is therefore
inappropriate to systemic oxygen requirements.
Polycythaemia of renal disease
As the kidney is the principal site of EPO production, it is not sur-
prising that renal disorders may be associated with erythrocytosis or
anaemia. Patients with hypertension and renal artery stenosis have
a higher incidence of erythrocytosis than similarly hypertensive pa-
tients without renal artery disease. Other benign kidney diseases
associated with an increase in EPO production and erythrocytosis
include polycystic kidney disease (acquired or familial) and renal
cysts. Unusual patients with glomerulonephritis may also occasion-
ally present with an elevated haematocrit. An uncommon cause of
polycythaemia is Bartter’s syndrome, a hereditary tubular disorder
characterized by hypokalaemia secondary to renal potassium loss
in association with elevated plasma renin activity and aldosterone
secretion. Post-renal transplant polycythaemia is defined as a per-
sistently elevated haematocrit higher than 51% after renal trans-
plantation without an elevation of the white blood cell or platelet
count. Between 5 and 13% of patients have been reported to de-
velop erythrocytosis 8 to 24 months after renal transplantation. It
has been postulated that the excessive response to EPO from the
donor kidney in a patient with previously low EPO levels could
cause this elevation in haematocrit in these patients. Approximately
60% of patients with post-​transplant erythrocytosis experience
headaches, plethora, lethargy, and dizziness and approximately 10
to 20% develop thromboembolic complications. Retention of the
native kidney is essential for the development of post-​transplant
erythrocytosis with the native kidney overproducing EPO leading
to the development of erythrocytosis. Frequently the erythrocytosis
resolves with the removal of the kidney. Angiotensin-​converting en-
zyme inhibitors have proved useful in controlling post-​transplant
polycythaemia, but phlebotomy may still be required in patients
with haematocrit levels over 55 to 60% to rapidly decrease the risk of
thrombotic complications.
Tumour-​associated polycythaemia
A number of tumours are associated with an inappropriately in-
creased production of EPO, including benign and malignant
tumours of the kidney, hepatomas, cerebellar haemangioblastomas,
parathyroid adenomas, Leydig tumours, paraganglioma, men-
ingioma, and phaeochromocytomas. Polycythaemia occurs in 1%
of patients with renal carcinomas, 9 to 20% of patients with cere-
bellar haemangioblastomas, and 10% of patients with hepatomas.
Resection of the tumour, if feasible, may be associated with re-
gression of the polycythaemia. Therapeutic phlebotomy is recom-
mended in patients with extreme increases in the haematocrit.
Two plasma cell dyscrasias associated with erythrocytosis
are POEMS (polyneuropathy, organomegaly, endocrinopathy,
monoclonal protein, and skin changes) syndrome and TEMPI
(telangiectasias, erythrocytosis with elevated EPO levels, mono-
clonal gammopathy of unknown significance, perinephric fluid
collections, and intrapulmonary shunting) syndrome. Treatment
is directed at the underlying disease. In TEMPI syndrome, the
most effective treatment has been bortezomib which suggests that
the TEMPI syndrome is a consequence of the abnormal plasma
cell clone.
Endocrine disorders
Phaeochromocytomas and aldosterone-​producing adenomas have
been associated with increased levels of EPO. Mild forms of poly-
cythaemia have also been observed in some patients with Cushing’s
syndrome, probably related to marrow stimulation by steroid hor-
mones. Recombinant human EPO has been abused by athletes and
its use can lead to erythrocytosis. The recombinant form of EPO can
be distinguished from native EPO by its electrophoretic mobility.
Drug associated
In addition to exogenous administration of EPO-​stimulating agents,
androgen administration, either illicit or for hypogonadism, can
be associated with erythrocytosis. Cessation of the offending agent
will resolve the erythrocytosis. The development of tyrosine kinase
inhibitors in cancer, particularly renal cell carcinoma, has resulted
in the observation of erythrocytosis associated with sunitinib,
sorafenib, and axitinib, usually with normal or high EPO levels.
While the erythrocytosis associated with these agents may resolve
with time, in other cases, phlebotomy may be needed if the agents
require continuation.
Primary polycythaemia
Primary familial and congenital polycythaemia
Primary familial and congenital polycythaemia is an inherited form
of polycythaemia caused by mutations in the EPO receptor, thereby
resulting in the hypersensitivity of erythroid progenitor cells to EPO
and low serum EPO levels. Most of the disease-​causing mutations
result in truncation of the cytoplasmic C-​terminal portion of the
EPO receptor, which leads to constitutive activation of the receptor
due to loss of its negative regulatory domain. In the autosomal dom-
inant form of the disease, family members have plethora, headaches,
dizziness, nosebleeds, and exertional dyspnoea. These symptoms
resolve with phlebotomy and reduction of the haematocrit. Unlike
those with polycythaemia vera, primary familial and congenital
polycythaemia patients have normal platelet counts, are JAK2 V617F
negative, lack splenomegaly, and do not progress to acute leukaemia.
Not all cases of primary familial and congenital polycythaemia can
be attributed to the mutations of the EPO receptor (which occur in
22.3.5  The polycythaemias
5233
c.10–​20% of cases), suggesting that other genetic defects can lead to
a similar phenotype. An up-​to-​date database with primary and sec-
ondary causes of congenital erythrocytosis can be found at http://​
www.erythrocytosis.org as well as information regarding reference
laboratories that perform mutational analyses, which can be used to
establish these diagnoses.
Polycythaemia vera
Polycythaemia vera is a clonal, chronic progressive haematological
malignancy characterized by excessive proliferation of erythroid,
myeloid, and megakaryocytic elements in the bone marrow. The
other myeloproliferative neoplasms include essential thrombocyth-
aemia and primary myelofibrosis. The hallmark of polycythaemia
vera is an absolute form of erythrocytosis usually associated with
leucocytosis, thrombocytosis, splenomegaly, hypersensitivity of
haematopoietic progenitor cells, and the JAK2 V617F mutation. This
mutation accounts for the cytokine hypersensitivity, which charac-
terizes haematopoietic progenitors from patients with this disorder.
The use of recently developed molecular tools to detect JAK2 V617F
has already revolutionized the diagnosis of polycythaemia vera
(Table 22.3.5.1). This mutation is a recognized molecular target
for therapy. In contrast to those with other haematological malig-
nancies, patients suffering from polycythaemia vera may enjoy pro-
longed survival, provided that the excessive production of red cells
and platelets is controlled. The development of myelofibrosis and/​or
acute leukaemia occurs occasionally.
Epidemiology
Polycythaemia vera was thought to be a rare disorder, with an es-
timated yearly incidence in the Western world between 5 and 17
cases per million population. Recent data suggest that its preva-
lence might be higher than previously expected with rates of at
least 300 cases per million population being reported. The very
high association between JAK2 V617F and polycythaemia vera
will probably have an impact on future studies of the incidence
and prevalence of this disease, and could change our knowledge
of the average age of diagnosis. A very low incidence of two cases
per year per million population has been reported in Japan. These
differences suggest that environmental as well as genetic factors
might be important.
Polycythaemia vera is slightly more common in men than in
women, with a male/​female ratio of 1.2:1. The average age at diagnosis
is 60 years; it is very rare in individuals younger than 30 years of age.
Only a handful of cases have been reported during childhood.
Biological and molecular aspects
The exaggerated production of red cells, granulocytes, and platelets
in polycythaemia vera suggests that the fundamental defect occurs
at the level of the pluripotent haematopoietic stem cell. The clonal,
and thereby malignant, nature of polycythaemia vera was first es-
tablished by the cellular analysis of blood cell production in African
American women heterozygous for X-​linked glucose-​6-​phosphate
dehydrogenase isoenzymes. These results have been confirmed
using restriction fragment length polymorphisms of the active X
chromosomes.
In patients with polycythaemia vera, EPO concentrations often
fall below the levels observed in normal individuals. These low
levels persist even after repeated phlebotomies, suggesting that
excessive production of EPO is not a critical component in the
pathogenesis of this disorder. Using in vitro cell-​culture systems,
polycythaemia vera bone marrow cells can form erythroid col-
onies in the absence of EPO and are hypersensitive to EPO (en-
dogenous colony formation). Polycythaemia vera progenitor cells
are also hypersensitive to other cytokines such as stem-​cell factor,
interleukin-​3, and granulocyte–​macrophage colony-​stimulating
factor. The presence of JAK2 V617F mutations can now explain
both the hypersensitivity to EPO and to multiple other growth
factors which use the JAK2–​STAT5 signalling pathway. A valine
to phenylalanine substitution occurs within exon 14 of the JH2
domain of the JAK2 tyrosine kinase gene. In the normal kinase,
this domain has an inhibitory effect over the catalytic domain,
JH1. A mutation in the JH2 domain disrupts this autoinhibitory
effect and renders the kinase constitutively active, leading to a
constant downstream phosphorylation and substrate activation.
STATs are intracytoplasmic molecules that initiate gene tran-
scription. One of the target genes of STATs is BCL2L1, which ex-
presses an antiapoptotic protein, overexpressed in polycythaemia
vera and believed to play an important role in increased survival
of erythroid precursors and megakaryocytes. Overexpression of
JAK2 V617F in cell lines is associated with increased cellular pro-
liferation and cytokine hypersensitivity and its transfection into
marrow cells which are then transplanted into mice is associated
with the development of a clinical phenotype similar to polycy-
thaemia vera. Progression of the disease has been also shown to
Table 22.3.5.1  The 2016 WHO diagnostic criteria for polycythaemia vera
Major criteria
1
Evidence of elevated red cell mass:
Hgb >165 g/​litre (men), >160 g/​litre (women), or
Hct >49% in men, >48% in women, or
Increased red cell mass (>25% above mean normal predicted value)
2
Bone marrow biopsy showing hypercellularity for age with trilineage growth (panmyelosis) including prominent erythroid,
granulocytic, and megakaryocytic proliferation with pleomorphic, mature megakaryocytes (differences in size)
3
Presence of JAK2 V617F or JAK2 exon 12 mutation
Minor criterion
1
Subnormal serum erythropoietin level
Hct, haematocrit; Hgb, haemoglobin.
The diagnosis of polycythaemia vera is made in the presence of all three major criteria or the first two major criteria and the minor criterion. Bone marrow biopsy many not be
required in cases with sustained haemoglobin levels >185 g/​litre in men (haematocrit 55.5%) or 165 g/​litre in women (haematocrit 49.5%) if major criterion 3 and the minor criterion
are present.
Source data from Arber DA, et al. (2016). The 2016 revision to the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia. Blood, 127, 2391–​405.
SECTION 22  Haematological disorders
5234
correlate with the level of allele chimerism. Retrospective analyses
have revealed that patients with a low burden of the mutated al-
lele can evolve over time to a higher burden of the mutated allele.
Loss of heterozygosity on the short arm of chromosome 9 (the
location of the JAK2 gene) is a consequence not of gene deletion
but rather uniparental disomy or mitotic recombination. Even in
patients with a low burden of JAK2 V617F, erythroid progenitors
that are homozygous for JAK2 V6717F are usually present. This
finding is characteristic of polycythemia vera but occurs less fre-
quently in primary myelofibrosis and rarely in essential thrombo-
cythaemia. The burden of JAK2 V617F is correlated with disease
progression and the development of complications in polycy-
thaemia vera patients. Additional mutations of JAK2 that are as-
sociated with erythrocytosis have also been identified. Somatic
gain-​of-​function mutations involving exon 12 rather than exon
14 have been identified in patients with isolated erythrocytosis
and low serum EPO levels. All erythroid colonies cloned from
the haematopoietic cells of such patients are heterozygous for
this mutation. JAK2 exon 12 mutations could therefore identify
a distinctive myeloproliferative disorder that affects patients who
currently carry a diagnosis of idiopathic erythrocytosis. Finally,
mutations in LNK, primarily in exon 2, have been described in
patients with polycythaemia vera and other myeloproliferative
neoplasms lacking JAK2 mutations. LNK, or lymphocyte-​specific
adaptor protein, functions to inhibit wild-​type and mutant JAK2
phosphorylation. The exact mechanism by which mutations in
LNK result in polycythaemia vera is under investigation.
Pathobiology
Patients with polycythaemia vera have an increased thrombotic
tendency resulting from the expansion of the red cell mass which
represents the main cause of mortality in these patients. There is a
direct relationship between the risk of thrombosis and age, with the
incidence of cardiovascular complications being higher in patients
over 65 years of age. Younger individuals are also at risk for throm-
botic episodes, many of them life-​threatening, such as Budd–​Chiari
syndrome, cerebrovascular thrombosis, cerebral sinus thrombosis,
acute myocardial infarction, and pulmonary embolism. The main
rheological abnormality is elevation of the total blood viscosity.
Cerebral blood flow is reduced in patients with polycythaemia vera
and a haematocrit of 53 to 62%. Reductions in blood flow are cor-
rectable by phlebotomy. Even small reductions in the haematocrit
result in significant reductions in blood viscosity and increased
cerebral blood flow, thereby reducing the likelihood of thrombus
development. Thrombocytosis and functional platelet abnormalities
and increased white blood cell count (found in 50% of the patients)
are frequently present, and may play a role in the development of
thrombosis.
Patients with polycythaemia vera are also at an increased risk
of developing life-​threatening haemorrhagic complications.
Abnormalities in platelet function and number have been impli-
cated. Qualitative platelet abnormalities include defective platelet
aggregation in vitro, acquired storage pool disease, and dysregulated
thromboxane A2 metabolism. Acquired von Willebrand’s disease
has been described in patients who have very high platelet counts
(>1000 × 109/​litre), in association with life-​threatening bleeding
episodes. No laboratory test has proven useful for the a priori
identification of patients at an increased risk of developing haemor-
rhagic or thrombotic events.
The progression to a post-​polycythaemic-​related myelofibrosis
phase of the disease is a common cause of morbidity. This stage is
characterized by cytopenias, marrow fibrosis, and extramedullary
haematopoiesis. There are conflicting data on the incidence of this
complication. Some studies reported a very low incidence after 10 to
20 years; others reported that up to 25 to 50% of patients with polycy-
themia vera may develop polycythemia vera-​related myelofibrosis.
The fibroblastic component represents a reactive event, and may be
due to the local release of growth factors, particularly transforming
growth factor-​β, by haematopoietic cells. The association between
the treatment modality and the development of myelofibrosis is as
yet unclear. There is, however, an established association between
the treatment type (alkylating agents and radioactive phosphorus
(32P)), and the development of acute leukaemia. It must be empha-
sized, however, that even those patients treated with phlebotomy
alone have a leukaemogenic risk significantly higher than that ex-
pected in the general population.
Clinical manifestations
The clinical manifestations of polycythaemia vera are the direct
consequence of the excessive production of cellular elements of the
various haematopoietic cell lineages.
The routine and widespread use of laboratory screening tests
during medical evaluations has led to an increased detection of
asymptomatic patients. In contrast, symptomatic patients may
present to their physician with a large array of nonspecific com-
plaints including headache, weakness, pruritus, dizziness, exces-
sive sweating, visual disturbances, paraesthesias, joint symptoms,
and epigastric distress. Approximately one-​third of patients will
have lost 10% of their body weight by the time they come to med-
ical attention, presumably due to the associated hypermetabolism.
Joint disease is usually the manifestation of gout, due to the in-
creased production of uric acid. The most important signs on phys-
ical examination include ruddy cyanosis, conjunctival plethora,
splenomegaly, hepatomegaly, and hypertension.
Patients left without appropriate treatment are at a particu-
larly high risk of developing thrombotic or haemorrhagic events.
In fact, thrombosis may be the cause of death in up to 30 to 40%
of patients. Thrombosis may occur in the deep venous system
of the lower extremities, or present as a pulmonary embolism.
Cerebrovascular, coronary, and peripheral vascular occlusions are
not rare. Thromboses at unusual sites are also characteristic of poly-
cythaemia vera. They include occlusion of the splenic, portal, hep-
atic, and mesenteric veins. Cardiac valve abnormalities affecting
the aortic or the mitral valves are commonly seen, frequently in the
form of leaflet thickening or frank vegetations. These lesions are as-
sociated with the occurrence of arterial thromboembolism. Hepatic
venous or inferior vena caval thrombosis is known as Budd–​Chiari
syndrome and is characterized by hepatosplenomegaly, ascites, oe-
dema of the peripheral extremities, jaundice, abdominal pain, and
distension of superficial abdominal veins as a result of portal hyper-
tension. The prevalence of myeloproliferative neoplasms in patients
with splanchnic vein thrombosis was estimated to be as high as
49% for hepatic vein thrombosis and 23% for portal vein throm-
bosis. Often these patients will present with normal haemoglobin
22.3.5  The polycythaemias
5235
and haematocrit levels. This phenomenon is regarded as ‘inapparent
polycythaemia vera’ and requires a full evaluation for the presence
of myeloproliferative neoplasms. Detection of the JAK2 V617F muta-
tion will help identify and appropriately treat these patients earlier.
Up to 34% of patients with portal vein thrombosis and up to 58% of
patients with Budd–​Chiari syndrome may have a myeloproliferative
neoplasm, identified by the presence of JAK2 V617F. Iron defi-
ciency may also mask the expected erythrocytosis in some patients
with polycythaemia vera. ‘Masked’ polycythaemia vera may exist
even in those without Budd–​Chiari syndrome or iron deficiency
and describes those patients with erythrocytosis who do not meet
the cut-​offs set forward by diagnostic criteria but who have bone
marrow biopsies consistent with polycythaemia vera. They often
have thrombocytosis, and discriminating these patients from essen-
tial thrombocythemia is important as the thrombotic risks differ be-
tween these two diagnoses. The 2016 WHO diagnostic criteria have
set lower thresholds for haemoglobin and haematocrit compared to
the 2008 WHO diagnostic criteria, which should reduce the number
of patients that fall into this category. Leucocytosis, thrombocytosis,
and splenomegaly are usually present.
Neurological abnormalities occur in up to 60 to 80% of patients.
They include transient ischaemic attacks, cerebral infarction, cere-
bral haemorrhage, confusional states, fluctuating dementia, and in-
voluntary movement syndromes. Dizziness, paraesthesiae, tinnitus,
visual problems, and headaches are common symptoms attributed
to the hyperviscosity state. Small infarcts in the basal ganglia region,
also known as lacunae, may explain some of the transient neuro-
logical manifestations. Symptoms of carotid, vertebral, or basilar ar-
tery insufficiency occur frequently. Peripheral vascular insufficiency
may be manifested by intense redness or cyanosis of the digits,
burning, classical erythromelalgia, digital ischaemia with palpable
pulses, or thrombophlebitis. Erythromelalgia consists of a burning
pain in the digits of either the lower and/​or upper extremities, an ob-
jective sensation of increased temperature, and relief by cooling. If
left untreated it may evolve into gangrene. Antiplatelet aggregation
therapy rapidly reverses the symptoms. Peripheral pulses are usually
normal in these patients, as this phenomenon is due to changes in
the microcirculation related to arteriolar activation and aggregation
of platelets in vivo.
Haemorrhagic complications are the cause of death in 2 to 10%
of patients with polycythaemia vera; 30 to 40% of patients will ex-
perience a haemorrhagic event sometime during the course of their
disease. Peptic ulcer disease occurs frequently and contributes to the
gastrointestinal tract being the most common source of bleeding.
Oesophageal varices are another common site of bleeding in patients
with intra-​abdominal thromboses. The bleeding diathesis may relate
to abnormalities in platelet function, and thus occurs frequently after
the ingestion of anti-​inflammatory agents. Spontaneous bleeding is
rare. Recent data suggest that low-​dose aspirin might not increase
the frequency of life-​threatening haemorrhages.
Generalized pruritus affects 50% of all patients, but its aetiology
is unknown. Increased blood and urine histamine levels have been
implicated. Pruritus triggered by water contact is characteristic, and
very poorly tolerated. There is no relationship between the severity
of the disease and the intensity of the pruritus. Up to 20% of pa-
tients experience persistent pruritus even after normalization of
their counts.
The risk of postoperative complications is high in patients with
polycythaemia vera. Bleeding, thrombosis, or a combination of both
can occur. The risk is higher for those patients who undergo surgery
with uncontrolled erythrocytosis. Inadequately controlled disease
may be associated with almost an 80% risk of complications. The
duration of controlled blood counts is also important: the longer this
duration is, the less the risk of complications (as low as 5%).
Polycythaemia vera evolves to polycythaemia vera-​related
myelofibrosis in up to 50% of the patients 10 to 20 years after the
initial diagnosis. It is characterized by increased splenomegaly, tear-
drop red cells, a leucoerythroblastic blood picture, marrow fibrosis,
and a normal or decreasing red cell mass. Fatigue, dizziness, weight
loss, anorexia, progressive anaemia, and thrombocytopenia associ-
ated with bleeding are common. Patients with progressive anaemia
should be evaluated for folate and iron deficiency. Occasional pa-
tients will respond to iron supplementation with resurgence of
erythropoiesis. Severe hyperuricaemia may induce gout or uric acid
nephropathy. Polycythemia vera-​related myelofibrosis portends
a grave prognosis, with over two-​thirds of patients dying within
3 years. In the appropriate setting, strong consideration should be
given to allogeneic stem cell transplantation, which offers an oppor-
tunity for cure.
The evolution to acute leukaemia is probably the natural con-
sequence of the malignant nature of polycythaemia vera, which
can be accentuated by therapeutic interventions commonly
used for its treatment, such as alkylating agent or 32P. In a recent
study, older age and prior exposure to 32P and busulfan, but not
hydroxycarbamide therapy, was associated with an increased risk
of leukaemia. Between 30 and 50% of patients who develop leu-
kaemia have previously developed myelofibrosis whereas 50%
progress directly from the erythrocytotic phase. A  significant
number of patients experience a myelodysplastic interval before
transforming to acute leukaemia. Patients should be treated with
either decitabine or standard acute myeloid leukaemia induction
in preparation for taking these patients rapidly to allogenic stem
cell transplantation.
Laboratory evaluation
Laboratory evaluation of patents with erythrocytosis involves the
careful use of a battery of diagnostic tests. The diagnosis of poly-
cythaemia vera has been dramatically simplified with the advent
of the molecular tests for the JAK2 V617F mutation. The use of
analyses for the JAK2 V617F and JAK2 exon 12 mutations has
proven particularly useful in diagnostically challenging cases and
in patients with inapparent erythrocytosis and a serious thrombotic
event occurring especially at a younger age. Allele-​specific poly-
merase chain reaction methods can be used to detect JAK2 V617F
in at least 95% of patients with polycythemia vera. Those pa-
tients who fulfil clinical criteria for polycythemia vera but are
JAK2 V617F negative should be evaluated for JAK2 exon 12 mu-
tations. Approximately 10% of such patients who are negative for
JAK2 V617F harbour an exon 12 JAK2 mutation. Rare mutations
in calreticulin (CALR) and LNK have been described. Quantitation
of the red cell mass to document absolute erythrocytosis remains
a useful diagnostic test in characterizing patients without JAK2
mutations or erythrocytosis, though its availability is limited.
Approximately two-​thirds of the patients present with leucocytosis
SECTION 22  Haematological disorders
5236
and approximately 50% have thrombocytosis. Red cell morph-
ology usually reflects an underlying iron-​deficiency state pre-
sent in the great majority of patients: microcytosis, hypochromia,
polychromatophilia, poikilocytosis, and anisocytosis are frequently
seen. White blood cell morphology is usually normal. Increased
numbers of basophils, eosinophils, and immature myeloid cells
are observed. Megathrombocytes are often seen in the peripheral
blood smear. Platelet counts are usually less than 1000 × 109/​litre,
but higher counts may be seen. The progression to polycythaemia
vera-​related myelofibrosis is characterized by the appearance of a
leucoerythroblastic blood picture with the presence in the periph-
eral blood of teardrop red cells (dacrocytes), immature myeloid
cells, megathrombocytes, and nucleated red blood cells. Bleeding
time and platelet aggregation studies are frequently, but not always,
abnormal. Prolongation of prothrombin and partial thrombo-
plastin times are frequently encountered, usually reflecting a la-
boratory artefact due to erythrocytosis (the volume of plasma in
the collection tube might be too small relative to the amount of
citrate anticoagulant present in these tubes). At diagnosis, serum
EPO levels are either reduced or within the lower limits of normal.
Low levels persist in two-​thirds of patients after normalization of
the haematocrit.
In patients with extreme thrombocytosis, acquired von Willebrand’s
disease occurs, characterized by a significant decrease in large von
Willebrand’s factor multimers due to their adsorption to platelets and
megakaryocytes. This acquired defect occurs mainly in patients with
very high platelet counts (>1000 × 109/​litre) and resembles type 2 von
Willebrand’s disease. The defect is corrected by normalization of the
thrombocytosis. Elevations in leucocyte alkaline phosphatase (70%),
serum vitamin B12 levels (40%), and serum vitamin B12 binding pro-
teins (70%) are common, as are hyperuricaemia and increased hista-
mine levels.
Bone marrow examination reveals a hypercellular marrow with
trilineage growth (panmyelosis) with prominent erythroid, granulo-
cytic, and megakaryocytic proliferation with pleomorphic, mature
megakaryocytes. Iron stores are usually absent prior to treatment.
Reticulin is often seen, but is not predictive of evolution into the
myelofibrotic phase. Cytogenetic abnormalities have been observed
in 25% of patients, but none is characteristic. A recent study using
fluorescence in situ hybridization analyses has shown that abnor-
malities involving chromosome 9 rearrangements are common,
being present in up to 53% of patients with polycythaemia vera.
A gain in 9p is the most frequent genomic abnormality in polycy-
thaemia vera. JAK2 is located on 9p and the duplication of 9p in
polycythaemia vera is thought to be the consequence of homolo-
gous recombination. Other chromosomal abnormalities involving
chromosomes 1, 5, 7, 8, 12, and 13 have been associated with disease
progression.
Diagnostic criteria for polycythaemia vera
The diagnosis of polycythaemia vera has been greatly simplified
by JAK2 molecular testing. The 2016 World Health Organization
(WHO) diagnostic criteria for polycythaemia vera require the
presence of all three major criteria, (an increased haemoglobin,
haematocrit, or red cell mass; presence of a JAK2 mutation; and
consistent bone marrow biopsy findings) or the first two major cri-
teria plus a subnormal serum EPO level (Table 22.3.5.1). It is im-
portant to note that a bone marrow biopsy remains an important
procedure except in cases with a JAK2 mutation and subnormal
EPO level as well as significant erythrocytosis (haemoglobin levels
185 g/​litre in men (haematocrit 55.5%) or 165 g/​litre in women
(haematocrit 49.5%)).
Approach to the patient with polycythaemia
It is wise to avoid the temptation of diagnosing polycythaemia on
the basis of a single blood count unless extremely high haematocrit
levels are observed. A rational diagnostic approach is required to
avoid unnecessary emotional distress to the patient as well as ex-
pensive and unnecessary evaluations (Fig. 22.3.5.1).
Dehydration from any cause can produce a spurious elevation in
the blood counts. Heavy smokers with mild polycythaemia should
be asked to stop smoking and their counts repeated after a few weeks.
Once a genuine elevation of haemoglobin or haematocrit has been
established, the next step is to decide whether this represents an ab-
solute increase in total red cell mass, or merely a relative phenom-
enon. A blood volume study with direct quantitation of both red cell
mass and plasma volume can be helpful in making this distinction if
available. In patients with extreme degrees of erythrocytosis (haem-
atocrit >55% in men and >49.5% in women) one can be assured that
the red cell mass is elevated. If absolute polycythaemia is confirmed,
it is essential to elucidate whether it is the consequence of a primary
myeloproliferative disorder such as polycythaemia vera or a sec-
ondary condition.
The determination of EPO levels may be useful in differentiating
between polycythaemia vera and secondary polycythaemia. An ele-
vated serum EPO level is indicative of the presence of a secondary
polycythaemia and a low level supports the diagnosis of polycy-
thaemia vera, but a normal EPO value does not exclude hypoxia-​
induced causes of erythrocytosis or the autonomous production of
EPO leading to erythrocytosis. Normal values may also be encoun-
tered in some cases of polycythaemia vera.
The presence of leucocytosis, thrombocytosis, or splenomegaly
is suggestive of polycythaemia vera as the cause for the elevated
red cell mass. Arterial blood gases and the direct determination
of oxygen saturation in arterial blood, if decreased, may aid in
the recognition of a chronic pulmonary or congenital cardiovas-
cular abnormality. If blood oxygen saturation is normal, the quan-
tification of haemoglobin’s oxygen affinity (P50o2) may indicate
the presence of high-​affinity haemoglobin variant. Otherwise,
causes for a physiologically inappropriate polycythaemia should
be sought.
Molecular methods to detect JAK2 V617F and exon 12 JAK2
mutations provide diagnostic tools for the evaluation of patients
suspected of having polycythaemia vera. JAK2 mutation analysis
provides a direct means of identifying the overwhelming number of
patients with polycythemia vera.
There is a small but definite group of patients in whom a specific
cause for polycythaemia remains elusive, despite appropriate diag-
nostic testing. Examining close relatives might disclose the presence
of a familial form of polycythaemia, a rare condition caused by an
abnormality in EPO receptor or defects in hypoxia sensing (Chuvash
polycythemia). Regular, continued surveillance is recommended for
all noncategorized patients, as some of them develop polycythaemia
vera in the future.
22.3.5  The polycythaemias
5237
Management of polycythaemia vera
The two main goals in the management of patients with polycy-
thaemia vera involve the confirmation of the diagnosis and reduc-
tion of the red cell mass. The untoward effects of an increased red
cell mass on tissue blood flow occur independently from the spe-
cific cause of the polycythaemia. It is thus reasonable to recom-
mend that all patients with uncorrectable erythrocytosis be offered
phlebotomy.
The main therapeutic goals are the maintenance of well-​being
and the prevention of complications for as long as possible. Several
therapeutic strategies have resulted in dramatic increases in the
survival of patients. Historical evidence suggests a median survival
of approximately 18  months in untreated patients with polycy-
thaemia vera whereas with appropriate management, survival of
over 10 years is now common. The main therapeutic objective is the
reduction of the haematocrit to a normal level. This is usually ac-
complished by the implementation of repeated phlebotomies. Every
possible effort should be made to discourage patients with polycy-
thaemia vera from smoking. A regimen of phlebotomies should be
prescribed as soon as the diagnosis has been clearly established. It
is often feasible to remove between 350 and 500 ml of blood every
other day until the desired haematocrit level is attained. The re-
moval of smaller aliquots might be necessary in older patients. In
the landmark Cytoreductive Therapy in Polycythemia Vera (CYTO-​
PV) trial, stringent control of the haematocrit at less than 45%
versus more permissive control of 45 to 50% was associated with a
decreased risk of thrombosis, making lower than 45% the standard
of care. Many haematologists still target 42% for women, though this
is not based on prospective data.
Once the target haematocrit level is achieved, a maintenance
regimen should be instituted. Venesection is preferred in those
younger individuals without critical elevations in their platelet
counts. Myelosuppressive therapy should be considered in elderly
patients who are intolerant of phlebotomies, and in younger individ-
uals with repeated thrombotic episodes and extremely high platelet
counts. There is controversy regarding what represents the optimal
myelosuppressive agent. A  major concern has been the possible
­association between exposure to some of these agents and the devel-
opment of leukaemia.
Hydroxycarbamide is useful for the management of patients
with polycythaemia vera and represents the first-​line therapy espe-
cially in older patients in whom phlebotomy alone is insufficient or
intolerable, due to its minimal leukaemogenic potential. It should,
however, be used with great caution in patients formerly treated
with radioactive 32P or alkylating agents as the risk of leukaemia
is higher.
Low-​dose aspirin (81–​100 mg/​day) administered to patients
with polycythaemia vera has been shown to decrease the risk of ar-
terial and venous events. The European Collaboration on Low dose
Aspirin for Polycythaemia Vera (ECLAP) study was a randomized
study comparing low-​dose aspirin with placebo in reducing throm-
bosis in polycythaemia vera. A clinically significant reduction in rate
of thrombosis was seen in favour of low-​dose aspirin compared to
Peripheral blood mutation screening for JAK2 V16F and serum erythropoietin measurement; send JAK2
exon 12 if JAK2 V16F is negative
JAK2 mutation (+) AND serum EPO
level ↓
Hgb 16.5–18.5
g/dL in men, 16–
16.5 g/dL in
women OR Hct
49–55.5% in
men, 48–49.5%
in women
Bone marrow
biopsy required
for definitive PV
diagnosis
Diagnostic for
PV: Bone
marrow boipsy
not required
for diagnosis
Bone marrow biopsy
required for
definitive PV
diagnosis
If bone marrow not
diagnostic, consider
congenital
polycythaemia with
EPOR mutation
Consider secondary
polycythemia
including congenital
polycythaemia with
VHL mutation
Hgb >18.5 g/dL
in men, >16.5
g/dL in women
OR Hct >55.5%
in men, >49.5%
in women
PV likely
PV unlikely
Bone marrow biopsy
required for
definitive PV
diagnosis
PV possible
JAK2 mutation (+)
AND serum EPO level
normal/↑
JAK2 mutation (–)
AND
serum EPO level ↓
JAK2 mutation (–)
AND
serum EPO level ↑
Fig. 22.3.5.1  Diagnostic algorithm for patients with erythrocytosis. EPO, erythropoietin; EPOR, erythropoietin receptor;
Hct, haematocrit; Hgb, haemoglobin; PV, polycythaemia vera; VHL, von Hippel–​Lindau.
SECTION 22  Haematological disorders
5238
placebo. Although there is no overall survival benefit, no significant
increased risk of bleeding was reported.
In younger patients, given their potential long-​term survival,
strong consideration should be given to the use of phlebotomy
therapy in combination with low-​dose aspirin, as well as with other
apparently nonleukaemogenic interventions such as interferon-​α
and anagrelide. In two recent studies, one in polycythaemia vera
patients and another in essential thrombocythaemia patients,
hydroxyurea has been shown to be nonleukaemogenic, and can
be used as an alternative to phlebotomy or in combination with
it. Interferon-​α2a and pegylated forms of interferon have been
shown to be effective in controlling the blood counts and symp-
toms (especially pruritus). The use of pegylated forms of inter-
feron has been reported to be associated with a reduction of the
percentage of cells bearing the JAK2 V617F mutation, suggesting
its activity at the level of an early haematopoietic stem cell/​pro-
genitor cell. In patients with a history of thrombosis where un-
controlled thrombocytosis is a problem, anagrelide, an inhibitor of
megakaryocytic maturation, has proven effective. Ruxolitinib has
been evaluated in the phase III RESPONSE trial, which compared
best available therapy to ruxolitinib, with primary co-​endpoints
of haematocrit control (<45%) and reduction in spleen volume
of at least 35% by week 32. Best available therapy included thera-
peutic phlebotomy as well as various pharmacological interven-
tions including hydroxycarbamide at tolerable doses, interferon-​α,
immunomodulators, and anagrelide. In total, 20.9% of patients as-
signed to ruxolitinib treatment achieved the composite primary
endpoint (38.2% for spleen volume reduction; 60% for haemato-
crit control) compared to 0.9% of patients assigned to best avail-
able therapy (0.9% for spleen volume reduction and 19.6% for
haematocrit control) (P <0.001). Ruxolitinib therapy resulted in
a significantly increased proportion of patients with reduction in
symptom scores, measured by several validated tools including the
Myeloproliferative Neoplasm Symptom Assessment Form (MPN-​
SAF). It is unclear what impact ruxolitinib will have on thrombotic
risk and transformation to myelofibrosis or acute leukaemia, which
should become clearer with long-​term follow-​up and population-​
based studies. Refractory aquagenic pruritus can be managed in
the majority of cases with ruxolitinib.
At present, oral melphalan should be used only in patients who
are incapable of complying with the other forms of therapy or are
unwilling or unable to return for follow-​up.
Elective surgery should only be undertaken after adequate and
sustained control of the blood counts has been achieved. When
emergency surgery is required, the patient should be phlebotom-
ized rapidly until a normal haematocrit is achieved, and platelets
should be available in case excessive operative bleeding occurs.
Patients should be mobilized promptly, and the use of prophy-
lactic doses of low molecular weight heparin should be considered
unless contraindicated. Dental extractions are associated with an
increased bleeding risk and should only be pursued in patients
with good haematological control. A particularly high-​risk inter-
vention is splenectomy, which has an operative mortality rate of
approximately 9%. The haematocrit should be normalized and
platelet count maintained below 400 × 109/​litre before splen-
ectomy. After splenectomy there is a particular risk of extreme
thrombocytosis. Although no prospective studies have been done,
prophylactic low molecular weight heparin is probably warranted
in all patients undergoing splenectomy during the perioperative
period. Due to the high probability of expanding extramedullary
haematopoiesis with rapid development of liver enlargement after
splenectomy, patients should receive hydroxycarbamide therapy
postoperatively.
In patients with polycythaemia-​related myelofibrosis, the
management is quite similar to that for primary myelofibrosis.
Allogeneic stem cell transplantation is now a curative option for
patients with myelofibrosis, both primary and secondary to other
myeloproliferative neoplasms, and should be considered in appro-
priately selected patients.
Pregnant patients with polycythaemia vera experience an in-
creased incidence of fetal loss, with 30% of pregnancies culminating
in spontaneous abortions. Unexpectedly, pregnancy in patients with
polycythaemia vera is frequently associated with a gradual normal-
ization of blood values, and it is not unusual for a woman who has
required extensive therapy for control of her disease to no longer
require phlebotomies during pregnancy. The European Leukaemia
Net (ELN) has published recommendations about the management
of pregnant patients with myeloproliferative neoplasms, which in-
cludes risk stratification and treatment considerations. Based on
available data, the ELN classifies high-​risk pregnant patients with
essential thrombocythaemia/​polycythaemia vera as those with his-
tory of thrombosis or myeloproliferative neoplasm-​related haem-
orrhage as well as those with previous pregnancy complications,
such as intrauterine growth restriction, stillbirth or intrauterine
death, peripartum haemorrhage, severe pre-​eclampsia, multiple
first-​trimester miscarriages, placental abruption, and platelet
counts over 1500 × 109/​litre. Both low-​ and high-​risk essential
thrombocythaemia/​polycythaemia vera pregnant patients should
be treated with therapeutic phlebotomy to maintain the haemato-
crit at less than 45% or midgestation-​specific range, low-​dose as-
pirin, and provided with low molecular weight heparin for 6 weeks
postpartum. For women with previous thrombosis or pregnancy-​
related complications, low molecular weight heparin should be
extended throughout the pregnancy unless there were previous
bleeding issues. Additionally, interferon-​α has been considered the
most appropriate cytoreductive agent during pregnancy for women
with a history of thrombosis. The role of ruxolitinib in pregnancy
is not defined.
Prognosis
The outcome of patients with secondary polycythaemia is usually re-
lated to the prognosis of the underlying disorder. In polycythaemia
vera, the nature and severity of the complications during the clinical
course of the disease are the most important determinants of out-
come. Disease duration is also important, as long-​term survival is
strongly associated with progression to myelofibrosis or acute leu-
kaemia. As previously emphasized, prompt and appropriate therapy
results in dramatic improvements in survival. Young patients should
be initially managed with phlebotomy and low doses of aspirin.
Supplemental therapy with interferon, anagrelide, or hydroxyurea
might be required in patients with serious haemorrhagic or throm-
botic episodes. The use of either hydroxycarbamide or melphalan
appears warranted in the treatment of elderly patients who, because
of their age, have a limited survival.

22.3.6 Thrombocytosis and essential thrombocythaem
22.3.6 Thrombocytosis and essential thrombocythaemia 5239 Daniel Aruch and Ronald Hoffman
22.3.6  Thrombocytosis and essential thrombocythaemia
5239
FURTHER READING
Arber DA, et  al. (2016). The 2016 revision to the World Health
Organization (WHO) classification of myeloid neoplasms and acute
leukemia. Blood, 127, 2391–​405.
Barbui T, et al. (2014). Rethinking the diagnostic criteria of polycy-
themia vera. Leukemia, 28, 1191–​5.
Falanga A, et  al. (2005). Pathogenesis of thrombosis in essential
thrombocythemia and polycythemia vera: the role of neutrophils.
Semin Hematol, 42, 239–​47.
Hoffman R, et al. (2005). The polycythemias. In: Hoffman R, et al. (eds)
Hematology:  basic principles and practice, pp. 1209–​45. Churchill
Livingstone, Philadelphia.
Hoffman R, et  al. (2007). Philadelphia chromosome-​negative
myeloproliferative disorders:  biology and treatment. Biol Blood
Marrow Transplant, 13 Suppl 1, 64–​72.
James C, et  al. (2005). A JAK2 mutation in myeloproliferative dis-
orders:  pathogenesis and therapeutic and scientific prospects.
Trends Mol Med, 11, 546–​54.
James C, et al. (2005). A unique clonal JAK2 mutation leading to con-
stitutive signalling causes polycythemia vera. Nature, 434, 1144–​8.
Landolfi R, et al. (2004). Efficacy and safety of low-​dose aspirin in poly-
cythemia vera. N Engl J Med, 350, 114–​24.
Lasho T, et  al. (2010). LNK mutations in JAK2 mutation–​negative
erythrocytosis. N Engl J Med, 363, 1189–​90.
Levine RL, et al. (2007). Role of JAK2 in the pathogenesis and treat-
ment of myeloproliferative disorders. Nat Rev Cancer, 7, 673–​83.
Marchioli R, et al. (2013). Cardiovascular events and intensity of treat-
ment in polycythemia vera. N Engl J Med, 368, 22–​33.
Papayannopoulou T, et al. (2005). Biology of erythropoiesis, eryth-
roid differentiation, and maturation. In:  Hoffman R, et  al. (eds)
Hematology:  basic principles and practice, pp 267–​88. Churchill
Livingstone, Philadelphia.
Scott LM, et al. (2007). JAK2 exon 12 mutations in polycythemia vera
and idiopathic erythrocytosis. N Engl J Med, 356, 459–​68.
Silver RT (2006). Treatment of polycythemia vera. Semin Thromb
Hemost, 32, 437–​42.
Skoda R, Prchal JT (2005). Lessons from familial myeloproliferative
disorders. Semin Hematol, 42, 266–​73.
Spivak JL (2018). Polycythemia Vera. Curr Treat Options Oncol, 19(2), 12.
Spivak JL (2019). How I treat polycythemia vera. Blood, 134(4), 341–52.
Stein BL, et al. (2015). Polycythemia vera: an appraisal of the biology
and management 10 years after the discovery of JAK2 V617F. J Clin
Oncol, 33, 3953–​60.
Tefferi A, Barbui T (2015). Polycythemia vera and essential thrombo-
cythemia: 2015 update on diagnosis, risk-​stratification and manage-
ment. Am J Hematol, 90, 162–​73.
Tefferi A, Vardiman JW (2008). New insights into the pathogenesis
of JAK2 V617F-​positive myeloproliferative disorders and conse-
quences for the management of patients. Semin Thromb Hemost, 32,
341–​51.
Tefferi A, et al. (2013). Survival and prognosis among 1545 patients
with contemporary polycythemia vera:  an international study.
Leukemia, 27, 1874–​81.
Vainchenker W, Kralovics R (2017). Genetic basis and molecular
pathophysiology of classical myeloproliferative neoplasms. Blood,
129(6), 667–79.
Vannucchi AM, et al. (2015). Ruxolitinib versus standard therapy for
the treatment of polycythemia vera. N Engl J Med, 372, 426–​35.
Zhao ZJ, et al. (2005). Role of tyrosine kinases and phosphatases in
polycythemia vera. Semin Hematol, 42, 221–​9.
22.3.6  Thrombocytosis and
essential thrombocythaemia
Daniel Aruch and Ronald Hoffman
ESSENTIALS
The term thrombocytosis refers to a platelet count elevated above
450 × 109/​litre, which can be (1)  primary—​including essential
thrombocythaemia, chronic myeloid leukaemia, polycythaemia
vera, and myelodysplastic syndromes; or (2) secondary—​including
iron deficiency, infection, blood loss, and malignancy.
Normal megakaryocytopoiesis
Platelets are released from megakaryocytes, whose development is
principally regulated by thrombopoietin. This is chiefly produced
in the liver and binds to its receptor (the thrombopoietin receptor,
MPL) to cause activation via the JAK–​STAT signalling pathway at dif-
ferent levels along the platelet production pathway, ranging from the
proliferation and survival of haematopoietic stem cell/​progenitor
cells to megakaryocyte maturation. Thrombopoietin production
is increased by a wide variety of stimuli, which explains the many
causes of secondary thrombocytosis.
Essential thrombocythaemia
Aetiology—​the JAK2 V617F missense mutation typical of polycy-
thaemia vera (see Chapter 22.3.5) is found in about 50% of cases.
In addition, 10% of patients have a mutation in the thrombopoietin
receptor gene, MPL, and 30% have a mutation in calreticulin (CALR).
Approximately 10% of patients have none of these mutations and are
referred to as having ‘triple negative’ essential thrombocythaemia.
Clinical features—​many patients (usually in late middle age) are
asymptomatic at diagnosis, but common manifestations include
(1) thrombotic episodes: (a) venous thromboses, including of the hep-
atic veins; (b) arterial thromboses, including stroke, myocardial infarc-
tion, transient ischaemic attacks, erythromelalgia (redness and burning
pain in the extremities), and (occasionally) frank arterial thrombosis with
gangrene; (2) bleeding episodes; and (3) moderate splenic enlargement.
Diagnosis—​this requires all of the following four major criteria: (1)
platelet count greater than 450 × 109/​litre; (2)  bone marrow bi-
opsy showing proliferation mainly of the megakaryocyte lineage
with increased numbers of enlarged, mature megakaryocytes with
hyperlobulated nuclei without a significant increase or left shift in
neutrophil granulopoiesis or erythropoiesis and very rarely minor
(grade 1) increase in reticulin fibres; (3) failure to meet the criteria for
other myeloproliferative neoplasms; and (4) presence of JAK2, CALR,
or MPL mutations. Alternatively, diagnosis can be met when the first
three major criteria are present and the one minor criterion, namely
the presence of another clonal marker or absence of evidence for
reactive thrombocytosis.
Treatment—​this requires risk stratification based on the age of the
patient and any prior history of thrombosis, with treatment being
reserved for those at a high risk of developing complications and
not introduced simply on the basis of platelet counts alone un-
less there is extreme thrombocytosis (>1500 × 109/​litre). Therapies
SECTION 22  Haematological disorders
5240
include (1) low-​dose aspirin—​should be considered in patients with
platelet counts less than 1000 × 109/​litre and without evidence
of acquired von Willebrand’s disease; and (2)  cytoreduction—​
hydroxycarbamide effectively reduces platelet counts and throm-
botic episodes in high-​risk patients; interferon-​α, anagrelide, and
other agents are also used.
Prognosis—​most patients will survive more than 10  years from
diagnosis. Most deaths result from thrombotic complications.
Introduction
Thrombocytosis refers to a platelet count elevated above the ac-
cepted normal range (>450 × 109/​litre). The widespread use of auto-
mated cell counters has made the identification of platelet count
abnormalities a relatively common event. The clinical consequences
of elevated platelet counts are usually determined by the cause of
the thrombocytosis, ranging from the uneventful recognition of
a laboratory abnormality, to medical emergencies such as life-​
threatening thrombosis or haemorrhage.
Normal megakaryocytopoiesis
An understanding of disorders of platelet production requires
knowledge of the regulatory events that occur during normal
megakaryocytopoiesis. Megakaryocyte development is a complex
process in which a wide variety of regulatory signals work in con-
cert to direct a highly specific response to thrombopoietic demand.
A large number of cytokines including interleukins (IL-​3, IL-​6, and
IL-​11), stem cell factor, granulocyte–​macrophage colony-​stimulating
factor, thrombopoietin, and, possibly, erythropoietin have been
shown to stimulate megakaryocyte development. Thrombopoietin
and the thrombopoietin receptor (MPL) are the primary physio-
logical regulators of in vivo megakaryocytopoiesis. Thrombopoietin
is produced primarily by the liver, but its mRNA has also been found
in the kidney, muscle, and bone marrow. Thrombopoietin acts at dif-
ferent levels of megakaryocyte maturation ranging from the prolif-
eration and survival of haematopoietic stem cells/​progenitor cells
to megakaryocyte maturation, but does not significantly affect the
release of platelets from megakaryocytes; thrombopoietin levels are
regulated by the total mass of platelets and megakaryocytes, and
thrombopoietin is cleared by binding to receptors on the surface
of these cells. Like erythropoietin, thrombopoietin uses the JAK–​
STAT signalling pathway (see Chapter 22.3.5). Activation of the re-
ceptor MPL by thrombopoietin provokes a conformational change
of the JAK2 tyrosine kinase, phosphorylation of intracytoplasmic
residues, and downstream activation of the genes controlling cell
cycle status, differentiation, and apoptosis. A mutation in the 617
position of the JAK2 protein replacing the amino acid phenylalanine
with valine disrupts the autoinhibitory domain of JAK2 and renders
the kinase constitutively active. This mutation, JAK2 V617F, is pre-
sent in the vast majority of patients with polycythaemia vera and
in approximately 50% of patients with essential thrombocythaemia.
Additional mutations found in essential thrombocythaemia include
activating mutations in MPL and CALR. Gain-​of-​function deletions
or insertional mutations of CALR exon 9 result in a mutant CALR
protein preferentially associating with MPL that is bound to JAK2,
which drives its phosphorylation.
During times of thrombopoietic stress, there is increased pro-
duction of thrombopoietin by the spleen and bone marrow.
Inappropriately elevated levels of thrombopoietin may be observed
in essential thrombocythaemia. This is probably not due to exces-
sive production but rather impaired thrombopoietin clearance as-
sociated with decreased expression of the thrombopoietin receptor
by megakaryocytes and platelets. Molecular abnormalities in the
thrombopoietin gene, however, have been identified in several fam-
ilies with an autosomal dominant form of hereditary thrombocytosis
where serum thrombopoietin levels are significantly elevated. This
syndrome has been shown to be due to a mutation in a portion of
the thrombopoietin gene, which plays a crucial role in regulating its
expression.
Pathophysiology and classification
of thrombocytosis
Thrombocytosis can occur in response to many underlying clinical
conditions (secondary or reactive), or as a consequence of a pri-
mary abnormality in bone marrow function (primary). A classifi-
cation of the causes of thrombocytosis is provided in Box 22.3.6.1.
Reactive or secondary thrombocytosis accounts for over 80% of all
recognized cases of thrombocytosis, iron deficiency being the most
common cause. Short-​lived, secondary thrombocytosis may be ob-
served in situations such as trauma, acute bleeding, major surgery,
or after strenuous physical exercise. Longer-​term thrombocytosis
Box 22.3.6.1  Classification of the causes of thrombocytosis
•	 Autosomal dominant familial thrombocytosis
•	 Secondary thrombocytosis (reactive):
—​	 Iron deficiency
—​	 Infection
—​	 Postsplenectomy (or hyposplenism)
—​	 Malignancy
—​	 Trauma
—​	 Inflammation (noninfectious)
—​	 Blood loss
—​	 Major surgery
—​	 Exercise
—​	 Rebound from myelosuppression
•	 Primary thrombocytosis (nonreactive):
—​	 Essential thrombocythaemia
—​	 Chronic myeloid leukaemia
—​	 Polycythaemia vera
—​	 Primary myelofibrosis
—​	 Unclassified myeloproliferative neoplasms
—​	 Myelodysplastic syndromes
—​	 Refractory anaemia with ringed sideroblasts and thrombocytosis
(RARS-​T)
•	 Uncertain aetiology
Adapted from The American Journal of Medicine, Vol. 96, Buss DH, et  al.,
Occurrence, etiology, and clinical significance of extreme thrombocytosis:
A study of 280 cases, Pages 247–​53, Copyright © 1994, with permission
from Elsevier.
22.3.6  Thrombocytosis and essential thrombocythaemia
5241
is associated with the presence of chronic disorders such as malig-
nancy, inflammation, chronic infections, and iron deficiency an-
aemia. The pathophysiology underlying reactive thrombocytosis is
not fully understood, but probably involves the increased generation
of inflammatory cytokines such as IL-​6, which appear to mediate in-
creased transcription of thrombopoietin by the liver.
Primary thrombocytosis by contrast is associated with a group
of bone marrow disorders including chronic myeloid leukaemia,
essential
thrombocythaemia,
polycythaemia
vera,
primary
myelofibrosis, and the myelodysplastic syndromes. The level of ele-
vation of platelet numbers is not helpful in differentiating a reactive
from a primary process.
An abnormality of thrombopoietin production or of the
thrombopoietin receptor has been suggested as the basis of sev-
eral familial disorders associated with thrombocytosis. In several
families, a point mutation of the thrombopoietin gene leads to
overproduction of thrombopoietin resulting in elevated levels of
thrombopoietin and thrombocytosis. Patients with this autosomal
dominant form of familial thrombocytosis have a benign course
which is not complicated by thrombosis or haemorrhage or the de-
velopment of acute leukaemia or myelofibrosis. A second familial
form of thrombocytosis has been attributed to a mutation in the
transmembrane mutation of MPL leading to its constitutive activa-
tion. These familial forms of thrombocytosis are the consequence
of germ-​line mutations while the myeloproliferative neoplasms are
the consequence of acquired somatic mutations. A third familial
form of thrombocytosis has been described with germline JAK2
mutations; some, but not all, of these mutations not only result in
thrombocytosis but additionally result in vascular events. Lastly,
gelsolin, which is a protein involved in actin assembly and disas-
sembly, is currently being evaluated as another familial cause of iso-
lated thrombocytosis; the mechanism by which a mutation in this
gene results in thrombocytosis is unknown.
For the most part, the underlying medical disorder leading to re-
active thrombocytosis can be identified by clinical criteria. A number
of laboratory tests can be useful in distinguishing primary from sec-
ondary thrombocytosis. C-​reactive protein synthesis in the liver is
mediated by IL-​6, with C-​reactive protein levels being high in those
patients with elevated IL-​6 levels. Elevated levels of both IL-​6 and C-​
reactive protein are strongly indicative of the elevated platelet count
being reactive in origin. Cytogenetic analyses and use of the poly-
merase chain reaction for the BCR-​ABL1 translocation are useful to
exclude the presence of a Philadelphia chromosome and a diagnosis
of chronic myeloid leukaemia in a patient with thrombocytosis.
Assays using probes for restriction fragment polymorphisms of
genes located on the X chromosome are helpful in identifying clonal
haematopoiesis in females with thrombocytosis. Clonal haemato-
poiesis occurs in patients with myeloid malignancies such as essen-
tial thrombocythaemia but not in cells of patients with secondary
forms of thrombocytosis. More recently, the presence or absence
of the JAK2 V617F mutation, calreticulin mutations, and/​or MPL
mutations (MPL W515L, MPL W515K, and MPL W515N) have
been used to differentiate secondary cases of thrombocytosis from
a Philadelphia-​negative myeloproliferative neoplasm leading to ele-
vated platelet numbers. The presence of these mutations in a patient
with thrombocytosis is diagnostic of a myeloproliferative neoplasm.
The natural history and prognosis of reactive thrombocytosis
is defined by its underlying cause. The thrombocytosis per se is
probably inconsequential and does not require specific therapy; it
usually resolves after the treatment of the underlying cause. In con-
trast, the thrombocytosis due to underlying myeloproliferative neo-
plasm can cause life-​threatening thromboembolic phenomena and
bleeding episodes, and frequently requires specific cytoreductive
therapy, emphasizing the need for accurate recognition.
Essential thrombocythaemia
Essential thrombocythaemia is a chronic myeloproliferative neo-
plasm characterized by marked bone marrow megakaryocytic
hyperplasia and peripheral blood thrombocytosis. The clinical
course is punctuated by episodes of thrombosis and/​or bleeding.
In 1951, Dameshek suggested that essential thrombocythaemia
represented a myeloproliferative disease. The myeloproliferative
neoplasms are currently thought to represent malignant stem cell
disorders.
Aetiology and pathogenesis
The causative factors which lead to essential thrombocythaemia
have become increasingly better understood. Its pathogenesis in-
volves the abnormal proliferation of a blood cell precursor that
differentiates mainly towards the megakaryocytic/​platelet lineage.
Current evidence suggests that hypersensitivity to stimulatory
cytokines such as thrombopoietin might provoke the expansion
of the megakaryocytic progenitor pool. The clonal origin of haem-
atopoiesis in patients with myeloproliferative neoplasms was ini-
tially established through biochemical isoenzyme characterization
of the blood cells of affected women who were heterozygous for
glucose-​6-​phosphate dehydrogenase. Analysis of X-​linked re-
striction fragment length polymorphisms in affected women has
confirmed a clonal pattern in some cases. There are, however, a
significant number of patients with polyclonal myelopoiesis. These
nonclonal cases may have a decreased risk for thrombosis. The
finding of the JAK2 V617F mutation in Philadelphia chromosome-​
negative myeloproliferative neoplasms has provided new insight
into the pathogenesis of this disease. Approximately 50% of pa-
tients with essential thrombocythaemia are JAK2 V617F positive.
The patients who are positive for the mutation almost uniformly
have a low burden of JAK2 V617F (<50%) as compared with poly-
cythaemia vera. Essential thrombocythaemia patients with a high
allele burden are older, have more symptoms (especially aquagenic
pruritus), a larger spleen volume, and significantly higher rate of
cardiovascular complications. Although polycythaemia vera pa-
tients almost always have haematopoietic progenitors that are
homozygous for the JAK2 V617F mutation, such homozygous pro-
genitor cells are only occasionally observed in essential thrombo-
cythaemia patients. These data indicate that the homologous
recombination step which leads to mutational homozygosity in
polycythaemia vera rarely occurs in essential thrombocythaemia.
Furthermore, the degree of mutant allelic chimerism remains
constant over time. Patients with JAK2 V617F-​positive essential
thrombocythaemia have a higher haematocrit, higher white blood
cell count, and higher rate of transformation to polycythaemia vera
than patients who are JAK2 V617F-negative. These findings have
led some investigators to suggest that JAK2 V617F-​positive essen-
tial thrombocythaemia represents a forme fruste of polycythaemia
SECTION 22  Haematological disorders
5242
vera. Acquired mutations of the thrombopoietin receptor MPL at
position 515 have been observed in 4 to 5% of patients with es-
sential thrombocythaemia and 9% of patients with JAK2 V617F-​
negative essential thrombocythaemia. A number of patients have
been shown to possess both the MPL and JAK2 V617F mutations.
Although the MPL mutations were first observed in patients with
primary myelofibrosis, they are now known to also occur in pa-
tients with essential thrombocythaemia. MPL mutant-​positive
patients have lower haemoglobin levels but higher platelet counts
than essential thrombocythaemia patients who do not have this
mutation. More recently, CALR mutations have been described in
about 30 to 40% of patients with essential thrombocythaemia. The
normal function of calreticulin is to ensure appropriate binding
of newly synthesized glycoproteins within the endoplasmic re-
ticulum and regular calcium homeostasis. Over 50 types of mu-
tations have been observed in essential thrombocythaemia and
primary myelofibrosis, which result in calreticulin losing its
calcium-​binding and endoplasmic reticulin retention domains.
The resultant mutant protein interacts directly with MPL, keeping
it active and thereby activating a downstream JAK–​STAT pathway.
Two primary types of CALR mutations have been described. Type
1 mutations account for 65% of those observed and are charac-
terized by a 52-​bp deletion while type 2 mutations account for
32% and are characterized by 5-​bp insertion. The remaining mu-
tations are referred to as type 3, and account for the remaining 3%.
Patients with type 1 mutations are associated with a higher risk of
myelofibrotic transformation while type 2 mutations have a more
indolent course as well as a lower risk of thrombosis despite very
high platelet counts.
CALR mutations are not typically found in patients with MPL
or JAK2 mutations. Together these three mutations account for
over 90% of essential thrombocythaemia cases. However, in 10%
of essential thrombocythaemia, the driver mutation is currently
unknown and these patients are referred to as ‘triple negative’.
Recently the disease-​causing mutations in such triple-​negative
cases have been examined using whole-​exome sequencing.
Noncanonical mutations of MPL and JAK2, outside of the exons
usually examined for diagnostic purposes, were present in 18.9%
of cases. All the newly identified JAK2 mutations lead to constitu-
tive activation of the JAK–​STAT signalling pathway. The inability
to detect mutations in the remainder of such triple-​negative pa-
tients could be due to the technical limitations of the whole-​exome
sequencing or the possibility that such individuals have a heredi-
tary form of thrombocytosis.
A recent study of essential thrombocythaemia patients in a
paediatric population revealed that haematopoiesis was more often
polyclonal and JAK2 V617F-negative. As compared with the adult
population, about 20% of such paediatric cases had monoclonal
haematopoiesis and JAK2 V617F positivity was significantly less
frequent than in adults (20% vs 50–​60%). CALR exon 9 mutations
are observed in an additional 10% of patients less than 18 years of
age. Although one may hypothesize that clonal haematopoiesis in
nonclonal essential thrombocythaemia patients may become more
apparent with age, no significant evidence has been provided so far
to support the transition from nonclonal to clonal haematopoiesis.
Furthermore, patients who are negative for JAK2 V617F or MPL
mutations at presentation have not been observed to acquire the
mutation over time.
Epidemiology
The true incidence of essential thrombocythaemia is unknown due
to the lack of large epidemiological studies. Several smaller studies
estimated the incidence of essential thrombocythaemia to be 1.5 to
2.4 patients per 100 000 population annually. Approximately 6000
new cases are identified each year in the United States of America.
There seems to be a slight female predominance and the usual age
at onset is between 50 and 60 years. Approximately 20% of all cases
occur in individuals younger than 40 years, but it is very rarely seen
during childhood.
Pathobiology
The characteristic clinical features are dominated by the thrombo­
cytosis and abnormalities in platelet function. The association be-
tween increased numbers of circulating platelets and ischaemic
episodes remains unclear, but the duration of thrombocytosis may
play a role. Microvascular thrombosis results in a variety of clinical
syndromes associated with digital and cerebrovascular ischaemia.
Abnormalities in platelet function occur in 35 to 100% of patients,
and prolongation of the bleeding time occurs in 7 to 19%. Despite
being common, these abnormalities are poor predictors of bleeding
and/​or thrombotic risk. This is in contrast to the acquired von
Willebrand’s disease and erythromelalgia, clinical entities not in-
frequently seen in association with essential thrombocythaemia. In
acquired von Willebrand’s disease, extreme thrombocytosis (>1000
× 109/​litre) induces the adsorption of larger von Willebrand’s
multimers on to platelet membranes, with their subsequent degrad-
ation, triggering a haemostatic defect quite similar to that observed
in type 2 von Willebrand’s disease.
Erythromelalgia occurs commonly in patients with essential
thrombocythaemia. Erythromelalgia refers to a syndrome charac-
terized by redness and burning pain in the extremities which results
from platelet-​mediated thrombosis of the arterial microvasculature.
If left untreated it may progress to frank gangrene. The exquisite
platelet response to cyclooxygenase inhibitors such as aspirin and
indomethacin suggests that prostaglandin endoperoxides produced
by the metabolism of arachidonic acid might play a major role in the
generation of platelet-​associated thrombosis.
Increased frequency of venous thrombosis in uncommon sites
such as the splanchnic vasculature leading to catastrophic intra-​
abdominal thromboses such as Budd–​Chiari syndrome have
recently been reported in JAK2 V617F-​positive patients who sub-
sequently go on to develop essential thrombocythaemia. Although
the increased thrombotic risk cannot be explained exclusively by the
presence of the JAK2 V617F mutation, it appears to contribute to the
increased risk of thrombosis in these patients.
Clinical manifestations
As many as two-​thirds of patients with essential thrombocythaemia
are asymptomatic at diagnosis. Most symptomatic patients pre-
sent with either a thrombotic episode or a minor bleeding episode.
Bleeding can occur spontaneously but is frequently associated with
the recent use of a nonsteroidal anti-​inflammatory drug (NSAID).
Common sites of haemorrhage include the gastrointestinal and the
genitourinary tracts; there is also easy bruising. Thrombosis leads to
the most common presenting symptoms and can occur in arteries
and veins, large or small. Occlusion of the splanchnic vessels and of
22.3.6  Thrombocytosis and essential thrombocythaemia
5243
the superficial and deep veins of the lower extremities is common.
Pulmonary emboli may also occur. An occasional patient presents
with thrombosis of the hepatic veins causing the Budd–​Chiari syn-
drome or with occlusion of the renal veins manifesting clinically as
nephrotic syndrome.
When the microcirculation is involved, a number of clinical syn-
dromes may occur. Palpable lesions with small areas of gangrene
indistinguishable from vasculitic lesions of rheumatoid arthritis or
systemic lupus erythematosus may be observed. Erythromelalgia
may occur in association with transient ischaemic attacks or acute
episodes of cardiac angina. Peripheral pulses are usually preserved;
this helps differentiate erythromelalgia from atherosclerotic-​related
ischaemia.
Neurological symptoms are common and include headaches
and paraesthesias of the extremities. Transient ischaemic attacks
may present with symptoms of unsteadiness, dysarthria, dysphoria,
motor hemiparesis, scintillating scotomas, amaurosis fugax, vertigo,
dizziness, migraine headaches, and seizures. On occasion, transient
ischaemic attacks may progress to established infarcts.
Myocardial ischaemia with normal angiograms occurs occasion-
ally. Splenic enlargement is observed in 40 to 50% of individuals and
20% have hepatic enlargement.
Laboratory evaluation
An elevated platelet count, often above 450 to 1000 × 109/​litre, is
characteristic. The absolute number of platelets, even if higher than
1000 × 109/​litre, is not diagnostic of essential thrombocythaemia, as
extreme elevations in platelet numbers may be observed in reactive
thrombocytosis. Marked changes in platelet morphology, which
include large and bizarre-​looking platelets sometimes forming ag-
gregates, are also characteristic and may be more useful in helping
distinguishing primary from reactive thrombocytosis. The bone
marrow is hypercellular with megakaryocytic hyperplasia. Clusters
of hyperlobulated megakaryocytes are often observed within the
marrow. Absent or diminished iron stores are seen frequently. This
may be an epiphenomenon of an underlying myeloproliferative neo-
plasm or a true expression of iron depletion in patients with chronic
bleeding. Reticulin fibrosis is present in one-​quarter of bone marrow
specimens but collagen is limited. Mild leucocytosis is common.
Molecular analysis for JAK2 V617F, CALR, and MPL muta-
tions is an important diagnostic tool in identifying patients with
myeloproliferative neoplasms. If thrombocytosis associated with
megakaryocytic hyperplasia, and a JAK2 V617F, CALR, or MPL
mutation is observed in the absence of the clinical or laboratory
features of one of the other myeloproliferative neoplasms such as
polycythaemia vera or primary myelofibrosis, a diagnosis of es-
sential thrombocythaemia is certain. Unfortunately, for the other
10% of the patients with essential thrombocythaemia who lack the
above-​mentioned mutations the diagnosis remains one of exclusion,
although haematopoietic cell clonality assays are frequently useful in
women. Such patients who are thought to have essential thrombo-
cythaemia based upon marrow histopathology in the absence of
the formerly mentioned three mutations are referred to as triple-​
negative essential thrombocythaemia.
Platelet function abnormalities are commonly found and include
defective platelet aggregation in response to adrenaline, ADP, and col-
lagen. Aggregation in response to arachidonic acid and ristocetin is
often normal. An acquired platelet storage pool disease also occurs
due to abnormalities in the content and release of α granules associated
with a state of increased platelet activation. Cytogenetic evidence for a
Philadelphia chromosome and/​or the molecular identification of the
BCR-​ABL1 fusion gene aids in distinguishing essential thrombocyth-
aemia from chronic myeloid leukaemia. The presence of dyspoietic
changes in bone marrow precursor cells and of characteristic chromo-
somal abnormalities suggests the diagnosis of myelodysplasia. In par-
ticular, the 5q–​ syndrome is associated with thrombocytosis. More
recently, mutations in splicing genes (e.g. SF3B1) have been described
in refractory anaemia with ringed sideroblasts and thrombocytosis
(RARS-​T), which has features of both a myelodysplastic syndrome
and myeloproliferative neoplasm and is classified as such in the World
Health Organization (WHO) classification. The diagnostic criteria
and management of the other myeloproliferative neoplasms associated
with thrombocytosis are outlined in other chapters. Distinguishing
essential thrombocythaemia from prefibrotic primary myelofibrosis
can be challenging; the presence of an elevated lactate dehydrogenase,
systemic symptoms, leucocytosis, or a leucoerythroblastic smear with
minimal fibrosis on bone marrow biopsy still may suggest prefibrotic
primary myelofibrosis rather than essential thrombocythaemia des-
pite the presence of thrombocytosis. Careful consideration of the diag-
nostic features by WHO criteria between essential thrombocythaemia
and prefibrotic myelofibrosis is important from a prognostic stand-
point, as retrospective data suggest the latter have a reduced survival
associated with increased risk for evolution to overt myelofibrosis
and acute leukaemia. Cytogenetic abnormalities occur in approxi-
mately 5% of patients with essential thrombocythaemia, and the most
common are 1q–​, 20q–​, 21q–​, and 1q+. Elevated vitamin B12 levels
occur in 25% of patients.
Diagnostic criteria and differential diagnosis
The revised WHO diagnostic criteria for essential thrombocyth-
aemia are given in Box 22.3.6.2. Essential thrombocythaemia was
Box 22.3.6.2  2016 WHO criteria for the diagnosis
of essential thrombocythaemia
Major criteria
1	 Platelet count of at least 450 × 109/​litre.
2	 Bone marrow biopsy showing proliferation mainly of the
megakaryocyte lineage with increased numbers of enlarged, mature
megakaryocytes with hyperlobulated nuclei. There should be no sig-
nificant increase or left shift in neutrophil granulopoiesis or erythro-
poiesis and very rarely minor (grade 1) increase in reticulin fibres.
3	 Not meeting WHO criteria for chronic myeloid leukaemia, polycy-
thaemia vera, primary myelofibrosis, myelodysplastic syndrome, or
other myeloid neoplasm.
4	 Demonstration of JAK2 V617F, CALR, or MPL mutations or other
clonal marker.
Minor criterion
1	 Presence of another clonal marker or absence of evidence for reactive
thrombocytosis.
Diagnosis of essential thrombocythaemia requires all four of the following
major criteria or presence of the first three major criteria and the one minor
criterion
Source data from Arber DA, et al. (2016). The 2016 revision to the World
Health Organization (WHO) classification of myeloid neoplasms and acute
leukemia. Blood, 127, 2391–​405.
SECTION 22  Haematological disorders
5244
previously a diagnosis of exclusion, but the advent of JAK2 V617F,
CALR, and MPL mutational analyses have greatly facilitated the
diagnosis in approximately 90% of cases. The presence of these mu-
tations in the setting of thrombocytosis without evidence of polycy-
thaemia vera is virtually diagnostic of essential thrombocythaemia.
These diagnostic criteria are, however, of less use in paediatric pa-
tients since many of these individuals are JAK2 V617F negative.
Thrombocytosis may be the consequence of primary bone marrow
disorders associated with increased platelet production (nonreactive
thrombocytosis), or a secondary response to an underlying disorder
(reactive thrombocytosis). Box 22.3.6.1 summarizes the most im-
portant causes of thrombocytosis: iron deficiency anaemia, infec-
tion/​inflammation, malignancy, trauma, and hyposplenism, are the
most commonly encountered disorders. The exclusion of an iden-
tifiable cause for reactive thrombocytosis, in particular iron defi-
ciency, is a necessary step.
Risk assessment
Essential thrombocythaemia is a heterogeneous disorder asso-
ciated with patients encountering a varied risk of developing life-​
threatening complications. Many patients enjoy survival fairly
similar to that of their unaffected peers but a subset of patients is at
a high risk of developing additional thromboses. Myelosuppressive
therapy should be reserved for patients at a high risk of developing
such thrombotic complications. A  risk-​based decision approach
to therapy is outlined in Table 22.3.6.1 to identify such patients.
Advanced age (≥60  years) and a previous history of thrombosis
clearly define a group at high risk for the development of life-​
threatening complications. The degree of thrombocytosis and
the presence of associated cardiovascular risk factors, particu-
larly smoking and obesity, are also taken into consideration when
making treatment decisions. The International Prognostic Score for
Thrombosis in Essential Thrombocythemia (IPSET) uses patient
age, thrombosis history, and cardiovascular risk factors but addition-
ally recognizes that the presence of a JAK2 V617F mutation predis-
poses additional thrombotic risk. The utility of this scoring system
to guide treatment decisions is unknown at this time since it has
not been validated in a prospective fashion. In addition, it does not
predict the risk for evolution to myelofibrosis or acute leukaemia.
Isolated thrombocytosis per se is not an indication for therapy; how-
ever, it is common practice to treat extreme thrombocytosis (platelet
count >1500 × 109/​litre) because of the increased risk of bleeding
rather than thrombotic complications. The use of CALR, JAK2, or
MPL mutation status is not currently incorporated into risk assess-
ment for therapeutic decision although increasing data suggest that
these are distinct clinicopathological entities.
Treatment
The present goal of therapy in essential thrombocythaemia is to con-
trol symptoms and prevent thrombotic and haemorrhagic compli-
cations. Should a decision be made to treat the patient based on risk
assessment, the platelet count should be reduced to 400 × 109/​litre.
Although no target platelet count has been determined to be optimal
to reduce the incidence of thrombotic episodes in rigorous clinical
trials, this is considered a safe level by most practising physicians in
the field.
A number of agents are effective in the treatment of essential
thrombocythaemia. Low-​dose aspirin (81–​100 mg/​day) has been
shown to be safe and may decrease the recurrence of microcirculatory
events (erythromelalgia/​transient ischaemic attacks) and prevent the
development of other thrombotic phenomena, especially in combin-
ation with myelosuppressive agents in high-​risk patients. In order to
minimize the risk of iatrogenic bleeding, only patients with platelet
counts less than 1500 × 109/​litre and without evidence of an acquired
von Willebrand’s disease should be considered for low-​dose aspirin
administration.
The use of hydroxycarbamide, an antimetabolite that interferes
with DNA repair, decreased the number of thrombotic events in a
randomized study of high-​risk patients when given at 15 mg/​kg ini-
tially, with subsequent adjustments based on initial response. In this
study, the target was a platelet count of less than 600 × 109/​litre. It
is unknown whether tighter control (<350–​400 × 109/​litre) is more
effective in reducing thrombotic and haemorrhagic complications.
The onset of action is usually 3 to 5 days and frequent side effects
include dose-​related neutropenia, nausea, stomatitis, hyperpigmen-
tation, rash, nail changes, leg ulcers, increased risk of squamous cell
carcinoma of the skin, and hair loss. The leukaemogenic potential of
hydroxycarbamide when given as a single agent is still a subject of
controversy although it is clearly less leukaemogenic than alkylating
agents. Recent data from at least two large studies, one in polycy-
thaemia vera and the other one in essential thrombocythaemia pa-
tients, failed to show an increased incidence of acute leukaemia in
patients treated with hydroxycarbamide.
Interferon-​α, a biological response modifier, is also useful in
treating patients with essential thrombocythaemia. Ninety per cent
response rates with median times to response of approximately
3 months have been seen when 3 to 5 million units are admin-
istered subcutaneously 3 to 5 days per week. It is nonmutagenic
and does not cross the placenta. Frequent side effects include
influenza-​like symptoms, fatigue, lethargy, and depression. The
long-​term use of interferon is associated with mild weight loss,
alopecia, autoimmune thyroiditis, autoimmune haemolytic an-
aemia, and neuropsychiatric effects. Its extensive toxicity profile
and the need for parenteral administration limit its use as initial
therapy, particularly in elderly patients. Pegylated forms of inter-
feron have a prolonged half-​life, can be administered weekly, and
are often better tolerated. Prolonged therapy with interferon has
the potential to reduce the allele burden of the patient’s driver
mutation (CALR, JAK2 V167F, or MPL) and in 20 to 30% of pa-
tients to result in a complete molecular remission. In addition,
a subset of patients clear the marker chromosomal abnormal-
ities associated with their disease. It remains unknown whether
the correction of these molecular correlates translate into im-
proved survival, decreased thrombotic events, or progression to
myelofibrosis or acute leukaemia. Many patients find it difficult to
continue to receive interferon for more than 1 to 2 years primarily
due to adverse events.
Anagrelide is another treatment option for patients with essen-
tial thrombocythaemia. This drug acts by selectively inhibiting
megakaryocytic maturation. Responses have been documented
in over 90% of treated patients with a median time to response of
2.5 to 4 weeks and an onset of action of 6 to 10 days. Anagrelide
is nonmutagenic and its use has not been associated with the de-
velopment of acute leukaemia. The UKMRC PT-​1 study com-
paring hydroxycarbamide with anagrelide in addition to aspirin
therapy found that patients treated with anagrelide plus aspirin
22.3.6  Thrombocytosis and essential thrombocythaemia
5245
had an increased rate of arterial thrombotic events, haemor-
rhage, and transformation to myelofibrosis as compared to the
hydroxycarbamide plus aspirin arm. There was no increase in the
incidence of acute leukaemia in the hydroxycarbamide arm. On the
other hand, the ANAHYDRET study demonstrated that anagrelide
was not inferior to hydroxycarbamide in the prevention of throm-
botic complications in patients with essential thrombocythaemia
and its use was not associated with an increase in transformation
to acute leukaemia or myelofibrosis. Anagrelide is a good second-​
line treatment option for patients intolerant to hydroxycarbamide.
Common side effects of anagrelide therapy include headaches, diz-
ziness, fluid retention, palpitations, nausea, abdominal pain, and
diarrhoea. Anagrelide can trigger episodes of tachyarrhythmias and
heart failure, especially in the elderly. For this reason, it should be
used carefully in older people and avoided in patients with known
heart disease.
Alkylating agents have been extensively used in the past to
treat essential thrombocythaemia. Within this group of agents,
melphalan has been shown to be quite effective and relatively
nontoxic, with predictable cytopenias as its major untoward effect.
The drug is useful in treating the very elderly or those with com-
pliance issues. It is usually prescribed at 4 mg/​day until a platelet
count of 400 × 109/​litre is reached. Therapy can then be stopped and
patients experience prolonged periods of normalization of platelet
numbers. Additional courses can be given if and when the platelet
count rises over 400 × 109/​litre.
Given the number of available therapeutic options and their dif-
ferent toxicity profiles, the choice of the appropriate cytoreductive
drug for a given individual requires the consideration of a number
of variables. These include age, childbearing potential, projected
life expectancy, comorbidities, and cost of treatment. Furthermore,
the overall low risk for the development of life-​threatening com-
plications that affect patients with essential thrombocythaemia
highlights the need for systematic, risk-​based approaches to thera-
peutic decision-​making (Table 22.3.6.1). All patients should stop
smoking. Indiscriminate use of high doses of NSAIDs should be
avoided; their excessive use is clearly associated with bleeding
episodes.
Low-​risk patients have a risk of thrombosis similar to that of an
age-​ and sex-​matched control population and a very low risk of life-​
threatening bleeding. These observations support close observa-
tion without cytoreductive therapy as the most sensible approach.
Although the use of aspirin therapy is common in this group, retro-
spective studies have failed to show a benefit. High-​risk patients are
those more than 60 years of age and with a prior history of throm-
bosis. According to the results presented in the PT-​1 trial, these pa-
tients should be treated with hydroxycarbamide as the cytoreductive
agent of choice, in addition to aspirin. Anagrelide or a pegylated
form of interferon should be offered to patients who are intolerant of
hydroxycarbamide or who have developed adverse effects to it. For
elderly patients with limited projected survival (<10 years) and who
either have problems with drug compliance or are too ill to comply
with the minimum follow-​up requirements during cytoreductive
therapy, intermittent melphalan might be appropriate. α-​Interferon
may be an acceptable option in younger patients of reproductive
age with a history of life-​threatening thrombotic episode. In pa-
tients at intermediate risk based on platelet numbers at or greater
than 1000 to 1500 × 109/​litre and/​or patients who have acquired von
Willebrand’s disease, platelet reduction therapy is indicated to avoid
the higher risk of complications.
Smokers and obese individuals, unless symptomatic, should be
managed by risk modification. Smoking has been proved to be an in-
dependent risk factor for developing arterial thrombotic complica-
tions. Patients with essential thrombocytopenia should be strongly
encouraged to stop smoking to decrease their thromboembolic risk.
In severe, life-​threatening episodes, rapid cytoreduction may
be achieved by plateletpheresis or by the administration of high
doses of hydroxycarbamide. In patients who present with a life-​
threatening episode of acute bleeding, the site of bleeding should
be promptly identified and any antiplatelet agent should be stopped.
Those suffering from an acquired von Willebrand’s disease can be
treated with desmopressin and von Willebrand’s factor concentrates.
If the bleeding is due to a platelet function abnormality, or if alter-
native haemostatic agents have failed, platelet transfusion therapy is
recommended. Cytoreductive therapy with hydroxycarbamide must
be promptly initiated.
Up to 10% of patients with essential thrombocythaemia will
evolve to secondary myelofibrosis, recognized by the development
of cytopenias, leucoerythroblastic blood picture, and worsening
splenomegaly. These patients have a very poor prognosis and should
undergo evaluation for allogeneic stem cell transplantation. The use
of reduced-​intensity conditioning regimens for allogeneic stem cell
transplantation has been shown recently to improve the outcome of
such patients with a relatively low mortality rate.
At this time, JAK inhibitors do not have a role in the treatment
of essential thrombocythaemia. The management of patients who
are or want to become pregnant requires special consideration.
The risk of fetal loss is quite high (c.40%). High-​risk pregnancy is
defined as one occurring in an individual with a previous throm-
bosis or major bleeding episode, platelet count more than 1500 ×
Table 22.3.6.1  Risk stratification-​based treatment of essential
thrombocythaemia
Risk categorya
Treatment
Low risk
Observation
Age <60 years, and
No history of thrombosis, and
Platelet count <1000 × 109/​litre, and
No cardiovascular risk factors (smoking, obesity)
High risk
Treatment
Age ≥60 years, or
Previous history of thrombosis
Intermediate risk
Treatmentb
Age <60 years, and
Platelet count >1000–​1500 × 109/​litre, or
Cardiovascular risk factors (smoking, obesity)
a Leucocytosis and JAK2 V617F mutation appear to confer a higher risk of thrombosis,
but no established treatment guidelines exist for these patients.
b The decision to treat is at the discretion of the clinician. We offer treatment to most of
our patients with platelets more than 1000–​1500 × 109/​litre. Risk modification is strongly
encouraged.
Adapted from Blood Reviews, Vol. 19, Finazzi G, Barbui T, Risk-​adapted therapy in
essential thrombocythemia and polycythemia vera, Pages 243–​52, Copyright © 2005,
with permission from Elsevier.
SECTION 22  Haematological disorders
5246
109/​litre, and previous severe complications such as fetal loss or
placental abruption. Patients with low or intermediate disease
risk should be managed with careful observation. Specific treat-
ment should be considered for high-​risk pregnancies as follows.
(1) If previous thrombosis or major complications during prior
pregnancies have occurred, patients should receive low molecular
weight heparin throughout pregnancy until 6 weeks postpartum.
(2) If there is a history of major bleeding, or if the platelet count is
above 1500 × 109/​litre, aspirin should be avoided and consideration
should be given to cytoreduction with interferon to decrease the
platelet count to normal levels. Despite the lack of endorsement by
the manufacturers of α-​interferon, it is the drug of choice during
pregnancy given its lack of mutagenic potential and its inability to
cross the placenta. Hydroxycarbamide, given its mechanism of ac-
tion, could theoretically cause fetal malformations, and anagrelide,
because of its small molecular size, probably crosses the placenta
and may cause life-​threatening thrombocytopenia and haemor-
rhage in the fetus. Despite these concerns, several reports have de-
scribed first-​trimester exposures to these two drugs resulting in
the delivery of normal newborns. We therefore do not consider
unintended exposures to hydroxycarbamide or anagrelide as ab-
solute indications for the termination of a pregnancy. Recently,
JAK2 V617F-​positive essential thrombocythaemia has been found
to be an independent adverse predictor of pregnancy outcome.
These pregnancy-​associated complications in patients with JAK2
mutations were not prevented by the use of aspirin therapy raising
the question of whether prophylactic anticoagulant therapy is war-
ranted for JAK2 V617F-​positive patients.
Prognosis
The probability that a patient with essential thrombocythaemia will
survive 10 years is 64 to 80%, not substantially different from that of
a control age-​ and sex-​matched population. The actual risk for the
development of a catastrophic thrombotic or haemorrhagic event
in an asymptomatic patient is quite low. Most deaths come from
thrombotic complications. Transformation to myelofibrosis and/​or
acute leukaemia has been reported with increasing frequency at a
rate of transformation of 3 to 10%. JAK2 V617F-​positive essential
thrombocythaemia not infrequently evolves into the clinical picture
of polycythaemia vera.
While the presence or absence of JAK2, MPL, or CALR muta-
tions has an impact on prognosis in primary myelofibrosis, the im-
pact in essential thrombocythaemia is less clear but evolving. Those
with CALR mutations, including type 1 and 2 mutations, appear
to have lower thrombotic risk than those with the JAK2 mutation.
Type 1 CALR mutations have a higher risk of thrombosis as well
as transformation to primary myelofibrosis than type 2 mutations.
Type 1 CALR mutations also have a higher risk of transformation
to primary myelofibrosis than those patients with JAK2 mutations,
who appear to have a similar risk as those with type 2 mutations.
For patients with MPL mutations, conflicting data exist regarding
thrombotic risk. Overall survival may be inferior in those patients
with MPL mutations and best in those with triple negative essen-
tial thrombocythaemia while JAK2 and CALR mutations are inter-
mediate. More recent analyses have suggested that mutational status
does not impact upon overall survival or leukaemia-free survival.
Therefore, more research is needed to clarify the clinical impact of
these individual mutations.
Future directions
A better understanding of the factors that contribute to the devel-
opment of thrombotic episodes and evolution to myelofibrosis in
essential thrombocythaemia patients is clearly required. Studies that
evaluate the effects of currently used therapeutic agents on patient
outcomes are needed before conclusions can be made on when to
treat with a particular agent. Since target platelet numbers following
treatment are not predictive of eliminating the risk of thrombosis,
additional biomarkers are needed to guide the management of such
patients.
FURTHER READING
Arber DA, et  al. (2016). The 2016 revision to the World Health
Organization (WHO) classification of myeloid neoplasms and acute
leukemia. Blood, 127, 2391–​405.
Barbui T, Finazzi G (2007). Therapy for polycythemia vera and essen-
tial thrombocythemia is driven by the cardiovascular risk. Semin
Thromb Hemost, 33, 321–​9.
Barbui T, et al. (2011). Philadelphia-​negative classical myeloproliferative
neoplasms:  critical concepts and management recommendations
from European LeukemiaNet. J Clin Oncol, 29, 761–​70.
Barbui T, et al. (2011). Survival and disease progression in essential
thrombocythemia are significantly influenced by accurate morpho-
logic diagnosis: an international study. J Clin Oncol, 29, 3179–​84.
Barbui T, et al. (2013). Problems and pitfalls regarding WHO-​defined
diagnosis of early/​prefibrotic primary myelofibrosis versus essential
thrombocythemia. Leukemia, 27, 1953–​8.
Beer PA, et al. (2011). How I treat essential thrombocythemia. Blood,
117, 1472–​82.
Ding J, et al. (2004). Familial essential thrombocythemia associated
with a dominant-​positive activating mutation of the c-​MPL gene,
which encodes for the receptor for thrombopoietin. Blood, 103,
4198–​200.
Dingli D, Tefferi A (2005). A critical review of anagrelide therapy
in essential thrombocythemia and related disorders. Leukemia
Lymphoma, 46, 641–​50.
Elala YC, et al. (2016). Calreticulin variant stratified driver mutational
status and prognosis in essential thrombocythemia. Am J Hematol,
91, 503–​6.
Elliott MA, Tefferi A (2005). Thrombosis and haemorrhage in polycy-
thaemia vera and essential thrombocythaemia. Br J Haematol, 128,
275–​90.
Finazzi G, Barbui T (2005). Risk-​adapted therapy in essential thrombo-
cythemia and polycythemia vera. Blood Rev, 19, 243–​52.
Finazzi G, Harrison C (2006). Essential thrombocythemia. Semin
Hematol, 42, 230–​8.
Fruchtman SM, Hoffman R (2005). Essential thrombocythemia.
In: Hoffman R, et al. (eds) Hematology: basic principles and practice,
pp. 1177–​296. Churchill Livingstone, Philadelphia.
Gisslinger H (2006). Update on diagnosis and management of essential
thrombocythemia. Semin Thromb Hemost, 32, 430–​6.
Gisslinger H, et al. (2013). Anagrelide compared with hydroxyurea
in WHO-​classified essential thrombocythemia: the ANAHYDRET
Study, a randomized controlled trial. Blood, 121, 1720–​8.
Harrison C (2005). Pregnancy and its management in the Philadelphia
negative myeloproliferative diseases. Br J Haematol, 129, 293–​306.
Harrison CN, et al. (2005). Hydroxyurea compared with anagrelide in
high-​risk essential thrombocythemia. N Engl J Med, 353, 33–​45.

22.3.7 Primary myelofibrosis 5247 Evan M. Braunste
22.3.7 Primary myelofibrosis 5247 Evan M. Braunstein and Jerry L. Spivak
22.3.7  Primary myelofibrosis
5247
Karow A, et  al. (2015). Mutational profile of childhood
myeloproliferative neoplasms. Leukemia, 29, 2407–​9.
Kaushansky K (2005). The molecular mechanisms that control
thrombopoiesis. J Clin Invest, 115, 3339–​47.
Levine RL, et al. (2007). Role of JAK2 in the pathogenesis and treat-
ment of myeloproliferative disorders. Nat Rev Cancer, 7, 673–​83.
Long MW, Hoffman R (2005). Thrombocytopoiesis. In: Hoffman R,
et al. (eds) Hematology: Basic Principles and Practice, pp. 303–​20.
Churchill Livingstone, Philadelphia.
McIntyre CJ, et  al. (1991). Essential thrombocythaemia in young
adults. Mayo Clin Proc, 66, 149–​54.
Nangalia J, et al. (2013). Somatic CALR mutations in myeloproliferative
neoplasms with nonmutated JAK2. N Engl J Med, 369, 2391–​405.
Pardanani AD, et al. (2006). MPL515 mutations in myeloproliferative
and other myeloid disorders: a study of 1182 patients. Blood, 108,
3472–​6.
Pietra D, et al. (2016). Differential clinical effects of different mutation
subtypes in CALR-​mutant myeloproliferative neoplasms. Leukemia,
30, 431–​8.
Rotunno G, et al. (2014). Impact of calreticulin mutations on clinical
and hematological phenotype and outcome in essential thrombo-
cythemia. Blood, 123, 1552–​5.
Tefferi A, Barbui T (2015). Polycythemia vera and essential thrombo-
cythemia: 2015 update on diagnosis, risk-​stratification and manage-
ment. Am J Hematol, 90, 162–​73.
Tefferi A, et al. (2014). Calreticulin mutations and long-​term survival
in essential thrombocythemia. Leukemia, 28, 2300–​3.
Van Genderen PJJ, et  al. (1996). Acquired von Willebrand disease
in myeloproliferative disorders. Leukemia Lymphoma, 22 Suppl
1, 79–​82.
Van Genderen PJJ, et al. (1997). Prevention and treatment of throm-
botic complications in essential thrombocythemia:  efficacy and
safety of aspirin. Br J Haematol, 97, 179–​84.
Wagstaff AJ, Keating GM (2006). Anagrelide: a review of its use in the
management of essential thrombocythaemia. Drugs, 66, 111–​31.
22.3.7  Primary myelofibrosis
Evan M. Braunstein and Jerry L. Spivak
ESSENTIALS
Myelofibrosis is a reactive process common to many malig-
nant and benign disorders. Primary myelofibrosis is a chronic
myeloproliferative neoplasm arising in a pluripotent haematopoi-
etic stem cell. It results in abnormalities in red cell, granulocyte,
and platelet production in association with marrow fibrosis and
extramedullary haematopoiesis.
Aetiology
Primary myelofibrosis is known to be a clonal disorder caused by
­acquired genetic mutations in haematopoietic stem cells leading to
activation of the JAK/STAT pathway. Many chromosomal abnormal-
ities have been found, and in about 58% of cases there is expres-
sion of the JAK2 V617F missense mutation typical of polycythaemia
vera (see Chapter 22.3.5). Mutations in CALR or MPL are present in
approximately 26 and 6% of cases respectively. None of these mu-
tations is specific for primary myelofibrosis however, and in 10% of
patients no initiating mutation can be identified.
Clinical features and prognosis
Many patients are asymptomatic at the time of diagnosis, but
common presenting manifestations include fatigue, weight loss,
night sweats, fever, dyspnoea, and abdominal discomfort due to
splenomegaly (which may be massive). The major complications
are the consequences of bone marrow failure and extramedullary
haematopoiesis, which most commonly occurs in the spleen and
liver, but can occur at any site and compromise organ or tissue func-
tion. About 20% of patients develop acute myeloid leukaemia as a
terminal event.
Investigation and diagnosis
Anaemia is the most consistent abnormality, with the blood film
showing evidence of a leucoerythroblastic reaction (presence of
metamyelocytes, myelocytes, promyelocytes, myeloblasts, nucle-
ated red cells, and teardrop-​shaped red cells) due to extramedullary
haematopoiesis. The presence of marrow fibrosis is essential for
diagnosis and usually results in the inability to aspirate marrow from
a properly placed needle (‘dry tap’). The presence of genetic markers
provides supportive diagnostic data.
Treatment
Treatment is aimed at improving symptoms. Splenomegaly is gen-
erally the most distressing complication, and the nonselective JAK2
inhibitor, ruxolitinib, is effective in reducing spleen size and allevi-
ating constitutional symptoms in a majority of patients. In addition,
it has been shown to improve survival in patients with advanced dis-
ease and lower the JAK2 V617F allele burden. Patients with good
performance status as well as those with advanced stage disease
who have a matched, related donor should be considered for allo-
geneic bone marrow transplantation. Other therapies found to be
effective include low-​dose interferon, low-​dose thalidomide and
prednisone, low-​dose busulfan, hydroxycarbamide, splenectomy,
and splenic irradiation. Folic acid supplementation is often given to
prevent deficiency in the context of increased folate requirements,
and hyperuricaemia should be treated with allopurinol.
Introduction
Primary myelofibrosis (also called myelofibrosis with myeloid meta-
plasia, agnogenic myeloid metaplasia, or primary myelosclerosis)
is a chronic myeloproliferative neoplasm, resulting from the ac-
quisition of somatic mutations in a multipotent haematopoietic
progenitor cell. This leads to abnormalities in red cell, white cell,
and platelet production in association with marrow fibrosis and
extramedullary haematopoiesis. Although myelofibrosis in asso-
ciation with leucoerythroblastosis and splenomegaly are the clin-
ical hallmarks of primary myelofibrosis, these abnormalities can
also be seen in other chronic myeloproliferative disorders such as
polycythaemia vera and chronic myeloid leukaemia, as well as in
a variety of benign and malignant disorders that involve the bone
marrow (Box 22.3.7.1). Since there is no specific clonal marker for
SECTION 22  Haematological disorders
5248
primary myelofibrosis, and since many of the disorders listed in Box
22.3.7.1 are responsive to specific therapies not effective in primary
myelofibrosis, the diagnosis of this disorder is one of exclusion.
Aetiology
Primary myelofibrosis is caused by the acquisition of somatic mu-
tations in haematopoietic cells. Analysis of glucose-​6-​phosphate
dehydrogenase isoenzyme expression, X-​linked gene inactivation
patterns in informative women, and a mutation in the N-​ras proto-​
oncogene have established that primary myelofibrosis is a clonal dis-
order with its origin in a pluripotent haematopoietic stem cell. In some
patients, T lymphocytes express the same clonal marker as B lympho-
cytes and myeloid cells, indicating involvement at the level of the pluri-
potent stem cell. Karyotype and comparative genomic hybridization
analysis of primary myelofibrosis patients has identified nonrandom
abnormalities on multiple chromosomes. The most frequent aber-
rations include deletions on 20q, 17q, and 7p; however, deletions on
5q, 11q, 12p, and 13q and gains on 1q and 9p are also common, as is
trisomy 8. Next-​generation sequencing analysis has identified mul-
tiple recurrent somatic mutations in genes located both within these
chromosome aberrations as well as outside these regions. Most not-
ably, a mutation in the gene for the tyrosine kinase JAK2 (producing
a change from valine to phenylalanine at the 617 position (V617F)),
which is located on 9p, is observed in approximately 58% of primary
myelofibrosis patients. JAK2 V617F is also present in over 95% of
polycythaemia vera patients and 59% of essential thrombocytosis
patients. This mutation leads to constitutive activation of the tyro-
sine kinase receptor, resulting in increased cytokine release and cell
proliferation. Recently, somatic insertions or deletions in exon 9 of
the calreticulin gene (CALR) on 19p have been found in approxi-
mately 26% of essential thrombocytosis and primary myelofibrosis
patients, most frequently occurring in patients who do not possess a
JAK2 mutation. Mutations in the thrombopoietin receptor MPL oc-
curs in approximately 5% of patients. Together, a mutation in one
of these three 'driver' genes can be found in ­approximately 90% of
­primary myelofibrosis patients and causes activation of the JAK/
STAT signalling pathway. Other frequently mutated genes include
ASXL1, TET2, DNMT3A, and EZH2. While the role of each of these
mutations in the pathogenesis of disease is not well understood, they
are important markers of clonal haematopoiesis in making a diag-
nosis of primary myelofibrosis. In addition, patients with primary
myelofibrosis tend to have a greater number of somatic mutations
than patients with either essential thrombocytosis or polycythaemia
vera, supporting their importance in disease progression.
Marrow fibroblasts in primary myelofibrosis, in contrast to the
haematopoietic stem cells, are polyclonal, suggesting that the marrow
fibrosis is a reactive process initiated by expansion of the malignant
clone. Marrow collagen is argyrophilic, so that changes in its distri-
bution and content can be analysed histochemically by silver staining
early in the disease and by Masson’s trichrome staining later on.
Under normal circumstances, the connective tissue stroma of the
bone marrow is composed of collagen types I, III, IV, and V together
with noncollagen proteins such as fibronectin, laminin, vitronectin,
and the proteoglycans. Collagen types I, III, and V form a delicate
and usually noncontinuous supporting network for haematopoietic
cells, while type IV collagen, laminin, and fibronectin are localized
in the basement membranes of arteries in a continuous fashion and
along marrow sinusoids in a discontinuous fashion. With increasing
marrow cellularity, the collagenous supporting network also increases.
With myelofibrosis, however, there is both an increase in the collagen
network and a change in its physical characteristics. Condensation
of the interstitial fibres results in the formation of thick, continuous,
and often wavy bundles in association with an increase in reticular or
fibroblastic cells. Sinusoidal basement membrane collagen becomes
continuous, leading to sinusoidal dilatation and obliteration with
an associated capillary neovascularization. The content of basement
membrane fibronectin as well as stromal fibronectin and vitronectin
also increases. Although best studied in primary myelofibrosis, the
types of collagen involved in marrow fibrosis in this condition do not
appear to differ from those involved in the marrow fibrosis associated
with the other disorders listed in Box 22.3.7.1.
The commonality of the types of collagen involved and the simi-
larity of the histological process, regardless of disease association,
implies that marrow fibrosis per se represents a final common
pathway involved in the response to diverse immunological, meta-
bolic, toxic, or infectious stimuli (Box 22.3.7.1). Megakaryocytic
hyperplasia, dysplasia, and clustering are characteristic features
of primary myelofibrosis. These cells produce cytokines such as
platelet-​derived growth factor (PDGF) and transforming growth
factor-​β (TGFβ) that promote fibroblast proliferation, and platelet
factor 4 (PF4), which, like TGFβ, inhibits collagenase. These find-
ings suggest that inappropriate release of these fibrogenic proteins
by dysfunctional megakaryocytes is the stimulus for myelofibrosis.
In support of this contention, increased concentrations of PDGF
and TGFβ as well as basic fibroblast growth factor have been
observed in the platelets and megakaryocytes from primary
myelofibrosis patients. Circulating levels of TGFβ are also increased
in primary myelofibrosis, as is the urinary excretion of basic fibro-
blast growth factor and calmodulin, another potential fibroblast
stimulant present in platelets. Finally, TGFβ promotes the syn-
thesis of osteoprotegerin, which impairs osteoclast proliferation
and promotes osteosclerosis. The observation that thrombopoietin
Box 22.3.7.1  Causes of marrow fibrosis
Malignant
•	 Acute lymphocytic leukaemia
•	 Acute myeloid leukaemia
•	 Chronic myeloid leukaemia
•	 Hairy cell leukaemia
•	 Primary myelofibrosis
•	 Lymphoma
•	 Plasma cell myeloma
•	 Metastatic carcinoma
•	 Polycythaemia vera
•	 Systemic mastocytosis
Nonmalignant
•	 HIV infection
•	 Hyperparathyroidism
•	 Renal osteodystrophy
•	 Systemic lupus erythematosus
•	 Thorium dioxide exposure
•	 Tuberculosis
•	 Vitamin D deficiency
22.3.7  Primary myelofibrosis
5249
overexpression in mice recapitulates the histological features of pri-
mary myelofibrosis supports the contention that megakaryocytes
are integrally involved in the development of marrow fibrosis in
primary myelofibrosis, as does the development of marrow fibrosis
in the grey platelet syndrome, in which impaired megakaryocyte α
granule synthesis results in release of cytokines and growth factors
into the marrow extracellular matrix.
Additional distinct histological abnormalities in primary
myelofibrosis include increased marrow angiogenesis due to an in-
crease in vascular endothelial growth factor production, marrow
sinusoidal dilatation, and intravascular haemopoiesis with an in-
crease in circulating CD34+ cells. Increased angiogenesis appears
to be an early feature of the disease and correlates better with
increased marrow cellularity than with marrow fibrosis. The in-
crease in circulating CD34+ cells appears to be a consequence of
elevations in neutrophil elastase and matrix metalloproteinases
with cleavage of the endothelial cell adhesion molecule VCAM-​
1 and down-​regulation of the chemokine receptor CXCR4 on the
CD34+ cells.
Recent work has also identified a noncell autonomous contri-
bution of the stem cell niche to disease pathogenesis in primary
myelofibrosis. Somatic mutations in haemopoietic progenitor cells
cause alterations in the bone marrow microenvironment, such as
loss of mesenchymal stem cells and Schwann cells that help regulate
haematopoietic stem cells. These changes are thought to occur via
aberrant cytokine release produced by clonal progenitor cells, and
ultimately promote disease progression.
Clinical features
Although considered to be an uncommon disorder with an inci-
dence of approximately 1:100 000 person-​years, clinical studies of
more than 1000 patients have been reported over the last 50 years.
In contrast to the other chronic myeloproliferative disorders, the
median age at diagnosis of primary myelofibrosis, 61 years (range
15–​94), is much older. Male predominance is the rule in contrast
to polycythaemia vera and essential thrombocytosis, and familial
clustering occurs in approximately 10% of cases. The presenting
manifestations depend on the stage of the illness but are often bland.
Many patients are asymptomatic at the time of discovery. Fatigue is
the commonest presenting complaint, followed by weight loss, night
sweats, fever, dyspnoea, and abdominal discomfort due to spleno-
megaly. Hearing loss due to otosclerosis is an interesting but often
nonelicited symptom. Easy bruising or bleeding and acute gout or
renal stones are other presenting manifestations that are reason-
ably common and directly related to the underlying disease process.
Rarely, periostitis may occur.
Splenomegaly is present in virtually every patient with primary
myelofibrosis at diagnosis. When it is absent, one should consider
other causes for the clinical abnormalities. The degree of spleno-
megaly varies but is frequently substantial. However, since the rate
of splenic enlargement is variable, spleen size cannot be used as an
indication of disease duration. Hepatomegaly, invariably of a lesser
extent than the splenomegaly, is present initially in approximately
50% of patients and is usually proportional to the degree of spleno-
megaly. Lymphadenopathy is uncommon. With substantial spleno-
megaly, cachexia may be prominent.
Laboratory studies
Due to its origin in a multipotent haematopoietic progenitor cell,
primary myelofibrosis affects all blood cell types but not in a pre-
dictable manner. Anaemia, usually mild, is the most consistent
abnormality. Indeed, a normal haemoglobin or haematocrit in
the presence of substantial splenomegaly should lead to imme-
diate consideration of polycythaemia vera, since the expanded
plasma volume associated with splenomegaly can mask a sub-
stantial increase in the red cell mass. The leucocyte and platelet
counts can be low, normal, or high without reference to spleen size.
Inevitably, due to extramedullary haematopoiesis, metamyelocytes,
myelocytes, promyelocytes, myeloblasts, and nucleated red cells
will be present in the circulation together with the teardrop-​shaped
red cells characteristic of this situation (Fig. 22.3.7.1). Although
this so-​called leucoerythroblastic reaction is not specific for pri-
mary myelofibrosis, its absence should challenge the clinical im-
pression. Abnormalities in liver function tests are not uncommon
but are usually mild and most often involve a reduction in serum
albumin and an elevation of liver alkaline phosphatase due to
extramedullary haematopoiesis, an abnormality that is usually
magnified by splenectomy. The lactate dehydrogenase level is usu-
ally mildly increased and correlates best with the leucocyte count.
Hyperuricaemia is not infrequent. The leucocyte alkaline phos-
phatase concentration can be low, normal, or high and cannot be
recommended as a diagnostic test. As mentioned, the JAK2 V617F
mutation and less commonly CALR or MPL mutations are present
in the majority of patients, but their absence does not exclude the
diagnosis. Cytogenetic abnormalities include 13q−, 20q−, trisomy
9, trisomy 8, 12p–​, and abnormalities of chromosomes 3, 5, 7, and
11. An increase in circulating CD34+ cells is also characteristic but
none of these abnormalities is diagnostic for primary myelofibrosis
or present in a majority of patients.
Perhaps the most intriguing laboratory abnormalities in pri-
mary myelofibrosis are those linked to autoreactivity, such as cir-
culating immune complexes, complement activation, elevations in
Fig. 22.3.7.1  Peripheral smear showing teardrop cells with
leucoerythroblastic picture.
Copyright ©2009 American Society of Hematology.
SECTION 22  Haematological disorders
5250
antinuclear antibody and rheumatoid factor titres, and a positive
Coombs’ test in the absence of an overt connective tissue disorder.
Although marrow fibrosis has been documented in patients with
systemic lupus erythematosus, the linkage between autoimmune ab-
normalities and marrow fibrosis is unclear. It does, however, provide
another therapeutic option as discussed in the following paragraphs.
The presence of marrow fibrosis is essential for a diagnosis of pri-
mary myelofibrosis and usually results in a ‘dry tap’ or the inability
to aspirate marrow from a properly placed needle. A  prefibrotic
phase of primary myelofibrosis has been described retrospect-
ively. However, given the similarity of the histopathology of poly-
cythaemia vera, essential thrombocytosis, and prefibrotic primary
myelofibrosis, prospective substantiation of the latter is not possible
in the absence of a specific clonal marker for the disease. Even
the presence of myelofibrosis, although mandatory, is not in itself
sufficient for diagnosis. This is because polycythaemia vera and
chronic myeloid leukaemia and other disorders such as hairy-​cell
leukaemia, myelodysplasia, and acute leukaemia can present with
myelofibrosis. Thus, it is essential to use the appropriate diagnostic
tests (cytogenetics, BCR-​ABL1 fluorescent in situ hybridization, flow
cytometry, and immunohistochemistry) to exclude these and the
other disorders listed in Table 22.3.7.1 that can cause myelofibrosis.
Marrow cellularity in primary myelofibrosis may be increased
with trilineage hyperplasia and erythroblastic and megakaryocytic
islands, decreased with scattered areas of hyperplastic marrow
embedded in a collagenous matrix, or hypoplastic with in-
tense osteomyelosclerosis and residual megakaryocytic islands
(Fig. 22.3.7.2). While there is a correlation between the degree of
fibrosis and osteosclerosis, there is no correlation between bone
marrow histology and disease duration, platelet count, or spleno-
megaly; marrow fibrosis does, however, appear to correlate with the
leucocyte count. In general, marrow fibrosis and extramedullary
haematopoiesis with myeloid metaplasia appear unrelated, and the
latter abnormalities cannot be considered as compensation for the
former. Increased marrow angiogenesis is a recently recognized
feature of primary myelofibrosis that correlates with increased cel-
lularity and extramedullary haematopoiesis independently of the
marrow fibrosis.
Table 22.3.7.1  Three scoring systems for predicting survival in patients with primary myelofibrosis at diagnosis and during the course
of the disease
Prognostic factor
IPSS
DIPSS
aaDIPSS
Score
Score
Score
0
1
2
0
1
2
0
1
2
Age (years)
<65
65
<65
65
White cell count × 1000
<25
25
<25
25
<25
25
Haemoglobin (g/​litre)
100
<100
≥100
<100
100
<100
Blood blast cells
<1%
1%
<1%
≥1%
<1%
≥1%
Constitutional symptoms
No
Yes
No
Yes
No
Yes
aaDIPSS, age-​adjusted Dynamic International Prognostic Scoring System; DIPSS, Dynamic International Prognostic Scoring System; IPSS, International Prognostic Scoring System.
(b)
(a)
Fig. 22.3.7.2  (a) Bone marrow biopsy: showing thick, irregular bone trabeculae with new bone formation, and hypercellular marrow with increased
megakaryocytes. (b) Bone marrow trephine biopsy: showing thick coarse reticulin fibres (grade 4 fibrosis).
Copyright ©2009 American Society of Hematology.
22.3.7  Primary myelofibrosis
5251
In conjunction with the most severe degree of marrow fibrosis,
osteosclerosis develops, with characteristic radiographic abnormal-
ities. These primarily involve the axial skeleton but can include the
skull, with thickening of the bony trabeculae and patchy or coalescent
sclerosis. With obliteration of the axial marrow cavities, extension of
the marrow into the long bones occurs. Interestingly, the increase in
trabecular bone in primary myelofibrosis is not accompanied by an
increase in either osteoblastic or osteoclastic activity. This feature
distinguishes the osteosclerosis of primary myelofibrosis from that
associated with metabolic causes of osteosclerosis.
Course and prognosis
Primary myelofibrosis is a chronic progressive disorder with a
median lifespan (5.5 years) that is much shorter than for poly-
cythaemia vera and essential thrombocytosis. However, the het-
erogeneity characterizing the initial clinical presentation is also
evident with respect to survival, which can range from less than
a year to more than 30 years. Death is usually a consequence of
bone marrow failure (haemorrhage, anaemia, or infection), trans-
formation to acute leukaemia, portal hypertension, heart failure,
cachexia, or myeloid metaplasia with organ failure. Retrospective
analysis of the adverse prognostic value of presenting manifest-
ations has identified a number of factors that may be useful for
both prognostic and therapeutic purposes. These include age
at onset (>65  years), anaemia (haemoglobin <100 g/​litre), con-
stitutional symptoms, leucocytosis (>25 × 109/​litre), thrombo-
cytopenia, circulating blast cells (≥1%), and certain cytogenetic
abnormalities (trisomy 8, 12p–​, 5/​5q−, 7/​7q−, 1(17q), inv(3),
11q23 rearrangement, alone or in combination). A  number of
scoring systems have been devised for identifying long-​ and short-​
term survivors based on the presence of more than one adverse
presenting manifestation. Three such scoring systems that are
useful in separating patients of any age with myelofibrosis into
low-​ and higher-​risk groups with respect to survival are shown in
Table 22.3.7.1. With the institution of these scoring systems, risk
stratification of primary myelofibrosis has been redefined using
a classification similar to that used for MDS. The IPSS is used to
stage primary myelofibrosis patients at diagnosis and the DIPSS
and age-​adjusted DIPSS are used during the course of the disease
with the difference being the greater weight accorded to the pres-
ence of anaemia not due to a correctable cause or therapy. The
DIPSS+ incorporates thrombocytopenia, transfusion number, and
cytogenetics. Unfavourable cytogenetics (trisomy 8, 12p–​, 5/​5q−,
7/​7q−, i(17q), inv(3), 11q23 rearrangement, alone or in combin-
ation) together with thrombocytopenia are also independent pre-
dictors of leukaemic transformation and as such should be applied
to the DIPSS and, in particular, to patients classified as low risk
since many of these will reclassified on the basis of these additional
risk factors to a higher risk group (Table 22.3.7.2). For example,
with normal or sole common cytogenetic abnormalities such as
+9, 13q−, or 20q−, median survival is 113  months; with more
than one cytogenetic abnormality or sole +8, 5/​5q−, 7/​7q− or
less common ones, median survival is 48 months; with complex
abnormalities, median survival is 34 months. This has particular
implications with respect to bone marrow transplantation since
those with high-​risk disease are most likely to benefit from trans-
plantation. A scoring system incorporating driver mutation status,
termed MIPSS, has also been developed.
Driver mutation status has also been shown to correlate with
clinical course and overall prognosis. Patients with CALR muta-
tions tend to have less incidence of anaemia and thrombocytopenia
and improved overall survival compared to patients without CALR
mutations, while patients without an identifiable CALR, JAK2, or
MPL mutation (the so-​called ‘triple negative’ cohort) appear to have
the worst overall survival.
Complications
The major complications of primary myelofibrosis are the conse-
quences of bone marrow failure and extramedullary haematopoi-
esis. Anaemia may be the result of ineffective erythropoiesis, but
haemodilution due to an expanded plasma volume associated with
splenomegaly, iron deficiency due to gastrointestinal blood loss,
folic acid deficiency due to the increased demands of haematopoi-
esis, haemolysis due to autoimmune phenomena or hypersplenism,
and, rarely, pyridoxine deficiency are also considerations. In some
patients, erythropoietin production may be inappropriately low for
the degree of anaemia but in this instance haemodilution also needs
to be excluded. Red cell survival and splenic sequestration studies
can be useful in determining the splenic contribution to anaemia.
Hyperuricaemia is a consequence of increased cell turnover
and can provoke acute gout or renal stone formation if left un-
treated. Splenic enlargement is inevitable and can lead to splenic
infarction, malnutrition due to easy satiety, plasma volume ex-
pansion, hypersplenism, portal hypertension, splanchnic vein
thrombosis, extreme discomfort due to its mass, and eventually
cachexia (Fig. 22.3.7.3). Hepatomegaly is associated with spleno-
megaly. Impaired hepatic function is usually a consequence of
extramedullary haematopoiesis, which can lead to hepatic fibrosis
and portal hypertension.
Although myeloid metaplasia due to exuberant extramedullary
haematopoiesis is most common in the spleen and liver, it can occur
at any site and compromise organ or tissue function. For example,
peritoneal involvement can lead to ascites; epidural involvement to
spinal cord compression; retroperitoneal involvement to obstructive
uropathy or portal hypertension; and pulmonary extramedullary
haematopoiesis to pulmonary hypertension., which can lead to sub-
stantial cachexia The reason why myeloid metaplasia is more aggres-
sive in some patients than in others is unclear.
Table 22.3.7.2  Survival in months according to risk status
in primary myelofibrosis
Risk status
IPSS
DIPSS
aaDIPSS
DIPSS+
Low
135
∞
∞
185
Int-​1
  95
170
106
  78
Int-​2
  48
  48
  56
  35
High
  27
  18
  27
  16
aaDIPSS, age-​adjusted Dynamic International Prognostic Scoring System; DIPSS, Dynamic
International Prognostic Scoring System; IPSS, International Prognostic Scoring System.
SECTION 22  Haematological disorders
5252
Approximately 20% of patients with primary myelofibrosis de-
velop acute leukaemia as a terminal event. Although some clinicians
do not distinguish acute leukaemia presenting with myelofibrosis
(malignant myelosclerosis) from primary myelofibrosis, they are
clinically distinct entities. The extent to which therapeutic interven-
tion with potentially mutagenic drugs such as hydroxycarbamide,
alkylating agents, or irradiation predisposes patients with primary
myelofibrosis to progress to acute leukaemia (as it does in patients
with polycythaemia vera or essential thrombocytosis) is unknown.
Again for unknown reasons, splenectomy also appears to be a
predisposing factor for the development of acute leukaemia.
Platelet dysfunction is a common feature of the chronic
myeloproliferative disorders and can lead to spontaneous haem-
orrhage as well as increased bleeding during surgical procedures.
Although abnormalities in platelet morphology, prolongation of
the bleeding time, and abnormal platelet aggregation are frequently
observed in patients with primary myelofibrosis, no consistent bio-
chemical abnormality has been identified and no platelet function
test is predictive for the risk of haemorrhage.
Therapy
There is no specific therapy for primary myelofibrosis. Treatment
should be individualized, based on the patient’s IPSS or DIPSS
risk group and age. Asymptomatic, low-​risk patients without
hyperuricaemia or a remedial cause of anaemia require no therapy,
although the oral administration of folic acid (1 mg/​day) appears
reasonable. Anaemia associated with an inappropriately low en-
dogenous erythropoietin level (<100 mU/​ml) may respond to re-
combinant erythropoietin therapy but the hormone can cause an
increase in splenomegaly or hepatomegaly. Patients most likely to
respond are those with an absent or low transfusion requirement.
Prednisone may also be effective if there is evidence of active inflam-
mation or autoimmune abnormalities; a role for danazol remains
to be established. Hyperuricaemia should be treated with allopur-
inol. Asymptomatic leucocytosis or thrombocytosis requires no
therapy. Patients less than 65 years of age and an intermediate-​2 or
high DIPSS score who have a matched donor should be considered
for allogeneic bone marrow transplantation. Unfortunately, the best
results have been obtained in patients with good prognosis disease
and the transplantation-​related mortality can be as high as 32% with
a 5-​year survival of 60%. Recently, reduced-​intensity conditioning
was found to decrease transplant-​related mortality and achieve re-
mission rates of more than 70% even in individuals greater than
60 years of age. However, prospective studies will be required to es-
tablish the most effective conditioning regimen, the optimal timing
for transplantation, and which patients will benefit most from this
procedure. Haploidentical transplantation is also under scrutiny.
In symptomatic patients in the absence of a suitable bone marrow
donor, treatment has been revolutionized by the introduction of the
nonselective JAK2 inhibitor, ruxolitinib, which appears to be ef-
fective in JAK2 V617F-​negative patients, as well as those positive
for this mutation. Currently approved for intermediate-​1 to -​2 and
high-​risk patients, the drug is effective in alleviating constitutional
symptoms and reducing spleen size in the majority of patients.
Importantly, significant relief of symptoms can be achieved with only
a small reduction in spleen size. In some patients, anaemia is exacer-
bated, but is usually transient and can be temporized with transfu-
sion. Thrombocytopenia, even if severe, may also be improved and
is not an absolute contraindication to the drug. Improvements in
survival have also been documented in higher-​risk patients, how-
ever, persistent molecular or pathogenic responses do not occur
with this treatment. Loss of response after 3–5 years of treatment is
common. The effects of ruxolitinib are durable as long as the drug
is administered, but will recur when the drug is stopped and thus
it should be discontinued by gradual dose reduction. Thalidomide
at low doses (50–​100 mg/​day) in combination with prednisone has
proved to be effective in ameliorating anaemia as well as thrombo-
cytopenia in approximately 60% of primary myelofibrosis patients,
and reducing spleen size in approximately 20%. Lenalidomide has
not proved to be more effective than thalidomide in small phase II
trials in the absence of del5q. Pegylated interferon-​α at low doses is
worth considering to reduce splenomegaly, and has been most ef-
fective early in the course of the illness, but can cause cytopenias.
Newer targeted therapies, such as selective JAK2 inhibitors, or anti-
sense oligonucleotides such as the telomerase-​inhibiting imetelstat
may lead to improved outcomes in the future.
Chemotherapeutic agents or irradiation therapy should be
used judiciously in the treatment of idiopathic myelofibrosis.
Hydroxycarbamide, while easy to use and with a low incidence of
acute toxicity, can cause marrow depression. Low-​dose alkylating
agent therapy has been demonstrated to reduce organomegaly, reverse
marrow fibrosis, and improve blood counts in primary myelofibrosis,
occasionally in a durable fashion. However, alkylating agents can also
cause severe bone marrow depression and are leukaemogenic.
Splenomegaly is the most distressing complication of pri-
mary myelofibrosis, leading to mechanical discomfort, inanition,
splenic infarction, portal and pulmonary hypertension, and blood
cell sequestration. Reduction in splenic size can be achieved with
ruxolitinib in most patients, but interferon, thalidomide, alkylating
agents, hydroxycarbamide, or splenic irradiation are also effective.
After ruxolitinib, either thalidomide or interferon are the ini-
tial treatments of choice, followed by chemotherapy with either
hydroxycarbamide or low-​dose busulfan. Splenic irradiation can be
effective at alleviating splenic pain and temporarily reducing spleen
Fig. 22.3.7.3  Patient with primary myelofibrosis and massive
splenomegaly who underwent splenectomy.
Copyright ©2009 American Society of Hematology.
22.3.7  Primary myelofibrosis
5253
size. However, its use should be restricted to inoperable patients
since there is an unpredictable risk of severe cytopenias as well as an
increased risk of haemorrhage, if irradiation precedes splenectomy.
Local irradiation is, of course, appropriate for the management of
symptomatic extramedullary haematopoiesis in tissues and organs
other than the spleen.
Splenectomy in primary myelofibrosis is a prodigious pro-
cedure, given the large size of the spleen and its vessels, the inevit-
able presence of adhesions, the haemorrhagic tendency of primary
myelofibrosis patients, and their often poor nutritional status.
Evaluation for portal hypertension should precede surgery and,
if necessary, parental hyperalimentation should be used to avoid
postoperative complications. ε-​Aminocaproic acid can be used if
bleeding is a problem.
Leucocytosis, thrombocytosis, and postoperative hepatic enlarge-
ment are the usual consequences of splenectomy, as is elevation of
liver alkaline phosphatase. Postoperative splenic and portal vein
thrombosis occur in approximately 10% of patients, most often in
the first few weeks after surgery and presumably due to the size of
the splenic vein remnant. However, there is no correlation between
splenic or portal vein thrombosis and the platelet count. Surveillance
by ultrasonography or CT may useful in identifying this complica-
tion with the intent of administering anticoagulants or thrombolytic
agents. For unknown reasons, the incidence of the transformation of
primary myelofibrosis to acute leukaemia is increased post splenec-
tomy. Postoperative ventral hernia can be a source of distress, par-
ticularly in women.
Finally, as mentioned earlier, autoimmune phenomena are fea-
tures of primary myelofibrosis. Corticosteroids may be beneficial
if autoimmune phenomena are clinically significant, if there are as-
sociated constitutional symptoms, and for the amelioration of an-
aemia. Tuberculosis was once a frequent complication of primary
myelofibrosis and, thus, constitutional symptoms in these patients
should not be attributed to the myeloproliferative disease without
first excluding an infectious process.
FURTHER READING
Arranz L, et al. (2014). Neuropathy of haematopoietic stem cell niche
is essential for myeloproliferative neoplasms. Nature, 512, 78–​81.
Brecqueville M, et al. (2014). Array comparative genomic hybridiza-
tion and sequencing of 23 genes in 80 patients with myelofibrosis at
chronic or acute phase. Haematologica, 99, 37–​45.
Cervantes F, et al. (2009). New prognostic scoring system for primary
myelofibrosis based on a study of the International Working Group
for Myelofibrosis Research and Treatment. Blood. 113, 2895–​2901.
Chagraoui H, Wendling F, Vainchenker W (2006). Pathogenesis of
myelofibrosis with myeloid metaplasia: insight from mouse models.
Best Pract Res Clin Haematol, 19, 399–​412.
Elliott MA, et al. (1998). Splenic irradiation for symptomatic spleno-
megaly associated with myelofibrosis with myeloid metaplasia. Br J
Haematol, 103, 505–​11.
Gangat N, et al. (2011). DIPSS Plus: a refined dynamic international
prognostic scoring system for primary myelofibrosis that incorp-
orates prognostic information from karyotype, platelet count, and
transfusion status. J Clin Oncol, 29, 392–​7.
Glew RH, Wolfgang HH, Mclntrye PA (1973). Myeloid metaplasia
with myelofibrosis. The clinical spectrum of extramedullary
hematopoiesis and tumor formation. Johns Hopkins Med J, 132,
253–​70.
Hussein K, et al. (2010). International Prognostic Scoring System –​ in-
dependent cytogenetic risk categorization in primary myelofibrosis.
Blood, 115, 496–​9.
Klampfl T, et  al. (2013). Somatic mutations of CALR in
myeloproliferative neoplasms. N Engl J Med, 369, 2379–​90.
Kroger N, et  al. (2009). Allogeneic stem cell transplantation after
reduced-​intensity conditioning in patients with myelofibrosis: a pro-
spective, multicenter study of the Chronic Leukemia Working Party
of the European Group for Blood and Marrow Transplantation.
Blood, 114, 5264–​70.
Kroger N, et  al. (2015). Impact of allogeneic stem cell transplant-
ation on survival of patients less than 65 years of age with primary
myelofibrosis. Blood, 125, 3347.
Lundberg P, et al. (2014). Clonal evolution and clinical correlates of
somatic mutations in myeloproliferative neoplasms. Blood, 123,
2220–​8.
McLornan DP, et al. (2012). Allogeneic stem cell transplantation for
myelofibrosis. Br J Haematol, 157, 413–​25.
Mesa R, et  al. (2000). Evaluation and clinical correlations of bone
marrow angiogenesis in myelofibrosis and myeloid metaplasia.
Blood, 96, 3374–​80.
Mesa R, et al. (2003). A phase 2 trial of combination low-​dose thalido-
mide and prednisone for the treatment of myelofibrosis with mye-
loid metaplasia. Blood, 101, 2534–​41.
Mesa R, et al. (2006). Palliative goals, patient selection, and periopera-
tive platelet management. Cancer, 107, 361–​70.
Pardanani AD, et al. (2006). MPL515 mutations in myeloproliferative
and other myeloid disorders: a study of 1182 patients. Blood, 108,
3472–​76.
Passamonti F, et  al. (2010). A dynamic prognostic model to pre-
dict survival in primary myelofibrosis: a study by the IWG-​MRT
(International Working Group for Myeloproliferative Neoplasms
Research and Treatment). Blood, 115, 1703–​8.
Passamonti F, et al. (2014). Impact of ruxolitinib on the natural his-
tory of primary myelofibrosis: a comparison of the DIPSS and the
COMFORT-​2 cohorts. Blood, 123,1833–​5.
Rondelli D, et al. (2005). Allogeneic hematopoietic stem-​cell trans-
plantation with reduced-​intensity conditioning in intermediate-​
or high-​risk patients with myelofibrosis with myeloid metaplasia.
Blood, 105, 4115.
Rumi E, et  al. (2014). Clinical effect of driver mutations of JAK2,
CALR, or MPL in primary myelofibrosis. Blood, 124, 1062–​9.
Silver RY, Vandris K, Goldman JJ (2011). Recombinant interferon
alpha may retard progression of early primary myelofibrosis: a pre-
liminary report. Blood, 1117, 6669–​72.
Tefferi A, et  al. (2008). Low JAK2V617F allele burden in pri-
mary myelofibrosis, compared to either a higher allele burden or
unmutated status, is associated with inferior overall survival and
leukemia-​free survival. Leukemia, 22, 756–​61.
Tefferi A, et  al. (2015). A pilot study of the telomerase inhibitor
imetelstat for myelofibrosis. N Engl J Med, 373, 908–​19.
Truong LD, Saleem A, Schwartz MR (1984). Acute myelofibrosis.
a report of four cases and review of the literature. Medicine,
63, 182–​7.
Tsiara SN, et al. (2007). Recombinant erythropoietin for the treatment
of anaemia in patients with chronic idiopathic myelofibrosis. Acta
Haematol, 117, 156–​61.
Verstovsek S, et al. (2012). A double-​blind, placebo-​controlled trial of
ruxolitinib for myelofibrosis. N Engl J Med, 366, 799–​87.

22.3.8 Eosinophilia 5254 Peter F. Weller
22.3.8 Eosinophilia 5254 Peter F. Weller
SECTION 22  Haematological disorders
5254
22.3.8  Eosinophilia
Peter F. Weller
ESSENTIALS
Eosinophilia (eosinophil count >0.45 × 109/​litre) is associated with
some infections, some allergic diseases, and a variety of other condi-
tions, sometimes neoplastic.
Infectious diseases
Parasitic diseases—​eosinophilia is a characteristic feature of infection
by multicellular helminth parasites (e.g. Strongyloides stercoralis) with
diagnosis typically based on geographical/​dietary history, serological
tests, and examination of stool or tissues for parasite forms.
Other diseases—​eosinophilia can be caused by the fungal disease
coccidioidomycosis, and modest eosinophilia (0.45–​1.5 × 109/​litre)
may accompany retroviral infections such as HIV and HTLV-​1.
Allergic, immunological, neoplastic, and other disorders
Common allergic diseases—​asthma, rhinitis, and atopic dermatitis are
associated with modest eosinophilia.
Drug reactions—​these are a frequent cause of eosinophilia, at
times in reactions characterized by rashes and pyrexia. More severe
reactions may also manifest with (1) pulmonary eosinophilia and
lung infiltrates; (2) interstitial nephritis; (3) hepatitis; (4) myocarditis;
(5)  drug-​induced hypersensitivity vasculitis; (6)  gastroenterocolitis;
and (7)  drug-​induced rash, eosinophilia, and systemic symptoms
(DRESS syndrome).
Other conditions—​these include (1) eosinophilic granulomatosis with
polyangiitis (EGPA, formerly Churg–​Strauss syndrome); (2) hyper-​IgE
syndromes—​comprising recurrent staphylococcal abscesses, derma-
titis, hyperimmunoglobulinaemia E, and eosinophilia; (3) chronic mye-
loid leukaemia, acute myeloid leukaemia, and lymphoma; (4) a variety
of pulmonary, skin, gastrointestinal, and endocrine diseases.
Hypereosinophilic syndromes
These are defined by (1) eosinophilia (>1.5 × 109/​litre) sustained over
a month, (2) lack of an identifiable cause precipitating a secondary
eosinophilia, and (3)  symptoms and signs of organ involvement.
About 30% of patients will have either a myeloproliferative condition
(chronic eosinophilic leukaemia) or hypereosinophilia mediated by
clonal expansion of specific T-cells producing interleukin-​5 (IL-​5).
Clinical features—​common manifestations are (1)  cardiac—​
endomyocardial damage typically leads to a restrictive cardio-
myopathy; (2)  neurological—​strokes related to thromboemboli,
encephalopathy, and polyneuropathy; (3)  dermatological (e.g.
angio-​oedema, urticarial lesions); and (4) pulmonary—​infiltrates of
eosinophils may be seen in any part of the lung, sometimes leading
to pulmonary fibrosis.
Treatment—​patients without organ damage do not require treat-
ment. Aside from supportive care: (1) chronic eosinophilic leukaemia—​
may respond to tyrosine kinase inhibitors (e.g. imatinib); and
(2) nonmyeloproliferative hypereosinophilic syndrome—​may respond
to high-​dose corticosteroids, with hydroxycarbamide, interferon-​α or
anti-​IL-​5 monoclonal antibody used in refractory cases.
Introduction
Eosinophilia is associated with distinct diseases that include hel-
minth parasitic infections, allergic diseases, and varied diseases of
often ill-​defined cause.
In comparison with other leucocytes, eosinophils are distin-
guished by their morphologies, constituents, products, and asso-
ciations with specific diseases. The cytokine interleukin-​5 (IL-​5),
specific in promoting the development, differentiation, and release
of bone marrow-​derived eosinophils, is principally responsible for
increases in eosinophilopoiesis. Eosinophils are normally tissue-​
dwelling cells primarily distributed in those tissues with an epithe-
lial interface with the environment, including the gastrointestinal
and lower genitourinary tracts. Eosinophils are distinguished mor-
phologically from neutrophils by their cytoplasmic granules, which
uniquely contain crystalloid cores visible by electron microscopy.
Within these granules are four specific cationic proteins: major basic
protein, eosinophil peroxidase, eosinophil cationic protein, and
eosinophil-​derived neurotoxin. The heavy content of these cationic
granule proteins, which bind acidic dyes such as eosin, are respon-
sible both for the identifying tinctorial properties of eosinophils and
for many of the functional properties of eosinophils. Eosinophils are
sources of over four dozen cytokines, and many, if not all, of these
are stored preformed within eosinophil granules and cytoplasmic
vesicles. In addition to their content of preformed proteins, eo-
sinophils synthesize lipid mediators, including the 5-​lipoxygenase
pathway-​derived eicosanoid, leukotriene C4. Eosinophils have roles
in normal homeostatic functioning, including adipocyte biology,
and in the pathogenesis of allergic diseases and in other immuno-
logical responses. The potential functional roles of eosinophils in
parasite–​host defence, although often assumed to be helminthotoxic
effector cells, are proving to be more complex and varied.
Eosinophils normally number less than 450/​μl in the blood with
a mild diurnal variation, being slightly higher in the morning and
falling as endogenous glucocorticosteroid levels rise. Blood eosino-
phil numbers do not, however, always reflect the extent of eosinophil
involvement in affected tissues in various diseases, and at times, as
in eosinophilic pneumonias, eosinophils may be recruited into in-
volved tissues without a concomitant increase in enumerable blood
eosinophils. Eosinopenia, diminished blood eosinophil levels, oc-
curs with corticosteroid administration and is frequent with active
bacterial and viral infections. Thus, even normal blood eosinophil
numbers in a febrile patient suggest that an illness is not simply due
to a bacterial or viral infection.
Some, but not necessarily all, patients with sustained blood eo-
sinophilia can develop organ damage, especially cardiac, as found
in hypereosinophilic syndromes (HES), and patients with sustained
eosinophilia should be monitored for evidence of cardiac disease.
Diseases associated with eosinophilia
Infectious diseases
Parasitic diseases
Eosinophilia is not elicited by infections with protozoan parasites
(with the exceptions of the intestinal parasites Isospora belli and
22.3.8  Eosinophilia
5255
Dientamoeba fragilis), but rather characteristically by multicel-
lular helminth parasites. Magnitudes of eosinophilia tend to par-
allel the extent of tissue invasion, especially by helminth larvae.
Eosinophilia may be absent in established infections which are
well contained within tissues or are solely intraluminal within
the gastrointestinal tract (e.g. ascaris, tapeworms). Even with
helminth diseases, superimposed bacterial infections (e.g. in dis-
seminated strongyloidiasis) can suppress expected eosinophilia.
In patients with eosinophilia, geographical and dietary histories
are pertinent in suggesting potential exposures to helminth para-
sites. Stool examinations for diagnostic ova and larvae should
be obtained, and for strongyloidiasis, in which stool exams lack
sensitivity, serological testing should be performed. In addition,
for several helminth parasites that cause eosinophilia, diagnostic
parasite stages are never present in faeces. Hence, negative stool
specimens do not necessarily exclude a helminth aetiology for
eosinophilia, and examination of appropriate blood or tissue bi-
opsies and/​or serological tests, as guided by clinical findings and
exposure histories, may be needed to diagnose specific tissue-​ or
blood-​dwelling infections, including trichinellosis and filarial
infections.
Other infectious diseases
The characteristic response in acute bacterial and viral infections
is eosinopenia. The fungal disease, coccidioidomycosis, either
following primary infection, at times with progressive dissem-
inated disease, or with central nervous system infection (with
cerebrospinal fluid eosinophilia), may be associated with eo-
sinophilia. Cerebrospinal fluid eosinophilia may be present with
cryptococcosis. Basidiobolomycosis may be associated with
eosinophilia.
HIV and retroviral infections
Eosinophilia may be associated with HIV infections as a result of
adverse reactions to medications or adrenal insufficiency in pa-
tients with AIDS from cytomegalovirus and other infections. In
addition, eosinophilia, often modest, is observed in some HIV-​
infected patients and may accompany eosinophilic folliculitis in
HIV infection. Eosinophilia frequently develops with HTLV-​1
infections.
Allergic and immunological disorders
Common allergic diseases, including allergic rhinitis, asthma, and
atopic dermatitis, are accompanied by tissue eosinophil infiltration
and usually modest blood eosinophilia. The occurrence of marked
blood eosinophilia suggests the presence of other diseases, such
as EGPA.
Medication-​related eosinophilias
Therapeutic agents, including herbal or ‘natural’ therapies, can elicit
eosinophilia. Eosinophilia may develop without other manifest-
ations of adverse drug reactions, such as rashes or drug fevers. In
the absence of organ involvement, blood eosinophilia by itself need
not mandate cessation of drug therapy, if such is medically indi-
cated. Drug-​induced blood eosinophilia, however, should prompt
an evaluation of whether organs, including the lungs, kidneys, and
heart, are involved in the eosinophil-​associated drug reaction. If
organ involvement develops, cessation of drug administration is
necessary.
Some cytokines are potential causes of eosinophilia. Granulocyte–​
macrophage colony-​stimulating factor (GM-​CSF) and IL-​2 can
cause prominent blood and tissue eosinophilia and, less commonly,
eosinophil-​associated diseases, including eosinophilic pneumonia
and eosinophilic endomyocardial fibrosis. Diverse agents, including
many antimicrobial agents and nonsteroidal anti-​inflammatory
agents, may elicit pulmonary eosinophilia. Blood eosinophilia
is usually, but not always, present; if blood eosinophilia is absent,
sputum or bronchoalveolar lavage eosinophilia is necessary to help
make the diagnosis.
In drug-​induced acute interstitial nephritis, eosinophilia is
common in the involved kidneys, urine, and at times, the blood. In
addition to eosinophilia, fever, rash, and arthralgia support the diag-
nosis, but these are commonly absent in cases of drug-​induced acute
interstitial nephritis. Eosinophiluria is not uniformly present in all
with drug-​induced interstitial nephritis.
Acute necrotizing eosinophilic myocarditis is a serious but un-
common type of hypersensitivity myocarditis, with reactions to
medications, responsible in some cases. A  syndrome of hepa-
titis with eosinophilia can be a manifestation of drug reactions.
Other medication-​related eosinophilic responses include drug-​
induced hypersensitivity vasculitis, the DRESS syndrome (drug-​
induced rash, eosinophilia, and systemic symptoms), and forms of
gastroenterocolitis.
Immunological disorders
The several genetic hyper-​IgE syndromes are characterized by
recurrent staphylococcal abscesses of the skin, lungs, and other
sites, pruritic dermatitis, hyperimmunoglobulinaemia E, and
blood, sputum, and tissue eosinophilia. Eosinophilia is charac-
teristic of Omenn’s syndrome, combined immunodeficiency with
hypereosinophilia, and is present in about 25% of subjects with
IgG4-​related disease
Eosinophil infiltration accompanies rejection of lung, kidney, and
liver allografts. Tissue and blood eosinophilia occur early in the re-
jection process.
Myeloproliferative and neoplastic diseases
The HES are considered in a later section. Eosinophilia may ac-
company chronic myeloid leukaemia and the M4Eo subtype of
acute myeloid leukaemia. Blood eosinophils may be elevated
in the nodular sclerosing form of Hodgkin lymphoma. Some
patients with carcinomas, especially of mucin-​producing epi-
thelial cell origins, have blood eosinophilia. Eosinophilia may
accompany angio-​immunoblastic lymphadenopathy, mycosis
fungoides, Sézary syndrome, lymphomatoid papulosis, and sys-
temic mastocytosis.
Pulmonary syndromes
Diverse eosinophilic pulmonary syndromes are noted in Box 22.3.8.1.
Skin and subcutaneous diseases
Various cutaneous diseases can be associated with a heightened
level of blood eosinophils (Box 22.3.8.1). In episodic angio-​oedema
with eosinophilia, recurrences are marked by prominent blood
SECTION 22  Haematological disorders
5256
eosinophilia, significant angio-​oedema, at times with excessive
weight gain due to fluid retention, and less frequently by fever.
Gastrointestinal diseases
Eosinophilia is common with eosinophilic gastroenteritis and eo-
sinophilic oesophagitis, and tissue eosinophils are found in inflam-
matory bowel diseases and collagenous colitis.
Rheumatological diseases
The principal eosinophil-​related vasculitis is EGPA. Most of these
are antineutrophil cytoplasmic antibody test negative and may be
difficult to distinguish from a HES.
Cutaneous necrotizing eosinophilic vasculitis with hypo­
complementaemia and eosinophilia, a distinct vasculitis of small
dermal vessels which are extensively infiltrated with eosinophils,
may occur in patients with connective tissue diseases. Eosinophilia
may uncommonly accompany rheumatoid arthritis itself but is
more commonly due to adverse reactions to medications or con-
comitant vasculitis.
Endocrine diseases
Loss of normal adrenoglucocorticosteroid production causes in-
creased blood eosinophilia.
Other disorders
The syndrome of atheromatous cholesterol embolization can be as-
sociated with eosinophilia and eosinophiluria. Rare kindreds with
hereditary eosinophilia have been recognized. Irritation of serosal
surfaces, as in eosinophilic pleural effusions and during chronic
peritoneal dialysis, can be associated with eosinophilia.
Hypereosinophilic syndromes
Patients with pronounced and prolonged eosinophilia not associ-
ated with other clinical diseases noted previously had been classified
previously as having the idiopathic HES. More recently, it has be-
come clear that the HES include a clinically heterogeneous diverse
group of eosinophilic disorders. For myeloproliferative and lympho-
cytic variants of HES, aetiologies are now known, yet for most others
with HES underlying aetiologies remain to be delineated.
Definition
Chusid and colleagues proposed three defining criteria for the then
named idiopathic HES that may can now be modified and updated.
Contemporary criteria include the following:
Box 22.3.8.1  Diseases and disorders associated with eosinophilia
Infectious diseases
•	 Helminth parasites
•	 Coccidioidomycosis
•	 Other infections—​infrequent, but includes HIV-​1 and HTLV-​1
Allergic and immunological disorders
•	 Allergic rhinitis, asthma
•	 Medication-​related eosinophilias
•	 Immunological diseases: hyperIgE syndromes, Ommen’s syndrome, and
IgG4-​related diseases
•	 Transplant rejections
Myeloproliferative and neoplastic disorders
•	 Hypereosinophilic syndromes
•	 Leukaemia, notably M4Eo subtype of acute myeloid leukaemia
•	 Lymphoma-​ and tumour-​associated, notably with nodular sclerosing
Hodgkin lymphoma
•	 Systemic mastocytosis
Pulmonary syndromes
•	 Parasite-​induced eosinophilic lung diseases:
—​	 Transpulmonary passage of developing larvae (Löffler syndrome):
patchy migratory infiltrates, especially ascaris
—​	 Tropical pulmonary eosinophilia: miliary lesions and fibrosis; height-
ened immune responses to lymphatic filariae with increased IgE and
antifilarial antibodies
—​	 Pulmonary parenchymal invasion: paragonimiasis
—​	 Heavy haematogenous seeding with helminths: disseminated stron-
gyloidiasis, trichinellosis, schistosomiasis, larva migrans
—​	 Allergic bronchopulmonary aspergillosis
•	 Chronic eosinophilic pneumonia:  dense often peripheral infil-
trates, fever; blood eosinophilia may be absent; may be antecedent
to EGPA
•	 Acute eosinophilic pneumonia—​acute presentation, often without
blood eosinophilia; diagnosed by bronchoalveolar lavage or biopsy
•	 EGPA vasculitis:  small-​ and medium-​sized arteries; perivascular eo-
sinophilia early and granulomas and necrosis later; asthma often
antecedent; extrapulmonary, for example, neurological, cutaneous,
cardiac, or gastrointestinal vasculitic involvement likely
•	 Drug-​ and toxin-​induced eosinophilic lung diseases
•	 Other:  neoplasia, hypereosinophilic syndromes, bronchocentric
granulomatosis
Skin and subcutaneous diseases
•	 Skin diseases—​atopic dermatitis, blistering diseases, including bullous
pemphigoid, urticarias, drug reactions
•	 Diseases of pregnancy: pruritic urticarial papules and plaques syndrome,
herpes gestationis
•	 Eosinophilic pustular folliculitis
•	 Eosinophilic cellulitis (Wells’ syndrome)
•	 Kimura’s disease and angiolymphoid hyperplasia with eosinophilia
•	 Shulman’s syndrome (eosinophilic fasciitis)
•	 Episodic angio-​oedema with eosinophilia—​recurrent periodic epi-
sodes with fever, angio-​oedema, and secondary weight gain; may be
longstanding without untoward cardiac dysfunction
Gastrointestinal diseases
•	 Eosinophilic gastroenteritis—​(1) blood eosinophilia; (2) eosinophil cell
infiltrates in the mucosa, muscularis, or serosa; (3) oedema of stomach
or intestines; and (4) absence of extraintestinal involvement
•	 Inflammatory bowel disease and collagenous colitis—​eosinophils in
tissue lesions
Rheumatological diseases
•	 EGPA vasculitis
•	 Cutaneous necrotizing eosinophilic vasculitis
Endocrine disease
•	 Hypoadrenalism: Addison’s disease, adrenal haemorrhage, hypopituitarism
Other causes of eosinophilia
•	 Atheromatous cholesterol embolization
•	 Hereditary
•	 Serosal surface irritation, including peritoneal dialysis and pleural
eosinophilia
22.3.8  Eosinophilia
5257
•	Eosinophilia in excess of 1.5 × 109/​litre of blood. The eosinophilia
needs to be sustained over 1 month, but does not require the older
prior 6-​month duration, especially if therapies are needed.
•	Lack of an identifiable parasitic, allergic, or other aetiologies for
secondary eosinophilia. As noted later (see ‘Aetiologies’), some
forms of HES now have identified aetiologies.
  Among parasitic aetiologies of eosinophilia, it is especially im-
portant to exclude Strongyloides stercoralis, which may persist for
decades and be difficult to diagnose solely by stool examinations,
not only because of its capacity to cause marked eosinophilia
mimicking HES, but also because it, unlike other helminthic
causes of marked eosinophilia, can develop into a disseminated,
often fatal, disease (hyperinfection syndrome) in patients given
immunosuppressive corticosteroids.
•	Evidence by symptoms and signs of organ involvement. This
older criterion has been modified to include patients with eosino-
philia who do not yet exhibit evidence of organ involvement. Not
all patients with prolonged eosinophilia develop organ involve-
ment and many have benign courses. These patients are often not
reported or subjected to evaluation at referral centres due to the
absence of eosinophil-​associated disease. Blood eosinophilia per
se does not warrant therapy in the absence of evidence of con-
comitant organ involvement.
Aetiologies
HES encompass a spectrum of hypereosinophilic disorders, for
which aetiologies are now recognized for a couple of HES variants:
•	Myeloproliferative variants—​these represent forms of chronic
eosinophilic leukaemia. The most common form arises from an
interstitial deletion on chromosome 4q12 that leads to fusion of
the FIPL1 (FIP1-​like 1) and PDGFRA genes that generates a pro-
tein with constitutively active receptor tyrosine kinase activity.
Rearrangements of the PDGFRB and FGFR1 genes are additional
causes of chronic eosinophilic leukaemia. In addition, clonal ab-
normalities in the eosinophil lineage have been detected in a small
number of women using X-​linked polymorphisms.
•	Lymphocytic variants—​these represent causes of HES mediated
by clonal expansions of specific T-cells. The most common are
due to CD3–​CD4+ T-​cell subsets and less frequently CD3+CD4–​
CD8–​ or other T-​cell subsets. These clonal T-cells elaborate
eosinophil-​stimulating IL-​5 and often other Th2-​associated cyto-
kines, including IL-​4.
•	Other—​over half of patients with HES currently have still un-
defined aetiologies for their eosinophilia.
Clinical features
With the evolving recognition of variant forms of HES, some
clinical features are more common with myeloproliferative,
lymphocytic, or other variants of HES. Myeloproliferative HES
is rare in women, whereas other variants of HES have no sex
bias. HES tend to occur between the ages of 20 and 50 years, al-
though cases have developed in children. Initial manifestations
may be due to sudden cardiac or neurological complications, but
tend to be more insidious and present over months or longer.
Eosinophilia may be detected only incidentally. Other frequent
presenting symptoms include tiredness, cough, breathlessness,
muscle pains, angio-​oedema, rash, sweating, pruritus, or retinal
lesions. Patients with HES do not exhibit a propensity to bacterial
or other infections.
Haematological manifestations
The defining haematological abnormality is sustained eosinophilia.
Total leucocyte counts are usually less than 25 × 109/​litre, with be-
tween 30 and 70% eosinophils, but extremely high leucocyte counts
(>90 × 109/​litre) develop in some patients and are associated with
a poor prognosis. Eosinophils in the blood may be mature or less
commonly can include numbers of eosinophilic myeloid precursors.
Eosinophils often exhibit morphological abnormalities including
diminished granule numbers, cytoplasmic vacuolization, and nu-
clear hypersegmentation.
Some patients with HES will have an absolute neutrophilia along
with their eosinophilia. Serum vitamin B12 and tryptase levels are
often elevated in myeloproliferative variant HES and should be as-
sayed. Anaemia is present in some patients.
Bone marrow findings demonstrate increased numbers of
eosinophils, often 30 to 60%, with a shift to the left in eosino-
phil maturation. Increased numbers of myeloblasts are not usu-
ally seen. Myelofibrosis and splenomegaly are more frequent in
myeloproliferative variant HES.
Cardiac manifestations
In HES, the heart is a commonly affected organ due to the devel-
opment of endomyocardial damage leading to a restrictive car-
diomyopathy. This distinct form of cardiac involvement may also
complicate other varied diseases marked by sustained eosinophilia,
including eosinophilia with carcinomas or lymphomas, eosino-
philia from GM-​CSF or IL-​2 administration or drug reactions,
and less commonly eosinophilia from helminthic infections such
as trichinellosis, visceral larva migrans, and filariasis. However,
many patients with eosinophilia do not develop any evidence of
endomyocardial damage; hence in addition to increased num-
bers of eosinophils, the pathogenesis of eosinophil-​mediated car-
diac damage probably involves some, as yet ill-​defined, activating
events that promote eosinophil-​mediated endomyocardial damage.
Patients with sustained eosinophilia should be monitored by
troponin assays and echocardiography or cardiac magnetic reson-
ance imaging for evidence of cardiac disease.
Cardiac damage progresses through three stages, the first
involving acute necrosis in the early weeks, the second involving
the development of endocardial thrombi over many months, and
the final stage being the fibrotic stage after a couple of years of
disease.
The risks of developing cardiac disease in two series of pa-
tients with HES were not related to the extent of eosinophilia
or duration of disease. Those who developed evident cardiac
disease were more likely to be male and to have splenomegaly,
thrombocytopenia, elevated levels of vitamin B12, hypogranular
or vacuolated eosinophils, and abnormal early myeloid pre-
cursors in their blood. Cardiac involvement is more common in
the myeloproliferative than the lymphocytic variants of HES.
Neurological manifestations
Neurological complications may be of three types. The first
type is due to thromboemboli originating from the left ven-
tricle, which may occur before cardiac disease is demonstrable
SECTION 22  Haematological disorders
5258
by echocardiography and can be the presenting manifestation of
HES. The second type of neurological disease is primary central
nervous system dysfunction, presenting as an encephalopathy
including changes in behaviour, confusion, ataxia, and memory
loss, and exhibiting upper motor neuron signs with increased
muscle tone, deep tendon reflexes, and a positive Babinski sign.
Impaired cognitive abilities may persist for months. The patho-
logical basis for this form of diffuse central nervous system dis-
ease remains unknown. Peripheral neuropathies constitute the
third type of neurological dysfunction. Symmetric or asymmetric
polyneuropathies manifest by sensory deficits, painful paraesthe-
siae, or mixed sensory and motor deficits are most common, but
mononeuritis multiplex occurs with HES (as well as with EGPA),
as do radiculopathies and muscle atrophy due to denervation.
Biopsies of affected nerves generally show an axonal neuropathy
with varying degrees of axonal loss and no evidence of vasculitis
or contiguous eosinophil infiltration.
Cutaneous manifestations
The skin is one of the most frequently involved organs, especially
with lymphocytic variant HES. The most common skin manifest-
ations are of two types:  either angio-​oedematous and urticarial
lesions, or erythematous, pruritic papules, and nodules. Some
patients with angio-​oedema and eosinophilia have a syndrome
of episodic angio-​oedema and eosinophilia (Gleich’s syndrome)
and many of these have lymphocytic variant HES. Particularly
incapacitating mucocutaneous manifestations of HES are mu-
cosal ulcers that may occur in the mouth, nose, pharynx, penis,
oesophagus, stomach, and anus.
Pulmonary manifestations
Pulmonary involvement is reported in about 40% of HES pa-
tients, the commonest respiratory symptom being a chronic, per-
sistent, generally nonproductive cough. The basis for this may
be sequestration of eosinophils in pulmonary tissues, although
most symptomatic individuals have clear chest radiographs.
Pulmonary involvement in HES may also be secondary to con-
gestive heart failure, pulmonary emboli originating from right
ventricular thrombi, or primary infiltration of the lungs by eo-
sinophils. Infiltrates may be diffuse or focal without a predilection
for any region of the lungs, in contrast to the often peripheral in-
filtrates in chronic eosinophilic pneumonia (see Chapter 18.14.2).
Pulmonary fibrosis may develop over time, especially in those
with cardiac fibrosis.
Other manifestations
Arthralgias, large joint effusions, cold-​induced Raynaud’s phe-
nomenon, and digital necrosis of fingers or toes can occur with
HES. Although myalgias are frequent, focal myositis or poly-
myositis occur only uncommonly. Gastrointestinal tract in-
volvement can accompany HES, and 20% of patients at some
time may have diarrhoea. Eosinophilic gastritis, enterocolitis,
or colitis may be present. Pancreatitis and sclerosing cholangitis
occur rarely. Hepatic involvement with HES includes chronic
active hepatitis and the Budd–​Chiari syndrome from hepatic
vein obstruction.
Diagnosis
Patients with myeloproliferative HES due to the FIP1L1-​PDGFRA
fusion can be identified by polymerase chain reaction or by
evaluating the associated deletion of the CHIC2 gene by fluor-
escence in situ hybridization. Bone marrow biopsy with cyto-
genetics may identify less common myeloproliferative variants.
Lymphocytic variants of HES are diagnosed based on both per-
ipheral T-​cell phenotyping by flow cytometry and assessments of
clonal T-​cell receptor rearrangements.
Treatment
For patients with myeloproliferative HES, therapy for chronic eo-
sinophilic leukaemia is with imatinib or related tyrosine kinase in-
hibitors. For those eosinophilic patients without organ damage, no
therapy need be administered. There is no clear threshold value of
blood eosinophilia that predicts organ involvement or damage. For
HES patients without myeloproliferative variants requiring therapy,
prednisolone is the initial agent, administered at 60 mg/​day in adults.
For those not responsive to prednisolone or needing a steroid-​sparing
regimen, daily hydroxycarbamide or interferon-​α (1–​10 million
units/​day or three times a week) are options. Anti-​IL-​5 monoclonal
antibodies may also prove to be an additional therapy for HES.
Medical management of cardiac complications, including ar-
rhythmias and congestive heart failure, is important and effective
in the longer-​term management of HES, as is surgical replacement
of damaged valves. Although early reports emphasized the mor-
tality due to this disorder, many of the deaths were due to con-
gestive heart failure and complications of endomyocardial damage.
If the sequelae of organ damage, especially to the heart, can be
managed, many patients with HES can have a prolonged course.
FURTHER READING
Akuthota P, Weller PF (2012). Eosinophil pneumonias. Clin Microbiol
Rev, 25, 649–​60.
Gotlib J (2017). World Health Organization-defined eosinophilic dis-
orders: 2017 update on diagnosis, risk stratification, and manage-
ment. Am J Hematol, 92, 1243–59.
Klion AD, Weller PF (2014). Eosinophilia and eosinophil-​related dis-
orders. In: Adkinson NF, Jr et al. (eds) Allergy: principles and prac-
tice, 8th edition, pp. 1205–​23. Elsevier, London.
Lefevre G, et al. (2014). The lymphoid variant of hypereosinophilic
syndrome. Medicine, 93, 255–​66.
Ogbogu PU, et al. (2009). Hypereosinophilic syndrome: a multicenter,
retrospective analysis of clinical characteristics and response to
therapy. J Allergy Clin Immunol, 124, 1319–​25.
Valent P, et al. (2012). Pathogenesis and classification of eosinophil
disorders: a review of recent developments in the field. Expert Rev
Hematol, 5, 157–​76.
Williams, KW, Milner JD, Freeman AF (2015). Eosinophilia associ-
ated with disorders of immune deficiency or immune dysregulation.
Immunol Allergy Clin N Am, 35, 523–​44.
Wilson ME, Weller PF (2011). Eosinophilia. In:  Guerrant RL,
Walker DH, Weller PF (eds) Tropical infectious diseases: princi-
ples, pathogens and practice, 3rd edition. pp. 939–​49. Elsevier,
Philadelphia.

22.3.9 Histiocytosis 5259 Chris Hatton
22.3.9 Histiocytosis 5259 Chris Hatton
22.3.9  Histiocytosis
5259
22.3.9  Histiocytosis
Chris Hatton
ESSENTIALS
The histiocytoses are disorders derived from the dendritic cell and
monocyte/​macrophage lineages, with the classification of this group
of disorders relating to the underlying cell of origin.
Dendritic cell disorders
There has been much debate about the nature of these conditions,
and their status as neoplastic or primary inflammatory diseases; for
Langerhans’ cell histiocytosis in particular, there is increasing evi-
dence of their clonal nature, as manifest by recurrent BRAF mutations.
Clinical features and diagnosis—​these are highly variable and de-
pendent on the sites affected by histiocytic infiltration. Symptoms
and signs may include rashes, bony pain, lymphadenopathy, hep-
atomegaly and splenomegaly, cough and dyspnoea, features of
marrow failure, and endocrine presentations (classically diabetes
insipidus). Classical clinical presentations have previously given
rise to eponymous syndromes (Hand–​Schuller–​Christian syn-
drome among others). Diagnosis typically follows imaging and bi-
opsy, with the demonstration of a histiocytic infiltrate confirmed by
immunostaining.
Treatment and prognosis—​the rarity and heterogeneity of these
diseases has made it difficult to achieve a consensus on treatment.
For localized disease, curettage, steroid injections, or targeted radio-
therapy may be helpful. For more systemic disease, combination
chemotherapy is typically used. Treatment schedules differ between
adults and children. Prognosis is dependent mainly on the site(s) of
involvement. Our expanding appreciation of the molecular basis of
these conditions also provides some justification for the use of BRAF
inhibitors and other targeted small molecule therapies.
Macrophage-​related disorders
These include haemophagocytic lymphohistiocytosis, a collection of
macrophage-​activating syndromes which may be either reactive to
underlying inflammatory, infective, or neoplastic disease, or conse-
quent upon a primary genetic lesion affecting cytotoxic T-​cell killing
function. Rosai–​Dorfman disease is a separate macrophage prolifer-
ation syndrome, thought to be non-​neoplastic, which causes mas-
sive cervical lymphadenopathy, usually in children.
Introduction
Histiocytosis describes a group of varied disorders that are considered
to be derived from dendritic and monocyte/​macrophage lineages.
Dendritic cells are part of the adaptive immune system, their main
function being to present antigens to effector lymphocytes. The classifi-
cation of this group of disorders attempts to relate the disease categories
to the underlying cells of origin. It is likely that as our understanding of
the cellular and molecular biology of dendritic cells, their precursors,
and related cellular systems expands, the classification and nomencla-
ture will change. A simplified classification of these diverse conditions
is set out in Table 22.3.9.1. This chapter will focus on dendritic and
macrophage disorders; malignant histiocyte/​monocyte disorders are
described in detail in Chapter 22.3.3 as acute myeloid leukaemia.
Dendritic cell disorders
Langerhans’ cell histiocytosis
Langerhans’ cell histiocytosis (LCH) is now considered to be a dis-
order derived from myeloid dendritic cells and not, as first thought,
arising from Langerhans’ cells of the skin. There remains some de-
bate about the true neoplastic nature of LCH as clonality has not
been demonstrated in all cases, though more recent gene expression
profiling continues to lend weight to the concept that LCH is a clonal
disorder. A  high proportion of cases harbour cancer-​associated
proto-​oncogene mutations and more than 50% of cases have the
BRAF V600E mutation, often with additional mutations of the ERK
pathway. It is likely that mutations occurring in early, less differen-
tiated LCH cells will lead to widespread multisystem involvement,
whereas mutations arising in more differentiated LCH cells are tissue
restricted and lead to localized single system disease. Proliferations
of these abnormal cells infiltrate various tissues—​most commonly
bone, skin, lymph nodes, spleen and liver, and the oral mucosa. The
abnormal LCH cells express CD1a, S100, and CD207 (langerin),
which are all of diagnostic significance. The cells also have charac-
teristic Birbeck granules visible on electron microscopy. There is an
additional group of disorders that mimic LCH which do not express
these markers but express macrophage-​associated antigens. The
best-​known example of this subgroup is Erdheim–​Chester disease
briefly described later in this chapter.
Patterns of organ presentation were previously used to group
these disorders into the eponymously named Hand–​Schuller–​
Christian and Letterer–​Siwe diseases—​both describe multisystem
involvement with LCH. The third entity was eosinophilic granu-
loma, which described a localized lesion typically affecting bone.
It is now accepted that this clinical classification is largely artificial
and the terms are now redundant. Patients are better classified into
those who present with a single system involvement and those with
multisystem involvement.
In young children, LCH commonly presents as a widespread ec-
zematous rash not dissimilar to the rash found with candida infec-
tion. In both adults and children, bone lesions are very common
and typically lytic, and they may be accompanied by a soft tissue
mass. Any bone may be affected though special attention should be
Table 22.3.9.1  Histiocytic diseases in summary
Category
Disorders
Dendritic cell disorders
Langerhans’ cell histiocytosis
Follicular dendritic cell sarcoma
Interdigitating dendritic cell sarcoma
Juvenile xanthogranuloma
Erdheim–​Chester disease (non-​LCH)
Macrophage-​related disorders
Haemophagocytic Lymphohistiocytosis
Rosai–​Dorfman disease
Malignant histiocyte disorders
Monocytic acute myeloid leukaemia
Histiocytic sarcoma
SECTION 22  Haematological disorders
5260
paid to involvement of the skull or the jaw, as these sites confer risk
of central nervous system involvement. The presence of diabetes
insipidus should always be sought, as infiltration of the pituitary is
a characteristic and common finding in LCH. While posterior pitu-
itary involvement is well recognized, anterior pituitary failure may
also occur necessitating a full endocrine assessment. A rare form of
neurodegeneration affecting the dentate nucleus of the cerebellum
or basal ganglia may occur, giving rise to profound ataxia and cog-
nitive impairment.
Liver, spleen, and lymph node involvement is also common in
patients with multisystem disease. Liver and spleen involvement
confer a poor prognosis. Bone marrow infiltration is likely to be
common in patients with multisystem disease though this is not well
characterized. An unusual but well-​described presentation of LCH
is a discharging ear with associated mastoid bone erosion. A seem-
ingly distinctive entity—​pulmonary LCH—​causing lung infiltration
affects smokers. Cessation of smoking may lead to an improvement
in this condition.
The important consideration is that LCH lesions may be localized
(single system) or multisystem in nature and their clinical manage-
ment and treatment is based on this principle.
In the modern era, the finding of LCH on biopsy should trigger
a full staging protocol including a CT or positron emission tomog-
raphy (PET)/​CT scan, skeletal survey, bone marrow biopsy, endo-
crine screen, and renal and liver profile (Table 22.3.9.2).
Treatment
The management and treatment of LCH depends upon staging.
Patients who present with single system disease can be managed
with simple curettage or local injection of steroids. Radiotherapy
may be considered for adult patients presenting with single bone
lesions.
Patients presenting with multisystem disease will require sys-
temic chemotherapy. In children, a number of clinical trials have
attempted to improve on the standard induction regimen of vin-
blastine and prednisolone (vinblastine 6 mg/​m2 weekly for 6 weeks
and prednisolone 40 mg/​m2/​day for 4 weeks and tailed). No def-
inite benefit has been found for the addition of methotrexate or
etoposide. Maintenance treatment with further courses of vin-
blastine and prednisolone and 6-​mercaptopurine has been found to
prolong remissions.
In adults there are no randomized trials though a number of
chemotherapy agents have proved effective in inducing remission.
The combination of vinblastine and prednisolone appears to be less
effective in this older age group (20% attaining complete remission,
CR), and low-​dose continuous cytarabine (80% CR) or cladribine
(60% CR) are therefore often the preferred regimens.
LCH may be an aggressive disorder, so relapse is not uncommon.
Patients who achieved durable remissions with their first-​line induc-
tion regimens may be retreated with the same agents. For patients
who relapse early, a trial of one of the alternative drugs is reason-
able. Some centres have consolidated such higher-​risk patients
with haematopoietic stem cell transplantation, with some success.
Interestingly, BRAF mutation-​positive cases have been reported to
respond to vemurafenib—​6 out of 18 patients who were known mu-
tation positive benefited from the drug. The addition of a MEK in-
hibitor may be beneficial.
Non-​Langerhans’ cell histiocytosis
(Erdheim–​Chester disease)
The cells that give rise to this disorder appear to be macrophage
derived. They do not stain with S100 proteins or group 1 CD1a
glycoproteins, and electron microscopy of the cell cytoplasm does
not disclose Birbeck granules. The pathology is characterized by
xanthomatous or xanthogranulomatous infiltration with lipid-​laden
or foamy macrophages, usually surrounded by fibrosis. The pathog-
nomonic Touton giant cells may also be seen. Bone biopsy may offer
the greatest likelihood of reaching a diagnosis, and typically shows
infiltration with foamy macrophages staining positive for CD68
but negative for CD1a. Approximately 50% of cases have the BRAF
V600E mutation.
The disease is very rare in children; the mean age of onset is
52  years. The most common presenting symptom is bone pain
with characteristic symmetric infiltration of long bones causing
osteosclerotic lesions. The disorder mimics classical LCH and is
often multisystemic; patients are generally systemically unwell with
fever and weight loss. Central nervous system involvement with
diabetes insipidus occurs, though lymphadenopathy and splenic
involvement are less usual than in classical LCH. Bilateral exoph-
thalmos is a notable clinical feature. The management and treatment
are as for LCH.
A related disorder, at least pathologically, is juvenile
xanthogranuloma which occurs in infants and children and affects
the skin usually as a single nodule. This disorder is normally self-​
limiting and almost never multisystemic.
Follicular dendritic cell sarcoma
Follicular dendritic cell sarcoma (FDCS) is a rare malignant disorder
derived from dendritic cells residing in the follicles of lymph nodes.
The malignant cells are of mesenchymal origin and express markers
of follicular dendritic cell differentiation including CD21, CD35,
R4/​23, and KiM4. The disorder may present at any age and typically,
in the adult, manifests with painless lymphadenopathy often in the
cervical region. Multisystem involvement can occur, particularly in
children, with patients experiencing marked B symptoms of sweats,
fever, and weight loss.
Management requires full staging with CT or PET/​CT and for
those patients with localized disease, meticulous surgical resection
is indicated. Additional involved nodal irradiation may be con-
sidered though a clear benefit of this combined approach remains
unproven. Patients with multisystem disease are usually treated
with lymphoma type protocols such as CHOP (cyclophosphamide,
Table 22.3.9.2  Staging investigations for LCH
Investigation
Details
Laboratory
testing
Full blood count, liver enzymes, urea and creatinine, TSH,
LH, FSH, and glucose
Biopsy
Bone marrow biopsy if liver/​spleen involvement, and/​or
cytopenias
Imaging
CT or PET/​CT and MRI for CNS disease
Other
Water deprivation test
Endoscopy (if malabsorption)
CNS, central nervous system; CT, computed tomography; FSH, follicle stimulating
hormone; LH, luteinizing hormone; MRI, magnetic resonance imaging; TSH, thyroid
stimulating hormone.
22.3.9  Histiocytosis
5261
doxorubicin, vincristine, and prednisolone) or at relapse ifosfamide-​
and platinum-​containing regimens such as ICE (ifosfamide,
carboplatin, and etoposide). The prognosis relates more to the size of
the presenting tumour and the mitotic rate than the clinical stage of
the disease. Finally, there is an association with other disorders such
as Castleman disease and low-​grade lymphoma, the significance of
which is yet to be determined.
Interdigitating follicular cell sarcoma
This is a very rare tumour with a clinical picture that closely resem-
bles FDCS. Management is very similar to cases of FDCS although
outcome is dependent on stage, with early-​stage disease tending to
have a good prognosis.
Macrophage-​related disorders
Haemophagocytic lymphohistiocytosis
Haemophagocytic lymphohistiocytosis (HLH) is a group of dis-
orders characterized by activation of macrophages (macrophage
activation syndrome) leading to a progressive and often fatal char-
acteristic clinical presentation of pancytopenia, coagulopathy, de-
ranged liver function, and fever. The disorder may be familial or
acquired.
Familial HLH
In children and younger adults, the disorder is often caused by spe-
cific genetic mutations, inherited in an autosomal recessive fashion,
which result in disruption of cytotoxic T-​lymphocyte and NK cell
function. In this familial form of HLH, approximately 50% of cases
are found to have mutations in the PRF1 or UNC13D genes. The
PRF1 gene product perforin is one of the key molecules used by T
cells and NK cells to kill virally infected cells. The UNC13D gene
is implicated in the process of exocytosis—​critical for the delivery
of cytotoxic granules by T and NK cells. Other less common gene
mutations have been described which may also be diagnostic for
example in Hermansky-Pudlak and Griscelli syndromes affecting
lysosome-related organelles (see Chapter 12.8). A failure to produce
perforin or the production of an abnormal perforin protein, or a
failure of delivery of cytotoxic granules causes marked derangement
of both T-​cell and NK cell function.
Acquired HLH
In adults, HLH is more commonly secondary to a number of trig-
gering factors and predisposing diseases (Fig. 22.3.9.1) which
include infectious agents such as HIV, Epstein–​Barr virus, and cyto-
megalovirus infection, autoimmune disorders, and haematological
malignancy, most notably lymphoma. It is highly likely that acquired
defects in the immune system underpin these disorders.
Pathology of HLH
Laboratory investigation reveals a marked and progressive pan-
cytopenia. The bone marrow may be reactive and hypercellular in
the early stages but becomes progressively more hypocellular as
the disease advances. Careful inspection of the bone marrow as-
pirate usually reveals macrophages that have ingested elements of
the blood such as red cells, granulocytes, and platelets—​so-​called
haemophagocytosis. While haemophagocytosis is characteristic of
HLH it is by no means diagnostic, being found in a wide range of re-
active states. In nearly all patients there is marked derangement of
clotting with prolongation of the prothrombin and activated partial
thromboplastin time together with a marked hypofibrinogenaemia.
Liver function is usually deranged though renal function is normally
preserved at least in the early stages of the illness. Perhaps the most
useful additional tests in clinical practice are the finding of a mark-
edly raised ferritin (usually in the tens of thousands micrograms/​
litre) and high plasma triglycerides.
Clinical findings
Patients with HLH are systemically unwell with marked pyrexia and
weight loss. Patients may have lymphadenopathy and frequently
have hepatosplenomegaly on clinical examination. There may be
Bacteria
Other
Mycobacterium
tuberculosis
Rickettsia
spp
Escherichia coli
Staphylococcus spp
Toxoplasma
spp
Other
HIV
Other
herpes
Epstein-Barr
virus
Cytomegalovirus
Vaccination
or acute
injuries
Surgery
Drugs
Other
Plasmodium spp
Histoplasma spp
Other
Other
Idiopathic
Transplant
Other AOD
Other systemic
autoimmune diseases
Adult-onset still’s
disease
Systemic lupus
erythematosus
Other neoplasms
Other
haematological
malignancies
Hodgkin
lymphoma
Haemophagocytic lymphohistiocytosis
Leukaemia
Lymphoma
Other
Leishmania
spp
Non-infectious triggers
Viruses
Triggering factors
Parasites
Fungi
Predisposing diseases
Fig. 22.3.9.1  Prompts and predisposing diseases to haemophagocytic lymphohistiocytosis.
Reprinted from The Lancet, Vol. 383, Ramos-​Casal M, Brito-​Zerón P, López-​Guillermo A, Khamashta MA, Bosch X, Adult haemophagocytic syndrome, Pages 1503–​16.
Copyright © 2014, with permission from Elsevier.
SECTION 22  Haematological disorders
5262
clinical evidence of the associated coagulation derangement with
purpura and bleeding. Central nervous system symptoms and signs
reminiscent of encephalitis have been reported in a high proportion
of cases.
Investigations
Laboratory investigations in a patient suspected of haemophagocytic
syndrome include a full blood count, liver and renal function, lac-
tate dehydrogenase, serum ferritin, lipid profile, coagulation screen,
autoimmune serology, C-​reactive protein, viral screen to include
Epstein–​Barr virus and cytomegalovirus polymerase chain re-
action, bone marrow biopsy, and cerebrospinal fluid examin-
ation. Serology for possible infective causes should be considered.
Lymphadenopathy should be biopsied as underlying lymphoma is a
common cause. Mutation analysis looking for underlying HLH mu-
tations (mentioned previously) should be performed.
International collaborative groups led mainly by the Histiocyte
Society have developed clinical criteria for the diagnosis of HLH
(Box 22.3.9.1).
Treatment
HLH is a rapidly progressive and, without treatment, fatal condition.
Modern protocols have successfully treated patients using a combin-
ation of immune suppression and etoposide. A widely used protocol
involves the use of ciclosporin, steroids, and etoposide and, in re-
sponding patients, consideration should be given to consolidating
with stem cell transplantation. Allogeneic stem cell transplant is in-
dicated for all patients identified to have a pathogenic gene defect.
Patients with a macrophage activation syndrome secondary to auto-
immune disorders such as Still’s disease may respond to less aggres-
sive therapy with prednisolone and ciclosporin alone.
Rosai–​Dorfman disease
This is a rare disorder characterized clinically by massive cervical
lymphadenopathy. A proportion of patients have paranasal sinus in-
volvement which can lead to airway obstruction. Skin involvement
may also occur.
Pathologically, the disorder is characterized by the proliferation of
macrophages within the sinuses of involved lymph nodes. A notable
feature is the ingestion of lymphocytes by the proliferating macro-
phages, known as emperipolesis. The disorder is thought to be re-
active and not malignant, the lymph nodes slowly resolving over a
few months.
FURTHER READING
Aricò M (2016). Langerhans cell histiocytosis in children:  from the
bench to bedside for an updated therapy. Br J Haematol, 173, 663–​70.
Brisse E, Matthys P, Wouters CH (2016). Understanding the spectrum
of haemophagocytic lymphohistiocytosis:  update on diagnostic
challenges and therapeutic options. Br J Haematol, 174, 175–​87.
Brisse E, Wouters CH, Matthys P (2016). Advances in the pathogenesis
of primary and secondary haemophagocytic lymphohistiocytosis:
differences and similarities. Br J Haematol, 174, 203–​17.
Grom AA, Horne A, De Benedetti F (2016). Macrophage activation
syndrome in the era of biologic therapy. Nat Rev Rheumatol, 12,
259–​68.
Henter JI, et al. (2007). HLH-​2004: diagnostic and therapeutic guide-
lines for hemophagocytic lymphohistiocytosis. Pediatr Blood
Cancer, 48, 124–​31.
Jordan MB, et  al. (2011). How I  treat hemophagocytic lympho­
histiocytosis. Blood, 118, 4041–​52.
Ramos-​Casal M, et  al. (2014). Adult haemophagocytic syndrome.
Lancet, 383, 1503–​16.
Box 22.3.9.1  Diagnostic features for HLH
Molecular features
A	 Pathological mutations in PRF1, UNC13D, Munc18-​2, Rab27a, STX11,
SH2D1A, or BIRC4
or
B	 Five of the eight following criteria are fulfilled:
1	 Fever ≥38.5°C
2	 Splenomegaly
3	 Cytopenias (affecting at least two of three lineages in the peripheral
blood):
Haemoglobin less than 100 g/​litre
Platelets less than 100 × 109/​litre
Neutrophils less than 1 × 109/​litre
4	 Hypertriglyceridaemia (fasting, >265 mg/​dl) and/​or hypofibrino­
genaemia (<1.5 g/​litre)
5	 Haemophagocytosis in bone marrow, spleen, lymph nodes, or liver
6	 Low or absent NK cell activity
7	 Ferritin greater than 500 mcg/​litre
8	 Elevated sCD25 (α-​chain of soluble interleukin-​2 (sIL-​2) receptor)
Adapted from Jordan MB, et  al. (2011). How I  treat hemophagocytic
lymphohistiocytosis. Blood, 118, 4041–​52 and Henter JI, et  al. (2007).
HLH-​2004:  diagnostic and therapeutic guidelines for hemophagocytic
lymphohistiocytosis. Pediatr Blood Cancer, 48, 124–​31.

22.4 Lymphoid disease 5263 22.4.1 Introduction to
22.4 Lymphoid disease 5263 22.4.1 Introduction to lymphopoiesis 5263 Caron A. Jacobson and Nancy Berliner
22.4
Lymphoid disease
CONTENTS
22.4.1	 Introduction to lymphopoiesis  5263
Caron A. Jacobson and Nancy Berliner
22.4.2	 Acute lymphoblastic leukaemia  5269
H. Josef Vormoor, Tobias F. Menne, and Anthony V. Moorman
22.4.3	 Hodgkin lymphoma  5280
Vijaya Raj Bhatt and James O. Armitage
22.4.4	 Non-​Hodgkin lymphoma  5288
Vijaya Raj Bhatt and James O. Armitage
22.4.5	 Chronic lymphocytic leukaemia  5302
Clive S. Zent and Aaron Polliack
22.4.6	 Plasma cell myeloma and related monoclonal
gammopathies  5310
S. Vincent Rajkumar and Robert A. Kyle
22.4.1  Introduction to lymphopoiesis
Caron A. Jacobson and Nancy Berliner
ESSENTIALS
Lymphoproliferative disorders occur when the normal mechanisms
of control of proliferation of lymphocytes break down, resulting
in autonomous, uncontrolled proliferation of lymphoid cells and
typically leading to lymphocytosis and/​or lymphadenopathy, and
sometimes to involvement of extranodal sites (e.g. bone marrow).
Causes of lymphoproliferative disorder
These include (1) malignant—​clonal in nature, resulting from the un-
controlled proliferation of a single transformed cell (e.g. lymphoma);
(2)  nonmalignant—​polyclonal lymphoproliferative disorders may
result from conditions including (a)  infections—​lymphocytosis is
commonly caused by viral infections (e.g. Epstein–​Barr virus (EBV)),
lymphadenopathy is a common feature of a very wide variety of
infections; and (b)  reactive—​conditions such as systemic lupus
erythematosus and sarcoidosis frequently cause lymphadenopathy.
Clinical approach
Distinguishing between the lymphoproliferative disorders clinically
and pathologically is not always easy.
Clinical assessment—​when eliciting the history of a patient with
suspected lymphoproliferation, particular note should be taken
of their general health, the type and duration of any constitutional
symptoms, and any episodes of recent infection/​exposure to drugs/​
travel. Thorough examination of all lymph node sites is required, as is
careful examination of the oropharynx, tonsils, skin, spleen, and liver.
Investigation—​whenever a lymphoproliferative disorder is sus-
pected, the key initial investigation is the full blood count and
examination of the blood film, sometimes augmented by immuno-
cytochemistry and flow cytometry. Depending on clinical context,
other investigations may include (1)  serological studies for viral
pathogens; (2)  serological studies for rheumatological disease;
(3) imaging for mediastinal and intra-​abdominal lymphadenopathy;
(4) bone marrow examination; and—​if no diagnosis is apparent—​(5)
lymph node biopsy. However, there are many pitfalls in morpho-
logical interpretation of lymph node histology, which is a matter for
the specialist, who will often draw on supplementary information
from flow cytometry, cytogenetics, and immunoglobulin/​T-​cell re-
ceptor gene rearrangement studies to demonstrate the clonal nature
of malignant disease and provide data with prognostic and thera-
peutic significance, or to identify the presence of specific viruses such
as EBV and human herpesvirus 8.
Introduction
The human immune system has the capacity to identify and respond
specifically to invading pathogens. It can also ‘remember’ the ex-
posure, such that subsequent exposure to the same pathogen results
in a more rapid and potent immune response. Lymphocytes play the
key role in the adaptive immune response, mediating both specifi-
city and memory.
Lymphocytes
The lymphocytes can be divided into two morphologically indistin-
guishable types, which play different and complementary roles in
the immune system. Both are derived from lymphohaematopoietic
section 22  Haematological disorders
5264
stem cells that reside in fetal liver and in adult bone marrow. B cells
develop in the marrow (the human equivalent of the avian bursa of
Fabricius) and their principal role is to generate immunoglobulin
(antibodies). B cells represent about 20% of the lymphocyte popu-
lation in peripheral blood. T cells, which mature within the thymus,
orchestrate the immune response: they are capable of cell-​mediated
cytotoxicity, they generate inflammatory cytokines, and they pro-
vide help for B-​cell function. T cells account for approximately 80%
of the lymphocytes in the peripheral circulation. A much smaller
population of lymphoid-​appearing cells express neither B-​cell nor
T-​cell markers. These null cells, also known as natural killer (NK)
cells and large granular lymphocytes, are capable of cell-​mediated
cytotoxicity, especially against tumour cells and virally infected cells.
NK cells are a component of the innate immune response, as they do
not demonstrate immunological memory.
Lymph nodes
In their role in infection surveillance, lymphocytes circulate through
the body via a network of lymphatic and blood vessels. At strategic
locations, lymphoid cells are organized to allow direct interaction
among lymphocytes and other specialized cells of the immune
system.
These interactions permit the production of specific, functional
effector cells. The network includes approximately 500 to 600 dis-
crete lymph nodes, lymphoid populations in the oropharynx
(Waldeyer’s ring), bronchial tree, and gut, as well as in the thymus,
the bone marrow, and the spleen.
Within lymph nodes, lymphocytes are arranged in a central me-
dulla surrounded by an outer cortex contained within a connective
tissue capsule (Fig. 22.4.1.1). Afferent lymphatics penetrate the
cortex and lymphocyte-​rich fluid filters towards the medullary si-
nusoids and the efferent lymphatics at the hilum of the node. The
vascular supply to the lymph node includes specialized postcapillary
venules that allow the passage of peripheral blood lymphocytes into
the node. Lymphocytes are ultimately returned to the bloodstream
via the thoracic duct.
Roughly spherical follicles are found in the lymph node cortex
and predominantly comprise B cells. Primary follicles contain clus-
ters of naive, unstimulated B cells. Secondary follicles, with pale ‘ger-
minal centres’ surrounded by a darker ‘mantle’ zone, represent foci
of B cells proliferating and differentiating in the presence of antigen-​
bearing dendritic cells and activated ‘helper’ T cells (Th cells). The
interfollicular and paracortical zones of the lymph node are densely
populated by T cells. Macrophages, follicular dendritic cells, and
interdigitating dendritic cells all process and present antigens to the
lymphocytes within the node.
The design of the lymph node facilitates the process whereby
the subpopulation of lymphocytes capable of responding to a spe-
cific antigen is expanded. Antigens are delivered to the subcapsular
sinus of the node via afferent lymphatics, and are taken up by den-
dritic cells and presented on their surface in the context of the major
histocompatibility complex (MHC) proteins. Specific T-​lymphocyte
responses require that peptide antigens, derived from ‘foreign’ pro-
teins, appear on the surface of antigen-​presenting cells in close as-
sociation with a ‘self’ MHC molecule. B cells, on the other hand,
are capable of responding to some antigens in solution. Optimal B-​
cell responses require the ‘help’ of T cells both via direct cell–​cell
contact and in response to cytokines secreted by T cells. Only those
T cells and B cells that have been genetically preprogrammed to
interact with a specific antigen will proliferate and differentiate in
response to it.
Antigen receptors
Both B and T cells express transmembrane receptors on their cell
surfaces. These proteins bind antigens and define the antigenic
specificity of the cell. In the case of B cells, the immunoglobulin
molecule serves as the B-​cell receptor (Fig. 22.4.1.2). Each im-
munoglobulin molecule is a bivalent tetramer comprising a pair
of heavy chains bound to two light chains (of either κ or λ type).
Genetic recombination of approximately 400 immunoglobulin
gene segments (located on chromosomes 2, 14, and 22) generates
about 1015 distinct antibody specificities. The expression of recom-
bination activating genes (RAG1 and RAG2) early in B-​cell develop-
ment mediates the random rearrangement of variable (V), diversity
(D), and joining (J)  gene segments. Terminal deoxynucleotidyl
transferase (TdT) contributes to the diversity of immunoglobulin
molecules by inserting additional nucleotides during the splicing
of gene segments. This process gives rise to a vast repertoire of anti-
body molecules, each with a unique antigen-​binding cleft. All of
the progeny cells of a B cell that has rearranged its immunoglobulin
genes have the same antigenic specificity and are referred to as a
clone. Most protein antigens are complex and contain many dif-
ferent epitopes (structures capable of binding an antigen receptor).
Therefore, most pathogens stimulate many lymphocyte clones to
proliferate: that is to say, they result in polyclonal responses. As
B-​cell clones mature, the isotype of the antibodies they produce
‘switches’ from IgM/​IgD to IgG, IgA, or IgE.
In an analogous fashion, T-​cell precursors rearrange the T-​cell re-
ceptor (TCR) genes. The TCR consists of a heterodimer of α and
β chains, or γ and δ chains in a minority of T cells. The α and β
genes are encoded on chromosomes 14 and 7, respectively, while
the γ and δ chains are on chromosomes 7 and 14, respectively. T-​cell
Paracortex
Postcapillary
venule
Medulla
Follicular
centre
Lymphocyte
mantle
Efferent
lymphatic
Cortex
Medullary
cord
Subcapsular
sinus
Fig. 22.4.1.1  Functional architecture of a normal lymph node.
From Arno J (1980). Atlas of lymph node pathology, with permission.
22.4.1  Introduction to lymphopoiesis
5265
precursors randomly assemble V, J, and D gene segments to generate
a vastly diverse array of antigen-​specific T-​cell clones. When the T
cell encounters antigen to which it can productively bind, the cell
undergoes clonal expansion, and generates both activated effector
cells and long-​lived memory cells.
Lymphocyte ontogeny
As lymphocytes develop and mature from multipotent progenitors
to terminally differentiated effector cells, they express a sequential
pattern of surface proteins. Some of these cell surface molecules sub-
serve known, critical functions in the cells that bear them. Others
are of less clear biological significance, but are useful markers of
cell type and status of differentiation and activation. Malignant
lymphomas and lymphoid leukaemias are frequently classified and
understood on the basis of their expression of cell surface markers
(Fig. 22.4.1.3). In some cases, the stage of differentiation at which
malignant transformation occurred can be inferred from the pattern
of the surface antigens expressed by the malignant cells.
Lymphocytes develop from bone marrow-​derived haemopoi-
etic stem cells. Although the surface characteristics of these elu-
sive cells are not well understood, it is likely that human stem cells
express the cell surface glycoprotein CD34. The first recognizable
sign of commitment to the B-​lymphoid lineage is the expression of
TdT and the rearrangement of the immunoglobulin heavy chain.
(a)
H
Variable
region
(b)
(c)
(d)
(e)
Constant
region
Membrane
Heavy chain mRNA
J C
D
V
V
(150)
D
(1–25)
J
(1–6)
C
H
L
L
μ δ γ
μ
3γ1α1γ 2γ 4
α2
ε
Fig. 22.4.1.2  Immunoglobulin gene rearrangement. The top line
(A) represents the germ-​line pattern of the immunoglobulin heavy-​
chain locus found on human chromosome 14. B-​cell progenitors
express recombination activating genes that mediate the random,
sequential rearrangement of gene modules (lines B and C) such that
only one of several variable (V)1, diversity (D), and joining (J) segments is
expressed by a B-​cell clone (line D). As the gene components are spliced,
terminal deoxynucleotidyl transferase (TdT) randomly inserts additional
nucleotides at splice junctions. Diverse antigenic specificity is thus
somatically generated from a relatively small amount of genetic material.
The immunoglobulin molecule (line E) is a tetramer of two heavy and two
light chains that may be cell associated (as shown) or secreted. The region
of the molecule that interacts specifically with antigen is the variable
region. The constant region of the light chain is of either the κ or λ type.
The constant region of the heavy chain determines the isotype of the
antibody (IgM, IgD, IgG, IgA, or IgE).
Fig. 22.4.1.3  Simplified depiction of lymphocyte ontogeny. Lymphocytes derive from lymphoid progenitors in the bone marrow,
which in turn are derived from multipotent haemopoietic stem cells. B-​lymphoid progenitors are recognized by their expression of
terminal deoxynucleotidyl transferase (TdT) and the rearrangement of the immunoglobulin heavy-​chain locus. As B-cells mature, the
light chain is rearranged and immunoglobulin is expressed first within the cell cytoplasm, then on the cell surface, and is ultimately
secreted. T-​lymphoid progenitors migrate to the thymus where they express TdT and rearrange the β-​subunit followed by the α-​subunit
of the T-​cell receptor (TCR). An overlapping sequence of cell surface proteins are expressed as the cells differentiate, these have been
numerically classified using cluster of differentiation (CD) designations. The status of the immunoglobulin and TCR genes are represented
as follows: α, TCR-​α; β, TCR-​β; G, germ line; H, immunoglobulin heavy chain; L, light chain; R, rearranged.
section 22  Haematological disorders
5266
As differentiation progresses, B-​cell progenitors express class  II
MHC molecules (HLA DR) as well as CD19 and then CD10 (the
latter is also known as the ‘common acute lymphoblastic leu-
kaemia antigen’, CALLA). The immunoglobulin light chain is re-
arranged and the cells (now termed pre-​B cells) express the µ heavy
chain within their cytoplasm. As the cells progress to the early B-​
cell stage, CD34, TdT, and CD10 expression are extinguished, and
CD19, CD20, and CD21, as well as IgM, are expressed on the sur-
face of the cells. Mature B cells express surface IgM and/​or IgD, in
addition to CD19 and CD20. Plasma cells, the end result of B-​cell
differentiation, produce cytoplasmic as well as secreted immuno-
globulin, but do not express surface immunoglobulin. They lack
CD19 and CD20 expression.
Similarly, as T cells mature, they progress through an orderly cas-
cade of genetic and cell surface events. CD34-​positive progenitors
that are destined for a T-​lymphoid fate migrate from the marrow to
the thymus and express TdT as well as CD7 and CD2. The TCR genes
are then rearranged and subsequently expressed on the surface of the
T cell (thymocyte) in association with a protein complex, the CD3
molecule. CD3 is used to identify T cells by immunohistochemistry
and flow cytometry. Distinct populations of mature T cells emerge
from the thymus: those that express CD4 and function as cytokine-​
secreting ‘helper’ cells and those that express CD8 and function as
cytotoxic ‘killer’ cells. Rare ‘double positives’ (CD4+CD8+) and
‘double negatives’ (CD4–​CD8–​) also exist. The CD4 molecule medi-
ates the binding of T cells to MHC class II molecules, whereas CD8
binds MHC class I proteins.
The third descendant of the lymphoid stem cell, the NK cell, is
characterized by its expression of CD7, CD2, CD16, and CD56, in
addition to other surface proteins. NK cells are distinguished from
T cells by the fact they do not express CD3 (and therefore the TCR).
Lymphoproliferative disorders
A variety of conditions spanning the spectrum of benign, reactive
processes to frank malignant transformation results in the expan-
sion of lymphocyte populations. The lymphoproliferative dis-
orders are a loosely defined group of malignant and nonmalignant
entities characterized by the autonomous, poorly controlled prolif-
eration of lymphoid cells. Lymphoproliferation is typically mani-
fested by lymphocytosis and/​or lymphadenopathy. In addition,
lymphoproliferation may involve extranodal sites, including bone
marrow, liver, skin, and soft tissues. Distinguishing among the
lymphoproliferative disorders clinically and pathologically is not
always easy. Malignant tumours are clonal in nature; they result
from the uncontrolled proliferation of a single transformed cell. In
contrast, nonmalignant lymphoproliferation contains polyclonal
lymphocyte populations. Lymphoproliferative disorders may result
from chronic antigenic stimulation, certain viral infections, or from
an imbalance among interacting lymphocyte populations, as may
occur in congenital or acquired immunodeficiency syndromes. In
addition, lymphocytes are prone to the acquisition of chromosomal
translocations, particularly involving the immunoglobulin and TCR
genes, and such changes may contribute to malignant transform-
ation (Table 22.4.1.1).
Lymphocytosis
Normal peripheral blood usually contains approximately 1000 to
5000 lymphocytes/​µl, accounting for approximately 40% of the
circulating leucocytes. Infants and young children typically have
higher absolute lymphocyte counts. Increased numbers of cir-
culating lymphocytes (lymphocytosis) and/​or the appearance of
Table 22.4.1.1  Causes of lymphadenopathy
Clinical features
Histological characteristics
Infectious
Bacterial
Regional, often tender
Suppurative
Mycobacterial (tuberculosis, leprosy)
Regional or generalized
Suppurative granulomas
Viral (EBV, CMV, HIV)
Often generalized
Follicular hyperplasia
Fungal (Histoplasma, Coccidioides spp.)
Often hilar
Suppurative granulomas
Parasitic (Toxoplasma, Chlamydia spp.)
Usually regional (cervical, inguinal)
Suppurative granulomas
Reactive
Rheumatological conditions (SLE, RA)
Often generalized
Follicular hyperplasia
Sarcoidosis
Especially hilar
Epithelioid granulomas
Drugs (e.g. phenytoin)
Generalized
Paracortical expansion
Castleman’s disease
Localized/​multicentric
Follicular hyperplasia (hyaline vascular or plasma cell)
Rosai–​Dorfman disease
Usually cervical
Sinus hyperplasia
Neoplastic
Leukaemia/​lymphoma
Often generalized, ‘rubbery’
Effacement of nodal architecture
Metastatic (carcinoma, melanoma)
Regional, rock hard
Subcapsular expansion, effacement of nodal architecture
Other
Storage diseases (e.g. Gaucher disease)
Generalized
Paracortical or sinusoidal lipogranulomas
EBV, Epstein–​Barr virus; CMV, cytomegalovirus; HIV, human immunodeficiency virus; SLE, systemic lupus erythematosus; RA, rheumatoid arthritis.
22.4.1  Introduction to lymphopoiesis
5267
abnormal (or atypical) lymphocytes in the blood are usually caused
by either viral infection or lymphoid malignancy. The appearance
of the circulating lymphocytes on a peripheral blood smear may
provide clues to the pathogenesis of the elevated lymphocyte count.
For example, infectious mononucleosis results from primary in-
fection with the Epstein–​Barr virus (EBV), and gives rise to large
numbers of ‘atypical’ lymphocytes with abundant cytoplasm in the
peripheral blood. Chronic lymphocytic leukaemia (CLL) leads to
an increase in circulating normal-​appearing ‘mature’ lymphocytes.
CLL is also frequently associated with the appearance of ‘smudge’
cells in the peripheral smear, a preparation artefact caused by the
destruction of the fragile CLL cells. Follicular lymphoma may be
associated with the circulation of characteristic cells with a cleaved
nucleus, while hairy cell leukaemia and splenic marginal zone
lymphoma can present with an abundance of circulating atypical
lymphocytes with villous projections from their cell surface. Adult
T-​cell leukaemia/​lymphoma is a mature T-​cell malignancy caused
by human T-​lymphotropic virus infection and is often associated
with the detection of ‘flower’ cells on the peripheral blood smear
(Fig. 22.4.1.4).
Lymphadenopathy
Enlargement of one or more lymph nodes (lymphadenopathy)
is an extremely common clinical finding. With the exception of
inguinal nodes, normal lymph nodes are nonpalpable. Nodes that
are palpable and/​or exceed approximately 1 × 1 cm on imaging
studies are considered pathological. Lymph node enlargement
often results from the body’s normal and adaptive response to an
immunological challenge; however, it may signify a pathological
inflammatory or malignant disease. The causes of lymphaden-
opathy fall into three main categories: infectious, inflammatory
(reactive), and neoplastic (Table 22.4.1.1). Younger patients, es-
pecially children, are more likely to develop adenopathy as a result
of infection, while the likelihood of haematological or metastatic
malignancy increases with age.
Approach to the patient
with suspected lymphoproliferation
The evaluation of the patient with a suspected lymphoproliferative
disorder should take into account the age and general health of the
patient, the duration of the adenopathy, the coexistence of fever,
weight loss, night sweats, pruritus, and cough, as well as any recent
infections, medications, travel, and animal exposures. The physical
examination should make note of the location (generalized vs re-
gional), the texture (hard vs rubbery), and the mobility (fixed vs
mobile) of the lymph nodes, and the presence or absence of as-
sociated signs of inflammation (warmth, tenderness, erythema).
The skin and oropharynx should be examined and the size of the
liver and spleen should be assessed. Additional screening studies
may include a complete blood count, measurement of the erythro-
cyte sedimentation rate, and/​or C-​reactive protein. The level of
lactate dehydrogenase may be elevated. Serological studies for
certain viral pathogens and for rheumatological diseases can be
helpful. Radiographs of the chest should be obtained if medias-
tinal adenopathy is suspected. Ultrasound of enlarged nodes may
demonstrate central suppuration, which is characteristic of acute
lymphadenitis. Axial imaging (e.g. CT) is required to diagnose
intra-​abdominal adenopathy.
Biopsy
When a lymphoproliferative disorder is suspected, pathological
analysis of involved tissue is necessary to determine the spe-
cific diagnosis. In some cases, analysis of peripheral blood and/​
or bone marrow may yield a diagnosis. However, lymph node
biopsy is often needed. Before proceeding to biopsy, a trial of
observation with or without empirical antibiotics (usually an
antistaphylococcal agent) may be appropriate in some patients
with lymphadenopathy. However, empirical treatment with ster-
oids should be avoided because it may undermine the diagnosis
and proper therapy of lymphoid malignancy. If the lymphadenop-
athy does not improve within 2 weeks, then a lymph node biopsy
should be strongly considered. The largest accessible node is most
often selected for biopsy. A fine needle aspiration of lymph nodes
is adequate for diagnosis in a restricted set of clinical circum-
stances: for example, diagnosis of recurrent disease or metastatic
carcinoma or melanoma. Culture of a lymph node aspirate may
yield a microbiological diagnosis in infective lymphadenitis. Most
pathologists prefer an excisional biopsy, when possible, because
nodal architecture is preserved. A portion of the sample should be
reserved fresh (i.e. not fixed in formalin) for flow cytometry and
cytogenetic studies, if indicated.
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 22.4.1.4  Examples of peripheral blood smears for several
lymphoproliferative disorders. (a) Atypical lymphocytes seen in infectious
mononucleosis; (b) smudge cells seen in chronic lymphocytic leukaemia;
(c) small cleaved lymphocytes seen in follicular lymphoma; (d) hairy
lymphocytes seen in hairy cell leukaemia; (e) villous lymphocytes seen in
splenic marginal zone lymphoma; and (f) flower cells seen in human
T-​lymphotropic virus-​1-​associated adult T-​cell leukaemia/​lymphoma.
section 22  Haematological disorders
5268
Histological examination of lymph nodes is the mainstay of diag-
nostic studies, however nondiagnostic or nonspecific inflammatory
findings are frequently encountered. Reactive lymph nodes demon-
strate characteristic, but by no means specific, histological patterns
that involve the three functional domains of the lymph node: the fol-
licles, the paracortex, and the medullary sinuses.
An increase in the size and/​or number of lymphoid follicles
(which contain proliferating B cells) is termed ‘follicular hyper-
plasia’. The specific cause is rarely identified. This pattern of lymph
node reactivity is characteristic of rheumatological conditions,
HIV infection, Castleman’s disease, and IgG4-​related disease.
Castleman’s disease is a rare and poorly understood non-​neoplastic
cause of lymphadenopathy that occurs in localized and multicentric
forms. The multicentric form is a systemic illness without defined
therapy that is associated with infection with human herpesvirus-​
8 (HHV-​8, also known as Kaposi’s sarcoma herpesvirus). IgG4-​
related disease is a relatively newly recognized, nonmalignant
entity that involves a lymphoplasmacytic infiltration of tissue, most
commonly involving lymph nodes, lacrimal and salivary glands,
lung, pancreas, thyroid, and retroperitoneum. Its pathogenesis is
poorly understood but it is defined by a marked infiltration of IgG4-​
positive plasma cells and CD4+ T cells, usually associated with fi-
brosis and elevated serum levels of IgG4.
Paracortical expansion accompanies T-​cell proliferation and is
characteristic of certain viral causes of lymphadenopathy, such as
EBV infection. Paracortical expansion with granuloma formation
is typical of mycobacterial infections and sarcoidosis. In Kikuchi’s
disease and Kawasaki’s disease (mucocutaneous lymph node syn-
drome), paracortical necrosis is seen in involved lymph nodes.
Sinus hyperplasia is caused by an increased number of histiocytes
in the medullary sinuses. This pattern of lymph node reactivity is
seen in the histiocytic syndromes and in storage diseases. A rare
condition known as sinus histiocytosis with massive lymphaden-
opathy or Rosai–​Dorfman disease is characterized by an extreme
polyclonal proliferation of macrophages. This entity often involves
the cervical lymph nodes, but may occur in virtually any nodal or
extranodal site and is usually, but not always, self-​limited.
Involvement by a malignant lymphoma leads to effacement of
the lymph node structure to a greater or lesser degree. Histology
correlates with clinical behaviour and will be described in subse-
quent sections focused on the classification of lymphoma. Histology
alone may be inadequate to distinguish the malignant from the
nonmalignant lymphoproliferative disorders. Supplemental infor-
mation from flow cytometry, cytogenetics, and immunoglobulin/​
TCR gene rearrangement studies demonstrate the clonal nature of
malignant disease and provide data with prognostic and therapeutic
significance.
Immunohistochemistry and flow cytometry
Immunohistochemistry is used to characterize the pattern of surface
marker expression in fixed or frozen tissue samples. Flow cytometry
is performed on cells in suspension, such as peripheral blood or
bone marrow, or on cell suspensions prepared from a lymph node
or other solid tumour. For flow cytometry, solid specimens should
not be fixed or frozen but kept refrigerated until processing. Both
techniques detect the binding of monoclonal antibodies of known
specificity to the clinical sample. Using a panel of antibodies,
these studies demonstrate the types of cells present in the sample.
Nonhaemopoietic metastatic tumours can be identified. The lineage
of lymphoid malignancies can be revealed (e.g. B cell vs T cell vs NK
cell). In the case of B-​cell lymphoproliferation, the relative expres-
sion of κ and λ light chains can be measured. As described previ-
ously, B cells express either the κ or the λ light chain, but not both.
Predominant expression of either the κ or λ light chain by a popula-
tion of B cells, a phenomenon known as light-​chain restriction, sug-
gests a clonal process. Using flow cytometry, lymphoid neoplasms
can be placed within the hierarchy of normal lymphocyte ontogeny,
and clinical behaviour, such as response to cytotoxic therapy, can
often be predicted. These studies may be used to demonstrate the
presence of a surface antigen to which monoclonal antibody-​based
therapy has been developed (e.g. CD20 and rituximab; CD30 and
brentuximab). Sometimes, malignant cells demonstrate lineage in-
fidelity, with expression of a pattern of surface markers that does
not correspond to a normal cellular counterpart. This may fortuit-
ously provide an immunophenotypical fingerprint to detect small
amounts of disease, early relapse, or minimal residual disease after
therapy.
Genetic studies
The high proliferative rate of lymphocytes and their intrinsic gen-
etic instability set the stage for the development of chromosomal
translocations that are aetiologically linked to malignant trans-
formation. Increasingly, haemopoietic cancers are being defined
genetically by the presence of specific, nonrandom chromosomal
translocations. The detection and study of these translocations
has increased diagnostic precision, provided insights into the
molecular mechanisms of oncogenesis, and revealed molecular
targets for rational therapeutic design. Chromosomal transloca-
tions can be demonstrated using classical cytogenetic techniques.
Additionally, specific gene rearrangements may be detected using
the polymerase chain reaction (PCR) and/​or fluorescence in situ
hybridization. As experience with these specialized studies in
lymphoproliferative disorders has accumulated, certain genetic
abnormalities have become highly associated with specific clinical
entities. For example, rearrangement of the c-​MYC oncogene on
chromosome 8 with the immunoglobulin heavy-​chain gene locus
on chromosome 14 (t(8;14)) is detected in the majority of patients
with Burkitt lymphoma and its presence may be used to support
this diagnosis. The majority of cases of follicular lymphoma will
harbour a t(14;18) translocation, which involves the gene for
BCL2 on chromosome 18 and the immunoglobulin heavy-​chain
gene locus on chromosome 14. However, this translocation can be
found in a small proportion of nonmalignant B lymphocytes in pa-
tients without lymphoma. Other examples are discussed in detail
in subsequent chapters in this text. In some circumstances, these
techniques are applied to the detection of minimal residual disease
during and after therapy. In addition, molecular methods may be
used to identify the presence of specific viral sequences, such as
those encoded by EBV and HHV-​8.
As described earlier, the hallmark of lymphocyte differentiation is
the somatic rearrangement of the antigen receptor genes, immuno-
globulin in the case of B cells and the TCR in the case of T cells. Each
lymphocyte clone has a unique arrangement of the components of
the antigen receptor genes, while cells of nonlymphocyte lineages
preserve the germ-​line structure of these genes. Lymphoproliferative
malignancies are composed of clonal proliferations arising from a

22.4.2 Acute lymphoblastic leukaemia 5269 H. Josef
22.4.2 Acute lymphoblastic leukaemia 5269 H. Josef Vormoor, Tobias F. Menne, and Anthony V. Moorman
22.4.2  Acute lymphoblastic leukaemia
5269
single cell with a rearranged antigen receptor locus. The pattern of
gene rearrangement helps to characterize the lineage and stage of
differentiation of the tumour. For example, pre-​B-​cell acute lympho-
blastic leukaemia cells usually contain rearranged heavy-​chain
genes with germ-​line light-​chain genes, whereas B-​CLL cells usually
have a rearrangement of both heavy-​ and light-​chain genes and ex-
press surface immunoglobulin. Analysis as to whether the malig-
nant lymphocytes have undergone somatic hypermutation also has
prognostic significance in some diseases, such as CLL. The degree of
genetic difference between the immunoglobulin heavy-​chain gene
in the malignant clone from germ-​line sequence correlates with
outcome in this disease, with an unmutated sequence being associ-
ated with a poor prognosis. Furthermore, since clonal populations
of lymphocytes all contain the same antigen receptor rearrange-
ment, these cells possess a ‘molecular signature’ that is unique to the
malignant clone.
Consequently, antigen receptor rearrangements have become the
target of DNA diagnostic techniques for diagnosing and following
lymphoproliferative malignancies. Antigen receptor rearrange-
ments can be detected largely by PCR-​based techniques. For these
studies, PCR is performed using oligonucleotide primers based
on conserved sequences within the immunoglobulin heavy-​chain
locus; approximately 70 to 90% of rearrangements can be detected
by this approach. To detect minimal residual disease with maximal
sensitivity, such rearrangements are then subjected to sequence ana-
lysis to determine the antigen-​specific sequences unique to the tu-
mour rearrangement. An allele-​specific oligonucleotide can then
be synthesized and used in a PCR analysis that can detect residual
clonal populations representing as few as 1 in 105 cells.
FURTHER READING
Cooper MA, et al. (2006). Lymphocyte biology. In: Young NS, Gerson
SL, and High KA (eds) Clinical hematology, pp. 71–​88. Elsevier,
Philadelphia.
Delves P, Roitt I (2000). Advances in immunology 1. N Engl J Med,
343, 37–​49.
Delves P, Roitt I (2000). Advances in immunology 2. N Engl J Med,
343, 108–​17.
Foon KA, Todd RF 3rd (1986). Immunologic classification of leukemia
and lymphoma. Blood, 68, 1–​31.
Look A (1997). Oncogenic transcription factors in human acute leuke-
mias. Science, 278, 1059–​64.
MacIntyre EA, Delabesse E (1999). Molecular approaches to the diag-
nosis and evaluation of lymphoid malignancies. Semin Hematol, 36,
373–​89.
Rao DS (2017). Overview and compartmentalization of the immune
system. In: Hoffman R, et al. (eds) Hematology: basic principles and
practice, 7th edition, pp. 199–​209. Elsevier, Philadelphia.
Rose MG, Degar BA, Berliner N (2004). Molecular diagnostics of ma-
lignant disorders. Clin Adv Hematol Oncol, 2, 650–​60.
Sell S (1996). Immunology, immunopathology, and immunity. Appleton
& Lange, Stamford, CT.
Strauchen J (1998). Diagnostic histopathology of the lymph node.
Oxford University Press, New York.
Wickremasinghe R, Hoftbrand A (1999). Biochemical and genetic
control of apoptosis: relevance to normal hematopoiesis and hema-
tological malignancies. Blood, 93, 3587–​600.
22.4.2  Acute lymphoblastic
leukaemia
H. Josef Vormoor, Tobias F. Menne,
and Anthony V. Moorman
ESSENTIALS
Acute lymphoblastic leukaemia (ALL) is a malignant proliferation
of lymphoid blasts, most commonly of B-​lineage origin. The clin-
ical symptoms and signs are either a consequence of bone marrow
failure (infections, bruising, petechiae, pallor, and tiredness) or a
consequence of the uncontrolled proliferation of the blasts (lymph-
adenopathy, hepatosplenomegaly, and cranial nerve palsies). Its peak
incidence is in young children but ALL occurs at all ages.
More than 80% of all affected children are cured with modern
chemotherapy, but unfortunately the outcome of adults is much
worse despite some improvements led by the introduction of
paediatric-​inspired protocols and tyrosine kinase inhibitors in
BCR-​ABL1-​positive ALL.
Standard chemotherapy for ALL consists of several months of
intensive multidrug induction, consolidation and intensification
chemotherapy (including steroids, vincristine, asparaginase and
anthracyclines), intrathecal methotrexate to target blasts in the central
nervous system, and low-​intensity maintenance therapy (with oral 6-​
mercaptopurine and methotrexate) for up to 3 years. Treatment is
stratified according to the response and other prognostic biomarkers
(including genetics). Allogeneic haematopoietic stem cell transplant-
ation is used predominantly in the relapse setting for children but
in frontline therapy for adult patients to consolidate chemotherapy.
Novel targeted small molecules and, in particular, immunotherapy
are promising to offer new treatment options for patients with high-​
risk or relapsed disease.
Introduction
The treatment of acute lymphoblastic leukaemia (ALL) in children
constitutes one of the success stories of modern medicine. A  le-
thal disease in the 1960s, now over 80% of affected children are
cured. Initially, single agents, such as methotrexate, pioneered by
Sidney Farber, were used to achieve temporary remissions. The big
breakthrough, however, came with the introduction of multidrug
chemotherapy to prevent the evolution of chemotherapy-​resistant
subclones, combined with therapy targeting leukaemia cells in the
brain (craniospinal irradiation). These approaches were pioneered
and optimized by the first generation of physicians specializing in
childhood leukaemia, including Donald Pinkel in the United States
of America and Hansjörg Riehm in Europe. Both initially faced
major resistance in the medical community, which widely believed
that children with cancer should be palliated rather than exposed
to experimental therapies. Paediatric haematology and oncology
underwent a steep learning curve as the intense treatment caused
toxic side effects, in particular life-​threatening infections. Learning
to manage these side effects and developing supportive care played
section 22  Haematological disorders
5270
a key role in facilitating the improvement in outcomes for children
with ALL. More recently, stratifying therapy according to predictive
biomarkers (genetics) and response to treatment has allowed further
tailoring of management. This personalized or stratified approach
has been the basis for further improvement in outcomes for slow re-
sponders and a reduction of toxicity in low-​risk patients.
Treatment of ALL in adults has proven to be more challenging,
partly as the leukaemias are more resistant to chemotherapy and
partly as there is a reduced treatment tolerance particularly in eld-
erly patients. However, the introduction of paediatric-​inspired
protocols for young adults, the introduction of tyrosine kinase in-
hibitors in the treatment of BCR-​ABL1-​positive ALL, and improved
supportive care have all led to significant improvements in the sur-
vival and cure rate for adult patients. Novel targeted therapies and,
in particular, the successful introduction of immunotherapy into
the treatment of ALL promise to further improve the treatment of
this condition.
This spectacular progress in the treatment of ALL is underpinned
by decades of research that have increased our understanding
of the origin, key drivers, and genomic complexity of the different
types of ALL.
Epidemiology and pathogenesis
ALL occurs at all ages but is more prevalent among children than
adults and, importantly, accounts for a much higher proportion of
cancers in that age group. In the United Kingdom, approximately
600 new patients are diagnosed each year with over 50% of cases
occurring in patients less than 15 years old. The incidence of ALL
among younger children aged between 1 and 4 years is at 6 to 7 per
100 000 per year, and this is often referred to as the ‘childhood peak’.
The incidence in infants and older children is 2 to 3 per 100 000/​year
and drops to 1 to 2 per 100 000/​year among adults.
Although the precise aetiology of ALL remains unknown, it is now
clear that several factors play a role in causation, including chance,
exogenous exposures, endogenous exposures, and the individual’s
genetic background. Epidemiological and laboratory-​based studies
indicate a two-​stage process. The initial step is the development of
a preleukaemic clone which arises when a normal cell acquires a
genetic abnormality, often a chromosomal translocation, as the re-
sult of a random error in DNA replication. The preleukaemic clone
can lie dormant for several years but is susceptible to the acquisition
of additional genetic or epigenetic abnormalities that promote de-
velopment of the disease and the onset of symptoms. In childhood
ALL, it has been demonstrated by elegant twin and back-​tracking
studies that this first stage occurs in utero in the majority of cases.
A number of exogenous (e.g. infections, chemicals, and radiation)
and endogenous (e.g. inflammation) factors have been postulated
as potential triggers in the development of full-​blown leukaemia.
However, conclusive proof remains elusive because individuals har-
bouring a silent preleukaemic clone will be more susceptible to these
risk factors compared to other individuals. The role of infections has
been explored extensively due to the existence of the childhood in-
cidence peak and spatiotemporal leukaemia clusters. Two major hy-
potheses exist in this area: Kinlen’s population mixing and Greaves’
delayed infection hypotheses. The key concept underlying these
hypotheses is the unusual or abnormal response to an infection or
infections not previously encountered by the individual due to mi-
gration or the modern lifestyle.
There is increasing evidence that an individual’s inherited genetic
background plays a role in the development of ALL. Patients with
Down syndrome are approximately 40 times more likely to develop
ALL between the ages of 0 to 4 years compared to other children.
More recently, genome-​wide association studies (GWAS) have iden-
tified several allelic variants that are significantly associated with an
increased likelihood of developing ALL, including single nucleotide
polymorphisms (SNP) in the IKZF1, ARID5B, CEPBE, CDKN2A/​
B, and ETV6 genes. Of note most of these genes are also somatically
altered in ALL. For example, 25% of children with ALL have ETV6-​
RUNX1 fusion (Table 22.4.2.1) while approximately 40% of cases
harbour CDKN2A/​B deletions. A causative role for an individual’s
genetic make-​up is further supported by observations that more than
50% of Down syndrome-​ALL patients harbour CRLF2 deregulation
which occurs in just 5 to 10% of non-​Down syndrome ALL children;
patients with a Robertsonian der(15;21)(q10;q10) translocation are
over 2700 times more likely to develop intrachromosomal amplifi-
cation of chromosome 21 (iAMP21) ALL and approximately 40% of
individuals with low hypodiploidy have a germline mutation in the
TP53 gene.
The pathogenesis of ALL is directed by the accumulation of one
or more driver mutations each of which enhances the tumour cap-
ability of the preleukaemic clone until eventually the clone expands
and develops into overt acute leukaemia. The vast majority of pa-
tients will harbour a single primary genetic abnormality, which
initiates the oncogenic process, coupled with one or more cooper-
ating mutations. Primary chromosomal abnormalities are typically
translocations or gross aneuploidy and frequently affect leukaemia-​
specific pathways such as B/​T-​cell development and differentiation.
Table 22.4.2.1 describes the key primary genetic abnormalities that
have been described in ALL. In childhood ALL, ETV6-​RUNX1 and
high hyperdiploidy together account for approximately 50% of cases
whereas the most prevalent abnormality in adult ALL is BCR-​ABL1
fusion accounting for approximately 25% (Fig. 22.4.2.1). These pri-
mary genetic lesions are clonal (i.e. present in all leukaemic cells)
and pathognomonic of ALL; hence their identification at diagnosis
is key for effective treatment stratification (see later).
Despite this key role, few of these abnormalities are sufficient to
cause clinical disease and are usually accompanied by one or more
additional abnormalities. The spectrum of additional aberrations is
broad and over 30 genes have been implicated. In contrast to the pri-
mary abnormality, secondary lesions are typically deletions, often
microdeletions, or sequence mutations. The most prevalent sec-
ondary aberrations are deletions of CDKN2A/​B, IKZF1, and PAX5,
and mutations of RAS pathway genes (e.g. KRAS and NRAS). There
is a strong correlation between specific primary abnormalities and
the spectrum of secondary lesions indicating epistatic interactions
between the affected genes. Examples of cooperation between pri-
mary and secondary lesions include (1) loss of the nontranslocated
ETV6 allele in 60 to 70% of cases with ETV6-​RUNX1 fusion; (2) acti-
vating RAS pathway mutations in approximately 50% of cases with
high hyperdiploidy; (3) deletion of IKZF1 in 80% of cases with BCR-​
ABL1 fusion; and (4) CDKN2A/​B deletion in 80 to 90% of T-​cell ALL
(T-​ALL) cases. Secondary genetic abnormalities are, by definition,
subclonal and hence are present in less than 100% of leukaemic cells.
Minor subclones accounting for 1% of leukaemic cells have been
22.4.2  Acute lymphoblastic leukaemia
5271
detected with highly sensitive assays and ultra-​deep next-​generation
sequencing may well identify even smaller subclones. The number
of subclones present in each patient is difficult to determine accur-
ately but all the available evidence indicates widespread genetic het-
erogeneity with the majority of patients harbouring several (two
to four) subclones and some with numerous subclones (more than
five). Importantly, cell-​based and sequencing studies have demon-
strated that the same gene can be mutated/​deleted independently
in different subclones; further supporting the notion that secondary
lesions actively cooperate with the primary abnormality to drive
leukaemogenesis.
Although the vast majority of patients with ALL (>90%) will
achieve a complete remission, relapse does occur and is the leading
cause of death. Therefore, identifying the genetic drivers of relapse
is a key clinical and scientific challenge. Numerous genomic studies
have attempted to determine the clonal origins of relapsed disease by
comparing the genomic landscape relapse and initial presentation.
Several key concepts have arisen from these studies: (1) the primary
genetic abnormality is almost always retained at relapse; (2)  the
dominant leukaemic clone at relapse is usually present at diagnosis
as a major or minor subclone; and (3) the spectrum of abnormal-
ities present at relapse is similar to that observed at initial diagnosis.
Therefore, for the majority of relapsed patients it is likely that the
relapse emerges via the natural selection of a pre-​existing clone or
clones under the evolutionary pressure of chemotherapy. The rare
exceptions to this scenario are very late recurrences which are gen-
etically unrelated to the presentation clone. These ‘relapses’ are likely
to represent secondary ALL and have a strong germline compo-
nent. Whether or not all genetic abnormalities observed at relapse
are also present at initial diagnosis—​as opposed to being induced by
therapy—​is a matter of debate. Currently our knowledge is limited
by the sensitivity of the available assays. There are no truly relapse-​
specific mutations or abnormalities. However, mutations in TP53,
CREBPB, NT5C2, and NR3C1 are more prevalent at relapse than ini-
tial diagnosis. Moreover, they have been linked to resistance to spe-
cific therapies: NT5C2 and nucleoside analogues, NR3C1/​CREBPB,
and glucocorticoids. There is a growing body of evidence suggesting
a link between the presence of specific mutations at relapse and the
Table 22.4.2.1  Immunophenotypic and genetic classification of acute lymphoblastic leukaemia
Primary/​class-​defining abnormality
Description
Frequency
Relative prognosis
B-​cell precursor ALL
High hyperdiploidy
51–​65 chromosomes
30% children, 10% adults
Very good
ETV6-​RUNX1
t(12;21)(p13;q22)
25% children, 1% adults
Very good
TCF3-​PBX1
t(1;19)(q23;p13)
3–​6%
Intermediate/​good
KMT2A (MLL) translocations
Gene fusions involving different partner genes including AFF1
(AF4), MLLT1 (ENL), MLLT4 (AF6), MLLT3 (AF9), MLLT10 (AF10)
80% infants, 2% children,
10% adults
Poor/​very poor
BCR-​ABL1
t(9;22)(q34;q11.2)
2% children, 25% adults
Very poor unless
treated with tyrosine
kinase inhibitors
TCF3-​HLF
t(17;19)(q22;p13)
<1%
Extremely poor
Near haploidy
<30 chromosomes
1–​2% children
Extremely poor
Low hypodiploidy/​near triploidy
30–​39/​60–​78 chromosomes
1–​2% children, 5% adults
Extremely poor
iAMP21
Intrachromosomal amplification of chromosome
21q22.11–​21q22.12
2–​3% children
Very poor unless
treated as high risk
Emerging subgroups of B-​other ALL
ERG deletions
Expression of ERG isoforms
3–​5%
Very good
IGH translocations
Overexpression of various partner genes including members of
the CEBP gene family and ID4
3–​5%
Intermediate/​poor
JAK–​STAT pathway activation
IGH-​CRLF2, P2RY8-​CRLF2, CRLF2 mutations, JAK2 mutations and
translocations, IGH-​EPOR
5–​10%
Intermediate/​poor
ABL-​class fusions
Rearrangements of ABL1, ABL2, PDGFRB and CSF1R with
numerous partner genes, e.g. EBF1-​PDGFRB
1–​2%
Poor
T-​cell ALL
TAL/​LMO
Overexpression of TAL1, LM02 and related genes. Common
rearrangements—​SIL-​TAL1, t(1;14)(p32;q11)/​TRAD-​TAL1
c.40%
Good
TLX3
Overexpression of TLX3. Common translocation—​t(5;14)
(q35;q32)/​BCL111B-​TLX3
c.25%
Variable
TLX1
Overexpression of TLX1, NKX2-​1, and related genes. Common
rearrangements: t(10;14)(q24;q11)/​TCRAD-​TLX1
c.15%
Good
HOXA
Rearrangements resulting in HOXA deregulation e.g. KMT2A
(MLL) translocation, CALM-​AF10 fusion
c.10%
Variable
Immature
Rearrangements and mutations associated with an immature T-​
cell phenotype, e.g. MEF2C fusions
c.10%
Poor
section 22  Haematological disorders
5272
frontline treatment protocol indicating that some protocols may
preferentially select or induce subclones carrying specific mutations.
Clinical features/​differential diagnosis
The clinical features of ALL are mainly a consequence of the failure
of normal haematopoiesis. In addition, local infiltration and expan-
sion of the leukaemic blasts can cause pathology. Patients usually
present with the symptoms and signs shown in Box 22.4.2.1.
This clinical picture of bone marrow failure and the consequences
of local leukaemic infiltration explain the differential diagnosis
which consists of other types of leukaemia (including acute mye-
loid leukaemia, chronic lymphoid leukaemia, and chronic mye-
loid leukaemia); other causes of bone marrow failure (including
myelodysplastic syndromes, drug-​induced, aplastic anaemia,
and inherited bone marrow failure syndromes); rarely other ma-
lignancies with widespread bone marrow infiltration (sarcomas,
desmoplastic round cell tumours, and non-​Hodgkin lymphoma or
neuroblastoma,); other causes of bony pain (in particular, rheuma-
toid arthritis and osteomyelitis); and idiopathic thrombocytopenic
purpura.
Clinical investigations
If leukaemia is suspected, the investigations listed in Box 22.4.2.2
are routinely performed to establish or exclude the diagnosis and
presence of leukaemia-​related complications.
If these investigations confirm or suggest a diagnosis of leukaemia
is likely then a haematological opinion is required to discuss further
investigations and initial management.
Further leukaemia-​specific investigations are listed in Box 22.4.2.3
and illustrated in Fig. 22.4.2.2.
100%
90%
80%
70%
60%
50%
40%
30%
Estimated frequency
20%
10%
0%
<1
1–9
10–14
15–19
Age at diagnosis (years)
20–24
25–39
40–59
60+
B-other
IGH translocation
TCF3-PBX1
BCR-ABL1
Low hypodiploidy/near
haploidy
iAMP21
KMT2A/MLL
translocation
High hyperdiploidy
ETV6-RUNX1
Fig. 22.4.2.1  Age-​specific frequency of key primary chromosomal abnormalities in B-​cell precursor acute lymphoblastic leukaemia.
Box 22.4.2.1  Symptoms and signs of acute
lymphoblastic leukaemia
•	 Weakness, lethargy, pallor—​due to insufficient production of red cells
(anaemia).
•	 Petechiae, nose and gum bleeding—​due to insufficient production of
platelets (thrombocytopenia).
•	 Febrile infections, sometimes with an unusual or prolonged course—​
due to insufficient production of immune cells, including granulocytes
(neutropenia).
•	 Lymphadenopathy (particularly cervical), hepatosplenomegaly—​due
to leukaemic expansion in lymphoid organs.
•	 Bone pain—​most likely as a consequence of osseous and periosteal
leukaemic infiltration although the mechanism is poorly understood.
•	 Painless testicular swelling—​a consequence of leukaemic infiltration.
•	 Superior vena cava syndrome with distension of the external jugular
veins, facial and neck swelling, shortness of breath, cough—​due to
thymic enlargement (T-​cell leukaemia).
•	 Abdominal pain—​due to hepatosplenomegaly (tension of the liver
capsule) or abdominal lymph node enlargement.
•	 Cerebral nerve palsies (including facial numbness), headache,
meningism—​due to CNS leukaemia (rare).
•	 Other rare clinical presentations—​hypoxia, confusion due to leucostasis
(very high WCC >>100 × 109/​litre); visual disturbances due to retinal
infiltration or haemorrhage.
22.4.2  Acute lymphoblastic leukaemia
5273
Treatment
Initial management
The initial management steps should include transfer of the patient
to the nearest specialist centre. Immediate treatment with broad-​
spectrum antibiotics should be considered (after taking appropriate
samples for culture) in febrile patients or those with a suspected
infection. Transfusion may be required, depending on symptoms
and blood counts, but caution is advised in patients with espe-
cially high white cell counts (WCCs) due to the risk of developing
hyperviscosity syndrome.
The presence of a large mediastinal mass is a clinical emergency as
it can cause superior vena cava syndrome and/​or tracheal obstruction
very quickly. These patients should not be sedated or anaesthetized
and—​if possible—​biopsy of the mass should be avoided: diagnostic
samples can often be obtained from pleural or pericardial effusions,
which are commonly present. Bone marrow examination may also
confirm a diagnosis of T-​lineage ALL. If the patient is very symp-
tomatic, then it may be necessary to commence steroid therapy prior
to establishing a diagnosis.
The development of tumour lysis syndrome needs to be avoided;
hence, it is important to monitor electrolytes and kidney function
while giving appropriate hydration and allopurinol. In patients with
a high WCC or with a large mediastinal mass, the administration
of rasburicase, a recombinant urate oxidase, should be considered.
All patients with ALL who undergo intensive chemotherapy will
require central venous line access, mostly port catheters in children
and Hickman lines or peripherally inserted central catheters in adults.
Current multidrug chemotherapy
Multidrug combination chemotherapy including steroids, vincris-
tine, asparaginase, daunorubicin or doxorubicin, cytarabine, cyclo-
phosphamide, methotrexate, 6-​mercaptopurine, and etoposide is
the backbone of any successful therapy for patients with ALL. ALL
treatment protocols take 2 to 3 years and are some of the most com-
plex chemotherapy regimens in haemato-​oncology. Most modern
treatment protocols are risk stratified, based on WCC at presenta-
tion, genetics, age, and response to therapy assessed by morpho-
logical bone marrow appearance and minimal residual disease
(MRD) status. Modern treatment protocols will utilize these risk
factors to assign patients to risk groups each of which will receive
different treatment. The precise definitions of risk groups vary from
one protocol to the next.
The key elements of modern ALL therapy can be divided into sev-
eral discrete phases and most ALL patients will be treated following
standardized national treatment protocols (Fig. 22.4.2.3).
Induction phase
The main aim of this 4-​ to 6-​week long chemotherapy block is to
achieve morphological and molecular remission. This treatment
cycle includes steroids, vincristine, and asparaginase, which are the
three most useful induction agents for ALL. Based on recent evi-
dence, dexamethasone appears to be more effective in inducing
apoptosis of B-​cell blasts compared to prednisolone and has a higher
penetrance into brain tissue. While dexamethasone is still given
continuously for 28 days in most childhood ALL protocols, in adult
ALL protocols pulsed courses of dexamethasone are becoming the
standard of care to reduce steroid toxicity. Asparaginase significantly
reduces intracellular and circulating asparagine levels in ALL blasts
leading to apoptosis of the ALL blasts. At the end of the induction
block, nearly all children will achieve complete remission whereas in
adults the rate is around 80 to 90%. Importantly, this is not a cure and
the disease will nearly always relapse if no further therapy is given.
Consolidation phase/​intensification phase
To prevent relapse in the central nervous system (CNS) and to consoli-
date remission, patients undergo several more intensive combination
chemotherapy phases lasting in total 6 to 8 months. In each of the blocks
the chemotherapy varies and new drugs are introduced aiming to pre-
vent development of chemoresistance in the remaining leukaemic blasts.
Maintenance therapy
After patients have completed the intensive chemotherapy phases,
treatment is commenced with low-​dose maintenance therapy con-
sisting of daily oral mercaptopurine and weekly oral methotrexate
for 18 to 30 months. Patients might also receive periodic intravenous
vincristine and short courses of oral glucocorticoids. The aim of this
phase is to suppress and kill any persisting leukaemic blasts. Significant
shortening of maintenance length or reduced compliance with daily
administration of 6-​mercaptopurine leads to higher relapse rates.
Box 22.4.2.2  Routine investigations in patients
with possible leukaemia
•	 Careful history and clinical examination, including testicular examination
•	 FBC and blood film—​to check for presence of circulating leukaemia
blasts
•	 Clotting screen
•	 Liver function tests, lactate dehydrogenase, and C-​reactive protein
•	 Urea, creatinine, K, Ca2+, PO4
3−, uric acid—​to check for evidence of
tumour lysis syndrome
•	 Chest radiography—​to exclude presence of mediastinal mass
Box 22.4.2.3  Specific investigation of patients with leukaemia
•	 Bone marrow aspirate and biopsy—​for cytology and histology.
•	 Genetic analysis including G-​banding cytogenetics and fluorescence
in situ hybridization (FISH) on bone marrow aspirate or peripheral
blood sample analysis—​to identify low-​ and high-​risk patients (Table
22.4.2.1).
•	 Flow cytometric immunophenotyping of the leukaemia—​to determine
lineage and maturation stage (e.g. B-​cell precursor ALL, T-​ALL, mature
leukaemia).
•	 Leukaemia-​specific immunoglobulin or T-​cell receptor gene re-
arrangements may be identified by a PCR-​based method or the
leukaemia-​associated immunophenotype is identified using multi-
colour flow cytometry—​for future minimal/measurable residual
­disease (MRD) monitoring of disease response.
•	 Lumbar puncture—​to exclude central nervous system (CNS)
involvement.
•	 Magnetic resonance imaging of the head and/​or spine—​if the patient
has neurological symptoms at presentation.
•	 If mediastinal mass is present—​staging computed tomography (CT)
or positron emission tomography (PET)/​CT scan of neck, chest, ab-
domen and pelvis is indicated.
section 22  Haematological disorders
5274
Fig. 22.4.2.2  Patient with infant B-​cell ALL presenting with a high WCC. (a) Blood film showed presence of lymphoid blasts with high
nucleocytoplasmic ratio. (b) Immunophenotyping plots shows presence of CD19+​, CD79a+​, terminal deoxyribonucleotidyl transferase (TdT)+​,
CD15+ and CD117– cells in keeping with B-​ALL. (c) Cytogenetic analysis depicts a t(4:11) translocation with the long arm of chromosome 4 being
translocated to chromosome 11 leading to the generation of MLL-​AF4 fusion protein. (d) Fluorescence in situ hybridization confirms rearrangement
of MLL (red arrows cells with rearranged MLL; yellow arrow cells with wildtype MLL). Break Apart Rearrangement Probe used to check for MLL
rearrangement (see inset).
ALL protocol for children and young adults
ALL protocol for adults
Standard-risk patients
Intermediate-risk patients
High-risk patients
High-risk patients
0
5
10
15
20
25
30
35
weeks
Standard-risk patients
Fig. 22.4.2.3  Outline of ALL treatment protocols for clinical risk group as per current UKALL2011 trial for children and young adults and UKALL14
trial for adults.
22.4.2  Acute lymphoblastic leukaemia
5275
Central nervous system-​directed prophylaxis/​therapy
In the 1950s and 1960s, relapses in the CNS after completion of ALL
therapy were common and were explained by the fact that the blood–​
brain barrier prevented chemotherapy reaching adequate concentration
levels in cerebrospinal fluids. Therefore, cranial or craniospinal irradi-
ation was introduced and was administered to all patients with ALL. This
step dramatically improved overall survival in the 1960s and 1970s, but
unfortunately came at a cost of significant long-​term toxicity including
cognitive impairment, decreased growth, endocrinopathies, and sec-
ondary brain tumours. Over the last two decades, most ALL protocols
have gradually eliminated cranial irradiation and have replaced it suc-
cessfully with more intensive CNS prophylaxis including frequent ad-
ministration of intrathecal methotrexate. As methotrexate crosses the
blood–​brain barrier effectively, many ALL protocols will include several
high-​dose methotrexate infusions followed by folinic acid ‘rescue’ to
avoid increased methotrexate-​related toxicity to normal tissue.
Treatment of BCR-​ABL-​positive ALL and
BCR-​ABL-​like ALL
The management of BCR-​ABL-​positive ALL was very challenging prior
to the discovery of BCR-​ABL1 tyrosine kinase inhibitors (TKIs) such as
imatinib and related agents. Since the introduction of TKIs, complete
remission rates post induction close to 100% have been reported in
adults and children even with reduced cytotoxic therapy, and the overall
survival has doubled. Consequently, in children, the need for allogeneic
haematopoietic stem cell transplantation in first remission has dimin-
ished. This is in contrast to adult patients where allogeneic haemato-
poietic stem cell transplantation is still advocated in first remission.
Many patients with BCR-​ABL-​positive ALL will relapse and frequently
acquire point mutations in the BCR-​ABL1 oncoprotein especially the
T315I mutation within the ABL kinase domain. These mutations confer
resistance to first-​ and second-​generation TKIs. Currently, ponatinib, a
third-​generation TKI, which has significant activity against wild-​type
and most mutant forms of BCR-​ABL1 tyrosine kinase has been ap-
proved for patients who have failed second generation TKI or who have
acquired a T315I mutation. BCR-​ABL-​like ALL has a gene expression
profile very similar to BCR-​ABL-​positive ALL and is also associated
with a poor prognosis. This entity is present in approximately 10 to 15%
of children and adults. In these cases, the leukaemic blasts harbour gen-
etic alterations, which activate kinase and JAK–​STAT signalling (Table
22.4.2.1). Importantly, ongoing studies, in vitro data, and anecdotal re-
ports suggest that patients with ABL-​class fusions and JAK–​STAT ab-
normalities can be targeted by TKIs and JAK2 inhibitors.
Treatment of relapsed/​refractory ALL
Until recently the outlook for relapsed ALL in adult patients was poor
with standard fludarabine based salvage chemotherapy. However, in
the last few years, significant inroads have been made and several novel
compounds have been approved in the USA and Europe for the treat-
ment of relapsed/refractory ALL. These include antibodies targeting
B-lymphoid-specific antigens such as CD19 or CD22, and modified
T cell therapies. Two antibodies have been shown to be superior to
standard salvage chemotherapy in large phase 3 studies.  Inotuzumab
ozogamicin, an antibody–drug conjugate targeting CD22, delivers the
potent cytotoxic agent calicheamicin to CD22 expressing cells indu-
cing DNA damage and apoptosis. The bispecific T-engager (BiTE)
blinatumomab binds to CD3 and CD19 simultaneously, bringing CD3
expressing T cells in close proximity to CD19 expressing B-ALL cells
leading to activation of T cells with subsequent cytotoxic activity on
the target cells. Both of these compounds achieve good  responses and
measurable residual disease negativity. Furthermore, blinatumomab is
used to treat patients with B-ALL, who have achieved complete remis-
sion but still have presence of measurable/minimal residual disease.
In contrast, children who have a late first relapse have a high re-
mission rate after salvage therapy and an overall survival of 60%.
Patients who relapse early have a 50 to 70% chance of regaining re-
mission and have an overall survival of 20 to 30%. Most patients in
second complete remission will undergo myeloablative allogeneic
haematopoietic stem cell transplantation to consolidate their remis-
sion. MRD monitoring can be used to help with the decision process.
In 2017, tisagenlecleucel a chimeric antigen receptor CAR-T cell
therapy targeting CD19-positive cells was approved by the FDA for
patients up to 25 years of age with B-ALL that is refractory or in
second or later relapse. Autologous T cells from the patient are gen-
etically modified to kill leukaemic blasts and then reinfused into
the patient. CAR T cell therapies have significant toxicities (causing
cytokine release syndrome and neurotoxicity). The initial response
rates are very good and durability of remissions are very promising
but relapses with novel immunological relapse mechanisms have
been observed (e.g. loss of the epitope in the CD19 protein that
is recognized by both the currently used antibodies and chimeric
T-cell receptors). Relapses post CAR-T cell infusions are more diffi-
cult to manage and survival figures become poor.
Role of allogeneic haematopoietic stem
cell transplantation
In children, allogeneic haematopoietic stem cell transplantation is usu-
ally not recommended in first remission but is of significant benefit in
patients who attain a second complete remission after relapse. The role
of allogeneic haematopoietic stem cell transplantation in patients with
high-​risk disease (e.g. BCR-​ABL-​positive ALL and infant ALL with
KMT2A rearrangements) in first remission remains controversial.
In adults, the poor outcome after relapse has changed the strategy
over the last few years and patients regarded to be at high risk for re-
lapse (e.g. patients with high-risk genetics such as BCR-ABL positive
ALL and those with complex cytogenetics, or MRD positivity after
the first or second block of chemotherapy) will typically undergo
an allogeneic haematopoietic stem cell transplant in first remis-
sion. This improves overall survival by 10 to 20%. Reduced-intensity
conditioning is used in adults older than 40 years, whereas a
myeloablative allogeneic haematopoietic stem cell transplantation is
preferred in children and younger adults. The antileukaemic benefit
of allogeneic haematopoietic stem cell transplantation is attributed
to two factors: high-dose chemotherapy with or without total-body
irradiation conditioning regimens and the graft-versus-leukaemia
effect of donor cells against recipient malignant cells.
Management of ALL in elderly patients
The treatment of elderly patients is difficult and very little progress
has been made over the last decades. The expected 5-​year overall
survival is between 5 and 10%. This poor outcome reflects the pro-
portion of elderly patients with ALL who have either BCR-​ABL1
positivity or other high-​risk genetics, but also highlights the diffi-
culty elderly patients have in tolerating intensive ALL chemotherapy
protocols, Hence, the focus over the last years has become more se-
lective. Elderly patients who are fit enough will be considered for a
section 22  Haematological disorders
5276
more intensive approach including allogeneic haematopoietic stem
cell transplantation. However, the majority of elderly patients will
receive specifically designed less intense chemotherapy protocols,
which have a chance of cure but at the same time are not too toxic;
thus, the treatment can be administered in outpatient settings with
reduced hospitalization in order to provide improved quality of life.
Prognosis/​outcome
Without treatment, ALL is fatal and its treatment is one of the
success stories of modern medicine (Fig. 22.4.2.4). There are five
key risk factors predicting treatment response: age, sex, immuno-​
genetic subtype, initial disease burden, and response to therapy.
Age is one of the main risk factors and, with the exception of infants
(<1 year), risk increases with age. Traditionally, males have had a
greater risk of relapse compared to females, largely due to the risk
of testicular relapse. However, modern protocols which sometimes
prolong the length of maintenance therapy for boys now rarely re-
port major differences in outcome by sex. Initial disease burden is
also an indicator of relapse risk and overall survival. High levels
of burden are indicated by a high WCC, the presence of CNS dis-
ease, and the infiltration of the lymph nodes, spleen, and liver. Age
and disease burden form the basis for the widely used paediatric
National Cancer Institute (NCI) risk score (NCI risk: age <1 year
(infant leukaemia), age ≥10 years, and a high WCC ≥50 × 109/​litre).
Immunophenotype and acquired genetic abnormalities, which
are tightly related, are key predictors of outcome (Table 22.4.2.1). In
childhood B-​cell precursor (BCP) ALL, there are two well-​established
good-​risk abnormalities—​ETV6-​RUNX1 and high hyperdiploidy—​
along with six high-​risk abnormalities—​KMT2A translocations, BCR-​
ABL1, TCF3-​HLF, near haploidy, low hypodiploidy, and iAMP21
(Fig. 22.4.2.1). The age-​specific frequency of these abnormalities
likely explains some of the risk associated with age (Figs. 22.4.2.5
and 22.4.2.6). Among children, T-​ALL is associated with a poorer
outcome whereas in adults it is associated with a lower risk of relapse.
The underlying genetics of T-​ALL is less age related than in BCP-​ALL
and this is likely to explain the difference in relative risk. In adult ALL,
although the distribution of genetic abnormalities is different, there
are also strong prognostic genetic markers.
The most important risk factor is the response to initial treat-
ment. Traditionally, this has been measured by morphological
examination of early bone marrow samples to assess the propor-
tion of blasts remaining after treatment has been initiated or by
clearance of blasts in the peripheral blood after 1 week of steroid
treatment. However, this is a crude measurement and has now been
superseded by MRD monitoring, which utilizes polymerase chain
reaction (PCR) or flow cytometry assays to quantify the size of
the leukaemic clone at predefined time points after therapy. PCR
can be used to track clone-​specific physiological rearrangements
of the Ig or TCR gene loci or leukaemia-​specific gene fusions (e.g.
BCR-​ABL1). Alternatively, flow cytometry can be used to track
the presence and abundance of a leukaemia-​associated aberrant
immunophenotypic profile. These assays are able to detect very low-​
level leukaemic clones (up to 1 leukaemic cell out of 10 000–​100 000
bone marrow cells). The greater the rate of disease clearance at key
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
UKALLVIII
(1980–85)
UKALLX
(1985–90)
UKALLXI
(1990–97)
UKALLXII
(1993–2006)
UKALLXA
(1985–1992)
ALL97/99
(1997–2002)
ALL2003
(2003–10)
EFS
OS
Fig. 22.4.2.4  Improvement in event-​free (EFS) and overall survival (OS) for patients with acute lymphoblastic leukaemia
treated on successive United Kingdom clinical trials.
22.4.2  Acute lymphoblastic leukaemia
5277
120
100
80
60
40
20
0
00 to
01
01 to
04
05 to
09
10 to
14
15 to
19
20 to
24
25 to
29
30 to
34
35 to
39
40 to
44
45 to
49
Age at diagnosis
Rate per 100 000
Average number of cases per year
50 to
54
55 to
59
60 to
64
65 to
69
70 to
74
75 to
79
80 to
84
85 to
89
90+
0
1
2
3
4
5
6
Male Cases
Female Cases
Male Rates
Female Rates
Fig. 22.4.2.5  United Kingdom incidence of ALL in 2010 to 2012: average number of new cases per year and age-​specific incidence rates
per 100 000 population. The main incidence peak is between years 2 to 4.
Data from Cancer Research UK, ‘Acute lymphoblastic leukaemia (ALL) incidence statistics’, available from: http://​www.cancerresearchuk.org/​health-​professional/​
cancer-​statistics/​statistics-​by-​cancer-​type/​leukaemia-​all/​incidence#heading-​One (accessed December 2015).
1.00
0.75
0.50
Event-free survival
0.25
0.00
0
1
2
3
Years from diagnosis
4
5
BCP-ALL Cytogenetic Good Risk
BCP-ALL Cytogenetic Intermediate Risk
BCP-ALL Cytogenetic High Risk
T-ALL
Fig. 22.4.2.6  Event-​free survival of children and young adults (1–​24 years) treated on UKALL2003
stratified by immuno-​genetic subgroup.
Data courtesy of Dr Nick Goulden and Professor Ajay Vora, UKALL2003 trial coordinators.
section 22  Haematological disorders
5278
1.00
End of induction MRD level
0%
<0.01%
<0.1%
<1%
<10%
10%
0.75
0.50
0.25
0.00
0
1
2
3
Years from diagnosis
Relapse rate
4
5
Fig. 22.4.2.7  Relapse rate of children and young adults (1–​24 years) treated on UKALL2003
stratified by the end of induction minimal residual disease (MRD) level.
Data courtesy of Dr Nick Goulden and Professor Ajay Vora, UKALL2003 trial coordinators.
1.00
0.75
0.50
0.25
0.00
0
2
4
6
Years
Overall survival
8
10
Good risk
Intermediate risk I
Intermediate risk II
High risk
t(9;22)/BCR-AB/1
Fig. 22.4.2.8  Outcome heterogeneity among adult patients (25–​59 years) with ALL by genetic
subtype. Definition of risk groups: Good risk, high hyperdiploidy; intermediate risk I, B-​other,
iAMP21; intermediate risk II, IGH translocations, CRLF2 rearrangements, IKZF1 deletions, TCF3-​
PBX1; high risk, BCR-​ABL1, KMT2A-​AFF1, low hypodiploidy, complex karyotype.
Data courtesy of Professors Adele Fielding, Tony Goldstone and Jacob Rowe, UKALLXII trial coordinator.
22.4.2  Acute lymphoblastic leukaemia
5279
time points (usually end of induction), the lower the likelihood of
relapse (Fig. 22.4.2.7).
Children with standard or low-​risk ALL are usually defined as
those less than 10 years old, with low WCCs (<50 × 109/​litre), low
levels of postinduction MRD, and good-​risk cytogenetics. These
children will have an excellent chance of a cure and many protocols
have reported overall survival rates of greater than 90% for this sub-
group. In contrast, children with high-​risk cytogenetics, high levels
of postinduction MRD and high WCCs (>100 × 109/​litre) will have
relapse rates up to 50% and will typically be treated with more inten-
sive protocols and possibly stem cell transplantation. Overall survival
rates for adults aged 25 to 59 years range from 25 to 60% depending
on the genetic subtype (Fig. 22.4.2.8) but average approximately
40% and are even lower (<15%) for patients over 60 years.
Complications/​long-​term follow-​up
Approximately 2 to 4% of children and 5 to 10% of adults will die
from direct toxic effects of treatment, mostly due to intractable
bacterial or fungal infections. Infants, patients with Down syn-
drome, and elderly patients have a higher risk of dying from toxic
complications.
Short-​term toxic effects
The commonest of these is febrile neutropenia due to bacterial and
fungal infections, which can be life-​threatening as patients have sig-
nificant treatment-​induced immunosuppression. Broad-​spectrum
antibiotics should be commenced immediately as per hospital
policy. In most adult ALL protocols, patients will receive antifungal
prophylaxis. Pneumocystis (carinii) jirovecii pneumonia is now a rela-
tively rare complication as it can be prevented by antipneumocystis
prophylaxis (standard of care while on chemotherapy for ALL).
Varicella zoster infections can be severe.
Common noninfective short-​term toxic effects include nausea
and vomiting, hair loss (intermittent), and transfusion reactions.
Specific side effects of particular drugs are listed in Box 22.4.2.4.
Long-​term toxic effects
Anthracycline-​induced cardiomyopathy is rare, but if it occurs it can
cause severe cardiac failure, often decades after the original treatment.
As with all chemotherapy regimens, the risk of secondary malignan-
cies is increased and is in low single figures. A few patients will become
infertile; this occurrence is significantly higher if patients undergo
allogeneic haematopoietic stem cell transplantation and reaches
nearly 100% if patients receive total-​body irradiation as part of the
conditioning protocol. There is also an increased risk of premature
ovarian failure in women who received chemotherapy in childhood.
A significant number of long-​term survivors of childhood ALL appear
to have an increase in neurocognitive impairment during later life.
Future developments
Until recently, no new curative drug agents had become available
for the management of ALL since the early 1970s. However, in
the last few years, significant inroads have been made and sev-
eral novel compounds have been approved in relapsed/refractory
B-ALL. Currently the main focus of clinical ALL research is how
to combine blinatumomab or inotuzumab with frontline B-ALL
chemotherapy and use of CAR-T cells or similar immune effector
cells in the treatment of relapsed/refractory adult B-ALL patients.
In adult patients with de novo ALL, rituximab, an antibody against
CD20, was shown to improve overall survival by nearly 10% when
added to standard therapy in a large French randomized trial.
With the advent of immunotherapy and a pipeline of novel tar-
geted drugs, treatment of ALL will continue to improve and become
more effective, particularly for high-risk and relapsed patients, as
well as being less toxic for the many patients who are cured on cur-
rent treatment protocols.
FURTHER READING
Eswaran J, et al. (2015). The pre-​B-​cell receptor checkpoint in acute
lymphoblastic leukaemia. Leukemia, 29, 1623–​31.
Fielding AK (2015). Treatment of Philadelphia chromosome-​positive
acute lymphoblastic leukemia in adults: a broader range of options,
improved outcomes, and more therapeutic dilemmas. Am Soc Clin
Oncol Educ Book, 2015, e352–​9.
Guerra VA, et al. (2019). Novel monoclonal antibody-based treatment
strategies in adults with acute lymphoblastic leukemia. Ther Adv
Hematol, doi: 10.1177/2040620719849496.
Hunger SP, Mullighan CG (2015). Acute lymphoblastic leukemia in
children. N Engl J Med, 373, 1541–​52.
Hunger SP, Mullighan CG (2015). Redefining ALL classification:
toward detecting high-​risk ALL and implementing precision medi-
cine. Blood, 125, 3977–​87.
Inaba H, Greaves M, Mullighan CG (2013). Acute lymphoblastic leu-
kaemia. Lancet, 381, 1943–​55.
Jabbour E, et al. (2015). Monoclonal antibodies in acute lymphoblastic
leukemia. Blood, 125, 4010–​16.
Maude SL, et al. (2015). CD19-​targeted chimeric antigen receptor T-​cell
therapy for acute lymphoblastic leukemia. Blood, 125, 4017–​23.
Box 22.4.2.4  Side effects of drugs used to treat ALL
Steroids
•	 Mood and behavioural changes (emotional instability, aggressiveness,
depression)—​can be severe and very stressful for the families
•	 Cushing syndrome and obesity
•	 Diabetes mellitus
•	 Osteonecrosis (avascular necrosis) affecting major joints—​can occur
in 5 to 10% of children, sometimes requiring surgical procedures
including joint replacement
Vincristine
•	 Neuropathy (neuropathic pain, jaw pain, loss of deep tendon reflexes,
‘foot drop’, constipation)
Asparaginase
•	 Allergic reactions—​more common after intravenous as opposed to
intramuscular administration
•	 Coagulopathy
•	 Sinus venous thrombosis and other thromboembolic complications
•	 Pancreatitis
Methotrexate
•	 Seizures (following intrathecal and high-​dose methotrexate)
•	 Encephalopathy
•	 Mucositis

22.4.3 Hodgkin lymphoma 5280 Vijaya Raj Bhatt and
22.4.3 Hodgkin lymphoma 5280 Vijaya Raj Bhatt and James O. Armitage
section 22  Haematological disorders
5280
Moorman AV (2012). The clinical relevance of chromosomal and
genomic abnormalities in B-​cell precursor acute lymphoblastic leu-
kaemia. Blood Rev, 26, 123–​35.
Pui CH, et al. (2015). Childhood acute lymphoblastic leukemia: pro-
gress through collaboration. J Clin Oncol, 33, 2938–​48.
Soverini S, Bassan R, Lion T (2019). Treatment and monitoring of
Philadelphia chromosome-positive leukemia patients: recent ad-
vances and remaining challenges. J Hematol Oncol, 12, 39.
van Dongen JJM, et al. (2015). Minimal residual disease diagnostics in
acute lymphoblastic leukemia: need for sensitive, fast, and standard-
ized technologies. Blood, 125, 3996–​4009.
Vora A, et  al. (2013). Treatment reduction for children and young
adults with low-​risk acute lymphoblastic leukaemia defined by min-
imal residual disease (UKALL 2003): a randomised controlled trial.
Lancet Oncol, 14, 199–​209.
Vora A, et al. (2014). Augmented post-​remission therapy for a minimal
residual disease-​defined high-​risk subgroup of children and young
people with clinical standard-​risk and intermediate-​risk acute
lymphoblastic leukaemia (UKALL 2003): a randomised controlled
trial. Lancet Oncol, 15, 809–​18.
22.4.3  Hodgkin lymphoma
Vijaya Raj Bhatt and James O. Armitage
ESSENTIALS
Hodgkin lymphoma is derived from a profoundly defective B cell,
with the pathobiology, histology, and clinical features being both
characteristic and distinct from non-​Hodgkin lymphomas. Its cause
is unknown. Incidence is about 3 per 100 000 per year in Western
countries with a bimodal age distribution meaning that it is one of
the commoner lymphomas of young people. Most cases are highly
responsive to combination chemotherapy, with many patients being
cured of their disease.
Presentation and diagnosis
Patients with Hodgkin lymphoma may present with lymphaden-
opathy, a mediastinal (or other) mass, and systemic symptoms
including weight loss, fever, and night sweats (B-​symptoms).
Workup requires staging with positron emission tomography
(PET)-​CT imaging, and biopsy.
Classical Hodgkin lymphoma is defined by the presence of the
binucleate Reed–​Sternberg cell in an appropriate inflammatory
histological context. These cells have a characteristic immuno­
histochemical phenotype, showing CD15 and CD30 positivity, but
being immunonegative for classical B-​cell markers such as CD20.
Additional histological subtypes have been defined, with the most
clinically significant being lymphocyte-​predominant Hodgkin
lymphoma (with its differing clinical profile and management).
Treatment and prognosis
Patients with localized disease (Ann Arbor stage I or nonbulky stage
II, without B-​symptoms) may be treated with chemotherapy with
or without radiotherapy. Those with more advanced-​stage disease
require combination chemotherapy, with radiotherapy to sites of
initial bulk disease. While relapse is uncommon in patients with
early-​stage disease, some 20% of patients with advanced-​stage dis-
ease may need ‘salvage’ chemotherapy regimens for relapsed dis-
ease, when the aim is to attain PET negativity before embarking on
high-​dose therapy with stem cell rescue.
Since many patients have a good outcome, it is essential to min-
imize the long-​term sequelae of treatment, such as pulmonary
fibrosis and the development of secondary cancers. Risk stratifi-
cation, including through the use of PET-​CT imaging, is a major
focus of clinical trials in this area to determine optimal treatment
strategies.
Introduction
Epidemiology and risk factors
Unlike non-​Hodgkin lymphoma, the incidence of Hodgkin
lymphoma has been stable over the period 1950 to 2000, with about
three new cases per 100 000 population/​year in Western countries.
Approximately 9000 new cases are diagnosed each year in the United
States of America and 1800 in the United Kingdom. The condition
displays a peculiar bimodal distribution of occurrence, with peaks
in young adulthood and older age (Fig. 22.4.3.1). This dual peak has
led some to propose that Hodgkin lymphoma actually represents
two illnesses, with the earlier peak being related to an infectious aeti-
ology and the latter representing a true malignancy, but there is little
evidence to support this hypothesis.
An association has been demonstrated between the occurrence
of Hodgkin lymphoma and infection by the Epstein–​Barr virus
(EBV). Monoclonal or oligoclonal proliferation of EBV-​infected
cells is found in 20 to 40% of patients with Hodgkin lymphoma.
175
150
125
100
75
50
25
Number of cases
0–4
5–15
15–24
25–34
35–44
45–54
55–64
65–74
75–84
85+
Age groups
Incidence: 3.29/100 000 males
2.07/100 000 females
Females
Males
Fig. 22.4.3.1  Age distribution of Hodgkin lymphoma expressed as new
cases registered in England and Wales in 1973.
5281
22.4.3  Hodgkin lymphoma
Patients infected by HIV are at an increased risk for Hodgkin
lymphoma in addition to non-​Hodgkin lymphoma.
The subtypes of Hodgkin lymphoma vary geographically and
by age group. Patients in Western countries who develop Hodgkin
lymphoma in young adulthood usually have the nodular scler-
osis subtype. Patients from developing countries, elderly patients,
and those infected with HIV, frequently have mixed cellularity or
lymphocyte-​depleted classical Hodgkin lymphoma.
Hodgkin lymphoma is approximately 100 times more likely in
an identical twin of an infected patient. Numerous instances of case
clustering have been described. Although these might be taken as
evidence of a genetic or infectious aetiology, the cause of Hodgkin
lymphoma remains unknown.
Pathology
In 1832, Thomas Hodgkin of Guy’s Hospital, London, reported
seven patients who died from a disorder involving lymph node
and spleen enlargement. Then, early in the 20th century, Reed and
Sternberg independently described the characteristic giant cells
that now bear their name. The diagnosis of Hodgkin lymphoma
­requires the identification of Reed–​Sternberg cells in a charac-
teristic cellular background. The World Health Organization
(WHO) classification for Hodgkin lymphoma is presented in Box
22.4.3.1. Hodgkin lymphoma is unique in that the tumour cells
constitute a minority of the total cellular population of involved
lymph nodes. The tumour cell (i.e. the Reed–​Sternberg cell) has
been shown to be of B-​cell origin. In the WHO classification,
Hodgkin lymphoma is divided into classical Hodgkin lymphoma
(95% of cases), and nodular lymphocyte-​predominant Hodgkin
lymphoma (5% of cases). Classical Hodgkin lymphoma is sub-
divided into nodular sclerosis, which is characterized by bands of
fibrosis that are often visible to the naked eye; mixed cellularity,
where a larger number of Reed–​Sternberg cells in a mixed cel-
lular background are typical; and lymphocyte-​depletion Hodgkin
lymphoma, which has either a large number of Reed–​Sternberg
cells and atypical mononuclear cells, or a background of diffuse
fibrosis with occasional Reed–​Sternberg cells. The diagnosis of
lymphocyte-​depletion Hodgkin lymphoma should always raise
the possibility that an unusual diffuse large B-​cell lymphoma
is being confused with Hodgkin lymphoma. Although diffuse
lymphocyte-​rich (or predominant) Hodgkin lymphoma is listed as
a diagnosis, in practice this is exceedingly rare. Most patients with
lymphocyte-​predominant Hodgkin lymphoma have the entity
nodular lymphocyte-​predominant Hodgkin lymphoma, which is
a more indolent illness but usually treated in a manner similar to
classical Hodgkin lymphoma.
Pathobiology of lymphoma
Increased understanding of the biology of the immune system has
allowed the improved classification of lymphomas, and provided
new prognostic information and new potential targets for therapy.
Lymphomas are malignancies of lymphocytes in which the surface
proteins involved in cell recognition and intracellular signalling
are important in diagnosis, predicting clinical course, and therapy.
Although the genetics of lymphomas are complicated, they too are
beginning to be unravelled. Information gleaned from all these
studies is likely to further change both the classification and therapy
of the lymphomas.
Immunology
The recognition of new surface antigens has improved the ability
to recognize specific subtypes of lymphoma. For example, dis-
covery of the Ki-​1 (CD30) antigen by investigators in Germany pro-
vided a marker for the Reed–​Sternberg cells in classical Hodgkin
lymphoma. However, it was soon discovered that this antigen was
found on the surface of cancers that were previously felt to be un-
differentiated carcinomas and malignant histiocytosis. This obser-
vation allowed the description of anaplastic large cell lymphoma as a
diagnostic entity and, more importantly, allowed some patients with
lymphoma to receive appropriate therapy. However, it is important
to remember that not all cases of a particular type of lymphoma will
have exactly the characteristic immunophenotype, and this does not
invalidate the diagnosis.
The Reed–​Sternberg cells in classical Hodgkin lymphoma express
CD15 and CD30 but downregulate B-​cell markers including CD20.
The Reed–​Sternberg cells in nodular lymphocyte-​predominant
Hodgkin lymphoma express the leucocyte common antigen and
other B-​cell markers including CD20 but do not express CD15 and
CD30 (Table 22.4.3.1).
Genetics
Clonal immunoglobulin gene rearrangement is observed in the
Reed–​Sternberg cells in essentially all cases of Hodgkin lymphoma.
The presence of somatic hypermutation in the variable region of the
immunoglobulin heavy chain genes indicates a germinal centre B-​
cell origin of the Reed–​Sternberg cells. Deregulation of transcrip-
tion factors including nuclear factor kappa-​B (NF-​κB) and Janus
kinase/​signal transducers and activators of transcription (JAK/​
STAT) signalling pathway is associated with antiapoptotic changes
and proliferation of the neoplastic cells. Neoplastic cells often dem-
onstrate overexpression of p53, aneuploidy, and hypertetraploidy.
Comparative genomic hybridization has highlighted recurrent
gains on chromosomal arms 2p, 9p, and 12q and distinct high-​level
amplifications on chromosomal bands 4p16, 4q23 to q24, and 9p23
to p24. Gene expression profiling has demonstrated that Hodgkin
Box 22.4.3.1  World Health Organization classification
of Hodgkin lymphoma
•	 Nodular lymphocyte-​predominant Hodgkin lymphoma (5%)
•	 Classical Hodgkin lymphoma (95%):
—	 Nodular sclerosis
—	 Mixed cellularity
—	 Lymphocyte depletion
—	 Lymphocyte rich
Table 22.4.3.1  Immunological markers useful in the diagnosis
of Hodgkin lymphoma
Subtype
Characteristic immunophenotype
Classical Hodgkin lymphoma
CD15+ CD20– ​CD30+
Nodular lymphocyte-​predominant
Hodgkin lymphoma
CD15– ​CD20+ CD30–​
section 22  Haematological disorders
5282
lymphoma resembles primary mediastinal lymphoma more closely
than germinal centre B-​cell-​like diffuse large B-​cell lymphoma. For
example, the chromosome 9p region that contains regulators of T-​
cell responses, programmed death ligand (PDL)-​1 and PDL2, is amp-
lified in Hodgkin lymphoma and primary mediastinal lymphoma.
More recently, flow-​sorting and exome sequencing have revealed
alterations in genes involved in antigen presentation, chromosome
integrity, transcriptional regulation, and ubiquitination.
Clinical features
Patients with classical Hodgkin lymphoma usually present with
palpable nontender lymphadenopathy. In most patients, lymph
nodes are discovered in the cervical, supraclavicular, and axillary
regions. More than half the patients have mediastinal lymphaden-
opathy at diagnosis, and symptoms from a large mediastinal mass
such as superior vena cava obstruction are often the initial presenta-
tion. Subdiaphragmatic presentation of Hodgkin lymphoma is un-
usual, and more common in older men. Approximately one-​third
of patients with classical Hodgkin lymphoma present with systemic
symptoms such as fevers, night sweats, pruritus, and/​or weight loss.
These systemic symptoms are believed to be the result of the release
of cytokines by normal or malignant cells. Patients might present
with cytopenia secondary to either bone marrow involvement or
autoimmune destruction of the formed elements of the blood.
Hodgkin lymphoma can present as a fever of unknown origin.
This is more likely in older patients, those with mixed-​cellularity
or lymphocyte-​depletion subtypes, and those who present with
lymphoma below the diaphragm. Fevers associated with Hodgkin
lymphoma occasionally persist for days to weeks, followed by
afebrile periods, with subsequent reoccurrence of the fever. This
pattern is known as Pel–​Ebstein fever. Unusual presentations of
Hodgkin lymphoma include severe and unexplained pruritus, cen-
tral nervous system involvement, paraneoplastic cerebellar degen-
eration, nephrotic syndrome, immune haemolytic anaemia and/​or
thrombocytopenia, hypercalcaemia, and pain in lymph nodes with
alcohol ingestion. The possible presentations of lymphomas are so
varied that the diagnosis should be considered in many patients, and
not just those presenting with lymphadenopathy or splenomegaly.
Diagnosis and evaluation
The diagnosis of Hodgkin lymphoma is based on a review of an
adequate biopsy by an expert haematopathologist. Fine needle as-
piration or small biopsies should be avoided as the basis for diag-
nosing lymphoma whenever possible. The differential diagnosis that
the pathologist considers when diagnosing a lymphoma includes
benign proliferations of lymphoid tissue, malignancies of myeloid
cells, nonhaemopoietic malignancies, viral infections, and unusual
disorders such as Castleman’s disease and giant lymph node hyper-
plasia. Having tissue available for immunological studies and/​or
genetic studies will help to confirm the diagnosis.
Once the diagnosis of a type of lymphoma has been established,
a series of studies should be carried out to determine the stage of
disease and to allow prognostication (Box 22.4.3.2). The anatomical
spread of Hodgkin lymphoma is expressed as an Ann Arbor stage
(Table 22.4.3.2). This staging system divides patients into those
with lymphoma confined to one lymphatic site, multiple lymphatic
sites on one side of the diaphragm, lymphatic involvement on both
sides of the diaphragm, and those with bone marrow involvement,
liver involvement, or other extensive extranodal lymphoma. The
Ann Arbor stage also includes a suffix A or B indicating the absence
(A) or presence (B) of unexplained fevers above 38°C, weight loss
of more than 10% of the body weight in the preceding 6 months, or
drenching night sweats. Additional factors can also have an impact
on a patient’s response to therapy and survival.
A revised classification includes stage II bulky lymphoma de-
fined as a single nodal mass of 10 cm, or greater than a third of the
transthoracic diameter at any level of thoracic vertebrae. Functional
images, today obtained using fluorodeoxyglucose positron emission
tomography (FDG-​PET) scans, identify areas of abnormal glucose
metabolism that are present in all patients with Hodgkin lymphoma.
At the completion of therapy, repeat CT will often show only par-
tial regression of mediastinal or retroperitoneal masses because of
a sclerotic reaction to the tumour. In these patients, reversion of a
previously abnormal PET scan to normal can confirm a complete
Box 22.4.3.2  Staging evaluation for a new patient
with lymphoma
•	 Complete history and physical examination.
•	 Haematological studies:
—	 Full blood count, erythrocyte sedimentation rate (ESR).
•	 Chemistry studies to measure normal organ function:
—	 Serum creatinine, liver function studies, serum albumin.
—	 Serum lactate dehydrogenase.
•	 Viral serology to include HIV, hepatitis B and C.
•	 Imaging studies:
—	 Chest radiograph.
—	 PET/​CT scan or contrast-​enhanced CT of the neck, chest, ab-
domen, and pelvis.
•	 Bone marrow biopsy.
•	 Pregnancy test in women of child-​bearing age.
•	 Fertility counselling.
•	 Other studies as appropriate to evaluate specific complaints and to
follow up abnormal results found from the studies previously listed:
—​	 Not appropriate in all patients.
—​	 At one time both bipedal lymphangiography and staging lapar-
otomy were popular studies in evaluating new patients with
Hodgkin’s lymphoma, but they are now rarely—​if ever—​indicated.
Newer imaging techniques have made clinical, as opposed to sur-
gical, staging appropriate for essentially all patients. PET/​CT is the
current imaging modality of choice.
—​	 Bone marrow biopsy may be omitted, particularly in early-​stage
lymphoma, if a PET is performed; this is because PET is highly sen-
sitive in detecting marrow involvement by Hodgkin lymphoma.
—​	 Other studies can be useful in particular situations. MRI studies
are particularly useful in evaluating suspected bone or central ner-
vous system sites of involvement. Cerebrospinal fluid cytology and
flow cytometry may be necessary in evaluating suspected central
nervous system sites of involvement. Hepatitis B serology should
be performed if rituximab is being considered as a part of therapy
in patients with nodular lymphocyte-​predominant Hodgkin
lymphoma, as the use of rituximab can increase the risk of hepatitis
B reactivation. Echocardiography and pulmonary function tests are
often performed before initiation of anthracycline and bleomycin
respectively.
5283
metabolic response to therapy. Determining how much improve-
ment in a PET scan was required to document a complete remis-
sion limited the utility of this procedure until the development of
the Deauville, or 5-​point, score. This test takes advantage of the fact
that there is always PET uptake in the blood pool of mediastinum
and the liver, and the uptake in the liver is always greater than that
in the mediastinum. The Deauville score is as follows: 1—​no up-
take consistent with the possible presence of lymphoma; 2—​uptake
in areas of previously known lymphoma, but less than the uptake
in the mediastinum; 3—​uptake in areas of previous or suspected
lymphoma with an intensity between the mediastinum and the liver;
4—​uptake in areas of previous or suspected lymphoma greater than
the liver; and 5—​new areas of lymphomatous involvement and/​or
a dramatic increase in the level of uptake in previous or suspected
areas of lymphoma. In many trials, and increasingly in routine clin-
ical practice, a score of 3 or less at the end of therapy is considered
complete remission.
Interim PET/​CT scanning (i.e. early scans done after two cycles of
therapy) is becoming increasingly widely used. Using interim PET/​
CT it may be possible to escalate therapy in patients with a Deauville
score greater than 3, or, potentially reduce therapy in patients who
have responded favourably. So-​called risk-​adapted therapy is the
subject of a number of current clinical trials.
Nodular lymphocyte-​predominant Hodgkin lymphoma, as noted
previously, is a different clinical entity from classical Hodgkin
lymphoma. These patients represent less than 5% of all patients
found to have Hodgkin lymphoma. The evaluation of such pa-
tients is carried out in a similar way to that for classical Hodgkin
lymphoma. However, nodular lymphocyte-​predominant Hodgkin
lymphoma tends to follow a chronic, relapsing course and some-
times transforms to diffuse large B-​cell lymphoma.
Prognostic factors
The major factors determining treatment outcome for patients with
Hodgkin lymphoma include the Ann Arbor stage, bulky lymphoma,
the presence or absence of systemic symptoms, age, and gender.
Patients with asymptomatic, localized lymphoma who are young
and female have the best outlook. Histological subtypes do not ap-
pear to have major independent prognostic significance. Patients
with nodular sclerosing Hodgkin lymphoma are less likely to have
adverse prognostic factors than those with mixed-​cellularity or
lymphocyte-​depleted subtypes. The results of several laboratory
studies can predict outcome in patients with Hodgkin lymphoma.
Adverse results include anaemia, a greatly elevated ESR, a low al-
bumin level, and a low lymphocyte count. The ESR is sometimes
used to follow the course of patients with Hodgkin lymphoma as it
reverts to normal with successful treatment.
An International Prognostic Index for Hodgkin lymphoma has
been developed (Table 22.4.3.3). This index uses seven adverse
prognostic factors that determine the treatment outcome. These in-
clude age of at least 45 years, stage IV, male sex, white cell count of
at least 15 × 109/​litre, lymphocyte count less than 0.6 × 109/​litre
or less than 8% of all white cells, albumin less than 40 g/​litre, and
haemoglobin less than 105 g/​litre. The adverse prognostic factors
present in an individual patient are summed. In a large study, pa-
tients with no adverse prognostic factors had a 5-​year freedom from
progression of 84%, whereas for patients with five or more factors
it was only 42%.
The most important factor in predicting outcome for patients with
Hodgkin lymphoma is their response to therapy. Patients who have
a prompt, complete response to chemotherapy and/​or radiotherapy
have the best outlook and are most likely to be cured. Normalization
of a PET scan after two cycles of chemotherapy (interim PET) may
be used to omit radiotherapy in early-​stage favourable disease or to
deescalate therapy in advanced-​stage disease.
Patients who relapse after initial successful treatment for
Hodgkin lymphoma can sometimes be effectively treated with
further chemotherapy or radiotherapy. The chances for successful
treatment depend, in part, on the duration of initial remission in
addition to other prognostic factors present at relapse. Patients
with a longer initial remission are more likely to be successfully
retreated.
Table 22.4.3.2  The Ann Arbor staging system
Stage
Characteristics
I
1 nodal site involved
IE
1 site of localized extranodal involvement
II
2 or more nodal sites involved, but only on 1 side of the diaphragm
IIE
1 site of localized extranodal involvement plus regional nodes
involved—​all on 1 side of the diaphragm
III
Nodal involvement (i.e. spleen counts as a nodal site) on both sides
of the diaphragm
IV
Bone marrow, liver, or other extensive extranodal involvement
(e.g. multiple pulmonary nodules)
A
Absence of unexplained fever (i.e. >38°C), drenching night sweats,
or weight loss (i.e. ≥10% in 6 months)
B
Presence of unexplained fever (i.e. >38°C), drenching night sweats,
or weight loss (i.e. ≥10% in 6 months)
Table 22.4.3.3  Prognostic factors for advanced Hodgkin
lymphoma
Adverse prognostic factors
Age
≥45 years
Stage
IV
Sex
Male
White blood count
≥15 x 109/​litre
Lymphocyte count
<0.6 x 109/​litre or <8% of all white cells
Albumin
<40 g/​litre
Haemoglobin
<105 g/​litre
Outcome according to prognostic score
Number of factors
5-​year freedom from
progression (%)
5-​year overall
survival (%)
0
84
89
1
77
90
2
67
81
3
60
78
4
51
61
5
42
56
22.4.3  Hodgkin lymphoma
section 22  Haematological disorders
5284
General principles of lymphoma treatment
Types of treatment
Those treatments effective in the management of patients with
cancer include surgery, radiotherapy, cytotoxic chemotherapy, and
a variety of new approaches developed through increasing under-
standing of the biology of the immune system. The latter include
cytokines, antibodies, and attempts to direct an immune reaction
against cancer.
As few patients with lymphoma have truly localized lymphoma,
surgery has not been a major treatment modality. Since its utiliza-
tion in medicine in the first part of the 20th century, radiotherapy
has been a major treatment modality for patients with lymphoma,
but is limited in its application by toxicity. Its curative potential de-
pends upon being able to achieve a tumouricidal dose (typically 30–​
40 Gy) without irreversibly injuring normal organs. Thus, the site of
involvement by a lymphoma, as well as the number of sites involved,
can limit the effectiveness of this treatment, since toxicity increases
with the volume of tissue irradiated. Cure rates are often higher when
chemotherapy precedes the radiation. Cytotoxic chemotherapeutic
agents were first discovered in the 1940s when mechlorethamine
(i.e. the nitrogen mustard gas used in warfare), and subsequently
methotrexate, were found to cause regressions in immune system
malignancies. A wide variety of agents have since been shown to be
able to cause lymphoma regression in many patients with lymph-
omas. Unfortunately, early studies showed that regressions induced
by single agents were almost invariably followed by regrowth of
the tumour and eventual death of the patient. In an attempt to cir-
cumvent this, combinations of chemotherapeutic agents were first
utilized in the 1960s and early 1970s. The drugs were combined by
attempting to choose agents with different mechanisms of action
and nonoverlapping toxicities to allow the administration of doses
that were near to the maximum tolerated dose with an individual
agent. In both childhood acute leukaemia and Hodgkin lymphoma,
this approach was validated by the cure of a significant number of
patients. Today, several combination chemotherapy regimens with
acceptable toxicity have been shown to be effective and are widely
used worldwide (Table 22.4.3.4).
Increasing knowledge of the immune system has further led to the
recognition that a number of biologically active molecules can cause
regression of lymphomas and, in some cases, impact survival. The
first such agent to be widely used was interferon-​α, which has some
activity in both non-​Hodgkin lymphoma and Hodgkin lymphoma.
The ability to produce monoclonal antibodies has provided new
therapeutic molecules. Rituximab, which targets the CD20 antigen,
has been shown to be active in a variety of B-​cell lymphomas
including nodular lymphocyte-​predominant Hodgkin lymphoma.
The antibody conjugate brentuximab vedotin has shown signifi-
cant responses in CD30-​expressing tumours including Hodgkin
lymphoma.
Very high doses of cytotoxic chemotherapeutic agents with or
without radiotherapy and biologically active molecules have been
utilized in the treatment of patients with lymphomas as part of the
haematopoietic stem cell transplantation procedure. This involves
the administration of very high doses of antilymphoma therapy in
an attempt to overcome presumed treatment resistance. Patients are
rescued from the toxicity of treatment by the reinfusion of haemo-
poietic stem cells. The patient’s own haemopoietic stem cells (an au-
tologous transplant) or those from another individual with identical
Table 22.4.3.4  Popular combination chemotherapy regimens used in treating patients with Hodgkin lymphoma
Regimen
Drug
Dose (mg/​m2)
Route
Schedule
ABVD
Doxorubicin
25
IV
D1 and 15
28-​day cycles
Bleomycin
10 unit/​m2
IV
D1 and 15
Vinblastine
6
IV
D1 and 15
Dacarbazine
375
IV
D1 and 15
Stanford V
Doxorubicin
25
IV
D1 and 15
28-​day cycles
Vinblastine
6
IV
D1 and 15
Mechlorethamine
6
IV
D1
Vincristine
1.4
IV
D8 and 22
Bleomycin
5 unit/​m2
IV
D8 and 22
Etoposide
60
IV
D15 and 16
Prednisone
40
PO
Every other day, taper D15
of cycle 2 or 3
BEACOPP (escalated)a
Doxorubicin
35
IV
D1
21-​day cycles
Cyclophosphamide
1250
IV
D1
Etoposide
200
IV
D1–​3
Prednisone
40
PO
D1–​14
Procarbazine
100
PO
D1–​7
Bleomycin
10 unit/​m2
IV
D8
Vincristine
1.4
IV
D8
D, day(s); IV, intravenously; PO, orally.
a Baseline BEACOPP has lower doses of doxorubicin (25 mg/​m2), cyclophosphamide (650 mg/​m2), and etoposide (100 mg/​m2).
5285
HLA genes (an allogeneic transplant) can be utilized. Cells for this
procedure can be obtained from either bone marrow or peripheral
blood. Autologous transplantation has been widely used for patients
with lymphoma and shown to be able to cure patients with relapsed
Hodgkin lymphoma. Allogeneic transplantation, while apparently
curative, has a higher mortality rate and is reserved for younger,
fitter patients with multiply relapsed lymphoma or after failure of
autologous transplant.
General strategy of treatment
A number of factors need to be taken into account when formu-
lating a treatment recommendation for a patient with lymphoma
(Box 22.4.3.3). This decision should be made in conjunction with
the patient, and requires good judgement in addition to technical
knowledge. The aggressiveness of the treatment that is finally chosen
will often depend upon the physician’s interpretation of the chances
for cure. More toxicity is likely to be acceptable if the goal is cure ra-
ther than palliation.
For most patients, the goal of therapy is to achieve a complete
remission. This implies the disappearance of all symptoms and ob-
jective evidence of lymphoma. In practice, a complete remission is
documented by repeating all abnormal staging studies after sev-
eral cycles of therapy or at the completion of the planned therapy.
Documentation of complete remission is important. Patients who
achieve a complete remission have a chance for cure; those who do
not achieve a complete remission with initial therapy will often go
directly to second-​line treatments.
Patients who fail to be cured with initial therapy, either because
they do not achieve an initial remission or because they relapse
from remission, are candidates for what has been termed ‘salvage
therapy’. These second-​line regimens can cause tumour regression
in most patients with lymphoma and can occasionally produce
long-​term, lymphoma-​free survival. However, for most patients, the
only curative approach in this setting is haematopoietic stem cell
transplantation.
The toxicity of haematopoietic stem cell transplantation limits its
use to patients under 70 to 75 years of age, who have a good perform-
ance status, without serious compromise of major organ function;
and to patients who do not have bulky/​chemotherapy-​refractory
lymphoma.
Primary therapy
Patients with localized Hodgkin lymphoma (i.e. stage I or nonbulky
stage II) are usually treated with combined chemotherapy and
radiotherapy or chemotherapy alone. To minimize late compli-
cations, limiting the radiation dose and field size are increasingly
utilized. When radiotherapy alone is utilized, a dose of 30 to 36
Gy is usually administered in fractions of 1.75 to 2.00 Gy daily to
known sites of involvement and frequently to adjacent lymph node-​
bearing areas. However, radiotherapy alone is now rarely used ex-
cept in nodular lymphocyte-​predominant Hodgkin lymphoma
and possibly as a palliative care approach in relapsed or refrac-
tory Hodgkin lymphoma in unfit patients. When radiotherapy is
used as a consolidation therapy after chemotherapy, a dose of 20
to 36 Gy is usually administered. Although 90% of patients who
achieve a complete metabolic response following chemotherapy
will remain in remission at 3 years, there remains an additional
benefit for involved nodal irradiation in further reducing the
risk of relapse in some trials. Combined modality treatment will,
however, increase the risk of secondary malignancies and, if the
mediastinum is involved, of cardiovascular disease. It is hoped
that the reduction in radiation field using modern radiotherapy
techniques should help minimize these concerns. In very low-​risk,
early-​stage patients, the German Hodgkin Group have shown it is
possible to achieve excellent outcomes with two courses of chemo-
therapy (ABVD) followed by low-​dose (20 Gy) nodal irradiation.
In patients with higher-​risk disease (based on the number of nodal
sites and prognostic markers discussed previously), three or four
courses of chemotherapy followed by radiotherapy remains the
standard of care. In routine clinical practice, it seems reasonable to
individualize treatment decisions based on the outcome of interim
scanning and the risks and benefits of consolidative radiotherapy
in consultation with the patient.
Patients with otherwise localized Hodgkin lymphoma who pre-
sent with a large mediastinal mass pose special therapeutic problems.
A large mediastinal mass is often defined as one with a maximum
diameter greater than one-​third of the maximum thoracic diameter.
Treatment with radiotherapy alone, or chemotherapy alone, is as-
sociated with a high relapse rate. Large mediastinal masses are one
indication for combined-​modality therapy.
Patients who present with B-​symptoms or stage III or IV
lymphoma are best treated initially with a combination chemo-
therapy regimen. If complete remission is documented after com-
pleting a course of chemotherapy, the majority of patients will be
cured. Patients who have large masses often receive adjuvant radio-
therapy to the sites of previous bulky lymphoma after completing
the chemotherapy regimen. The most popular regimen for treating
Hodgkin lymphoma is currently ABVD (Table 22.4.3.4). ABVD
has been shown to be equivalent to more complicated regimens
that include the same drugs plus alkylating agents, and superior
to alkylator-​based regimens alone. Other treatment regimens in-
clude BEACOPP and Stanford V. BEACOPP involves higher doses
of drugs given in a very dose-​intensive fashion, while the drugs
are administered weekly for 12 weeks in the Stanford V regimen.
Excellent results have been reported with both of these approaches.
Comparative studies of the Stanford V regimen and ABVD have
found them to be of equivalent efficacy but the simplicity of the
ABVD regimen has made it the most popular.
Box 22.4.3.3  Factors to consider in therapy for a patient
with lymphoma
•	 Specific type of lymphoma
•	 Age
•	 Performance status
•	 Presence of other lymphomas
•	 Stage
•	 Systemic symptoms
•	 Pace of lymphoma
•	 Potential side effects
•	 Likelihood of cure
•	 Patient’s concerns about specific treatments
•	 Convenience
•	 Patient’s immediate and long-​term goals
•	 Quality of life
22.4.3  Hodgkin lymphoma
section 22  Haematological disorders
5286
The BEACOPP regimen is very intensive and associated with a
high rate of side effects. It cannot be given safely to older patients.
However, it appears to have a higher durable remission rate in
younger patients than ABVD. One study from Europe compared
BEACOPP with ABVD. The study demonstrated improved initial
tumour control with BEACOPP but no difference in the overall
survival between these two approaches. Patients who failed initial
therapy underwent autologous stem cell transplantation. BEACOPP
is associated with infertility and the risk of treatment-​related leu-
kaemia. These potential risks need to be balanced against its excel-
lent cure rate in younger patients.
Risk-​adapted strategies using interim PET/​CT are being
increasing used allowing escalation of therapy is those patients
failing to achieve remission after two courses of therapy. Other in-
vestigators start with a more intensive regimen and deescalate once
an early response has been achieved.
Elderly, pregnant, and HIV-​positive patients pose special thera-
peutic problems. In Hodgkin lymphoma, elderly patients have a
much worse prognosis: patients over 60 years of age at the time of
diagnosis have a survival rate less than half that of younger patients.
Elderly patients with localized lymphoma seem to benefit from
radiotherapy in a manner comparable to younger patients. However,
older patients tolerate aggressive chemotherapy regimens much less
well and, even if the drugs can be administered, older patients have
a higher relapse rate.
Since it occurs frequently in young adults, Hodgkin lymphoma
is sometimes diagnosed in pregnant women. Exposure to PET or
CT scanning should be avoided in pregnancy. Alternative imaging
modalities may include ultrasonography, magnetic resonance
imaging (MRI) and chest radiography with the use of an abdom-
inal shield. It is now clear that Hodgkin lymphoma can be treated
with chemotherapy at any point during pregnancy with a chance
of a good treatment outcome and a surviving infant. However, the
risks are higher in the first trimester. Most physicians would favour
delaying therapy past the first trimester, if possible. Beyond the
first trimester, the risks of chemotherapy to the fetus seem to less
than originally feared and ABVD appears to be a safe regimen with
outcomes comparable between pregnant and nonpregnant women
with Hodgkin lymphoma. The decision to defer chemotherapy
until after delivery is a reasonable strategy in patients with lower-​
risk disease providing careful monitoring with MRI is performed.
Pregnant patients should not be treated with radiotherapy. If the
decision is made to treat a pregnant patient with chemotherapy, it
must be remembered that the fetus will be myelosuppressed in a
manner similar to the mother, and this must be taken into account
when planning delivery of the baby.
The risk of classical Hodgkin lymphoma and non-​Hodgkin
lymphoma is significantly elevated in HIV-​positive patients. Unlike
non-​Hodgkin lymphoma, the risk of Hodgkin lymphoma is not re-
duced with the use of antiretroviral therapy. HIV-​positive patients
with Hodgkin lymphoma should be managed in consultation with
an infectious lymphoma specialist with expertise in antiretroviral
therapy. Early initiation of antiretroviral therapy and attention to
drug interactions are important. Otherwise, HIV-​positive patients
should be managed similarly to those without HIV infection. The
presence of HIV is not a contraindication for autologous transplant-
ation. With aggressive treatment, the outcomes of HIV-​positive pa-
tients are largely comparable to those without.
The optimal treatment for nodular lymphocyte-​predominant
Hodgkin lymphoma is unclear. Some clinicians favour no initial
therapy in asymptomatic patients. However, involved-​field radio-
therapy, or a brief course of chemotherapy plus radiation, can
produce durable remissions in some patients with this subtype
of Hodgkin lymphoma. Rituximab is an active agent in nodular
lymphocyte-​predominant Hodgkin lymphoma. The clinician must
be alert for transformation to diffuse large B-​cell lymphoma.
Treatment of relapse
Approximately 25 to 35% of patients treated with chemotherapy
for stage III or IV Hodgkin lymphoma will suffer a relapse after
achieving a remission, and a few patients will fail to enter initial
complete remission. Patients who fail to attain complete remis-
sion or who relapse within one year of completing therapy have a
poor prognosis with further standard chemotherapy. Autologous
haematopoietic stem cell transplantation can be curative in 25 to
50% of such patients, and is the treatment of choice. Patients who
have an initial remission of longer than one year pose a more com-
plicated therapeutic problem. These patients are likely to achieve
a second remission with a standard chemotherapy regimen.
However, long-​term follow-​up has demonstrated that most of
these remissions are not durable, and many physicians would rec-
ommend autologous haematopoietic stem cell transplantation
to such patients. The occasional patient with a localized relapse
after chemotherapy can sometimes be cured with radiotherapy.
Two new drugs are active in patients with relapsed Hodgkin
lymphoma and are likely to make a major impact in second-​line
therapy and, probably, incorporation into primary therapy. These
include brentuximab vedotin, and anti-​programmed death (PD)-​
1 antibodies, nivolumab and pembrolizumab. It is increasingly
clear that Hodgkin lymphoma cells, and some of the cells in the
tumour microenvironment, can express PDL-​1. When this inter-
acts with PD-​1 on activated T cells, the cells are ‘shut off’ and are
prevented from killing Reed–​Sternberg cells. Initial studies with
brentuximab vedotin as well as anti-​PD-​1 antibodies suggest a
very high response rate in patients who failed chemotherapy and
stem cell transplantation, and some of the responses are ongoing
after many months. The use of brentuximab vedotin as a consoli-
dation therapy after autologous transplantation significantly im-
proves lymphoma control. The eventual place of these agents in the
treatment of Hodgkin lymphoma will become apparent over the
next few years. Other single agent drugs available for treatment of
relapsed Hodgkin lymphoma include everolimus, lenalidomide,
and bendamustine.
Treatment complications
The treatment of Hodgkin lymphoma is associated with both short-​
term and long-​term complications. Prominent short-​term compli-
cations include hair loss, emesis, fatigue, anaemia, and infection
due to chemotherapy-​induced neutropenia. Hair loss is usually
transient. Emesis can be prevented in almost all patients by using
5-​hydroxytryptamine antagonists. Anaemia and fatigue do not usu-
ally limit the administration of therapy. Chemotherapy-​induced
neutropenia is a major problem, and neutropenic fever needs to be
managed aggressively with intravenous antibiotics after cultures are
obtained. Even so, treatment for Hodgkin lymphoma is adminis-
tered entirely on an outpatient basis.
5287
Delayed toxicity from the treatment of Hodgkin lymphoma has
become a major problem for young patients who are cured of the
lymphoma and have been followed for extended periods. In fact, for
patients with good-​prognosis Hodgkin lymphoma, long-​term com-
plications might lead to a higher mortality rate than the Hodgkin
lymphoma itself.
Most of the serious complications of radiotherapy appear after
long follow-​up. In the first few months after treatment, some pa-
tients will develop an electric shock sensation down the spine and
into the legs on flexion of the neck. This represents Lhermitte’s syn-
drome and needs to be recognized so that further evaluation can
be avoided. It is usually transient. In some patients, delayed pul-
monary fibrosis or cardiac injuries are associated with thoracic
radiotherapy. Modern radiotherapy techniques have minimized
the risk of these problems, but accelerated coronary artery disease
is a significant problem and leads to a number of treatment-​related
deaths. Follow-​up of these patients should emphasize reducing risk
factors for coronary artery disease. The major delayed problem with
radiotherapy is the development of secondary cancers. This risk be-
gins to appear beyond 10 years after therapy, and by 20 years after
therapy leads to a significant number of deaths. Patients treated with
thoracic radiotherapy for Hodgkin lymphoma should be strongly
encouraged not to smoke, to reduce the risk of lung cancer. Young
women, who have received chest or axillary radiotherapy between
the ages of 10 and 30 years should have screening mammography
and breast MRI instituted 8 to 10 years after completing treatment
or at the age of 40 years.
Patients who receive radiotherapy to the neck have a high risk
of developing subsequent hypothyroidism. Follow-​up in such pa-
tients should include periodic quantitation of their thyrotropin
levels to anticipate this problem. Some patients treated with ei-
ther radiotherapy or chemotherapy will develop herpes zoster.
This diagnosis does not necessarily signify a relapse of Hodgkin
lymphoma.
Long-​term problems associated with chemotherapy include
treatment-​related leukaemia, infertility, and aseptic necrosis of
bone. Infertility is most likely in patients who receive alkylating
agent-​containing regimens. In women, the risks of infertility are
age related. Women over 30 years of age are much more likely to be
permanently infertile than those under 30 years. However, in any
patient, resumption of fertility is possible and the patient should
be aware of this. Infertility is less of a problem in patients who re-
ceive the ABVD regimen. Men who wish to retain fertility should
be offered semen storage and women should be offered egg storage.
Treatment-​related leukaemia is most frequent in patients who
receive chemotherapy regimens containing alkylating agents and
who are treated on more than one occasion. Young patients treated
with only one chemotherapy sequence are unlikely to develop leu-
kaemia. The incidence of leukaemia rises dramatically in patients
over 40 years of age, and in those who receive alkylating agents on
more than one occasion. Leukaemia is unusual in patients treated
with ABVD. The combination of chemotherapy and radiotherapy
seems to increase the risk of leukaemia. The leukaemias that occur
in this setting usually present with myelodysplasia and typic-
ally have genetic abnormalities involving chromosomes 5, 7, and
8. Etoposide can lead to the development of acute leukaemia that
involves abnormalities on chromosome 11 without a preceding
myelodysplasia.
Patients who receive corticosteroid treatment as part of a com-
bination therapy are at risk for aseptic necrosis of the femoral heads,
and those who develop hip pain on follow-​up should be evaluated
for this possibility.
FURTHER READING
Advani RH, et  al. (2015). Randomized phase III trial comparing
ABVD plus radiotherapy with the Stanford V regimen in patients
with stages I  or II locally extensive, bulky mediastinal Hodgkin
lymphoma:  a subset analysis of the North American Intergroup
E2496 Trial. J Clin Oncol, 33, 1936–​42.
Advani RH, Hoppe RT (2013). How I treat nodular lymphocyte pre-
dominant Hodgkin lymphoma. Blood, 122, 4182–​8.
Ansell SM, et al. (2015). PD-​1 blockade with nivolumab in relapsed or
refractory Hodgkin lymphoma. N Engl J Med, 372, 311–​19.
Armitage JO (2010). Early-​stage Hodgkin lymphoma. N Engl J Med,
363, 653–​62.
Barrington SF, et al. (2014). Role of imaging in the staging and re-
sponse assessment of lymphoma:  consensus of the International
Conference on Malignant Lymphomas Imaging Working Group. J
Clin Oncol, 32, 3048–​58.
Cheson BD, et  al. (2014). Recommendations for initial evaluation,
staging, and response assessment of Hodgkin and non-​Hodgkin
lymphoma: the Lugano classification. J Clin Oncol, 32, 3059–​68.
Gordon LI, et al. (2013). Randomized phase III trial of ABVD versus
Stanford V with or without radiation therapy in locally extensive and
advanced-​stage Hodgkin lymphoma: an intergroup study coordin-
ated by the Eastern Cooperative Oncology Group (E2496). J Clin
Oncol, 31, 684–​91.
Hasenclever D, Diehl V (1998). A prognostic score for advanced
Hodgkin lymphoma. International prognostic factors project on ad-
vanced Hodgkin lymphoma. N Engl J Med, 339, 1506–​14.
Hutchings M, et al. (2014). In vivo treatment sensitivity testing with
positron emission tomography/​computed tomography after one
cycle of chemotherapy for Hodgkin lymphoma. J Clin Oncol, 32,
2705–​11.
Moskowitz CH, et al. (2015). Brentuximab vedotin as consolidation
therapy after autologous stem-​cell transplantation in patients with
Hodgkin lymphoma at risk of relapse or progression (AETHERA): a
randomised, double-​blind, placebo-​controlled, phase 3 trial. Lancet,
385, 1853–​62.
Radford J, et al. (2015). Results of a trial of PET-​directed therapy for
early-​stage Hodgkin lymphoma. N Engl J Med, 372, 1598–​607.
Viviani S, et al. (2011). ABVD versus BEACOPP for Hodgkin lymphoma
when high-​dose salvage is planned. N Engl J Med, 365, 203–​12.
22.4.3  Hodgkin lymphoma

22.4.4 Non- Hodgkin lymphoma 5288 Vijaya Raj Bhatt
22.4.4 Non- Hodgkin lymphoma 5288 Vijaya Raj Bhatt and James O. Armitage
section 22  Haematological disorders
5288
22.4.4  Non-​Hodgkin lymphoma
Vijaya Raj Bhatt and James O. Armitage
ESSENTIALS
Non-​Hodgkin lymphomas comprise precursor lymphoid neo-
plasms, mature B-​cell neoplasms, and mature T-​cell neoplasms.
The aetiology of most cases is unknown, but increased risk is as-
sociated with immune deficiencies, agricultural chemicals, auto-
immune disorders, treated Hodgkin disease, and some infectious
agents (e.g. Helicobacter pylori, human T-​cell lymphoma/​leu-
kaemia virus-​1, HIV, Epstein–​Barr virus, and human herpesvirus-​8).
Incidence varies from 10 to 22 cases per 100 000 per year in dif-
ferent populations.
Presentation and diagnosis
Patients with non-​Hodgkin lymphoma most commonly present with
lymphadenopathy, but other presentations include systemic symp-
toms or those attributable to mediastinal or retroperitoneal masses
or involvement. Diagnosis is typically based on expert evaluation of
an adequate lymph node biopsy. Staging depends largely on deter-
mining the anatomical extent of disease, with FDG positron emission
tomography/​CT scanning generally the preferable imaging modality.
Treatment and prognosis
For most patients, the goal of therapy is to achieve a complete remis-
sion. Patients with definitely curable lymphomas, such as diffuse large
B-​cell lymphoma and Burkitt lymphoma, are almost always treated
promptly with intensive regimens, for example, chemotherapy with
CHOP (cyclophosphamide, doxorubicin, vincristine (Oncovin), and
prednisone) plus the anti-​CD20 monoclonal antibody rituximab. By
contrast, follicular lymphoma is often not curable and the best treat-
ment is not clear, with many physicians favouring no initial therapy in
an asymptomatic patient.
Patients who are not cured with initial therapy are candidates for
what has been termed ‘salvage therapy’. For most patients, the only
curative approach in this setting is haematopoietic stem cell trans-
plantation, the toxicity of which means that it is only sensibly offered
to carefully selected patients. Various new agents, such as small mol-
ecule kinase and BCL-​2 inhibitors, and immune checkpoint inhibi-
tors, offer hope for the future.
Introduction
Lymphomas are malignancies of lymphoid cells and almost al-
ways present as solid tumours. They frequently respond to avail-
able therapies, and a significant subset of patients who develop
lymphomas can be cured. Lymphomas are usually divided into
Hodgkin lymphoma and non-​Hodgkin lymphoma (NHL). NHL
consists of indolent lymphomas that grow slowly and are often
asymptomatic until they reach an advanced stage, and aggres-
sive lymphomas that can be life-​threatening if not treated on a
timely fashion.
Epidemiology
NHL is much more frequent than Hodgkin lymphoma, with
more than 70 000 new cases being diagnosed in the United States
of America each year, and about 12 000 in the United Kingdom.
NHL increased in incidence at a higher rate than almost all other
malignancies from 1950 to 2000, but recent data suggests that the
incidence is stabilizing. In much of the world, it appears that the
incidence of NHL is increasing, but the incidence still varies widely
between countries. The incidence appears to be approximately 10
cases per 100 000 per year worldwide, 22 per 100 000 per year in
the United Kingdom, and more than 19 per 100 000 per year in the
United States of America. In the United States, the disease increased
in frequency in patients of all ages, but more strikingly in elderly
people, by approximately 4% per year between 1950 and the mid
1990s, although recent data suggest that the rate of increase may be
stabilizing.
The specific types of NHL vary in occurrence between coun-
tries. For example, follicular lymphoma is more common in North
America than in Europe or Asia. T-​cell lymphomas have been seen
more frequently in Asia, and certain types of T/​natural killer (NK)-​
cell lymphomas such as angiocentric nasal lymphomas are common
only in a few countries in Asia and Latin America. The explanation
for this geographical difference is unclear.
Aetiology
The aetiology of most NHL is unknown. Various aetiological fac-
tors, either proven or suggested to be associated with the develop-
ment of NHL, are listed in Box 22.4.4.1. It is now clear that exposure
to certain agriculture chemicals does increase the risk of this dis-
ease. A  variety of immune deficiencies, such as those associated
with immunosuppression following organ transplantation and
various hereditary immune deficiencies, are also associated with
an increased risk of developing NHL. Patients with disorders of the
immune system such as rheumatoid arthritis and systemic lupus
erythematosus also appear to be at increased risk.
A variety of infectious agents have been shown to be associated
with the development of NHL. Gastric Helicobacter pylori infection
is associated with the development of gastric mucosa-​associated
Box 22.4.4.1  Factors predisposing to the development of NHL
•	 Immune deficiencies:
—​	 Organ transplantation
—​	 Inherited immune deficiencies
—​	 AIDS
•	 Agricultural chemicals
•	 Autoimmune disorders:
—​	 Rheumatoid arthritis
—​	 Lupus erythematosus
•	 Treated Hodgkin lymphoma
•	 Infectious agents:
—​	 Viruses: EBV, HTLV-​1, HIV, HHV-​8, HCV
—​	 Bacteria: Helicobacter pylori, Chlamydia psittaci, Borrelia burgdorferi,
Campylobacter jejuni
EBV, Epstein–​Barr virus; HCV, hepatitis C virus; HHV-​8, human herpesvirus-​8;
HTLV-​1, human T-​cell leukaemia virus-​1.
5289
22.4.4  Non-Hodgkin lymphoma
lymphoid tissue (MALT) lymphoma, and eradication of the infec-
tion by antibiotics can lead to regression of the lymphoma. Human
T-​cell lymphoma/​leukaemia virus-​1 (HTLV-​1) appears to be the
cause of a specific type of NHL, seen predominantly in southern
Japan and the Caribbean, called adult T-​cell lymphoma/​leu-
kaemia. Epstein–​Barr virus (EBV) has been associated with Burkitt
lymphoma in Africa, the development of aggressive B-​cell lymph-
omas in immunosuppressed patients, and certain aggressive T-​cell
lymphomas. Human herpesvirus-​8 (HHV-​8) has been closely asso-
ciated with a rare diffuse large B-​cell lymphoma called primary effu-
sion lymphoma that is most frequently seen in immunosuppressed
patients. HIV infection can lead to the development of aggressive
B-​cell lymphomas that are often EBV positive. An association be-
tween hepatitis C virus (HCV) infection and the development of
splenic or large B-​cell lymphomas has been suggested. Similarly, the
association of Chlamydia psittaci and ocular adnexal lymphomas
has been reported. Other bacteria that have been associated with
MALT lymphomas include Campylobacter jejuni (i.e. small bowel)
and Borrelia burgdorferi (skin).
Pathology
The classification of NHL changed several times during the 20th cen-
tury. The first popular classification proposed by Gall and Mallory
divided lymphomas into giant follicular lymphoma, reticulum cell
sarcoma, and lymphosarcoma. Both the lack of adequate clinical cor-
relation and clear definitions of the entities led to further proposals.
Henry Rappaport recognized the importance of growth pattern in
the prognosis of NHL, and put forward his system that divided pa-
tients into those with nodular (i.e. follicular) or diffuse lymphomas
and those with large or small cell lymphomas. However, this system
was proposed before the recognition that lymphomas were all ma-
lignancies of lymphocytes and before the discovery of the existence
of subtypes of lymphocytes. The advent of modern immunology led
to new classification systems proposed by Lennert and colleagues in
Europe and Lukes and Collins in the United States of America. The
Kiel classification proposed by Lennert and colleagues became the
most widely used system in Europe. An attempt to unify the clas-
sifications of lymphomas led to the development of the Working
Formulation. This is a compromise system taking major elements
from the Rappaport classification, the Kiel classification, and the
Lukes/​Collins classification. It became widely used in the United
States of America but less so in Europe.
In the 1990s, a group of haematopathologists from Europe, North
America, and other parts of the world proposed a new system not
just based on morphology and immunophenotyping, but taking
into account other genetic and biological information that had be-
come available. In the 1990s, a number of ‘new’ lymphomas were
discovered that did not fit into previous classification systems. These
included mantle cell lymphoma, anaplastic large cell lymphoma, and
MALT lymphomas. The Revised European/​American Lymphoma
(REAL) classification classified lymphomas based on clinical
pathological syndromes (in other words, ‘real’ diseases) rather
than simply morphology. This system was tested in a large inter-
national study and shown to be more accurate than previous sys-
tems and to have high clinical relevance. Leaders in the fields of both
haematopathology and clinical haematology/​oncology agreed on a
modified REAL classification to be endorsed by the World Health
Organization (WHO) and published as the WHO classification
(Box 22.4.4.2). This, with some minor modifications published in
2016, is likely to be the major lymphoma classification for at least the
next decade. The incidence of major lymphoma subtypes according
to the WHO classification is listed in Table 22.4.4.1. Knowledge of
Box 22.4.4.2  WHO classification of NHL (2016)
•	 Precursor lymphoid neoplasms:
—​	 B-​lymphoblastic leukaemia/​lymphoma
—​	 T-​lymphoblastic leukaemia/​lymphoma
•	 Mature B-​cell neoplasms:
—​	 Chronic lymphocytic leukaemia/​small lymphocytic lymphoma
—​	 Splenic marginal zone lymphoma
—​	 Lymphoplasmacytic lymphoma
—​	 Extranodal marginal zone lymphoma of MALT
—​	 Nodal marginal zone lymphoma
—​	 Follicular lymphoma
—​	 Primary cutaneous follicle centre lymphoma
—​	 Mantle cell lymphoma
—​	 Diffuse large B-​cell lymphoma (DLBCL) of ABC and GC type:
•	 T-​cell/​histiocyte-​rich DLBCL
•	 Primary cutaneous DLBCL leg type
•	 Intravascular DLBCL
•	 Plasmablastic lymphoma
•	 Primary effusion lymphoma
—​	 Primary mediastinal (thymic) DLBCL
—​	 Burkitt lymphoma
•	 Mature T-​cell neoplasms:
—​	 Adult T-​cell leukaemia/​lymphoma
—​	 Extranodal NK/​T cell lymphoma, nasal type
—​	 Enteropathy associated T-​cell lymphoma
—​	 Hepatosplenic T-​cell lymphoma
—​	 Subcutaneous panniculitis-​like T-​cell lymphoma
—​	 Mycosis fungoides
—​	 Sézary’s syndrome
—​	 Primary cutaneous CD30-​positive T-​cell lymphoproliferative
disorders
—​	 Peripheral T-​cell lymphoma, not otherwise specified (NOS)
—​	 Angioimmunoblastic T-​cell lymphoma/​T-​follicular helper cell
lymphomas
—​	 Anaplastic large cell lymphoma, ALK positive
—​	 Anaplastic large cell lymphoma, ALK negative
Table 22.4.4.1  Worldwide relative frequency of occurrence
of major subtypes of NHL
Type of NHL
Percentage of all NHL
Diffuse large B cell
31
Follicular
22
Small lymphocytic/​chronic lymphocytic leukaemia
6
Mantle cell
6
Peripheral T cell
6
MALT
5
Anaplastic large cell lymphoma
2
Lymphoblastic
2
Burkitt
<1
MALT, mucosa-​associated lymphoid tissue.
section 22  Haematological disorders
5290
10 to 12 specific subtypes of NHL will allow a clinician to care for
almost all patients with NHL.
Pathobiology of lymphoma
Increased understanding of the biology of the immune system has
allowed the improved classification of lymphomas, and provided
new prognostic information and new potential targets for therapy.
Lymphomas are malignancies of lymphocytes in which the surface
proteins involved in cell recognition and intracellular signalling
are important in diagnosis, predicting clinical course, and therapy.
Although the genetics of lymphomas are complicated, they too are
beginning to be unravelled. Information gleaned from all these
studies is likely to further change both the classification and therapy
of the lymphomas.
Immunology
The recognition of new surface antigens has improved the ability
to recognize specific subtypes of lymphoma. For B-​cell NHL it is
possible to use immunophenotyping to help identify the cell of
origin of the lymphoma. For example, Burkitt lymphoma, follicular
lymphoma, and some diffuse large B-​cell lymphomas arise from
germinal centre B cells. Other diffuse large B-​cell lymphomas arise
from postgerminal centre B-​cells, demonstrating the biological
variability of tumours that can be morphologically similar. Further
insights into such phenomena are presented in the later section on
genetics of lymphomas.
The recognition of specific antigens by standardized antibodies
has improved the accuracy of diagnosis. Some of the more com-
monly recognized antigens are presented in Table 22.4.4.2. A char-
acteristic pattern of occurrence can be a key factor in making an
accurate diagnosis. Some types of lymphoma, such as follicular
lymphoma, can be diagnosed accurately without immunological
studies. Others such as all T-​cell lymphomas, diffuse large B-​cell
lymphoma, and mantle cell lymphoma can only be accurately diag-
nosed when immune markers are combined with traditional histo-
logical evaluation.
Genetics
A theme common to malignant disorders is the abnormal ex-
pression of specific genes. The search for these genes was facili-
tated by the frequent occurrence of chromosomal abnormalities
detectable by cytogenetic studies. These abnormalities include
chromosomal deletions or deletions of parts of a chromosome,
chromosomal duplications, and translocation of genetic material
from one chromosome to another. Chromosomal translocations,
through studying the sites of chromosome breakage, led to the
discovery of a number of genes that appear to be important in
lymphomagenesis or in determining the character of a particular
lymphoma. The best-​documented chromosomal abnormalities
associated with lymphomas, along with the involved oncogenes,
are presented in Table 22.4.4.3.
Specific chromosomal translocations are highly associated with
certain subtypes of lymphoma and thus are useful in diagnosis.
These include the t(2;5) and anaplastic large cell lymphoma; the
t(14;18) in follicular lymphoma; the t(8;14), t(2;8), and t(8;22)
in Burkitt lymphoma; and the t(11;14) in mantle cell lymphoma.
Cytogenetic studies in most patients with NHL display a large
number of chromosomal abnormalities. However, only a few have
been shown to be of diagnostic or prognostic significance.
Genetic abnormalities determine the nature of a lymphoma by
leading to the overexpression, underexpression, or abnormal expres-
sion of specific genes. The genes involved, frequently termed ‘onco-
genes’, are typically those that regulate the cell cycle, differentiation,
rate of proliferation, and apoptosis. Since the work of genes is done
by the proteins for which they code, the under-​, over-​, or abnormal
translation of specific proteins is an increasing subject for study.
In some cases, protein translations might be abnormal despite no
obvious translocation. For example, diffuse large B-​cell lymphoma
displays the t(14;18) in approximately 30% of patients. This trans-
location involves the BCL2 gene on chromosome 18, whose protein
product is involved in suppressing apoptosis (i.e. the mechanism of
cell death usually triggered by chemotherapeutic agents). Tumours
can overproduce the BCL-​2 protein with or without the t(14;18).
Overproduction of BCL-​2 protein might be expected to lead to
the increased survival of lymphoma cells when they are exposed to
therapeutic agents. In patients with diffuse large B-​cell lymphoma,
poorer outcome has been associated with overproduction of the
BCL-​2 protein, rather than with the t(14;18). Patients with diffuse
large B-​cell who have both the t(8;14) and the t(14;18) are com-
monly referred to as double-​hit lymphomas. These patients have a
poor outlook with currently available treatments, Approximately 5
to 10% of patients with diffuse large B-​cell lymphoma, a rare patient
with follicular lymphoma, and more than 50% of patients with the
Table 22.4.4.2  Immunological markers and their targets useful
in the diagnosis or management of lymphomas
Marker
Target
CD3
T cells
CD4
Helper/​inducer T cells
CD5
T cells, early B cells
CD8
Cytotoxic/​suppressor T cells and NK cells
CD10
CALLA
CD15
Lewis-​X
CD19
B4(leuk 12)
CD20
B cells
CD23
IgE receptor
CD25
IL-​2 receptor
CD30
Ki-​1
CD57
HNK-​1
Characteristic immunophenotype of selected lymphomasa
Subtype
Characteristic immunophenotype
Diffuse large B cell
CD5− CD10+/​− CD20+ CD23+/​−
Follicular
CD5− CD10+ CD20+ CD23+/​− BCL6+
Small lymphocytic/​chronic
lymphocytic leukaemia
CD5+ CD10− CD20+ (dim) CD23+
Mantle cell
CD5+ CD10− CD20+ CD23− cyclic D1+
CALLA, common acute lymphoblastic leukaemia antigen; IL-​2, interleukin-​2; NK,
natural killer.
a It is important to remember that not all cases of a particular type of lymphoma will
have exactly the characteristic immunophenotype, and this does not invalidate the
diagnosis.
22.4.4  Non-Hodgkin lymphoma
5291
unusual subtype high-​grade B-​cell lymphoma with features inter-
mediate between diffuse large B-​cell and Burkitt lymphoma have
‘double hits’.
The discovery of genetic abnormalities in lymphomas can be ac-
complished with cytogenetic analysis, fluorescent in situ hybridiza-
tion (FISH), by gene arrays, and more detailed genome sequencing.
Cytogenetic studies require fresh tissue. FISH studies can be done
on fixed tissue, but only specific abnormalities for which probes
are available can be investigated. Gene array studies are currently
a research technique that allows identification of genes that are
over-​ or underexpressed in specific specimens. They allow the ana-
lysis of thousands of genes simultaneously and have shown that
histologically identical groups of lymphomas can be subdivided
into clinically relevant subgroups on the basis of their gene ex-
pression patterns. For example, diffuse large B-​cell lymphoma can
be subdivided into at least three subgroups using gene expression
patterns that have different clinical characteristics and/​or treat-
ment outcome. Similar studies have been done in other subtypes of
lymphoma. At least one report of patients with follicular lymphoma
suggested that survival might be more affected by the gene expres-
sion pattern of infiltrating normal immune cells (i.e. a pattern char-
acteristic of T lymphocytes versus one characteristic of macrophage/​
dendritic cells) than by the gene expression pattern in the tumour
cells themselves.
Clinical features
Patients with lymphoma most commonly present with lymphaden-
opathy, but a variety of presentations are possible. These include
systemic symptoms such as fevers, night sweats, weight loss, and
pruritus, which are believed to be the result of the release of cytokines
by normal or malignant cells. Patients can present with symptoms
secondary to a mediastinal or retroperitoneal mass such as superior
vena cava obstruction, pleural effusion, pericardial tamponade, ab-
dominal or back pain, intestinal obstruction or perforation, gastro-
intestinal bleeding, or renal failure from urethral obstruction.
Central nervous system (CNS) presentations include primary brain
tumours, signs of meningeal involvement and spinal cord compres-
sion. Patients might present with cytopenia secondary to either bone
marrow involvement or autoimmune destruction of the formed
elements of the blood. Symptoms secondary to the overproduction
of a monoclonal immunoglobulin or hypogammaglobulinaemia
can be seen. In short, the possible presentations of lymphomas are so
varied that the diagnosis should be considered in many patients, and
not just those presenting with lymphadenopathy or splenomegaly.
Diagnosis and evaluation
The diagnosis of lymphoma should always be based on evaluation by
an expert haematopathologist of an adequate biopsy of a lymph node,
or of an extranodal tumour mass if lymph nodes are unavailable. It
is important not to handicap the haematopathologist by providing
inadequate material. Needle aspirates or small biopsies should be
avoided as the basis for diagnosing lymphoma whenever possible.
The differential diagnosis that the pathologist considers when diag-
nosing a lymphoma includes benign proliferations of lymphoid
tissue, malignancies of myeloid cells, nonhaemopoietic malignan-
cies, viral infections, and unusual disorders such as Castleman dis-
ease and giant lymph node hyperplasia. Having tissue available for
immunological studies and/​or genetic studies will frequently help to
confirm the diagnosis.
Once the diagnosis of a type of lymphoma has been established, a
series of studies should be carried out to determine the extent of dis-
ease and prognosis (Box 22.4.4.3). The anatomical spread of disease
is usually expressed as an Ann Arbor stage (Table 22.4.4.4). This sta-
ging system was originally developed for Hodgkin lymphoma and
divides patients into those with disease confined to one lymphatic
site, multiple lymphatic sites on one side of the diaphragm, lymph-
atic involvement on both sides of the diaphragm, and those with
bone marrow involvement, liver involvement, or other extensive
extranodal disease. The Ann Arbor stage also includes a suffix A or
B indicating the absence (A) or presence (B) of unexplained fevers
above 38°C, weight loss of more than 10% of the body weight in
the preceding 6 months, or drenching night sweats. A more recent
staging system has suggested removing the classification A and B
for NHL.
A revised classification includes stage II bulky disease defined as a
single nodal mass of 10 cm or greater than a third of the transthoracic
diameter at any level of thoracic vertebrae. A fluorodeoxyglucose
positron emission tomography (FDG-​PET)/​CT scan is the
Table 22.4.4.3  Chromosomal translocations characteristic of NHL
NHL subtype
Translocation
Genes involved
Frequency
Diffuse large B-​cell
t(3q27)
BCL-​6
35%
t(14;18)(q32;q21)
IqH, BCL-​2
15–​20%
t(8;14)(q24;q32)
MYC, IgH
<5%
Burkitt
t(8;14)(q24;q32)
MYC, IgH
100% have one of these;
most commonly t(8;14)
t(8;22)(q24;q11)
MYC, IgL
t(2;8)(p12;q24)
IgK, MYC
Follicular
t(14;18)(q32;q21)
IgH, BCL-​2
c.90%
Mantle cell
t(11;14)(q13;q32)
BCL-​1, IgH
90%
ALCL
t(2;5)(p23;q35)
ALK, NPM
80% of ALK + ALCLs
MALT
t(11;18)(q21;q21)
API 2, MALT1
35%
t(14;18)(q21;q32)
IgH, MALT1
20%
t(1;14)(p22;q32)
BCL-​10, IgH
10%
section 22  Haematological disorders
5292
preferable imaging modality for staging most lymphomas (though
it is less useful in MALT and small lymphocytic lymphoma). If PET/​
CT is not available, a CT can be used. At the completion of therapy,
where a repeat CT may show a partial regression of a mediastinal
or retroperitoneal mass (because of a sclerotic reaction to the tu-
mour), a PET/​CT will often show a complete metabolic response.
Determining how much improvement in a PET scan was required to
document a complete remission limited the utility of this procedure
until the development of the Deauville, or 5-​point, score, which is
discussed in more detail in Chapter 22.4.3.
Prognostic factors
Knowledge of the specific subtype of NHL is only one of two pieces
of information necessary to plan the intelligent management of
patients with these disorders. The other that must be available in-
volves the delineation of the prognostic characteristics of the indi-
vidual patient. Although it is true that follicular lymphoma has a
higher median overall survival than diffuse large B-​cell lymphoma,
individual patients with follicular lymphoma might have a much
worse survival because of adverse prognostic characteristics than
an individual patient with diffuse large B-​cell lymphoma who has
good prognostic characteristics. Codification of these prognostic
characteristics into a practical clinical tool was accomplished by a
large international study that yielded the International Prognostic
Index (IPI) (Table 22.4.4.5). The IPI is a summation of a number
of specific adverse prognostic factors in an individual patient. The
important factors include age greater than 60  years, Ann Arbor
stage III/​IV, serum lactate dehydrogenase (LDH) level greater than
normal, reduced performance status, and multiple extranodal sites
of involvement by lymphoma. The IPI is useful in essentially all
types of NHL, although it was developed using patients with dif-
fuse large cell lymphoma of both T-​ and B-​cell origin. A new index
for use in patients with follicular lymphoma has been developed
and referred to as the FLIPI (i.e. follicular lymphoma International
Prognostic Index), and others are available for mantle cell and T-​
cell lymphomas. The treatment plan for any individual patient with
lymphoma must always include knowledge of the specific subtype
of lymphoma and the patient’s prognostic characteristics
It is increasingly apparent that the genetic abnormalities in lymph-
omas represent an important prognostic factor. It has been known
for more than a decade that diffuse large B-​cell lymphoma can be
subdivided into two different subtypes based on patterns of gene
expression, with the germinal centre B-​cell subtype having a better
prognosis than the activated B-​cell subtype in some studies. Specific
genetic abnormalities that can be recognized by FISH studies (i.e.
considerably easier to perform than gene profiling) also appear to
be important. For example, diffuse large B-​cell lymphoma tumours
that have both MYC and BCL-​2 (double–​hit) rearrangements have a
particularly poor prognosis with currently available therapies.
Box 22.4.4.3  Staging evaluation for a new patient
with lymphoma
•	 Complete history and physical examination.
•	 Haematological studies:
—​	 Full blood count.
•	 Chemistry studies to measure normal organ function:
—​	 Serum creatinine, liver function studies.
—​	 Serum lactate dehydrogenase, β2-​microglobulin and protein
electrophoresis.
•	 Imaging studies:
—​	 Chest radiograph.
—​	 PET/​CT scan or, if not available, contrast-​enhanced CT of the chest,
abdomen, and pelvis.
•	 Bone marrow biopsy.
•	 Pregnancy test in women of child-​bearing age.
•	 Fertility counselling.
•	 Other studies as appropriate to evaluate specific complaints and to
follow up abnormal results found from the studies previously listed:
—​	 Not appropriate if CT chest is performed. However, if performed,
a chest radiograph offers an easy way to follow mediastinal or pul-
monary involvement.
—​	 Bone marrow biopsy may be omitted in diffuse large B-​cell
lymphoma, particularly in early stage disease, if a PET is performed
because of the sensitivity of PET to detect marrow involvement.
—​	 Other studies can be useful in particular situations. Magnetic res-
onance imaging studies are particularly useful in evaluating sus-
pected bone or CNS sites of involvement. Cerebrospinal fluid
cytology and flow cytometry may be necessary in evaluating sus-
pected CNS sites of involvement. In some patients, abdominal
ultrasonography will provide a more economical way to follow
intra-​abdominal disease. The use of rituximab can increase the risk
of hepatitis reactivation. Hepatitis B and C serologies should be
performed if rituximab is being considered as a part of therapy.
Baseline echocardiogram if there is a history of cardiac disease and
is often performed before initiation of anthracycline.
Table 22.4.4.4  The Ann Arbor staging system
Stage
Characteristics
I
1 nodal site involved
IE
1 site of localized extranodal involvement
II
2 or more nodal sites involved, but only on 1 side of the diaphragm
IIE
1 site of localized extranodal involvement plus regional nodes
involved—​all on 1 side of the diaphragm
III
Nodal involvement (i.e. spleen counts as a nodal site) on both sides
of the diaphragm
IV
Bone marrow, liver, or other extensive extranodal involvement
(e.g. multiple pulmonary nodules)
Table 22.4.4.5  International Prognostic Index
Full index
Age adjusted (i.e. for patients <60 years)
Prognostic factors (APLES)
Prognostic factors (PLS)
Age >60 years
Performance status >1
Performance status ≥2
LDH >1 × normal
LDH >1 × normal
Stage III or IV
Extranodal sites ≥2
Stage III or IV
Risk category factors
Risk category factors
Low 0 or 1
Low 0
Low–​intermediate 2
Low–​intermediate 1
High–​intermediate 3
High 4 or 5
22.4.4  Non-Hodgkin lymphoma
5293
General principles of lymphoma treatment
Types of treatment
Those treatments effective in the management of patients with
cancer include surgery, radiotherapy, cytotoxic chemotherapy, and
a variety of new approaches developed through increasing under-
standing of the biology of the immune system. The latter include
cytokines, antibodies, and attempts to direct an immune reaction
against cancer.
As few patients with lymphoma have truly localized disease, sur-
gery has not been a major treatment modality, except for selected
patients with extranodal MALT lymphomas. Since its utilization
in medicine in the first part of the 20th century, radiotherapy has
been a major treatment modality for patients with lymphoma, but
is limited in its application by toxicity. Its curative potential depends
upon being able to achieve a tumouricidal dose (typically 30–​40
Gy) without irreversibly injuring normal organs. Thus, the ana-
tomical site of involvement, as well as the number of sites involved,
can limit the effectiveness of this treatment, since toxicity increases
with the volume of tissue irradiated. If a lymphoma is truly local-
ized, radiotherapy is often curative. However, most patients have oc-
cult metastatic disease and cure rates are often higher when a brief
course of chemotherapy precedes the radiation. Two approaches
have been utilized to make radiotherapy a ‘systemic’ treatment. One
involves radiation of the total body. When this is part of a haem-
atopoietic stem cell transplant regimen, a total dose of 10 to 12 Gy
can be administered. More recently, it has been demonstrated that
it is possible to give higher doses of radiotherapy to multiple areas
by attaching radioactive molecules to antibodies that home to sites
of involvement by lymphoma. For example, radioimmunotherapy
such as 131I-​tositumomab or 90Y-​ibritumomab tiuxetan have dem-
onstrated significant response in patients with follicular lymphoma.
Cytotoxic chemotherapeutic agents were first discovered in the
1940s when mechlorethamine (i.e. the nitrogen mustard gas used
in warfare), and subsequently methotrexate, were found to cause re-
gressions in immune-​system malignancies. A wide variety of agents
have since been shown to be able to cause disease regression in many
patients with lymphomas. Unfortunately, early studies showed that
regressions induced by single agents were almost invariably fol-
lowed by regrowth of the tumour and eventual death of the patient.
In an attempt to circumvent this, combinations of chemotherapeutic
agents were first utilized in the 1960s and early 1970s. The drugs
were combined by attempting to choose agents with different mech-
anisms of action and nonoverlapping toxicities to allow the admin-
istration of doses that were near to the maximum tolerated dose
with an individual agent. In both childhood acute leukaemia and
Hodgkin’s lymphoma, this approach was validated by the cure of a
significant number of patients. Today, several combination chemo-
therapy regimens with acceptable toxicity have been shown to be ef-
fective and are widely used worldwide (Table 22.4.4.6). All regimens
are not equally good for treating all types of lymphoma.
Very high doses of cytotoxic chemotherapeutic agents with or
without radiotherapy and biologically active molecules have been
utilized in the treatment of patients with lymphomas as part of the
haematopoietic stem cell transplantation procedure. This involves
the administration of very high doses of antilymphoma therapy in
Table 22.4.4.6  Common combination chemotherapy regimens used in treating patients with non-​Hodgkin lymphoma
Regimen
Drug
Dose (mg/​m2)
Route
Schedule
CVP-​R
21-​day cycles
Cyclophosphamide
750–​1200
IV
D1
Vincristine
1.4 (max. 2)
IV
D1
Prednisone
100 total dose (not by m2)
PO
D1–​5
Rituximab
375
IV
D1
CHOP-​R
21-​day cycles
Cyclophosphamide
750
IV
D1
Doxorubicin
50
IV
D1
Vincristine
1.4 (max. 2)
IV
D1
Prednisone
100 total (not by m2)
PO
D1–​5
Rituximab
375
IV
D1
BR
28-​day cycles
Bendamustine
90
IV
D1 and 2
Rituximab
375
IV
D1
FCR
28-​day cycles
Fludarabine
25
IV
D1–​3
Cyclophosphamide
250
IV
D1–​3
Rituximab
375
IV
D1
EPOCH-​R
21-​day cycles
Rituximab
375
IV
D1
Etoposide
50
IV infusion
D1–​4
Doxorubicin
10
IV infusion
D1–​4
Vincristine
0.4
IV infusion
D1–​4
Cyclophosphamide
750
IV
D5
Prednisone
60
PO
D1–​5
D, day(s); IV, intravenously; PO, orally.
section 22  Haematological disorders
5294
an attempt to overcome presumed treatment resistance. Patients are
rescued from the toxicity of treatment by the reinfusion of haemo-
poietic stem cells. The patient’s own haemopoietic stem cells (an
autologous transplant) or those from another individual with iden-
tical HLA genes (an allogeneic transplant) can be utilized. Cells for
this procedure can be obtained from either bone marrow or per-
ipheral blood. Autologous transplantation has been widely used for
patients with lymphoma and shown to be able to cure patients with
aggressive NHL. Transplantation is generally reserved for relapsed
or refractory lymphoma. In aggressive NHL, a possible increased
cure rate has been demonstrated by utilizing adjuvant autologous
transplantation following initially effective standard chemotherapy
in patients with a poor prognosis. Allogeneic transplantation, while
apparently curative, has a high mortality rate. Allogeneic transplant-
ation is considered in the management of high-​risk chronic lympho-
cytic leukaemia, or in other high-​risk NHLs frequently after the
failure of autologous transplantation.
Increasing knowledge of the immune system has further led to
the recognition that a number of biologically active molecules can
cause regression of lymphomas and, in some cases, impact survival.
In B-​cell NHL, antibodies directed against the CD20 molecule have
been incorporated into clinical practice. Rituximab, obinutuzumab,
and ofatumumab are commonly utilized anti-​CD20 monoclonal
antibodies that have been shown to be active in a variety of B-​cell
lymphomas. The combination of lenalidomide and rituximab has
demonstrated efficacy against a variety of B-​cell lymphomas. The
antibody conjugate brentuximab vedotin has shown significant re-
sponses in CD30-​expressing tumours.
Ibrutinib is an oral irreversible inhibitor of Bruton’s tyrosine kinase
(BTK), an integral component of the B-​cell receptor. Ibrutinib has
single-​agent activity against chronic lymphocytic leukaemia, mantle
cell lymphoma, and Waldenström macroglobulinaemia. Idelalisib
is an oral small-​molecule inhibitor of the phosphatidylinositol
3-​kinase (PI3Kδ) that is highly expressed in B-​cell lymphomas.
Idelalisib, combined with rituximab, is active against chronic
lymphocytic leukaemia/​small lymphocytic lymphoma and follicular
lymphoma. Venetoclax or ABT-​199 is a small molecule inhibitor of
BCL-​2 and active against chronic lymphocytic leukaemia/​small
lymphocytic lymphoma. Several other agents are in development.
Promising drugs include an antibody–​drug conjugate (polatuzumab
vedotin containing an anti-​CD79B monoclonal antibody), immune
checkpoint inhibitors, and anti-​CD19 chimeric antigen receptor
(CAR) T cells.
General strategy of treatment
A number of factors need to be taken into account when formu-
lating a treatment recommendation for a patient with lymphoma
(Box 22.4.4.4). This decision should be made in conjunction with
the patient, and requires good judgement in addition to technical
knowledge. The aggressiveness of the treatment that is finally chosen
will often depend upon the physician’s interpretation of the chances
for cure. It is obvious that more toxicity will be acceptable if the goal
is cure rather than palliation. For this reason, patients with defin-
itely curable lymphomas, such as diffuse large B-​cell lymphoma
and Burkitt lymphoma, are almost always treated promptly with in-
tensive regimens. By contrast, the best treatment for patients with
follicular lymphoma remains a point for intense debate. Since the
curability of this disease is less clear, many physicians would favour
no initial therapy in an asymptomatic patient, but—​as discussed
later—​this is not a simple decision.
For most patients, the goal of therapy is to achieve a complete
remission. This implies the disappearance of all symptoms and ob-
jective evidence of lymphoma. In practice, a complete remission is
documented by repeating all abnormal staging studies after sev-
eral cycles of therapy or at the completion of the planned therapy.
Documentation of complete remission is important. Patients who
achieve a complete remission have a chance for cure; those who do
not achieve a complete remission with initial therapy will often go
directly to second-​line treatments.
Patients who are not cured with initial therapy, either because
they do not achieve an initial remission or because they relapse
from remission, are candidates for what has been termed ‘salvage
therapy.’ These second-​line regimens can regularly cause tumour re-
gression in most patients with lymphoma and can occasionally pro-
duce long-​term, disease-​free survival. However, for most patients,
the only curative approach in this setting is haematopoietic stem
cell transplantation. The toxicity of haematopoietic stem cell trans-
plantation limits its use to patients less than 70 to 75 years of age,
who have a good performance status, without serious compromise
of major organ function; and to patients who do not have bulky/​
chemotherapy-​refractory disease.
Precursor B-​ and T-​cell lymphomas
Lymphoblastic lymphoma of B-​cell or T-​cell origin
Lymphoblastic lymphoma is a tumour of the precursor cells of T-​
and B-​lymphocytes. It is intimately related to the acute lymphoid
leukaemias, with the difference being the method of presentation.
Lymphoblastic lymphoma is diagnosed when the lymphoblasts are
undetectable in blood and consist of less than 25% of marrow cells.
Sometimes it is difficult to determine when a patient should be said
to have acute lymphoid leukaemia or lymphoblastic lymphoma,
since bone marrow involvement is frequent with a lymphomatous
presentation and lymphadenopathy and mediastinal mass are
common in patients who present with leukaemia.
Most patients with lymphoblastic lymphoma have tumours de-
rived from T lymphoblasts, but approximately 15% are B cell in
origin. The differential diagnosis of lymphoblastic lymphoma
includes a blastic variant of mantle cell lymphoma, acute mye-
loid leukaemia, and peripheral T-​cell lymphoma in children and
Box 22.4.4.4  Factors to consider in therapy for a patient
with lymphoma
•	 Specific type of lymphoma
•	 Age
•	 Performance status
•	 Presence of other diseases
•	 Stage
•	 Systemic symptoms
•	 Pace of disease
•	 Potential side effects
•	 Likelihood of cure
•	 Patient’s concerns about specific treatments
•	 Convenience
•	 Patient’s immediate and long-​term goals
•	 Quality of life
22.4.4  Non-Hodgkin lymphoma
5295
young adults. When a mediastinal mass is present in lympho-
blastic lymphoma, the differential diagnosis also includes Hodgkin
lymphoma, primary mediastinal B-​cell lymphoma, thymoma, and
germ-​cell tumour.
The median age of patients with lymphoblastic lymphoma is the
late twenties; most patients are male with widely disseminated dis-
ease and an elevated serum LDH level, and about 50% will have bone
marrow involvement. In young men, testicular involvement should
be ruled out with ultrasonography. Patients with lymphoblastic
lymphoma who present with stage IV disease, elevated LDH levels,
and bone marrow or CNS involvement have a poorer prognosis
than adult patients who do not have these adverse characteristics.
Patients with none of these adverse characteristics have a high cure
rate with regimens such as those used for acute lymphoblastic leu-
kaemia. Some patients with stage IV disease, elevated LDH levels,
and bone marrow and CNS involvement can also be cured, but this
is less likely. Treatment in both groups of patients should include
high-​dose multiagent induction, consolidation/​intensification and
maintenance chemotherapy, and CNS prophylaxis or treatment.
Adolescents and young adults treated with paediatric acute lympho-
blastic leukaemia regimens have improved survival than those
treated with adult or NHL regimens. Patients with high-​risk charac-
teristics, or those who relapse after initial therapy, are candidates for
allogeneic or autologous haematopoietic stem cell transplantation.
Mature B-​cell lymphomas
Diffuse large B-​cell lymphoma
Diffuse large B-​cell lymphoma is the most common type of NHL, rep-
resenting approximately one-​third of all patients. It most commonly
presents de novo, but can also develop after histological transform-
ation of an indolent lymphoma such as follicular, small lymphocytic,
or MALT lymphoma. This tumour can arise in lymph nodes or may
involve any extranodal site, including the CNS. Rare presentations
include pleural effusions from involvement of serosal surfaces (effu-
sion lymphoma) and multiple organ system dysfunction secondary
to endothelial involvement (intravascular lymphomatosis).
Almost all diffuse large B-​cell lymphomas display the CD20
antigen, and several cytogenetic abnormalities are frequently asso-
ciated (Table 22.4.4.2). The condition can be subdivided using gene
microarrays into the germinal centre B-​cell type, the activated B-​cell
type, and the mediastinal large B-​cell type. These have different
clinical characteristics and response to therapy (Table 22.4.4.7 and
Fig. 22.4.4.1).
Several other subtypes of diffuse large B-​cell lymphoma deserve
special mention because of unique treatment considerations. The
plasmablastic subtype of diffuse large B-​cell lymphoma is usu-
ally CD20 negative and thus does not benefit from treatment with
rituximab. Diffuse large B-​cell lymphoma with rearrangements of
both BCL-​2 and MYC is a ‘double-​hit’ lymphoma and has a very
poor prognosis. Very intensive regimens and bone marrow trans-
plantation are often used in management, but the prognosis is still
poorer than for other subtypes. Diffuse large B-​cell lymphomas ori-
ginating in the testes, CNS, and skin require different treatment ap-
proaches. Testicular diffuse large B-​cell lymphoma, that is, the most
frequent testicular tumour in men over the age of 60 years, has fre-
quent relapses in the CNS and the contralateral testis. CNS prophy-
laxis and irradiation to the opposite testis is required. Primary CNS
diffuse large B-​cell lymphoma does not benefit from standard re-
gimens and requires treatments that penetrate the CNS including
high-​dose methotrexate. Certain lymphomas that originate in the
skin, typically on the scalp or upper trunk, might be called diffuse
large B-​cell lymphoma but have an indolent course and only require
local therapy.
The differential diagnosis of diffuse large B-​cell lymphoma
includes undifferentiated carcinoma, acute myeloid leukaemia,
Hodgkin lymphoma, and extramedullary plasmacytoma. Occasional
patients with diffuse large B-​cell lymphoma have a large number
of infiltrating T cells, and can be confused with a peripheral T-​cell
lymphoma. Appropriate immunological studies and genetic studies
can usually resolve any confusion.
The clinical characteristics of patients with diffuse large B-​cell
lymphoma are presented in Table 22.4.3.8. The median age at pres-
entation of patients with diffuse large B-​cell lymphoma is approxi-
mately 64  years and there is a slight male predominance. About
one-​half of the patients will have stage I or II disease and about one-​
half will have a more widely disseminated lymphoma: approximately
Table 22.4.4.7  Recognized molecular subtypes of diffuse large B-​cell lymphoma
Characteristics
Germinal centre B-​cell type
Activated B-​cell type
Mediastinal large B-​cell lymphoma
% of all patients with diffuse large B-​cell lymphoma
c.60
c.30
7
Median age (years)
58
66
37
% female
50
40
70
% 5-​year survival
(i.e. with rituximab)
70–​90
60–​65
80-​90
1.0
0.8
0.6
0.4
0.2
0.0
0
2
4
6
8
10
Overall survival (years)
Probability
PMBL
ABC
DLBCL
GCB
DLBCL
5-year
survival
64%
59%
30%
Fig. 22.4.4.1  Survival of patients with subtypes of diffuse large B-​cell
lymphoma (DLBCL) treated largely in the pre-​rituximab era. ABC,
activated B cell; GCB, germinal centre B cell; PMBL, primary mediastinal
B-​cell lymphoma.
section 22  Haematological disorders
5296
two-​thirds will have some sign of extranodal involvement, one-​third
will have B-​symptoms at presentation, and one-​half have an elevated
LDH. Bone marrow involvement is seen in approximately 15%
of cases.
Since the early 1970s, it has been known that patients with dif-
fuse large B-​cell lymphoma could sometimes be cured with combin-
ation chemotherapy regimens alone—​even those with disseminated
disease. The most popular regimen in use today is CHOP (cyclo-
phosphamide, doxorubicin, vincristine (Oncovin), and prednisone)
plus the anti-​CD20 monoclonal antibody rituximab. However, a
large number of other regimens including ACVBP (doxorubicin,
cyclophosphamide, vindesine, bleomycin, and prednisone) plus
rituximab, and EPOCH (etoposide, prednisone, vincristine, cyclo-
phosphamide and doxorubicin) plus rituximab, are at least as ac-
tive. Today all regimens used to treat diffuse large B-​cell lymphoma
will include the antibody rituximab unless the patient’s individual
tumour has been shown to be CD20 negative. When a staging evalu-
ation shows disease confined to one site (i.e. stage I) or two nearby
sites (i.e. minimal stage II) a brief course of chemotherapy followed
by radiotherapy has been the most popular treatment approach, al-
though a complete course of CHOP plus rituximab might be equally
effective. In patients with disseminated disease, a complete course
of CHOP plus rituximab, or another active regimen plus rituximab,
is standard therapy. Local radiotherapy is sometimes added to very
bulky (i.e. >7.5 cm) sites of disease. In patients who present with the
multiple adverse risk factors listed in the IPI, adjuvant autologous
haemopoietic stem cell transplantation after achieving an initial re-
mission might benefit selected patients. Lenalidomide and ibrutinib
are active against diffuse large B-​cell lymphoma of the ABC genetic
subtype, and a combination of CHOP plus rituximab, and ibrutinib
or lenalidomide appear to improve responses in early studies.
Approximately 75 to 85% of patients with localized disease can be
cured with CHOP and rituximab with or without radiotherapy. More
than 50% of patients with more disseminated disease can be cured
with combination chemotherapy regimens including rituximab.
Patients who relapse from complete remission can sometimes be
cured with autologous haematopoietic stem cell transplantation.
Patients who remain chemotherapy sensitive after relapse have been
cured approximately 40% of the time (i.e. perhaps less often after the
patients has failed rituximab) while chemotherapy-​resistant patients
are cured only about 10% of the time.
Follicular lymphoma
The second most common type of NHL is follicular lymphoma.
The clinical characteristics of patients with this disease are listed
in Table 22.4.4.8. The differential diagnosis of follicular lymphoma
includes benign follicular hyperplasia and follicular variants of
other NHLs.
Patients with follicular lymphoma are subdivided based on the
number of large cells in the tumour into grade 1 (i.e. those with the
least large cells), grade 2, and grade 3 (i.e. those with the most large
cells). The method for assigning grade is controversial among patho-
logists, whose ability to reproduce grading is much lower than their
ability to reproducibly diagnose follicular lymphoma. In general, a
higher proportion of large cells is associated with a higher prolif-
erative rate, more rapid tumour progression, and perhaps a better
response to anthracycline-​containing combination chemotherapy
regimens. The natural history of follicular lymphoma involves a re-
duction in the degree of follicularity in the tumour over time and an
increase in the proportion of large cells. Transformation—​usually to
diffuse large B-​cell lymphoma—​is recognized in approximately 20 to
40% of patients and is associated with a poor prognosis in most cases.
Table 22.4.4.8  Clinical characteristics of the major subtypes of B-​cell non-​Hodgkin lymphoma
Diffuse large B-​cell
lymphoma
Follicular
lymphoma
MALT lymphoma
Small lymphocytic
lymphoma
Mantle cell
lymphoma
Median age (years)
64
59
60
65
63
Percentage male
55
42
48
53
74
Stage (%):
  I
25
18
39
  4
13
  II
19
15
28
  5
7
  III
13
16
  2
  8
9
  IV
33
51
31
83
71
B-​symptoms (%)
33
28
19
33
28
Elevated LDH (%)
53
30
27
41
40
Reduced performance status (%)
24
  9
15
11
21
Tumour mass >10 cm (%)
30
28
  8
13
81
Bone marrow involvement (%)
16
47
14
72
64
Gastrointestinal tract involvement (%)
18
  4
50
  3
50
IPI score (%)
  0–​1
35
45
44
23
23
  2–​3
46
48
48
64
54
  4–​5
19
  7
  8
13
23
IPI, International Prognostic Index.
22.4.4  Non-Hodgkin lymphoma
5297
Follicular lymphomas regularly display the CD20 antigen.
Most tumours will have the t(14;18) translocation and express the
BCL-​2 protein. Transformation to diffuse large B-​cell lymphoma is
frequently associated with additional cytogenetic abnormalities.
Treatment approaches commonly utilized in the management of
patients with follicular lymphoma are presented in Box 22.4.4.5.
Asymptomatic patients with nonbulky disease and no organ com-
promise are often managed with no initial therapy—​a strategy that is
sometimes called ‘watchful waiting’. When followed in this manner,
approximately 10 to 25% of patients will undergo at least a partial
spontaneous regression (i.e. what would be called a partial response
if a treatment had been utilized), although these regressions are gen-
erally not durable. Over time, almost all patients will progress and
require therapy. There is no ‘standard’ treatment for patients with
follicular lymphoma. The most often utilized initial treatments are
single-​agent chemotherapy including bendamustine, CVP (cyclo-
phosphamide, vincristine, and prednisone), and CHOP. With few
exceptions, each of these regimens will be combined with the anti-
body rituximab. An increasingly popular treatment approach, and
one that is sometimes utilized instead of watchful waiting, is single-​
agent therapy with rituximab which has been shown to delay time
to starting chemotherapy, but which does not improve overall sur-
vival. Recently, the combination of bendamustine and rituximab has
become popular as a therapy for patients with low-​grade follicular
lymphoma. Although an initial study suggested this approach might
be superior to CHOP and rituximab, more recent studies suggest
that the two regimens have comparable outcomes. Most patients re-
spond, although many will not achieve a complete remission. When
ongoing or ‘maintenance’ treatment with the antibody is continued
after the initial induction therapy, more patients achieve a com-
plete remission and the duration of remissions is prolonged. It is
now clear that when combinations of traditional chemotherapeutic
agents are combined with rituximab as initial therapy, then patients
have a longer survival than when combination chemotherapy alone
is utilized. Maintenance rituximab for patients achieving a remis-
sion with combined chemotherapy/​rituximab regimens has been
shown to prolong remission duration but not overall survival.
Particular treatment approaches are more appropriate for certain
subsets of patients with follicular lymphoma. The rare patients with
localized follicular lymphoma can be managed with radiotherapy
alone. These patients have an excellent outlook with a 10-​year sur-
vival of 70 to 90% in most series, and approximately 40 to 50% of
patients not relapsing after 10 years of follow-​up. Patients with grade
3 follicular lymphoma often respond to treatments used for diffuse
large B-​cell lymphoma, and many physicians would favour CHOP
plus rituximab (CHOP-​R) as the initial treatment for these patients.
Bone marrow specimens from patients with follicular lymphoma
frequently test positive for the BCL2 gene rearrangement using poly-
merase chain reaction (PCR) technology. Some but not all patients
who achieve a complete remission will revert to a BCL-​2-​negative
status. This test has not been widely utilized clinically since some
patients who remain positive do not relapse, some patients who be-
come negative do relapse, and circulating lymphoid cells in normal
patients sometimes have the BCL2 gene rearrangement. This pre-
sumably reflects that rearrangement of the BCL2 gene is a very early
step in lymphomagenesis.
Most patients with follicular lymphoma will eventually fail their
initial treatment regimen, with this being especially true for pa-
tients with follicular lymphoma grade 1 and grade 2. Subsequent
treatments have included single drugs such as chlorambucil or
fludarabine, a variety of combination chemotherapy regimens, and
new drugs such as bortezomib, interferon, monoclonal antibodies
(e.g. rituximab), radiolabelled antibodies, and both allogeneic and
autologous haematopoietic stem cell transplantation. Several new
drugs have activity in patients with follicular lymphoma. The new
B-​cell receptor inhibitors, idelalisib and ibrutinib, both have activity;
the combination of rituximab and idelalisib appears to be particu-
larly active, though with significant associated toxicities including
colitis and pneumonitis. Ibrutinib has also been associated with
bleeding, and this drug should be withheld in the context of sur-
gical intervention. There is a 10–15% risk of developing atrial fib-
rillation on ibrutinib, with older patients being more commonly
affected. Both autologous and allogeneic haematopoietic stem cell
transplantation can produce long-​term disease-​free survival in a
proportion of patients with follicular lymphoma. Autologous haem-
atopoietic stem cell transplantation is more effective when utilized
at first relapse. Allogeneic transplantation has a much lower relapse
rate than autologous transplantation, but is associated with a higher
mortality rate.
Overall survival of patients with follicular lymphoma is ap-
proaching 20 years. The addition of rituximab to standard chemo-
therapy regimens has been instrumental in effecting recent
improvements to outcomes for patients with this disease. Some pa-
tients survive free of disease for extended periods of time, and hope-
fully this proportion will increase with new treatments.
MALT lymphoma
This lymphoma, also known as the extranodal marginal-​zone B-​
cell lymphoma of MALT type, always presents in extranodal sites.
A nodal presentation of a similar lymphoma is referred to as nodal
marginal-​zone lymphoma (see ‘Less common B-​cell lymphomas’).
The differential diagnosis of MALT lymphoma includes benign
lymphocytic infiltration of extranodal organs and other small cell
B-​cell lymphomas.
MALT lymphomas are tumours of CD5-​negative and CD23-​
negative B cells that express CD20. The commonly seen cytogenetic
abnormalities are listed in Table 22.4.4.2. Gastric MALT lymph-
omas are associated with infection by Helicobacter pylori, thy-
roid MALT lymphomas are frequently associated with Hashimoto
thyroiditis, and orbital MALT lymphomas are sometimes asso-
ciated with Sjögren syndrome. MALT lymphomas can undergo
Box 22.4.4.5  Treatment regimens used for patients
with follicular lymphoma
•	 Close observation and no initial therapy
•	 Radiotherapy
•	 Single-​agent therapy: chlorambucil, cyclophosphamide, bendamustine—​
all with rituximab—​or rituximab alone
•	 Combination chemotherapy: CHOP-​R, CVP-​R
•	 Interferon-​α
•	 Radioantibodies: tositumomab, ibritumomab
•	 Haemopoietic stem cell transplantation: autologous, allogeneic
CHOP, cyclophosphamide, doxorubicin, vincristine, prednisone; CVP, cyclo-
phosphamide, vincristine, prednisone; R, rituximab.
section 22  Haematological disorders
5298
histological transformation to diffuse large B-​cell lymphomas. After
this transformation, the patient should be treated for diffuse large
B-​cell lymphoma.
MALT lymphomas have a slight female predominance with a me-
dian age at presentation of approximately 60 years. The symptoms of
the disorder are those associated with involvement of the extranodal
site. The disease is usually localized and the presence of systemic
symptoms or elevated LDH is unusual. The characteristics of pa-
tients with this lymphoma are listed in Table 22.4.4.8.
Gastric MALT lymphomas are the first example of a lymphoma
that can be treated by eliminating a chronic infection. If the tumour
does not transform to a large cell lymphoma, and has not deeply
invaded the stomach, then most patients will have their tumour re-
gress with the eradication of H. pylori using a combination of two
antibiotics (amoxicillin, clarithromycin or metronidazole), and
proton pump inhibitors. It appears that in some patients this treat-
ment might be curative. However, complete regression following
eradication of H. pylori infection may take up to a year. Other local
therapies are also effective. Patients with localized MALT lymph-
omas can be effectively treated with local radiotherapy or, in some
cases, surgery. These lymphomas also respond to rituximab, single-​
agent chemotherapy, and combination chemotherapy. Patients with
disseminated MALT lymphomas usually respond to therapy, but are
rarely curable.
Most patients with localized MALT lymphoma can be cured, and
the 5-​year survival in such patients is approximately 90%. However,
patients with disseminated disease have a more serious illness and
those with a high IPI score have a 5-​year survival of only 50%.
Small lymphocytic lymphoma/​chronic
lymphocytic leukaemia
Small lymphocytic lymphoma is the tissue manifestation of chronic
lymphocytic leukaemia, which is covered in greater detail in
Chapter 22.4.5. Patients who present predominantly with blood and
bone marrow involvement will be diagnosed with chronic lympho-
cytic leukaemia, and those who present with lymphadenopathy as
having small lymphocytic lymphoma, although the WHO classifica-
tion suggests that all these patients might have chronic lymphocytic
leukaemia. Patients with plasmacytoid differentiation and mono-
clonal IgM protein in the serum can present with the syndrome
of Waldenström’s macroglobulinaemia (see ‘Less common B-​cell
lymphomas’).
Small lymphocytic lymphoma makes up approximately 7% of
NHL worldwide, although is more often seen in Western countries.
The differential diagnosis includes other small B-​cell lymphomas,
and patients with small lymphocytic lymphoma can undergo histo-
logical transformation to diffuse large B-​cell lymphoma. This syn-
drome is seen in approximately 3 to 10% of patients and is called
Richter’s syndrome. It is associated with a poor prognosis.
The lymphoma cells are B cells that are CD5, CD20, and CD23
positive, although the concentration of CD20 on the surface of
the tumour cells is less than in most other B-​cell lymphomas.
Common cytogenetic abnormalities seen in chronic lympho-
cytic leukaemia/​small lymphocytic lymphoma include trisomy 12,
del(11q), del(17p), and del (13q), with del (17p) being particularly
associated with a very poor prognosis, and del(13q) as an isolated
abnormality being associated with the best prognosis. Other bio-
logical measurements to predict outcome in patients with chronic
lymphocytic/​small lymphocytic lymphoma have included the pro-
portion of cells that express CD38, those that express ZAP70, and
those with unmutated variable region of immunoglobulin heavy
chain (IgVH) genes.
The clinical characteristics of patients with small lympho-
cytic lymphoma are listed in Table 22.4.4.8. Patients with chronic
lymphocytic leukaemia/​small lymphocytic lymphoma some-
times have acquired immunological abnormalities including
hypogammaglobulinaemia, autoimmune thrombocytopenia, and
autoimmune haemolytic anaemia. When present, these immune
abnormalities should be treated specifically, in addition to any
treatment given for the lymphoma. Hypogammaglobulinaemia
accompanied by frequent episodes of serious infection should be
treated with intermittent immunoglobulin infusions.
Patients with chronic lymphocytic leukaemia/​small lymphocytic
lymphoma can be followed without therapy when they present with
no systemic symptoms or organ compromise. However, most pa-
tients will require treatment within the first few years of follow-​up.
The most popular treatments for patients with chronic lymphocytic
leukaemia/​small lymphocytic lymphoma contain fludarabine and
cyclophosphamide in combination with rituximab. Older, frailer
patients are frequently treated with chlorambucil or bendamustine
combined with an anti-​CD20 antibody such as rituximab,
ofatumumab, or obinutuzumab. The patients most likely to respond
to fludarabine, cyclophosphamide, and rituximab with long remis-
sions are those with the genetic abnormality of either a 13q deletion
or trisomy 12 and who have mutated IgHV genes.
A number of new agents are changing the treatment paradigm
for chronic lymphocytic leukaemia/​small lymphocytic lymphoma.
Ibrutinib, idelalisib, and venetoclax are all extremely active drugs.
Ibrutinib is usually administered alone, and idelalisib is often com-
bined with rituximab. Almost all patients respond to ibrutinib, al-
though with all drugs of this class, there is an initial lymphocytosis
followed by gradual disappearance of the abnormal cells. Most pa-
tients with an excellent response to ibrutinib stay in remission for
at least a few years as long as they continue taking the drug. For
patients with a 17p deletion—​the most adverse genetic finding—​
ibrutinib is the treatment of choice. However, patients treated with
any of these approaches will eventually progress and require further
treatment. Only a few patients are candidates for allogeneic haem-
atopoietic stem cell transplantation, but this can achieve long-​term
disease-​free survival in some cases.
Mantle cell lymphoma
The clinical characteristics of patients with mantle cell lymphoma
are listed in Table 22.4.4.8. This lymphoma was recognized as a
specific entity because of its characteristic cytogenetic abnormality,
these tumours regularly manifesting the t(11;14) translocation that
involves the BCL1 gene on chromosome 11 and leads to overpro-
duction of the BCL-​1 protein. Indeed, before the recognition of this
disorder, patients with mantle cell lymphoma were placed in many
other histological categories. The tumours were previously termed
centrocytic lymphoma under the Kiel classification. An expert
haematopathologist is important in making the diagnosis, since this
lymphoma can be confused with small lymphocytic lymphoma, fol-
licular lymphoma, and lymphoblastic lymphoma.
Extranodal sites of involvement by mantle cell lymphoma are not
unusual. Large-​bowel involvement can present as the syndrome of
22.4.4  Non-Hodgkin lymphoma
5299
lymphomatous polyposis. Patients with distal gastrointestinal tract
lymphoma often have Waldeyer’s ring involvement in addition.
Mantle cell lymphoma has been among the most difficult types
of NHL to treat. Using standard chemotherapy regimens such
as CHOP, the remission duration has been brief and the median
overall survival approximately 3 to 5 years. The addition of the anti-
body rituximab, combinations including bendamustine, and the
use of very intensive chemotherapy regimens incorporating high
doses of cytosine arabinoside have improved the response rate
and remission duration. Ibrutinib is a very active drug in mantle
cell lymphoma and is now approved for use for patients with re-
current disease in many countries. It is likely that this drug will
make a contribution to the primary therapy of patients with mantle
cell lymphoma. Other novel agents include idelalisib, bortezomib,
lenalidomide, and everolimus. The combination of lenalidomide
and rituximab appears to have high response rate for untreated
patients in early studies. Maintenance therapy with rituximab fol-
lowing the completion of initial therapy may improve survival.
However, the cure of these patients with standard chemotherapy re-
gimens is still uncertain, hence many patients receive autologous or
allogeneic transplantation in first remission. Allogeneic transplant-
ation can occasionally cure patients who have failed their primary
regimen. As most patients with mantle cell lymphoma are elderly,
these more aggressive treatment approaches are often inappropriate
or impractical.
Less common B-​cell lymphomas
Burkitt lymphoma was originally described by Denis Burkitt while
studying an aggressive lymphoma that occurred in the jaw of children
in Central Africa; the disease can also present as acute leukaemia.
An association has been demonstrated between EBV infection and
this lymphoma, which is much more frequent in children and young
adults and in patients infected by HIV. The condition is associated
with specific chromosomal translocations involving the heavy-​chain
immunoglobulin gene on chromosome 14 or the light-​chain im-
munoglobulin genes on chromosomes 2 and 22. In each case, the
associated oncogene is the MYC gene on chromosome 8 (namely,
t(8;14), t(2;8), and t(8;22)). The condition can frequently be cured
utilizing short courses of very intensive regimens that incorporate
high doses of cyclophosphamide, and with the EPOCH-​R regimen.
Hence the distinction between Burkitt lymphoma and diffuse large
B-​cell lymphoma is extremely important.
The WHO now recognizes that there are some lymphomas with
characteristics that are intermediate between diffuse large B-​cell
lymphoma and Burkitt lymphoma where this distinction cannot
be made. These lymphomas typically have a high proliferative rate
and are usually treated like Burkitt lymphoma. These lymphomas
often have a double hit (i.e. the presence of both the t(8;14) and the
t(14;18) translocations) and have a much poorer response to therapy
than either Burkitt lymphoma or diffuse large B-​cell lymphoma.
Nodal marginal-​zone lymphoma is immunologically related to
MALT lymphoma (mentioned earlier), but presents in a manner
similar to follicular lymphoma. These patients respond to therapy
and have an overall survival similar to those with follicular
lymphoma.
Splenic marginal-​zone lymphoma is a rare disorder, also known
as splenic lymphoma with villous lymphocytes. It is typically an in-
dolent condition involving the spleen, marrow, and blood without
palpable lymphadenopathy, and can be treated with splenectomy
but often also responds dramatically to single-​agent rituximab, now
the treatment of choice.
Primary mediastinal diffuse large B-​cell lymphoma varies from
other diffuse large B-​cell lymphomas in that it occurs at a younger age
and has a striking female predominance. Gene expression profiling
has shown that these tumours are genetically distinct and have some
similarities to nodular sclerosing Hodgkin lymphoma. However,
the treatment and response to therapy are similar to that seen in
the germinal centre B-​cell type of diffuse large B-​cell lymphomas
(Table 22.4.4.7 and Fig. 22.4.4.1). Primary mediastinal lymphomas
can usually be cured with treatment utilizing the CHOP-​R regimen
and radiotherapy to the large mediastinal mass, but the EPOCH-​R
regimen has a high rate of complete remission, and patients who
achieve a complete remission by PET appear to not require radio-
therapy. The WHO now also recognizes lymphomas whose char-
acteristics are intermediate between primarily mediastinal diffuse
large B-​cell lymphoma and Hodgkin lymphoma, termed grey zone
lymphoma.
Lymphoplasmacytic lymphoma, a subtype of small lympho-
cytic lymphoma (see ‘Small lymphocytic lymphoma/​chronic
lymphocytic leukaemia’), is the histological subtype of lymphoma
seen in lymph nodes biopsied in patients with Waldenström
macroglobulinaemia. The syndrome of Waldenström macro­
globulinaemia is characterized by excessive IgM monoclonal
gammopathy and features of hyperviscosity syndrome such
as blurred vision, headaches, dizziness, retinal vein engorge-
ment, epistaxis, dyspnoea, and paraesthesiae. All patients with
Waldenström’s macroglobulinaemia have a lymphoplasmacytic
lymphoma, but all patients with lymphoplasmacytic lymphoma will
not manifest the syndrome of Waldenström macroglobulinaemia.
Patients with lymphoplasmacytic lymphoma can have the t(9;14)
cytogenetic abnormality. More recently, a MYD88 mutation has
been noted in the majority of patients. Treatment in symptom-
atic patients often includes alkylator such as cyclophosphamide
or bendamustine, fludarabine-​based regimens or bortezomib,
frequently combined with rituximab, or rituximab as a single
agent. Ibrutinib has shown significant response, particularly in
patients harbouring a MYD88 mutation and wild-​type CXCR4.
Patients with symptoms from a very high IgM level will also re-
quire plasmapheresis.
Mature T-​cell lymphomas
T-​cell lymphomas are much less common than their B-​cell coun-
terparts, accounting for about 10% of NHL in Western countries.
Mature T-​cell lymphomas are frequently called ‘peripheral T-​cell
lymphoma’. This refers not to the site of origin of the disease, but
to the mature T-​cell immunophenotype. Pathologists have been
less accurate in diagnosing T-​cell lymphomas than B-​cell lymph-
omas, which in part might relate to the absence of a characteristic
immunophenotype for most diseases, and only a few subtypes
having consistent genetic abnormalities. For T-​cell lymphomas, but
not for NK cell lymphomas, demonstration of rearrangements of the
T-​cell receptor gene will sometimes help solve difficult diagnostic
dilemmas.
The differential diagnosis of peripheral T-​cell lymphomas
includes diffuse large B-​cell lymphoma and T-​cell hyperpla-
sias, such as are seen in viral infections and drug reactions. The
section 22  Haematological disorders
5300
characteristics of the mature T-​cell lymphomas are presented in
Table 22.4.4.9.
Nodal T-​cell lymphomas
The nodal T-​cell lymphomas recognized in the WHO classification in-
clude peripheral T-​cell lymphoma unspecified, angioimmunoblastic
T-​cell lymphoma, and anaplastic large cell lymphoma. Patients with
peripheral T-​cell lymphoma unspecified represent a heterogeneous
group of NHL and are the largest subgroup of peripheral T-​cell
lymphomas. These tumours are generally CD3 and CD4 positive, al-
though a few will be CD8 positive. Although cytogenetic abnormal-
ities are frequent, there is no consistent abnormality. Most patients
have widespread disease and systemic symptoms are frequent.
Angioimmunoblastic T-​cell lymphoma is the second most
common subtype. These patients typically present with widespread
disease, systemic symptoms, and frequently skin rashes, and features
of immune dysregulation such as haemolytic anaemia and poly-
clonal hypergammaglobulinaemia. This type of peripheral T-​cell
lymphoma seems somewhat more frequent in northern Europe.
Anaplastic large cell lymphoma has a characteristic histological ap-
pearance and consistently overexpresses the CD30 antigen. Many of the
tumours have the t(2;5) translocation and overproduction of the ALK
protein (in c.50% of cases), and are responsive to ALK inhibitors such as
crizotinib. Patients whose tumours are ALK positive are younger, pre-
dominantly male, and might have a better outlook than those whose
tumours are ALK negative. Some patients have lymphoma with the
histological appearance of anaplastic large cell lymphoma, but with the
disease confined to the skin: these are one part of an entity that has been
referred to as CD30-​positive cutaneous lymphoproliferative disorders.
The treatment of patients with nodal peripheral T-​cell lymph-
omas has been largely unsatisfactory. Localized disease is unusual
and patients with disseminated disease are treated with combination
chemotherapy regimens such as CHOP with or without etoposide.
There is no consistently effective approach for patients with periph-
eral T-​cell lymphoma unspecified and angioimmunoblastic T-​cell
lymphomas. Novel drugs such as pralatrexate (folate analogue) and
histone deacetylase inhibitors (e.g. romidepsin) have been devel-
oped and are effective for a subset of patients.
Patients with anaplastic large cell lymphoma are more likely to re-
spond to anthracycline-​containing combination chemotherapy regi-
mens. Young patients whose tumours overexpress the ALK protein
are cured in more than 50% of cases. Patients with anaplastic large
cell lymphoma that is either ALK positive or ALK negative usually
respond to the anti-​CD30 antibody conjugate brentuximab vedotin.
Patients with cutaneous anaplastic large cell lymphoma have a particu-
larly indolent course and often do not need to be treated aggressively.
Extranodal peripheral T-​cell lymphomas
Mycosis fungoides or cutaneous T-​cell lymphoma is an indolent
lymphoma of mature T cells predominantly involving the skin.
Patients who present with circulating, atypical cells (Sézary cells)
and erythroderma are said to have Sézary syndrome. The median
age is approximately 50 years and the disease is more common in
males and black individuals.
Mycosis fungoides often presents with eczematous or dermatitic
skin lesions for many years, and patients will often have several
skin biopsies before the diagnosis is confirmed. Lymphoma first
manifests itself as superficial lesions in the skin that thicken and
eventually ulcerate. In the late stages of the illness, lymphoma can
metastasize to lymph nodes and visceral organs.
Treatments utilized for mycosis fungoides include topical cor-
ticosteroids, topical nitrogen mustard, phototherapy, psoralen
ultraviolet A-​range (PUVA) therapy, electron-​beam radiation,
interferon, vorinostat, bexarotene, and systemic cytotoxic therapy
among others. Recently, brentuximab vedotin has shown improved
response rates compared with methotrexate of bexarotene. Some
patients with localized mycosis fungoides can be cured with radio-
therapy, but most will progress. In the end stages of this disease,
management of ulcerating cutaneous lesions may be difficult. The
median survival from diagnosis averages over 10 years.
Table 22.4.4.9  Clinical characteristics of T-​cell lymphomas
PTCL—​
unspecified
ATL
Angioimmunoblastic
ALCL
ALK+
ALCL ALK−
Nasal
NK/​T
Subcutaneous
panniculitis-​like
Hepatosplenic
Enteropathy
associated
Median age (years)
60
62
65
33
58
49
33
34
61
Percentage male
66
55
56
63
61
65
75
68
53
Stage (%):
  I
13
  5
  1
12
19
48
17
5
10
  II
17
  5
10
23
22
20
  0
0
21
  III
26
18
41
29
21
  4
  0
0
  5
  IV
43
73
48
36
38
33
83
95
64
B-​symptoms (%)
35
31
69
60
57
46
67
84
63
Elevated LDH (%)
47
40
62
36
44
49
75
84
32
IPI score (%):
  0/​1
27
19
13
49
41
44
42
5
25
  2/​3
57
64
58
37
44
50
42
47
62
  4/​5
15
16
28
14
14
  6
17
47
12
% 5-​year survival
31
14
32
70
49
32
64
7
20
ALCL, anaplastic large cell lymphoma; ALK+/​−, ALK protein positive/​negative; ATL, angioimmunoblastic T-​cell lymphoma; IPI, International Prognostic Index; PTCL, peripheral T-​cell
lymphoma; nasal NK/​T, angiocentric nasal NK cell lymphoma.
22.4.4  Non-Hodgkin lymphoma
5301
A number of distinctive but unusual clinical syndromes are grouped
in the category of extranodal peripheral T-​cell lymphomas. These
­include angiocentric nasal NK cell lymphoma, which often presents
with necrotic nasal or facial lesions. These patients are most often seen
in South-​East Asia and certain parts of Latin America. Radiotherapy is
often an important part of the management of this disease.
Enteropathy-​type T-​cell lymphoma is a rare disorder that some-
times occurs in patients with gluten enteropathy. Patients are
frequently malnourished, sometimes present with intestinal perfor-
ation, and have a particularly poor outlook.
Hepatosplenic γδ T-​cell lymphoma presents as a systemic illness
with sinusoidal infiltration of the liver, spleen, and bone marrow
by malignant T-​cells. These patients often present a diagnostic di-
lemma, and treatment results have been poor.
Subcutaneous panniculitis-​like T-​cell lymphoma is a rare disorder
that presents with subcutaneous nodules and is frequently confused
with panniculitis. This is true even on biopsy if the slides are not re-
viewed by an expert haematopathologist. This frequently has a more
indolent course than some other types of extranodal peripheral T-​
cell lymphoma.
Adult T-​cell lymphoma/​leukaemia
The two major manifestations of infection by HTLV-​1 are tropical spastic
paraparesis and adult T-​cell lymphoma/​leukaemia. Patients can be in-
fected with HTLV-​1 through sexual transmission, blood transmission,
or through breast milk. The risk of developing lymphoma in a patient
infected with HTLV-​1 is between 1 and 7% according to various studies.
The latency between infection and the development of lymphoma aver-
ages approximately 20 years. The diagnosis is established by review of an
adequate biopsy by an expert haematopathologist, demonstration of a
T-​cell immunophenotype, and demonstration of antibodies to HTLV-​
Most patients will have circulating tumour cells with a characteristic
pleomorphic histology (flower-​like or clover leaf cells).
Adult T-​cell lymphoma/​leukaemia is most frequently seen in
the southern islands of Japan and in the Caribbean. Most patients
seen in Europe and North America are immigrants from those re-
gions. Blood transfusion provides a possible source for infection, but
screening for HTLV-​1 has reduced the risk.
The clinical characteristics of patients with adult T-​cell lymphoma/​
leukaemia vary considerably. Some patients present with an indolent
disease manifested by lymphadenopathy and skin lesions and survive
for extended times without specific therapy. Others present with pro-
gressive lymphadenopathy, hepatosplenomegaly, skin infiltration,
hypercalcaemia, lytic bone lesions, and elevated LDH levels. Although
patients sometimes respond to combination chemotherapy regi-
mens, complete remissions are unusual and survival is poor. Newer
therapies include zidovudine and interferon, and mogamulizumab
(a humanized monoclonal antibody against chemokine receptor 4).
Lymphoma-​like disorders
Lymphadenopathy caused by infectious mononucleosis, drug re-
actions to diphenylhydantoin or carbamazepine, autoimmune
disorders such as rheumatoid arthritis and systemic lupus
erythematosus, and bacterial infections, such as cat-​scratch disease,
can all be confused on biopsy with lymphoma.
Castleman disease is a specific condition that can present with
localized or disseminated lymphadenopathy and systemic symp-
toms. The disease appears to be related to an overproduction of
interleukin-​6 (IL-​6) and is frequently associated with infection by
HHV-​8. The disseminated form of Castleman disease is frequently ac-
companied by anaemia and polyclonal hypergammaglobulinaemia.
Patients with localized disease can frequently be treated with local
therapy, while systemic disease sometimes responds to systemic
glucocorticoids, combination chemotherapy regimens, autologous
or allogeneic haematopoietic stem cell transplantation, rituximab,
and siltuximab (antibody against IL-​6).
Sinus histiocytosis with massive lymphadenopathy (Rosai–​
Dorfman disease) typically presents with bulky lymphadenopathy
in children or young adults. The disease is usually nonprogressive
and self-​limited.
Lymphomatoid papulosis is a cutaneous lymphoproliferative
disorder that can be confused with T-​cell lymphoma in the skin.
It is one of the CD30-​positive cutaneous lymphoproliferative dis-
orders. The cells in lymphomatoid papulosis stain for CD30 and
have a monoclonal T-​cell receptor gene rearrangement. The con-
dition is characterized by waxing and waning skin lesions that
usually heal leaving small scars. Although these patients have
an increased risk of developing lymphoma, aggressive therapy is
inappropriate.
FURTHER READING
Ardeshna K, et al. (2014). Rituximab versus a watch-and-wait ap-
proach in patients with advanced-stage, asymptomatic, non-bulky
follicular lymphoma: an open-label randomised phase 3 trial.
Lancet Oncol, 15(4), 424–35.
Armitage JO (2007). How I  treat patients with diffuse large B-​cell
lymphoma. Blood, 110, 29–​36.
Armitage JO (2015). The aggressive peripheral T-​cell lymph-
omas: 2015. Am J Hematol, 90, 665–​73.
Barrington SF, et al. (2014). Role of imaging in the staging and re-
sponse assessment of lymphoma:  consensus of the International
Conference on Malignant Lymphomas Imaging Working Group. J
Clin Oncol, 32, 3048–​58.
Cheson BD, et  al. (2014). Recommendations for initial evaluation,
staging, and response assessment of Hodgkin and non-​Hodgkin
lymphoma: the Lugano classification. J Clin Oncol, 32, 3059–​68.
Dunleavy K, et al. (2013). Dose-​adjusted EPOCH-​rituximab therapy in
primary mediastinal B-​cell lymphoma. N Engl J Med, 368, 1408–​16.
Dunleavy K, et al. (2013). Low-​intensity therapy in adults with Burkitt’s
lymphoma. N Engl J Med, 369, 1915–​25.
Pfreundschuh M, et  al. (2006). CHOP-​like chemotherapy plus
rituximab versus CHOP-​like chemotherapy alone in young patients
with good-​prognosis diffuse large B-​cell lymphoma: a randomized
controlled trial by the MabThera International Trial (MInT) Group.
Lancet Oncol, 7, 379–​91.
Rosenwald A, et al. (2003). Molecular diagnosis of primary medias-
tinal B cell lymphoma identifies a clinically favorable subgroup of
diffuse large B cell lymphoma related to Hodgkin lymphoma. J Exp
Med, 198, 851–​62.
Salles G, et al. (2011). Rituximab maintenance for 2 years in patients
with high tumour burden follicular lymphoma responding to
rituximab plus chemotherapy (PRIMA): a phase 3, randomised con-
trolled trial. Lancet, 377, 42–​51.
Swerdlow SH, et al. (2016). The 2016 revision of the World Health
Organization classification of lymphoid neoplasms. Blood, 127,
2375–​90.

22.4.5 Chronic lymphocytic leukaemia 5302 Clive S.
22.4.5 Chronic lymphocytic leukaemia 5302 Clive S. Zent and Aaron Polliack
section 22  Haematological disorders
5302
The Non-​Hodgkin’s Lymphoma Classification Project (1997). A
clinical evaluation of the International Lymphoma Study Group
Classification of Non-​Hodgkin’s Lymphoma. Blood, 89, 3909–​18.
Vose JM, Armitage JO, Weisenburger D (2008). International per-
ipheral T-​cell and natural killer/​T-​cell lymphoma study: pathology
findings and clinical outcomes. J ClinOncol, 26, 4124–​30.
22.4.5  Chronic lymphocytic
leukaemia
Clive S. Zent and Aaron Polliack
ESSENTIALS
Chronic lymphocytic leukaemia (CLL)/​small lymphocytic lymphoma
is the most prevalent lymphoid neoplasm in Europe and North
America. The ‘cell of origin’ is a mature B lymphocyte with a re-
arranged immunoglobulin gene. CLL cells express modest amounts
of surface immunoglobulin, and are characterized by defective
apoptosis. The cause of CLL is unknown.
Clinical features
Most patients show no specific clinical features of disease and are
diagnosed during evaluation of an incidental finding of peripheral
blood lymphocytosis, lymphadenopathy, or splenomegaly. A small
percentage of patients (<10%) present with symptomatic disease re-
sulting from (1) tissue accumulation of lymphocytes such as disfig-
uring lymphadenopathy, splenomegaly with abdominal discomfort,
profound fatigue, drenching night sweats, weight loss, and fever; or
(2) manifestations of marrow failure with cytopenias including an-
aemia and thrombocytopenia. All CLL patients have an increased
risk of infection, autoimmune cytopenias, and second haemato-
logical (e.g. diffuse large B-​cell lymphoma) and nonhaematological
malignancies.
Diagnosis and clinical staging
Diagnosis is usually made by analysis of the immunophenotype of
the monoclonal circulating cells in the peripheral blood. In patients
with the small lymphocytic variant of CLL without a detectable cir-
culating monoclonal B-​cell population, the diagnosis is made using
tissue from the bone marrow, lymph nodes, or spleen. Current
diagnostic criteria have an arbitrary requirement for (1) a mono-
clonal B-​lymphocyte count greater than 5 × 109/​litre, or (2) clinic-
ally detectable lymphadenopathy of at least 1 cm in diameter, or
(3) organomegaly, or (4) over 30% bone marrow involvement by
CLL cells. Staging is based on both clinical examination and blood
count evaluation.
Treatment and prognosis
Treatment—​there is no standard curative therapy and patients
should not be treated until they have progressive and symptom-
atic disease or develop anaemia or thrombocytopenia due to bone
marrow failure. If a decision is made to treat, then the best initial
treatment should be given, based on evaluation of the patient’s dis-
ease characteristics with specific attention to the integrity of TP53
(coding for p53) and patient fitness. Treatment options include
chemoimmunotherapy combining purine analogues, alkylating
agents, and anti-​CD20 monoclonal antibodies, targeted small mol-
ecules inhibiting tyrosine kinases in the B-​cell receptor pathway
and BCL-​2, and entry into clinical trials of experimental therapies
including immune modulating drugs.
Prognosis—​this is highly variable and depends on the clinical stage
of disease, intrinsic biological characteristics of the CLL cells, the gen-
eral health and performance status of the patient, and type of treat-
ment. The median survival at diagnosis is in excess of 10 years but
varies considerably depending on CLL biology and patient fitness.
Many patients with CLL who are appropriately managed can expect
good quality and duration of survival.
Introduction
The mature lymphocytic leukaemias were historically defined by
light microscopic cell morphology and this category included a
wide range of disorders derived from different classes of lympho-
cytes. Subsequent major advances and improvements in the under-
standing of lymphocyte biology resulted in the development of
better diagnostic methodologies. These methods are now routinely
used for more accurate diagnosis of CLL, in clinical management
decisions, and during the period of subsequent patient follow-​up.
Individual disorders can now be more accurately defined and diag-
nosed, at both cellular and molecular levels. The majority of patients
with mature lymphocytic leukaemias in Europe and North America
have chronic lymphocytic leukaemia (CLL)/​small lymphocytic
lymphoma. The other less frequently encountered entities, which
always need to be considered in the differential diagnosis of CLL,
include the leukaemic phase of B-​, T-​, and natural killer (NK) cell
lymphomas, prolymphocytic leukaemias, and the nonclassified
chronic lymphoproliferative disorders. This chapter concentrates
on CLL, with only limited reference to these rarer malignancies of
mature lymphocytes. CLL, the most prevalent of these lymphoid
neoplasms in Europe and North America, is a distinct B-​cell ma-
lignancy. This lymphoid malignancy has a protean clinical presen-
tation, marked variation in the rate of disease progression, and a
rapidly increasing number of effective treatment options. The first
and major challenge to the practitioner who encounters a case of this
nature is to make an accurate diagnosis, recognize the significance
of prognostic factors, as well as the indications for treatment, and to
be aware of how to manage the many potential complications of the
disease. Although CLL is incurable with standard therapy and will
eventually cause major morbidity and possibly mortality in most pa-
tients, good medical care can improve the quality of life and prolong
longevity for most patients.
Historical perspective
The mature lymphocytic leukaemias were first recognized in the
latter half of the 19th century. The subsequent recognition of the piv-
otal role of lymphocytes in the immune system led to the discovery
22.4.5  Chronic lymphocytic leukaemia
5303
and recognition of the different T, B, and NK subsets. This infor-
mation has now been combined with a better understanding of
lymphocyte biology, in order to develop newer more appropriate
classifications of lymphoid malignancies. The current World Health
Organization (WHO) classification is based on the presumptive
normal counterpart of the neoplastic cells and in this scheme, CLL
is considered to be a distinct malignancy of a mature B lymphocyte.
The clinical presentation of CLL has changed dramatically during
the past few decades as the widespread use of automated cell ana-
lysers, providing rapid and accurate blood lymphocyte counts, in-
creased the incidental finding of lymphocytosis. In patients with
persistent lymphocytosis, these cells can be characterized by flow
cytometers in clinical pathology laboratories. This is a sensitive
and specific method of diagnosing CLL and is now used routinely.
Among populations with access to these technologies, most patients
with CLL are accordingly diagnosed at an earlier stage with asymp-
tomatic disease while considerably fewer cases present with symp-
tomatic disease.
CLL and small lymphocytic lymphoma, previously considered
to be different diseases, were subsequently found to have the same
immunophenotype and pathophysiology and are now considered to
be variants of the same disease in the WHO classification. Clinical
presentation of the disease with predominance of adenopathy as op-
posed to high circulating numbers of leukaemic cells is likely to re-
flect differences in CLL cell trafficking and does not appear to have
major biological or clinical significance.
In the last three decades, there has been a marked improvement
in the availability of more treatment options for patients with pro-
gressive CLL. Therapy with single-​agent alkylators was superseded
by chemoimmunotherapy combining purine analogues, alkyl-
ating agents, and lymphocyte-​targeting monoclonal antibodies.
Subsequent developments have led to therapies using targeted small
molecule inhibitors of the B-​cell receptor pathway and BCL-​2 that
are highly effective in the management of high-​risk and relapsed/​
refractory CLL. These therapies have resulted in better and more
durable responses to therapy, associated with longer disease-​free
periods and increased overall survival together with improved
quality of life for these patients.
Aetiology, genetics, pathogenesis, and pathology
The aetiology of CLL remains essentially unknown. CLL is familial
in about 5 to 10% of patients who have first-​degree relatives with
CLL or other B-​cell lymphoproliferative disorders. Additional evi-
dence for a genetic predisposition for CLL is the marked ethnic
variation in the incidence of the disease, which remains relatively
unchanged after large population migrations. The highest incidence
rates of CLL are in patients of European descent, with a substantially
lower risk in people of South East Asian ancestry. The specific gen-
etic defects in susceptible patients have not yet been clearly defined.
The role of environmental factors in the aetiology of CLL is also
poorly understood. Epidemiological studies have raised concerns
about the increased risk in patients with exposure to industrial and
agricultural chemicals but radiation exposure is not an established
risk factor.
The current model of B-​cell lymphoid malignancies assumes
that distinct diseases evolve from malignant transformation of
lymphocytes at a specific stage of maturation. Although the ‘cell of
origin’ of CLL is not yet fully defined, there is a body of increasing
data showing that the physiological counterpart of CLL cells is the
mature CD5+ B cells (B1 compartment) which comprises only a
small fraction of normal B cells in adults. This cell is a mature B
lymphocyte that has rearranged its immunoglobulin gene but ex-
presses only small amounts of detectable surface immunoglobulin.
These CD5+ B-​CLL cells are also capable of undergoing somatic
hypermutation in response to antigen stimulation.
CLL cells have a low proliferation rate (0.1–​1% per day) and ac-
cumulate largely because of defective apoptosis which is a major
mechanism in this disease. However, the fundamental defect in
apoptosis is as yet undefined. CLL cells have disrupted BCL2 gene
family expression with higher levels of intracellular antiapoptotic
proteins (especially BCL-​2, MCL1, and XIAP) and lower levels of
proapoptotic proteins, and the mechanisms underlying these cel-
lular changes are now being actively investigated.
CLL cells are characterized by several recurrent genetic defects,
which are likely to be events that occur after the CLL cell becomes
malignant and are generally found in subclones of the neoplastic
cell population. Clonal evolution with development of additional
subclones bearing new mutations is characteristic of CLL. These
genetic lesions include interstitial deletions of chromosome band
17p13 (loss of one allele of TP53 coding for p53), dysfunctional mu-
tations of TP53, interstitial deletions of 11q22 (loss of one allele of
ATM, a critical DNA damage-​sensing protein in the DNA damage
repair pathway), dysfunctional mutations of ATM and SF3B1, and
activating mutations of NOTCH1. However the most frequently re-
current chromosomal defect in CLL detected by the commonly used
interphase fluorescent in situ hybridization (FISH) assay is intersti-
tial deletion of 13q14 resulting in the loss of the microRNA genes
MIR15A and MIR16-​1 which negatively regulate BCL-​2 synthesis.
CLL cells are identified by immunophenotyping of membrane
proteins in a light chain restricted (predominant expression of either
the κ or λ immunoglobulin light chains) monoclonal B-​cell popula-
tion. The characteristic immunophenotype of the CLL clone includes
the expression of CD19 (pan-​B-​cell marker), and co-​expression of
CD5, CD23, CD20 (dim), CD79b (dim), and light chain (dim). CLL
cells grow in proliferation centres in the lymphoid tissue and less
frequently in the bone marrow. Mature cells can accumulate in the
bone marrow, lymph nodes, and spleen, and traffic between these
sites via the blood and lymphatics. CLL cells are often found in vis-
cera, serosal fluids, and cerebrospinal fluid, even in early-​stage CLL,
and can infiltrate any site of inflammation.
Pathological effects of CLL cells can be both direct and indirect.
The most common direct effects of accumulation of CLL cells are
lymphadenopathy, splenomegaly, hepatomegaly, and bone marrow
failure. Lymphadenopathy is an early event in disease progression
followed by splenomegaly and hepatomegaly, and bone marrow in-
volvement resulting in cytopenia is usually a late event. In contrast,
the indirect effects of CLL are less predictable, and their mechan-
isms are not well understood. Patients with CLL have an early-​onset
defect in humoral immunity characterized by decreased produc-
tion of antibodies and a restricted antibody repertoire. This results
in decreased antibody levels and increased rates of infection espe-
cially with encapsulated bacteria. T-​cell function is better preserved
in early-​stage disease despite a skewed T-​cell repertoire. However,
T-​cell function can be considerably compromised with disease
section 22  Haematological disorders
5304
progression and particularly after toxic treatments such as purine
analogues which also affect T cells. CLL patients have an increased
risk (c.5%) of developing autoimmune cytopenias during the course
of their disease. These can present as autoimmune haemolytic an-
aemia (AIHA), immune thrombocytopenia (ITP), pure red blood
cell aplasia (PRCA), or rarely autoimmune granulocytopenia. In
addition, CLL is associated with a marked increase in the risk of
developing second malignancies of the haematopoietic tissue, solid
organs, and skin. The most common second haematological ma-
lignancy is diffuse large B-​cell lymphoma (DLBCL) (Richter syn-
drome), which is often, but not always, clonally related to the CLL.
The most common nonhaematological malignancies are squamous
and basal cell carcinomas and melanoma, which can behave in an
aggressive and rapidly progressive manner. The risk of most other
common cancers is also increased. The reason for this association
of second malignancies is unknown, but defective immune surveil-
lance could be a risk factor.
Epidemiology
CLL is the most prevalent lymphoid malignancy in Europe and
North America, with a lower prevalence in Africa and the lowest
prevalence in the Far East. In most patients with access to modern
medical care, CLL is an incidental diagnosis made during investi-
gation of leucocytosis and lymphocytosis and these patients usually
have early-​stage asymptomatic disease.
CLL is very rare in patients under the age of 30, and the median
age at diagnosis is about 72 years with a 2:1 male to female predom-
inance. In the past, most patients with CLL died of the disease or
its complications. However, the improvements in understanding the
disease complications and availability of more targeted therapies
could change this in the near future.
The improvement in diagnostic methods in recent years has re-
sulted in an increased recognition of small monoclonal B-​cell popula-
tions (< 5 × 109/​litre), which usually have a CLL immunophenotype,
in patients who may have normal lymphocyte counts and no other
evidence of CLL. This monoclonal B-​cell lymphocytosis (MBL) in-
creases in prevalence with age and can be detected in 3 to 5% of
Europeans over the age of 65 years. The natural history of MBL is
not yet completely defined, but retrospective data suggest that only
a very small minority of people with MBL without lymphocytosis or
lymphadenopathy will progress to CLL or other clinically relevant
lymphoid malignancies.
Prevention
There are no known preventive measures to decrease the risk
of CLL.
Clinical features
CLL has a highly variable clinical presentation and course. Most pa-
tients have an incidental diagnosis on investigation of lymphocytosis
without any overt clinical features of disease. These patients should
have as complete an evaluation of the prognostic biological charac-
teristics of their disease as possible. Although no treatment is indi-
cated outside of clinical trials, this information is still very important
for the planning of subsequent care and to allow these individuals to
adjust to their new diagnosis.
The clinical manifestations of CLL can be a direct consequence of
the tumour burden itself or be due to the indirect effects of the CLL
cells. The rate of progression of CLL is highly variable and only a
minority of patients have rapidly progressive disease requiring treat-
ment for symptoms or cytopenia within a few years of diagnosis. In
contrast, about 20 to 30% of patients will never require treatment
for their CLL. Progressive adenopathy can cause disfigurement and
some abdominal distension and discomfort while also occasion-
ally causing obstruction of the ureters or other viscera. Progressive
splenomegaly can cause abdominal discomfort in the left upper
quadrant, abdominal distension, and early satiety due to pressure on
the stomach. Rare splenic infarcts can cause severe abdominal pain.
Anaemia is usually the first manifestation of bone marrow failure
caused by progressive infiltration by CLL cells and often followed
by the development of thrombocytopenia. However, the differential
diagnosis of both anaemia and thrombocytopenia in patients with
cytopenia includes AIHA, ITP, PRCA, and other unrelated causes.
The lymphocyte count can increase to very high levels in patients
with CLL, but complications of extreme lymphocytosis are extremely
rare. In patients with progressive disease, increased tumour burden
can be associated with severe fatigue, drenching night sweats, fever,
and weight loss. These clinical features need to be carefully investi-
gated to ensure that they are due to CLL rather than other medical
conditions.
Immune dysfunction associated with CLL causes both immuno-
suppression and an increased risk of autoimmune cytopenia. Due
to the early suppression of humoral immunity, patients are at high
risk of infections with encapsulated bacteria, which can cause severe
infections. Further deterioration of immune function including T-​
cell-​mediated immunity increases the risk of viral reactivation and
opportunistic infections as the disease progresses or after therapies
that decrease the general immune status.
Autoimmune complications of CLL usually cause cytopenia. The
most common problems are AIHA and PRCA resulting in symp-
tomatic anaemia, and ITP which may cause bleeding. These abnor-
malities need to be carefully distinguished from cytopenias due to
varying degrees of bone marrow failure, which bear a poorer prog-
nosis and often require very different therapy.
Second malignancies are markedly increased in CLL. The most
common second lymphoid malignancy is DLBCL (Richter syn-
drome), which can cause dramatic weight loss, night sweats, fevers,
and rapid increases in the size of lymph nodes.
Differential diagnosis
The differential diagnosis of CLL depends on the disease pres-
entation. Most patients present with sustained lymphocytosis.
The differential diagnosis then includes benign aetiologies as-
sociated with lymphocytosis, which is sometimes atypical mor-
phologically, and other lymphoid malignancies in the leukaemic
phase (Fig. 22.4.5.1). Benign lymphocytosis is usually caused by
22.4.5  Chronic lymphocytic leukaemia
5305
chronic infections (e.g. hepatitis C). The other lymphoid malig-
nancies that most frequently present with lymphocytosis of mature
small lymphocytes are the leukaemic phase of other B-​cell-​derived
lymphoid neoplasms such as mantle cell, marginal zone, fol-
licular, and lymphoplasmacytic lymphoma, hairy cell leukaemia,
prolymphocytic leukaemia, and the unclassified chronic B-​cell
lymphoproliferative disorders. In patients presenting with lymph-
adenopathy and splenomegaly, the differential diagnosis includes
a wide range of benign and malignant causes, and when malignant
peripheral blood lymphocytes are not available for analysis, a bone
marrow or lymph node biopsy, or even splenectomy may be re-
quired to establish the diagnosis.
Clinical investigation
CLL is most easily diagnosed by analysis of the immunophenotype
of the malignant cells from the blood. In rare patients without a
detectable monoclonal B-​cell population in the peripheral blood,
lymphocytes from the bone marrow, lymph nodes, or spleen can
be examined. Staging is based on a clinical examination and blood
count evaluation and does not require imaging studies or a bone
marrow study (Box 22.4.5.1). Bone marrow examination is re-
quired to investigate the cause of cytopenias and is done prior to
initiation of therapy by many physicians to exclude other causes
of cytopenia and to determine tumour burden. Imaging studies
are not required routinely in all patients, and their use should be
limited to investigating specific clinical concerns. However, an
increasing number of treating physicians perform CT scans before
starting specific therapy as a baseline study to facilitate evaluation
of the eventual therapeutic outcome. The only real indication for
the use of positron emission tomography (PET)-​CT in patients
with CLL is clinical concern that a patient could have other con-
comitant malignancies including DLBCL or infection. Evaluation
of established prognostic factors is important and is detailed in the
section on staging CLL.
Criteria for diagnosis
The CLL cell typically coexpresses the B-​cell surface antigen CD19
with CD5 and CD23 and has low levels of expression of surface im-
munoglobulin (and CD79b) and CD20. These characteristics are
used for diagnosis by flow cytometric or immunohistochemical
techniques. Interphase FISH examination of CLL cells with an IGH
probe is very useful for excluding mantle cell lymphoma with its
characteristic t(11;14). The current criteria for diagnosis of CLL re-
quire a B-​cell lymphocytosis greater than 5 × 109/​litre, clinically de-
tectable adenopathy (at least 1 cm in diameter), organomegaly, or
greater than 30% bone marrow involvement by CLL cells.
Staging CLL
The clinical staging systems for CLL are based on readily avail-
able clinical data. The widely used Rai and Binet classifications use
Lymphocytosis
(mature cells)
Polyclonal
Monoclonal
- light chain restriction
- TCR analysis
- NK cell markers
T cell
NK cell
B cell
Leukaemic phase of lymphoma
- mantle cell
- marginal zone
- lymphoplasmacytic
- other
Hairy cell leukaemia
Chronic lymphocytic
leukaemia/small
lymphocytic lymphoma
Chronic B cell lymphoproliferative
disease (not otherwise specified)
Fig. 22.4.5.1  Evaluation of chronic lymphocytosis. TCR, T-​cell receptor.
Box 22.4.5.1  Clinical staging of CLL
Rai classification
0	 Lymphocytosis
I	 Lymphocytosis and lymphadenopathy
II	 Lymphocytosis and palpable liver or spleen enlargement
III	 Lymphocytosis and anaemia (haemoglobin <110 g/​litre)
IV	 Lymphocytosis and thrombocytopenia (platelets <100 × 109/​litre)
Modified Rai classification
Low risk	
Stage 0
Intermediate risk	 Stages I–​II
High risk	
Stages III–​IV
Binet classification
A	 Lymphocytosis and lymphadenopathy in less than three areasa
B	 Lymphocytosis and lymphadenopathy in three or more areasa
C	 Anaemia (<100 g/​litre) and/​or thrombocytopenia (<100 × 109/​litre)
a Areas are cervical, axillary, and inguinal nodes (unilateral or bilateral), liver,
and spleen (n = 5).
section 22  Haematological disorders
5306
clinical examination and the complete blood count to determine
tumour burden (Box 22.4.5.1). These simple methods are very ef-
fective at identifying patients with advanced-​stage disease who have
a poorer prognosis. However, clinical staging using the Rai or Binet
classifications does not provide any information on the risk of dis-
ease progression in the majority of patients who are diagnosed with
early-​ to intermediate-​stage CLL.
Improvements in the diagnosis and management of CLL in the
past few decades have increased the utility of determining prog-
nosis at diagnosis in earlier-​stage disease. This information can be
very useful to health care providers as well as patients in planning
medical care. There has been impressive progress in defining mo-
lecular determinants of risk in CLL patients. These are direct meas-
urements of critical biological parameters in the malignant CLL
cells rather than indirect measures of tumour progression. The best-​
studied novel parameters are immunoglobulin mutation sequence
analysis (somatic hypermutation status of IGHV), specific chromo-
somal defects detected by using interphase FISH, expression of the
intracellular protein ZAP-​70, and surface membrane protein CD38
(Table 22.4.5.1). IGHV mutation (≥2% difference from germline
sequence) has been shown to be associated with a significantly better
survival in multiple analyses. FISH analysis is currently the most
useful available clinical method of chromosome analysis in CLL and
usually includes probes for detection of deletions at chromosome
bands 13q14, 11q22, and 17p13, trisomy 12 (12+), and abnormal-
ities involving 14q32 (IGH locus). Deletion of 17p13 (17p13−) re-
sulting in loss of one allele of TP53 coding for p53 is associated with a
shorter time to initial treatment, response duration and overall sur-
vival. Deletion of 11q22 (11q22−), resulting in loss of one allele of
the ATM gene, is more commonly encountered in younger patients
and associated with more aggressive disease, bulky adenopathy, and
a poorer prognosis. Patients with 12+ or no detected abnormality
have an intermediate prognosis and patients with only deletion of
13q14 (13q14−) generally have the least aggressive disease. Targeted
sequencing approaches have also substantially improved the ability
to evaluate risk of progression and poor prognosis at diagnosis or
later in the course of CLL.
Patients with impaired DNA damage responses caused by ac-
quired mutations have a very high risk of disease progression, poor
response to chemotherapy using DNA-​damaging drugs (purine
analogues and alkylating agents), and a worse prognosis with
chemoimmunotherapy and BCR pathway inhibitor therapies. The
best validated poor prognostic genetic mutations are dysfunctional
mutations of TP53, ATM, and SF3B1 and activating mutations of
NOTCH1. ZAP-​70 is an intracellular signalling molecule expressed
at a high level by T lymphocytes but only in a small minority of
normal B cells. ZAP-​70 expression (≥20% positive cells measured
by flow cytometry on peripheral blood CLL cells) was originally
predicted to be a surrogate marker for unmutated IGHV status and
although this predictive capability of ZAP-​70 measurement was sub-
sequently found to be limited, ZAP-​70 expression is still regarded as
an independent marker of poor prognosis in CLL. Unfortunately, the
assay for ZAP-​70 expression is technically demanding and poorly
reproducible, which limits its routine clinical application. CD38 is
a cell membrane protein of uncertain function expressed by mature
B cells and plasma cells. Expression of CD38 by at least 30% of CLL
cells is an independent predictor of poor prognosis, but the CD38
expression can change during the course of the disease, and there is
still no true consensus on the clinical application of this parameter.
Additional biological markers of prognosis in early-​stage CLL in-
clude serum β2-​microglobulin, soluble CD23, thymidine kinase, and
the percentage of smudge cells seen on the peripheral smear (low
numbers predict for poorer prognosis). The challenge in the future
is to combine a selection of these factors and other markers of prog-
nosis into a practical prognostic formulation that will be easy to use
and accessible to most patients with CLL.
Treatment
Currently, there is no standard curative therapy for CLL. Patients
should not be treated until they have progressive and symptom-
atic disease or develop anaemia or thrombocytopenia due to bone
marrow failure. In this regard, early treatment of all patients has not
been shown to be of benefit and could even be detrimental to some
patients.
Initial treatment
Patients are treated for progressive disease as defined by the
International Workshop for CLL (IWCLL) modification of the
1996 National Cancer Center Working Group criteria published
in 2008. The recommended criteria for treatment of patients for
CLL are symptoms attributable to CLL (severe fatigue, drenching
night sweats, fever, >10% weight loss, discomfort from lymphaden-
opathy or splenomegaly), rapidly progressive disease based on in-
creases in adenopathy and organomegaly, and bone marrow failure
manifesting as anaemia (haemoglobin <110 g/​litre) or thrombo-
cytopenia (platelets <100 × 109/​litre).
Optimal initial therapy for CLL depends on the biology of the dis-
ease (and especially TP53 analysis) and the patient’s comorbidities
Table 22.4.5.1  Molecular prognostic factors and risk of disease progression in early-​stage CLL
Low risk
Intermediate risk
High risk
Very high risk
FISH
Gene sequencing
13q14− as sole abnormality
Nil, 12+
11q22−
NOTCH1 mutation
SF3B1 mutation
ATM mutation
17p13−
TP53 mutation
IGHV mutation
Mutated (≥2%) except for VH3-​21
Unmutated, VH3-​21 mutated
ZAP-​70 (≥20%)
Negative
Positive
CD38 (≥30%)
Negative
Positive
FISH, fluorescent in situ hybridization.
22.4.5  Chronic lymphocytic leukaemia
5307
and functional status (fitness). For fit patients with progressive CLL
without a known defect in p53 function, the current ‘standard of
care’ initial therapy is chemoimmunotherapy combining a purine
analogue and anti-​CD20 monoclonal antibody. The purine ana-
logues (fludarabine, pentostatin, and cladribine) in use since the late
1980s, achieve higher response rates, a longer duration of response,
and better overall survival than alkylating agents such as chloram-
bucil. Multiple randomized phase III studies have shown better out-
comes with the combination of fludarabine and cyclophosphamide
compared to fludarabine alone. Therapy of CLL has been further
improved by the introduction of the therapeutic anti-​CD20 mono-
clonal antibodies such as rituximab. Although rituximab has limited
efficacy as a single agent, the addition of this monoclonal antibody
has been shown, in a randomized phase III trial, to improve both
responses and survival of CLL patients treated with fludarabine and
cyclophosphamide. These chemoimmunotherapy regimens usually
induce a high response rate (>90%) with complete response rates
ranging from 40 to 60% and median durations of response of about
3 to 5 years.
Less fit patients, and especially those who are older with associated
comorbidities, are less likely to tolerate purine analogue-​containing
chemoimmunotherapy. In this population, chemoimmunotherapy
with an alkylating agent (e.g. chlorambucil or bendamustine) and
an anti-​CD20 monoclonal antibody (rituximab, ofatumumab, or
obinutuzumab) is tolerated better and can achieve high response
rates and even improved overall survival. Monotherapy and sup-
portive care are also reasonable options in an even frailer subgroup
of older patients.
Patients with progressive CLL and a defective p53-​mediated DNA
damage repair pathway (17p13−, TP53 mutations) have a poor re-
sponse to purine analogue-​based therapy. In this patient popula-
tion, targeted therapy with tyrosine kinase inhibitors has markedly
improved treatment outcome. The current available agents include
the irreversible Bruton’s tyrosine kinase (BTK) inhibitor ibrutinib
and the phosphatidylinositol-​4,5-​bisphosphate-​3-​kinase catalytic
subunit delta (PI3Kδ) idelalisib. These drugs inhibit B-​cell receptor
pathway signalling, causing CLL cells to stop growing and to migrate
from the lymphoid tissues into the peripheral circulation. Ibrutinib
and idelalisib are highly effective in the management of patients
with very high-​risk CLL (purine analogue refractory, TP53 dysfunc-
tional) and their use as monotherapy (ibrutinib) and in combination
with anti-​CD20 monoclonal antibodies (idelalisib) has become the
standard of care for these patient populations.
Relapsed and refractory disease
Patients with progressive disease after initial primary therapy usu-
ally do not require treatment until they once again fulfil the standard
criteria for initiating treatment. Patients with CLL treated initially
with chemoimmunotherapy who had a response lasting for at
least 2 years and have no evidence of acquisition of TP53 dysfunc-
tion, can be considered for retreatment with the same or similar
chemoimmunotherapy regimen they received as frontline therapy.
However, patients with initial treatment failure, earlier progression,
or clonal evolution with development of TP53 dysfunction, should
be treated with ibrutinib monotherapy, idelalisib, and rituximab, or
be considered for a clinical trial. Ongoing and recently completed
clinical trials have shown promising results for therapy of this popu-
lation of CLL patients with the BCL-​2 inhibitor venetoclax and a
new BTK inhibitor acalabrutinib. Furthermore, there is consider-
able ongoing research on additional novel small molecule inhibitor
therapies in CLL and the repertoire of treatment options is likely
to continue to increase in the near future. However, none of the
currently available therapies is likely to be curative, so the role of
immunotherapy for those patients who respond less well to these
treatments is still under investigation. Potential options include
immune-​modulatory therapy with drugs that disrupt the PD1
pathway and novel methods to re-​establish immune surveillance
against CLL cells, including the use of chimeric antigen receptor T
cells directed against CLL cell antigens such as CD19 (CART-​19).
Transplantation
Allogeneic stem cell transplantation can induce a therapeutic
graft-​versus-​leukaemia effect and is potentially curative in CLL.
Myeloablative allogeneic transplantation is, however, associated
with very high treatment-​related mortality (30–​40%) largely due to
infection, and is thus of limited value in patients with CLL. However,
reduced-​intensity conditioning allogeneic transplantation has a
lower initial morbidity and mortality and can be considered as an
option for younger fit patients with purine analogue refractory dis-
ease or TP53 and p53 pathway dysfunction who have achieved a low
CLL tumour burden on initial therapy. High-​dose chemotherapy
with autologous stem cell support has no proven role in the man-
agement of CLL.
Management of complications of CLL
Autoimmune cytopenia is responsible for about 20% of anaemia and
thrombocytopenia seen in patients with CLL. In those cases without
a large CLL tumour burden, treatment should first be directed at the
autoimmune cytopenia. The initial management of severe anaemia
or thrombocytopenia is usually therapy with corticosteroids but may
also require the use of intravenous immunoglobulin (IVIG). Patients
with ITP may benefit from splenectomy, but those with AIHA are
less likely to improve after this procedure. Many patients with AIHA
and ITP will also respond well and benefit from the use of anti-​CD20
monoclonal antibody therapy; however, because about half of the
cases of PRBCA involve a cellular immunity-​mediated mechanism,
anti-​CD20 monoclonal antibody therapy is less likely to be effective
treatment for this entity. Management of more advanced-​stage CLL
complicated by autoimmune cytopenia requires regimens that are
used in the treatment of both the autoimmune disorder and the
underlying CLL. Effective regimens for this include the combination
of alkylating agents, corticosteroids, and rituximab, and ibrutinib.
Purine analogue-​containing regimens should not be used in patients
with active autoimmune cytopenia, because in some instances they
can themselves induce autoimmune haemolysis.
Infection is the most common direct cause of death in CLL. Even
patients with early-​stage CLL have increased susceptibility to in-
fections with encapsulated bacteria which can progress rapidly and
prove fatal. Patients are also at an increased risk of developing sinus-
itis and other respiratory tract infections and should be educated
about this risk and advised to seek early medical evaluation for all fe-
brile illnesses. With subsequent disease progression and continuing
therapy against CLL, more severe defects in cellular immunity de-
velop, and individual patients become more susceptible to viral and
opportunistic infections. In this regard, patients must again be edu-
cated about the need for early antimicrobial treatment in the event
section 22  Haematological disorders
5308
of infection. Prophylactic treatment of patients with monthly IVIG
does decrease the risk of bacterial infection but has not been shown
to improve overall survival. Furthermore IVIG therapy is expensive
and tedious for the individual concerned; nevertheless, it is used
routinely in most centres for a selected subpopulation of patients
with low immunoglobulin levels and recurrent serious bacterial in-
fections. Use of prophylactic antiviral therapy in patients with recur-
rent herpes zoster and herpes simplex infections can be beneficial
and should be considered in patients receiving therapy with purine
analogues. Vaccination against influenza and pneumococcus is less
effective than in immunocompetent subjects but may still be useful
and is indicated routinely.
Second malignancies
Patients with CLL need to be followed carefully for the develop-
ment of second malignancies. This includes providing the patient
with information about the symptoms of transformation to DLBCL
including being aware of the significance of drenching night sweats,
fever, and involuntary weight loss. DLBCL (Richter’s syndrome) is
most frequently related to clonal evolution of the original CLL cells
and when evident generally has a poor prognosis. In a minority of pa-
tients with CLL (about 20%), DLBCL can occur as a de novo second
malignancy, which may then be more responsive to conventional
therapy used for DLBCL. In this respect, genetic testing to determine
the clonal relationship between coexistent CLL and DLBCL has clin-
ical utility. Patients with CLL also need to be informed about the high
risk of developing skin cancers (squamous and basal cell carcinomas
and melanoma). They need to be educated about skin care including
avoidance of sun damage, and should be carefully observed for the
development of skin cancers, which should be treated aggressively
when detected. An important measure to decrease the risk of other
secondary cancers is the cessation of smoking. Careful routine checks
for malignancy should be advised and can be beneficial.
Quality of life
A diagnosis of incurable CLL is stressful for the patient. This emo-
tional burden is exacerbated by a number of factors including the
lack of effective early intervention, the need for a prolonged ob-
servation period before treatment (active monitoring or watchful
waiting), and the current lack of curative therapy for this disease.
Measures that could be taken for alleviating this problem include
placing an emphasis on the importance of active monitoring for
early detection and management of potential complications, gaining
additional general knowledge on CLL, encouraging simple and
frank discussions on the disease and its standard complications, and
recognition and greater awareness on the part of treating physicians
regarding the validity of the patient’s concerns and fear of the future.
Prognosis
The prognosis of patients with CLL is highly variable and depends
on the clinical stage of disease, biological characteristics of the leu-
kaemic cells, the presence or absence of associated comorbidities,
and the degree of patient fitness. In previous years, patients with
advanced-​stage disease generally had a poor prognosis with a me-
dian survival of approximately 6 years, but recent data suggest that
this has been considerably improved by the introduction of newer
targeted therapies. In contrast, patients with early-​stage CLL have
a wider variation of overall survival with the lowest risk cohort
(mutated IGHV, 13q14 deletion as the sole abnormality on FISH
analysis, negative ZAP-​70 and low CD38 expression, and absence
of mutations of TP53, ATM, NOTCH1, and SF3B1) likely to have
a median survival that is not significantly different to age and sex
matched populations without CLL.
Areas of uncertainty or controversy
Biology
Topics of active research aimed at resolving some uncertain issues
include defining the normal counterpart of the CLL cell, investiga-
tion of the possible causes of CLL, and defining the role of genetic
susceptibility to the disease.
Prognostic markers
Evaluation of the biology of the CLL cell is the key to risk stratifi-
cation, individualized patient management, and more successful
therapy. Recent studies have identified putative driver mutations in
CLL that include previously defined pathways and extend this concept
to additional intracellular pathways. The pathways affected by driver
mutations include DNA damage response (e.g. p53, ATM), NOTCH
signalling, RNA and ribosomal processing (e.g. SF3B1, XPO1),
BCR signalling (e.g. BRAF, IRF4), inflammatory pathways (BIRC3,
MYD88), and chromatin modification (e.g. HIST1H1E). Ongoing
studies are likely to better define these pathways and their associated
defects leading to the development of better prognostic and predictive
markers in CLL and the identification of future therapeutic targets.
Treatment
The current standard of care is still to treat patients with CLL only when
they have progressive disease causing clinical problems. Frontline
treatment for most patients has improved considerably with the intro-
duction of chemoimmunotherapy. The current challenge is to deter-
mine the role of the novel oral targeted nonchemotherapy agents in
the frontline/​primary treatment of progressive CLL in patients without
TP53 dysfunction. Recent data suggest that ibrutinib could be suitable
initial therapy for older (≥65 years old) patients. A number of large
phase III clinical trials have showed a clear benefit for ibrutinib +/−
rituximab over standard immunochemotherapy. Recently the com-
bination of ibrutinib with venetoclax, the novel BCL2 inhibitor, has
shown high remission rates of almost 90% in high risk older patients
with CLL. Longer follow-up will be necessary to understand long term
toxicity, but it is likely that ibrutinib and newer BTK inhibitors will re-
place previous chemotherapy regimens. Treatment of relapsed/​refrac-
tory CLL has improved considerably with the introduction of targeted
therapies and patients are living longer with less adverse effects of treat-
ment. However, CLL still remains an incurable malignancy.
Likely future developments
Biology
Determining the genetic basis of monoclonal B-​cell lymphocytosis
and the factors responsible for the subsequent development of CLL
22.4.5  Chronic lymphocytic leukaemia
5309
could provide useful data in disease prevention. Identification of the
cell of origin of CLL will provide a baseline for determining which
characteristics of CLL are disease specific and provide additional
targets for treatment. Better definition of the effects of CLL on both
innate and adaptive immunity will improve our understanding of
the acquired immunodeficiency associated with CLL and improve
management of infections, autoimmune disease and possible pre-
vention of second malignancies and especially the process of trans-
formation to DLBCL (Richter syndrome).
Treatment
Ongoing use of molecular prognostic and predictive markers will
improve individualized prognosis at diagnosis and approaches to
treatment that should be more effective and less toxic. An improved
understanding of the driver mutations and molecular pathology of
CLL will result in the development of more and improved targeted
therapies. This will no doubt lead to the design of further novel
multidrug regimens that will target the different critical pathways
in CLL cells that may contribute to the eventual elimination of the
malignant CLL clone. The development of chimeric antigen receptor
therapy via manipulation of autologous T cells targeting CD19 cells
is the first step to be taken in developing effective immunotherapy for
CLL. These therapies could prevent recurrence of disease in patients
with a low tumour burden after initial frontline therapy and increase
the chances for long-​term disease-​free survival. Better understanding
of the acquired immune deficiency in CLL patients could also help to
develop interventions to repair the immune system and prevent the
inevitable complications of immune dysfunction including infection,
autoimmune complications, and second malignancies.
FURTHER READING
Binet JL, et  al. (1981). A new prognostic classification of chronic
lymphocytic leukemia derived from a multivariate survival analysis.
Cancer, 48, 198–​205.
Brewer JD, Habermann TM, Shanafelt TD (2014). Lymphoma-​
associated skin cancer: incidence, natural history, and clinical man-
agement. Int J Dermatol, 53, 267–​74.
Burger JA, et al. (2015). Ibrutinib as initial therapy for patients with
chronic lymphocytic leukemia. N Engl J Med, 373, 2425–​37.
Byrd JC, et al. (2015). Three-​year follow-​up of treatment-​naive and
previously treated patients with CLL and SLL receiving single-​agent
ibrutinib. Blood, 125, 2497–​506.
Byrd JC, et al. (2016). Acalabrutinib (ACP-​196) in relapsed chronic
lymphocytic leukemia. N Engl J Med, 374, 323–​32.
Calin GA, et al. (2002). Frequent deletions and down-​regulation of
micro-​ RNA genes miR15 and miR16 at 13q14 in chronic lympho-
cytic leukemia. Proc Natl Acad Sci U S A, 99, 15524–​9.
Damle RN, et al. (1999). Ig V gene mutation status and CD38 expres-
sion as novel prognostic indicators in chronic lymphocytic leu-
kemia. Blood, 94, 1840–​7.
Dohner H, et al. (2000). Genomic aberrations and survival in chronic
lymphocytic leukemia. N Engl J Med, 343, 1910–​16.
Dores GM, et al. (2007). Chronic lymphocytic leukaemia and small
lymphocytic lymphoma: overview of the descriptive epidemiology.
Br J Haematol, 139, 809–​19.
Dreger P, et al. (2014). Managing high-​risk CLL during transition to a
new treatment era: stem cell transplantation or novel agents? Blood,
124, 3841–​9.
Farooqui MZ, et al. (2015). Ibrutinib for previously untreated and re-
lapsed or refractory chronic lymphocytic leukaemia with TP53 ab-
errations: a phase 2, single-​arm trial. Lancet Oncol, 16, 169–​76.
Furman RR, et al. (2014). Idelalisib and rituximab in relapsed chronic
lymphocytic leukemia. N Engl J Med, 370, 997–​1007.
Goede V, et al. (2014). Obinutuzumab plus chlorambucil in patients
with CLL and coexisting conditions. N Engl J Med, 370, 1101–​10.
Hallek M, et al. (2008). Guidelines for the diagnosis and treatment
of chronic lymphocytic leukemia: a report from the International
Workshop on Chronic Lymphocytic Leukemia (IWCLL) updating
the National Cancer Institute-​Working Group (NCI-​WG) 1996
guidelines. Blood, 111, 5446–​56.
Hallek M, et al. (2010). Addition of rituximab to fludarabine and cyclo-
phosphamide in patients with chronic lymphocytic leukaemia:  a
randomised, open-​label, phase 3 trial. Lancet, 376, 1164–​74.
Hamblin T, et al. (1999). Unmutated Ig V(H) genes are associated with
a more aggressive form of chronic lymphocytic leukemia. Blood, 94,
1848–​54.
Herishanu Y, et al. (2015). Efficacy and safety of front-​line therapy
with fludarabine-​cyclophosphamide-​rituximab regimen for chronic
lymphocytic leukemia outside clinical trials: the Israeli CLL Study
Group experience. Haematologica, 100, 662–​9.
Hillmen P, et al. (2014). Rituximab plus chlorambucil as first-​line treat-
ment for chronic lymphocytic leukemia: final analysis of an open-​
label phase II study. J Clin Oncol, 32, 1236–​41.
Hillmen P, et al. (2015). Chlorambucil plus ofatumumab versus chlor-
ambucil alone in previously untreated patients with chronic lympho-
cytic leukaemia (COMPLEMENT 1):  a randomised, multicentre,
open-​label phase 3 trial. Lancet, 385, 1873–​83.
Jain N, et al. (2019). Ibrutinib and Venetoclax for First-Line Treatment
of CLL. N Engl J Med, 380(22), 2095–103.
Keating MJ, et  al. (2005). Early results of a chemoimmunotherapy
regimen of fludarabine, cyclophosphamide, and rituximab as ini-
tial therapy for chronic lymphocytic leukemia. J Clin Oncol, 23,
4079–​88.
Landau DA, et al. (2015). Mutations driving CLL and their evolution in
progression and relapse. Nature, 526, 525–​30.
Messmer BT, et  al. (2005). In vivo measurements document the
dynamic cellular kinetics of chronic lymphocytic leukemia B cells.
J Clin Invest, 115, 755–​64.
Parikh SA, et  al. (2013). Diffuse large B-​cell lymphoma (Richter
syndrome) in patients with chronic lymphocytic leukaemia (CLL):
a cohort study of newly diagnosed patients. Br J Haematol, 162,
774–​82.
Pepper C, et al. (2012). Defining the prognosis of early stage chronic
lymphocytic leukaemia patients. Br J Haematol, 156, 499–​507.
Porter DL, et al. (2015). Chimeric antigen receptor T cells persist and
induce sustained remissions in relapsed refractory chronic lympho-
cytic leukemia. Sci Transl Med, 7, 303ra139.
Puente XS, et  al. (2011). Whole-​genome sequencing identifies re-
current mutations in chronic lymphocytic leukaemia. Nature,
475, 101–​5.
Rai KR, et al. (1975). Clinical staging of chronic lymphocytic leukemia.
Blood, 46, 219–​34.
Roberts AW, et al. (2016). Targeting BCL2 with venetoclax in relapsed
chronic lymphocytic leukemia. N Engl J Med, 374, 311–​22.
Rossi D, et al. (2013). Integrated mutational and cytogenetic analysis
identifies new prognostic subgroups in chronic lymphocytic leu-
kemia. Blood, 121, 1403–​12.
Rossi D, et al. (2014). Clinical impact of small TP53 mutated subclones
in chronic lymphocytic leukemia. Blood, 123, 2139–​47.

22.4.6 Plasma cell myeloma and related monoclonal
22.4.6 Plasma cell myeloma and related monoclonal gammopathies 5310 S. Vincent Rajkumar and Robert A. Kyle
section 22  Haematological disorders
5310
Royle JA, et al. (2011). Second cancer incidence and cancer mortality
among chronic lymphocytic leukaemia patients: a population-​based
study. Br J Cancer, 105, 1076–​81.
Ruchlemer R, Polliack A (2013). Geography, ethnicity and ‘roots’ in
chronic lymphocytic leukemia. Leuk Lymphoma, 54, 1142–​50.
Seifert M, et al. (2012). Cellular origin and pathophysiology of chronic
lymphocytic leukemia. J Exp Med, 209, 2183–​98.
Stilgenbauer S, Furman RR, Zent CS (2015). Management of chronic
lymphocytic leukemia. Am Soc Clin Oncol Educ Book, 2015,
164–​75.
Tadmor T, et al. (2014). Richter’s transformation to diffuse large B-​cell
lymphoma: a retrospective study reporting clinical data, outcome,
and the benefit of adding rituximab to chemotherapy, from the
Israeli CLL Study Group. Am J Hematol, 89, E218–​22.
Tadmor T, et al. (2014). Hodgkin’s variant of Richter transformation in
chronic lymphocytic leukemia; a retrospective study from the Israeli
CLL study group. Anticancer Res, 34, 785–​90.
Visco C, et al. (2014). Autoimmune cytopenias in chronic lymphocytic
leukemia. Am J Hematol, 89, 1055–​62.
Woyach
JA,
et
al.
(2018).
Ibrutinib
Regimens
versus
Chemoimmunotherapy in Older Patients with Untreated CLL.
N Engl J Med, 379, 2517–28.
Zent CS, Burack WR (2014). Mutations in chronic lymphocytic leu-
kemia and how they affect therapy choice:  focus on NOTCH1,
SF3B1, and TP53. Hematology Am Soc Hematol Educ Program, 2014,
119–​24.
Zent CS, et  al. (2008). The prognostic significance of cytopenia in
chronic lymphocytic leukemia/​small lymphocytic lymphoma
(CLL). Br J Haematol, 141, 615–​21.
22.4.6  Plasma cell myeloma and
related monoclonal gammopathies
S. Vincent Rajkumar and Robert A. Kyle
ESSENTIALS
The monoclonal gammopathies, also referred to as paraproteinaemias,
are a group of neoplastic (or potentially neoplastic) diseases associ-
ated with the proliferation of a single clone of immunoglobulin-​se-
creting plasma cells.
Monoclonal gammopathy of undetermined
significance (MGUS)
MGUS is an asymptomatic clonal plasma cell disorder characterized
by a serum monoclonal (M)-​protein level less than 30 g/​litre, less
than 10% of monoclonal bone marrow plasma cells, and no evi-
dence of hypercalcaemia, renal insufficiency, anaemia, or bone le-
sions (CRAB) related to the plasma cell proliferative process, and no
evidence of any other myeloma-​defining events (MDEs).
Epidemiology and prognosis—​MGUS is found in 3% of people aged
over 50 years and in 5% of those over 70 years. The risk of progres-
sion to plasma cell myeloma or a related disorder is about 1% per
year, with factors increasing the risk being higher concentration of
the monoclonal protein, some particular types of monoclonal pro-
tein (IgA and IgM > IgG), or an abnormal serum free light-​chain ratio.
Management—​observation is the standard of care. The mono-
clonal protein in the serum and urine is monitored, together with
re-​evaluation of clinical and laboratory tests, to determine whether
myeloma or a related disorder has developed.
Plasma cell myeloma
Plasma cell myeloma (commonly referred to as multiple myeloma)
is a clonal plasma cell malignancy that accounts for about 10% of
haematological cancers. The cause is unknown. Fluorescence in situ
hybridization of bone marrow plasma cells reveals specific primary
translocations or trisomies in more than 90% of patients. The pres-
ence of del 17p, t(4;14), t(14;16), and t(14;20) occur in 20 to 25% of
patients, and indicate higher-​risk disease.
Clinical features—​myeloma is defined by the presence of 10% or
more clonal plasma cells in the bone marrow (or a biopsy-​proven
plasmacytoma) plus any one or more MDEs: CRAB features attrib-
utable to the plasma cell disorder, 60% or more clonal plasma cells
in the bone marrow, serum free light-​chain ratio of at least 100 (pro-
vided involved free light-​chain level is ≥100 mg/​litre), and/​or more
than one focal lesion on magnetic resonance imaging. The most
common symptoms are weakness, fatigue, and bone pain.
Investigations—​an M-​protein (paraprotein) is found in the serum
or urine at diagnosis in 97% of cases. The bone marrow usually con-
tains more than 10% clonal plasma cells. Monoclonal plasma cells
in myeloma and related monoclonal gammopathies are light-​chain
restricted to either κ or λ expression in their cytoplasm. This mono-
typic pattern can be identified on flow cytometry and is critical for
differentiating monoclonal from reactive (polyclonal) plasmacytosis.
Conventional radiographs show lytic lesions, osteoporosis, or frac-
tures in almost 80% of patients at diagnosis.
Treatment and prognosis—​first-​line options, aside from supportive
care, are (1) initial therapy with a triplet regimen such as bortezomib,
lenalidomide, and dexamethasone (VRD); (2) autologous stem cell
transplantation (ASCT) in eligible patients; and (3) consideration of
maintenance therapy with lenalidomide. With this approach, pa-
tients can stay in remission for approximately 3 to 4 years. At relapse,
therapy includes newer agents such as carfilzomib, pomalidomide,
ixazomib, daratumumab, and elotuzumab, administered usually in
two-​ or three-​drug combinations. The choice of therapy at relapse
is usually dictated by aggressiveness of the relapse, and drug avail-
ability. The median survival of myeloma is about 6 to 7 years.
Waldenström’s macroglobulinaemia
Waldenström’s macroglobulinaemia (WM) is characterized by the
presence of an IgM M-​protein, 10% or more lymphoplasmacytic
infiltration of the bone marrow, and symptoms such as anaemia,
lymphadenopathy, and hyperviscosity. Rituximab, a monoclonal
antibody directed against CD20, is used as initial therapy in conjunc-
tion with other active drugs. Ibrutinib is a new agent that is highly
active against WM. The median survival is longer than 5 years.
Immunoglobulin light-​chain amyloidosis
Immunoglobulin light-​chain (AL) amyloidosis (formerly referred to
as primary amyloidosis) is a clonal plasma cell disorder character-
ized by tissue deposition of fibrils consisting of monoclonal κ or λ
light chains. Weakness, fatigue, and weight loss are the commonest
22.4.6  Plasma cell myeloma and related monoclonal gammopathies
5311
initial symptoms. Characteristic findings include periorbital purpura
(15% of cases), macroglossia (10%), nephrotic syndrome/​renal failure
(25%), and congestive heart failure (20%), but virtually any organ
system can be affected.
Standard treatment is with bortezomib, cyclophosphamide, dexa-
methasone (VCD), and ASCT in selected patients. Prognosis varies
greatly depending on the presence and extent of organ involvement.
Other conditions
Other less common monoclonal gammopathies include smoul-
dering multiple myeloma; plasma cell leukaemia; POEMS syndrome—​
characterized by polyneuropathy, organomegaly, endocrinopathy,
M-​protein, and skin changes; solitary plasmacytoma; and heavy-​
chain diseases.
Introduction
The monoclonal gammopathies, also referred to as paraprotein­
aemias, are a group of clonal plasma cell disorders associated with
the proliferation of a single clone of immunoglobulin-​secreting
plasma cells. They include monoclonal gammopathy of undeter-
mined significance (MGUS); plasma cell myeloma (commonly
referred to as multiple myeloma); smouldering myeloma (SM);
Waldenström’s macroglobulinaemia (WM); heavy-​chain diseases;
solitary plasmacytoma of bone, extramedullary plasmacytoma,
plasma-​cell leukaemia, POEMS syndrome; and immunoglobulin
light-​chain (AL) amyloidosis. The disease definition for the major
plasma cell disorders is given in Table 22.4.6.1.
Monoclonal gammopathies are characterized by the secretion of
electrophoretically and immunologically homogeneous monoclonal
(M)-​proteins. Each M-​protein consists of two heavy (H) polypep-
tide chains of the same class and subclass and two light (L) poly-
peptide chains of the same type. The heavy polypeptide chains are
designated by Greek letters: γ in IgG, α in IgA, µ in IgM, δ in IgD, and
ε in IgE. The light-​chain types are κ (kappa) or λ (lambda).
Recognition of M-​proteins
Agarose gel electrophoresis is preferred for the detection of M-​
proteins. Immunofixation is necessary to confirm the presence of
an M-​protein and distinguish the immunoglobulin class and its
light-​chain type.
Serum protein electrophoresis should be performed when mye-
loma, WM, or AL amyloidosis is suspected. An M-​protein is charac-
terized by a narrow peak or spike in the densitometer tracing, or as
a dense, discrete band on agarose gel (Fig. 22.4.6.1). In contrast, an
excess of polyclonal immunoglobulins (having one or more heavy-​
chain types and both κ and λ light chains) produces a broad-​based
peak or broad band. It is important to differentiate an M-​protein
from a polyclonal increase because the former is associated with a
malignant process or a potentially neoplastic condition, whereas a
polyclonal increase in immunoglobulins is associated with a reactive
or inflammatory process. Serum protein immunofixation is the pre-
ferred technique for identifying the heavy-​ and light-​chain type of
the M-​protein, and in detecting small M-​proteins that may not be
apparent on electrophoresis. In myeloma and related disorders, free
(unbound) monoclonal immunoglobulin light chains may be se-
creted in addition to intact immunoglobulin M-​proteins. Less fre-
quently, free monoclonal immunoglobulin light chains are secreted
without any evidence of an intact immunoglobulin or heavy-​chain
component. These free light chains (FLCs) can be missed on serum
electrophoresis and immunofixation. They are detected by the
serum FLC assay or by electrophoresis and immunofixation of an
adequately concentrated 24-​h urine specimen. The serum FLC assay
measures free κ and λ light chains; an abnormal serum FLC ratio of
κ to λ light chains is indicative of a monoclonal process. Diseases
associated with an M-​protein in the serum and/​or urine, as found
recently in the authors’ practice, are shown in Fig. 22.4.6.2.
Monoclonal gammopathy
of undetermined significance
The term ‘monoclonal gammopathy of undetermined significance’
(MGUS) denotes the presence of an M-​protein in persons without
evidence of myeloma, WM, AL, or related disorders. MGUS is
characterized by a serum paraprotein concentration of less than
30 g/​litre, fewer than 10% plasma cells in the bone marrow, and
no evidence of end-​organ damage (hypercalcaemia, renal insuffi-
ciency, anaemia, or bone lesions (‘CRAB’)) related to the plasma
cell proliferative process or other myeloma-​defining events (MDEs)
(Table 22.4.6.1). The prevalence of MGUS is 3% in patients aged
50 years or older, 5% in those over 70 years of age, and is higher in
men than women (Fig. 22.4.6.3). The prevalence of MGUS is twice
as high in black patients compared with white patients.
Clinical course of MGUS
MGUS is detected incidentally during clinical workup of a variety of
symptoms and laboratory abnormalities. In a series of 241 patients
with MGUS seen at Mayo Clinic from 1956 to 1970 and followed for
up to 39 years, 27% developed myeloma, WM, or a related disorder.
The actuarial rate of progression to one of these disorders was 17% at
10 years, 34% at 20 years, and 39% at 25 years, a rate of approximately
1.5% per year. A separate population-​based study of 1384 patients
with MGUS from the 11 counties of southeastern Minnesota con-
firmed the clinical course of MGUS that was described in the original
Mayo Clinic referral population. The M-​protein was of the IgG type
in 70%, IgM in 15%, IgA in 12%, and biclonal in 3%. After 11 009
person-​years (median 15.4 years; range 0–​35 years) of follow-​up, 115
patients (8%) developed multiple myeloma, AL, WM, plasmacytoma,
chronic lymphocytic leukaemia, or related plasma cell disorders. The
rate of progression was 10% at 10 years, 21% at 20 years, and 26% at
25 years, which is approximately 1% per year (Fig. 22.4.6.4).
Differential diagnosis of MGUS from multiple myeloma
and related disorders
MGUS is differentiated from myeloma and related disorders using
the criteria listed in Table 22.4.6.1. Thus the presence or absence
of MDEs is critical to distinguish MGUS from myeloma. MGUS is
differentiated from SM based on the level of the M-​protein and the
extent of bone marrow involvement. No single test will distinguish
section 22  Haematological disorders
5312
the patient with MGUS who remains stable from those who may
eventually develop myeloma or related disorders. Thus most pa-
tients need to be observed indefinitely. The paraprotein level in the
serum and urine should be serially measured, along with periodic
re-​evaluation of clinical and other features to determine whether
myeloma or a similar disorder is present. The cytogenetic changes
detected by fluorescence in situ hybridization (FISH) in MGUS are
similar to those seen in multiple myeloma.
Although the main concern with MGUS is the risk of progression
to a malignancy such as myeloma or WM, there are other disorders
that are associated with an M-​protein that should be considered.
These include membranoproliferative glomerulonephritis, periph-
eral neuropathy, and certain skin diseases, among others.
Risk stratification and management
Patients with MGUS do not require therapy. The serum FLC assay
is of prognostic value in MGUS. Patients with an abnormal FLC
ratio (<0.26 or >1.65) have an approximately threefold higher risk
of progression. Other risk factors for progression of MGUS include
the size of the M-​protein, and the type of M-​protein. Patients with
serum M-​protein levels less than 15 g/​litre, IgG class, and normal
FLC ratio are considered as low-​risk MGUS (2% lifetime risk of
Table 22.4.6.1  International Myeloma Working Group diagnostic criteria for plasma cell myeloma and related plasma cell disorders
Disorder
Disease definition
Non-​IgM monoclonal
gammopathy of
undetermined
significance (MGUS)
All 3 criteria must be met:
Serum monoclonal protein (non-​IgM type) <30 g/​litre
Clonal bone marrow plasma cells <10%a
Absence of end-​organ damage such as hypercalcaemia, renal insufficiency, anaemia, and bone lesions (CRAB) that can be
attributed to the plasma cell proliferative disorder
Smouldering myeloma
Both criteria must be met:
Serum monoclonal protein (IgG or IgA) ≥30 g/​litre, or urinary monoclonal protein ≥500 mg per 24 h and/​or clonal bone marrow
plasma cells 10–​60%
Absence of myeloma defining events or amyloidosis
Plasma cell myeloma
Both criteria must be met:
Clonal bone marrow plasma cells ≥10% or biopsy-​proven bony or extramedullary plasmacytoma
Any one or more of the following myeloma defining events:
Evidence of end organ damage that can be attributed to the underlying plasma cell proliferative disorder, specifically:
Hypercalcaemia: serum calcium >0.25 mmol/​litre (>1 mg/​dl) higher than the upper limit of normal or >2.75 mmol/​litre (>11 mg/​dl)
Renal insufficiency: creatinine clearance <40 mL per minute or serum creatinine >177 μmol/​litre (>2 mg/​dl)
Anaemia: haemoglobin value of >20 g/​litre below the lower limit of normal, or a haemoglobin value <100 g/​litre
Bone lesions: one or more osteolytic lesions on skeletal radiography, computed tomography (CT), or positron emission
tomography-​CT (PET-​CT)
Clonal bone marrow plasma cell percentage ≥60%
Involved: uninvolved serum free light chain (FLC) ratio ≥100 (involved FLC level must be ≥100 mg/​litre)
1 focal lesions on magnetic resonance imaging (MRI) studies (at least 5 mm in size)
IgM monoclonal
gammopathy of
undetermined
significance (IgM MGUS)
All 3 criteria must be met:
Serum IgM monoclonal protein <30 g/​litre
Bone marrow lymphoplasmacytic infiltration <10%
No evidence of anaemia, constitutional symptoms, hyperviscosity, lymphadenopathy, or hepatosplenomegaly that can be
attributed to the underlying lymphoproliferative disorder.
Light-​chain MGUS
All criteria must be met:
Abnormal FLC ratio (<0.26 or >1.65)
Increased level of the appropriate involved light chain (increased kappa FLC in patients with ratio >1.65 and increased lambda FLC
in patients with ratio <0.26)
No immunoglobulin heavy-​chain expression on immunofixation
Absence of end-​organ damage that can be attributed to the plasma cell proliferative disorder
Clonal bone marrow plasma cells <10%
Urinary monoclonal protein <500 mg/​24 h
Solitary plasmacytoma
All 4 criteria must be met:
Biopsy-​proven solitary lesion of bone or soft tissue with evidence of clonal plasma cells
Normal bone marrow with no evidence of clonal plasma cells
Normal skeletal survey and MRI (or CT) of spine and pelvis (except for the primary solitary lesion)
Absence of end-​organ damage such as hypercalcaemia, renal insufficiency, anaemia, or bone lesions (CRAB) that can be attributed
to a lympho-​plasma cell proliferative disorder
Solitary plasmacytoma
with minimal marrow
involvementb
All 4 criteria must be met:
Biopsy-​proven solitary lesion of bone or soft tissue with evidence of clonal plasma cells
Clonal bone marrow plasma cells <10%
Normal skeletal survey and MRI (or CT) of spine and pelvis (except for the primary solitary lesion)
Absence of end-​organ damage such as hypercalcaemia, renal insufficiency, anaemia, or bone lesions (CRAB) that can be attributed
to a lympho-​plasma cell proliferative disorder
a A bone marrow can be deferred in patients with low-​risk MGUS (IgG type, M-​protein level <15 g/​litre, normal free light chain ratio) in whom there are no clinical features concerning
for myeloma.
b Solitary plasmacytoma with 10% or more clonal plasma cells is considered as plasma cell myeloma.
Reprinted from The Lancet, Vol 15, Rajkumar SV et al., International Myeloma Working Group updated criteria for the diagnosis of multiple myeloma, Pages e538–​48, Copyright ©
2014, with permission from Elsevier.
22.4.6  Plasma cell myeloma and related monoclonal gammopathies
5313
progression) (Table 22.4.6.2). In these patients, if the diagnosis of
MGUS is suspected and there is no concern for malignancy, baseline
bone marrow examination and skeletal radiography can be omitted.
They can be reassessed in 6 months, and if stable, further workup
is needed only if symptoms suspicious of myeloma or a related dis-
order develop. Patients with any one or more risk factors (e.g. M-​
protein level ≥15 g/​litre, non-​IgG class, or abnormal FLC ratio) need
a baseline bone marrow examination, and skeletal radiographs (or
similar bone examination). Such patients need to be reassessed in
6 months, and if stable, yearly thereafter.
Biclonal gammopathies
Occasionally, patients are found to have two M-​proteins (biclonal
gammopathies). Most such patients have biclonal gammopathy of
undetermined significance, with the remainder representing plasma
cell myeloma, WM, or another malignant lymphoproliferative dis-
order. Triclonal gammopathies may also occur.
Light-​chain MGUS
Approximately 1% of the general population over the age of 50 has
a light-​chain MGUS characterized by an abnormal serum FLC ratio
without any evidence of intact immunoglobulin heavy-​chain (IgH)
expression (Table 22.4.6.1). These patients are observed and may
remain stable for many years.
Smouldering (asymptomatic) myeloma
SM is an intermediate clinical stage between MGUS and plasma
cell myeloma. SM is characterized by a serum monoclonal protein
level of 30 g/​litre or more, a urine M-​protein concentration of 500
mg/​24 h or more, and/​or a level of bone marrow clonal plasma cells
of 10 to 60%, and no MDEs (Table 22.4.6.1). The subset of SM pa-
tients without an intact immunoglobulin, but with light-​chain only
presentation are considered to have light-​chain SM. In contrast to
MGUS, the risk of progression of SM to plasma cell myeloma or AL
is much higher at almost 10% per year for the first 5 years, approxi-
mately 3% per year for the next 5 years, and then 1 to 2% per year for
the following 10 years. Thus patients with SM need to be monitored
more closely, every 3 to 4 months. Selected patients with SM who are
considered to have high-​risk features (Table 22.4.6.3) may be candi-
dates for clinical trials testing preventive strategies.
Plasma cell myeloma
Plasma cell myeloma (multiple myeloma, myelomatosis) is a plasma
cell malignancy that commonly presents with osteolytic skeletal de-
struction, renal failure, anaemia, recurrent bacterial infections, or
hyperviscosity syndrome.
History
Myeloma was documented by the description of Sarah Newbury
which was followed by that of Thomas Alexander McBean a few
years later. The physician and clinical pathologist, Henry Bence
(a)
(b)
alb
γ
Fig. 22.4.6.1  (a) Monoclonal pattern of serum protein as traced by a
densitometer after electrophoresis on agarose gel; tall, narrow-​based
peak of γ mobility. (b) Monoclonal pattern from electrophoresis of
serum on agarose gel (anode on left); dense, localized band representing
monoclonal protein of γ mobility.
From Kyle RA, Katzmann JA (1997). Immunochemical characterization of
immunoglobulins. In: Rose NR, et al. (eds) Manual of clinical laboratory immunology,
5th edition, pp. 156–​76. ASM Press, Washington, DC. By permission of the American
Society for Microbiology.
Monoclonal gammopathies
Mayo Clinic 1960–2013
n = 48 853
SMM 4% (1908)
Solitary or extramedullary
2% (902)
Macro
3.0% (1322)
Other
4% (2187)
MGUS
57% (27 884)
Lymphoproliferative
3% (1434)
(a)
AL amyloidosis
9% (4593)
Plasma cell myeloma
18% (8619)
Monoclonal gammopathies
Mayo Clinic 2013
n = 1773
Smouldering myeloma 2.5% (45)
Solitary or extramedullary
0.5% (12)
Macro
4.5% (84)
Other
6% (101)
MGUS
56% (992)
Lymphoproliferative
1.5% (30)
(b)
AL amyloidosis
9% (159)
Plasma cell myeloma
20% (350)
Fig. 22.4.6.2  Types of monoclonal gammopathies at the Mayo Clinic (a) 1960–​2013 and (b) 2013. SMM, smouldering multiple myeloma.
section 22  Haematological disorders
5314
Jones described the unusual protein excreted by McBean in the mid
19th century. The term ‘multiple myeloma’ was introduced by J. Von
Rustizky in 1873. It was the case report of Dr Loos, published by
Professor Otto Kahler in 1889, that introduced multiple myeloma
to clinical practice. This has now been replaced by the World Health
Organization’s favoured term, plasma cell myeloma.
The introduction of melphalan in the late 1950s was a major step
forward in the treatment of myeloma. Autologous stem cell trans-
plantation (ASCT) was introduced in the 1980s and refined in
the 1990s. Later, thalidomide, bortezomib, and lenalidomide be-
came important agents for treatment of myeloma. More recently, a
number of new treatments have emerged that greatly prolong the
survival of the disease.
Epidemiology and aetiology
Myeloma accounts for 1% of all malignant diseases and slightly more
than 10% of haematological malignancies in the United States of
America. The annual incidence is 4 to 5 per 100 000/​year. The apparent
increase in the incidence of myeloma during the past few decades is
probably related to the increased availability and use of medical facil-
ities, especially in older persons. The incidence in African Americans
is twice that in white populations, but rates are lower in Asian popu-
lations. The median age at diagnosis is 65 to 70 years. Only 10% of
patients are younger than 50 years, and 2% are younger than 40 years.
The cause of myeloma is unknown, but radiation, benzene and other
organic solvents, herbicides, and insecticides may play a role. The dis-
ease is a rare complication of Gaucher disease (see Chapter 12.8); poly-
clonal gammopathy is frequent in the untreated disease and MGUS
appears to occur at greater frequency than in age-matched control
subjects. Myeloma has been reported in two or more first-​degree rela-
tives and in identical twins, suggesting a genetic element.
Biological aspects
Myeloma evolves from the premalignant stage of MGUS. The pre-
cise mechanisms of disease evolution are unclear, but the two key
steps are establishment of the premalignant clone (MGUS), and the
progression of MGUS to malignancy (myeloma). The development
of primary IgH translocations (40%), trisomies (40%), or a com-
bination of IgH translocations and trisomies (15%) are critical in
the pathogenesis of MGUS. The five most common primary IgH
translocations involve the IgH locus on chromosome 14q32, one
of five recurrent partner chromosome loci: 11q13 (CCND1 (cyclin
D1 gene)), 4p16.3 (FGFR-​3 and MMSET), 6p21 (CCND3 (cyclin
D3 gene)), 16q23 (c-​MAF), and 20q11 (MAFB). The progression of
MGUS to multiple myeloma is associated with secondary cytogen-
etic events, RAS mutations, overproduction of interleukin (IL)-​6,
IL-​1β, macrophage inflammatory protein-​α (MIP-​1α), and tumour
necrosis factor-​α (TNFα).
Clinical manifestations
Bone pain, frequently in the back or chest, is present at diagnosis in almost
two-​thirds of patients. This is secondary to osteolytic bone lesions that
are a prominent feature of most patients with myeloma. Loss of height
from multiple vertebral collapses may occur. Other common symptoms
of multiple myeloma include weakness and fatigue, which are often due
to anaemia. Anaemia is the result of bone marrow infiltration by clonal
plasma cells. Renal failure is present in approximately 20% of patients
at diagnosis. The two major causes of renal failure are light-​chain cast
nephropathy and hypercalcaemia. Light-​chain cast nephropathy occurs
in patients with high levels of circulating FLC (typically >500 mg/​litre)
and is characterized by the presence of dense, waxy, laminated casts in
the distal and collecting tubules. The casts consist mainly of monoclonal
light chains. Dilatation and atrophy of the tubules occur, and the entire
nephron becomes nonfunctional. Hypercalcaemia, present in 15% of
patients initially, is a major and treatable cause of renal insufficiency.
Other causes of renal dysfunction are dehydration, concurrent amyl-
oidosis, and hyperuricaemia. Extramedullary plasmacytomas are un-
common and are usually observed late in the course of the disease as
large, purplish, subcutaneous masses.
Other manifestations of myeloma include increased suscepti-
bility to infections. The incidence of bacterial infection is increased.
Impairment of antibody response, neutropenia, treatment with
8
6
4
2
0
10
50
60
Age (years)
Prevalence (%)
Females
Males
70
80
90
Fig. 22.4.6.3  Prevalence of MGUS according to age. The I-​bars
represent 95% confidence intervals. Years of age greater than 90 have
been collapsed to 90 years of age.
From Kyle RA (2006). Prevalence of monoclonal gammopathy of undetermined
significance. N Engl J Med, 354, 1362–​9. Copyright © 2006 Massachusetts Medical
Society. Reprinted with permission.
Progression
Cumulative probability (%)
Years
n = 1384
Pt at risk (no.)
30
21%
26%
10%
25
20
15
10
5
0
0
1384
867
423
177
56
17
5
10
15
20
25
Fig. 22.4.6.4  Probability of progression among 1384 residents
of southeastern Minnesota in whom monoclonal gammopathy of
undetermined significance (MGUS) was diagnosed, 1960–​1994. The
curve shows the probability of progression of MGUS to plasma cell
myeloma, IgM lymphoma, primary amyloidosis, macroglobulinaemia,
chronic lymphocytic leukaemia, or plasmacytoma (115 patients). The
bars show 95% confidence intervals.
From Kyle RA (2002). Prognosis in monoclonal gammopathy of undetermined
significance. N Engl J Med, 346, 564–​9. Copyright © 2002 Massachusetts Medical
Society. Reprinted with permission.
22.4.6  Plasma cell myeloma and related monoclonal gammopathies
5315
glucocorticoids, and reduction of normal immunoglobulins in-
crease the likelihood of infection. Organomegaly is uncommon;
the liver is palpable in about 5% of patients, and the spleen in 1%.
Radiculopathy is the most frequent neurological complication re-
sulting from bone disease in the spine, and usually involves the
thoracic or lumbosacral areas. Compression of the spinal cord from
extradural myeloma occurs in 5% of patients. Leptomeningeal in-
volvement is uncommon but is being recognized more frequently.
Laboratory findings
If myeloma is suspected, the laboratory tests listed in Box 22.4.6.1
should be performed. Anaemia is initially present in 70% of patients
but eventually is found in almost all. The serum protein electrophor-
etic pattern shows a spike or localized band in 80% of cases; serum
immunofixation increases the sensitivity to 93%. Addition of the
serum FLC assay and/​or 24-​h urine studies will detect an M-​protein
in 97 to 98% of patients with myeloma. Approximately 2% of pa-
tients with myeloma do not secrete any M-​protein (nonsecretory
multiple myeloma). The M-​protein type is IgG in about 50% of pa-
tients, IgA in 20%, FLC only in 15 to 20%, IgD in 2%, and biclonal
in 2%. Hypercalcaemia is initially present in almost 15%; about one-​
fifth have renal failure at diagnosis.
The bone marrow contains 10% or more plasma cells in 97% of
patients; the remaining patients must have evidence of a biopsy-​
proven plasmacytoma elsewhere to meet the criteria for myeloma
(Table 22.4.6.1). Monoclonal plasma cells in myeloma and related
monoclonal gammopathies are light-​chain restricted to either κ or
λ (but not both) expression in their cytoplasm. This monotypic pat-
tern can be identified on flow cytometry and is critical for differen-
tiating monoclonal from reactive (polyclonal) plasmacytosis that
can occur due to connective tissue disorders, metastatic carcinoma,
liver disease, or chronic infections, etc. Bone marrow samples must
be tested using FISH or similar methods for the presence of IgH
translocations, trisomies, deletion 17p, and gain 1q, all of which
help with prognostic assessment.
Radiography
Conventional radiographs show abnormalities consisting of lytic le-
sions, osteoporosis, or fractures in almost 80% of patients at diag-
nosis. The vertebrae, skull, thoracic cage, pelvis, and humeri and
femora are the most commonly involved sites. Bone disease can be
Table 22.4.6.2  Risk of progression of monoclonal gammopathy of undetermined significance to myeloma or related disorders
Risk group
Relative risk of
progression
Cumulative absolute
risk of progression at
20 years (%)a
Cumulative absolute risk of progression
at 20 years accounting for death as a
competing risk (%)b
Low risk: serum M-​protein level <15 g/​litre, IgG subtype,
normal free light chain ratio (0.26-​1.65)
1
  5
  2
Low-​intermediate risk: any 1 factor abnormal
5.4
21
10
High-​intermediate risk: any 2 factors abnormal
10.1
37
18
High risk: all 3 factors abnormal
20.8
58
27
a Estimates in this column represent the risk of progression assuming that patients do not die of other causes during this period.
b Estimates in this column represent the risk of progression calculated by using a model that accounts for the fact that patients can die of unrelated causes during this time.
Ig, immunoglobulin.
Adapted from Rajkumar SV, et al. (2005). Serum free light chain ratio is an independent risk factor for progression in monoclonal gammopathy of undetermined significance (MGUS).
Blood, 106, 812–​817. © The American Society of Hematology.
Table 22.4.6.3  Criteria for high-​risk smouldering myeloma
Patients who meet the International Myeloma Working Group Criteria for
Smouldering Myeloma who have bone marrow clonal plasma cells ≥10%
and any one or more of the following:
Serum M-​protein level ≥30g/​litre
IgA SM
Immunoparesis with reduction of two uninvolved immunoglobulin isotypes
Serum involved/​uninvolved free light chain ratio ≥8 (but less than 100)
Progressive increase in M-​protein level (evolving type of SM)a
Bone marrow clonal plasma cells 50–​60%
Abnormal plasma cell immunophenotype (≥95% of bone marrow plasma
cells are clonal) and reduction of one or more uninvolved immunoglobulin
isotypes
t (4;14) or del 17p or 1q gain
Increased circulating plasma cells
MRI with diffuse abnormalities or 1 focal lesion
PET-​CT with focal lesion with increased uptake without underlying osteolytic
bone destruction
M, monoclonal; MRI, magnetic resonance imaging; PET-​CT, positron emission
tomography computed tomography; SM, smouldering myeloma.
a Increase in serum monoclonal protein by ≥25% on two successive evaluations within a
6-​month period
Reproduced from Rajkumar SV, et al. (2005). Serum free light chain ratio is an
independent risk factor for progression in monoclonal gammopathy of undetermined
significance (MGUS). Blood, 106, 812–​817. © The American Society of Hematology.
Box 22.4.6.1  Baseline tests for patients in whom plasma cell
myeloma is suspected
•	 Complete blood count, creatinine, calcium
•	 Serum protein electrophoresis, immunofixation, quantitation of im-
munoglobulins and serum FLCs
•	 Serum albumin, β2-​microglobulin, C-​reactive protein, and lactate
dehydrogenase
•	 24-​h urine electrophoresis and immunofixation
•	 Bone marrow aspiration, biopsy, with FISH studies to detect
chromosome 14q32 translocations, trisomies, deletion 17p, and
amplification 1q
•	 Metastatic bone survey, including single views of humeri and femurs;
or preferably low-​dose whole-​body CT scan, or PET-​CT scan
•	 Peripheral blood circulating plasma cells by flow cytometry
section 22  Haematological disorders
5316
detected in a more sensitive manner by low-​dose whole-​body com-
puted tomography (CT) scanning, or positron emission tomog-
raphy (PET)-​CT scanning. Magnetic resonance imaging (MRI) is
useful in patients with suspected SM, where the finding of multiple
focal lesions will upstage the diagnosis to overt plasma cell mye-
loma. MRI scans are also useful in patients with suspected spinal
cord involvement.
Diagnosis
The diagnosis of myeloma requires the presence of 10% or more
clonal plasma cells in the bone marrow (or a biopsy-​proven
plasmacytoma) plus any one or more MDEs: CRAB features attrib-
utable to the plasma cell disorder, 60% or more clonal plasma cells
in the bone marrow, serum FLC ratio of at least 100 (provided in-
volved FLC level is ≥100 mg/​litre), and/​or more than one focal le-
sion on MRI. Plasma cell myeloma needs to be differentiated from
related plasma cell disorders based on the diagnostic criteria listed
in Table 22.4.6.1.
Prognostic features
The median duration of survival in myeloma is approximately 6 to
7 years, but there is considerable variability from one patient to an-
other. The prognosis is affected by the presence or absence of certain
cytogenetic abnormalities. Patients with trisomies, t(11;14) or t(6;14)
are considered to have standard risk myeloma. The presence of del
17p, t(4;14), t(14;16), and t(14;20) is associated with adverse prog-
nosis and is considered as high risk. Other simple markers such as
serum albumin, lactate dehydrogenase, and serum β2 microglobulin
levels also affect prognosis. These are incorporated into the Revised
International Staging System (RISS) (Table 22.4.6.4). In addition,
other markers of adverse prognosis include the presence of renal
failure, increased circulating plasma cells, high plasma cell prolifera-
tive rate, and extramedullary disease.
Treatment of newly diagnosed plasma cell myeloma
Patients with myeloma require therapy. The specific steps of therapy
include initial therapy, stem cell transplantation (if eligible), consoli-
dation/​maintenance therapy, and treatment of relapse. Patients must
be evaluated carefully to ensure that they meet criteria for myeloma,
and that they do not have MGUS or SM. The specific regimens used
in therapy vary depending on the availability of new drugs. The field
is advancing rapidly, and several new drugs are in development.
Improved treatment options have greatly improved the outcome
in myeloma. The most common regimens used in the treatment of
myeloma are listed in Table 22.4.6.5.
Patients eligible for autologous stem cell transplantation
Transplant eligible patients (typically patients less than 65–​70 years
with good performance status) receive approximately four cycles of
initial therapy followed by stem cell collection and ASCT. Selected
patients with standard risk myeloma who respond well to induction
can opt for delayed ASCT; in this strategy, stem cells are collected
after four cycles of initial therapy and cryopreserved for future use
(Fig. 22.4.6.1). If patients are considered eligible for ASCT, it is im-
portant to avoid melphalan which can damage the stem cells. ASCT
is not curative in myeloma, but prolongs overall survival (OS). ASCT
is safe, with a treatment-​related mortality of less than 1% if patients
are appropriately selected.
The initial therapy for transplant eligible patients typically con-
sists of a triplet regimen. We prefer bortezomib, lenalidomide, plus
dexamethasone (VRD) which is associated with a high response rate.
In a recent randomized trial conducted by the Southwest Oncology
Group (SWOG), progression-​free survival (PFS) and OS were sig-
nificantly superior with VRD compared with lenalidomide plus
dexamethasone (Rd). Other alternative regimens are bortezomib,
thalidomide, dexamethasone (VTD), and bortezomib, cyclophos-
phamide, and dexamethasone (VCD). With all these regimens, we
prefer the low-​dose dexamethasone regimen (40 mg once a week) to
minimize toxicity. In a randomized trial conducted by the Eastern
Cooperative Oncology Group (ECOG), the low-​dose dexametha-
sone approach was associated with superior OS and significantly
lower toxicity compared to the high-​dose pulsed dexamethasone
schedules. We also prefer the once-​weekly subcutaneous schedule
of bortezomib in all regimens. This approach greatly reduces the risk
of neurotoxicity. Higher doses of dexamethasone, and twice-​weekly
bortezomib can be considered if a rapid response is desired such as
patients with acute kidney injury due to cast nephropathy, extensive
extramedullary disease, plasma cell leukaemia, or impending cord
compression. Deep venous thrombosis occurs in approximately 15%
of patients given thalidomide or lenalidomide and so prophylaxis
with aspirin or warfarin or low-​dose heparin is needed in all patients
receiving these drugs.
Table 22.4.6.4  Revised International Staging System for plasma cell myeloma
Stage
Frequency
(% of patients)
5-​year survival rate
(%)
Stage 1
ISS stage I (serum albumin >35 g/L, serum beta-​2-​microglobulin <3.5 mg/L) and
No high-​risk cytogenetics
Normal LDH
28
82
Stage II
Neither stage I or III
62
62
Stage III
ISS stage III (serum beta-​2-​microglobulin >5.5 mg/L) and
High-​risk cytogenetics (t(4;14), t(14;16), or del(17p)) or elevated LDH
10
40
LDH, lactate dehydrogenase.
Derived from Palumbo A, et al. (2015). Revised International Staging System for Multiple Myeloma: A Report From International Myeloma
Working Group. J Clin Oncol, 33, 2863–​9.
22.4.6  Plasma cell myeloma and related monoclonal gammopathies
5317
ASCT is performed with melphalan 200 mg/​m2 as the prepara-
tive or conditioning regimen. This is followed by infusion of the
previously collected peripheral blood stem cells. In a randomized
trial performed by the French Myeloma Group, ASCT was associ-
ated with a higher rate of 5-​year, event-​free survival (28% vs 10%)
and OS (52% vs 12%) compared with standard dose chemotherapy.
Other trials have subsequently confirmed the value of ASCT in mye-
loma. The timing of ASCT has been studied, and OS appears similar
whether ASCT is performed as part of initial therapy or at first re-
lapse. Thus in standard-​risk myeloma, the choice of early versus
delayed ASCT can be offered to patients if adequate resources are
available for cryopreservation of stem cells.
A randomized trial of 399 previously untreated myeloma pa-
tients under 60 years of age from France found significantly im-
proved 7-​year event-​free survival (20% vs 10%) and OS (42% vs
21%) in recipients of double versus single ASCT. However this
trial was done prior to the arrival of new agents, and the benefit of
double ASCT was mainly seen in patients who had not achieved
very good partial response with the first transplant. Since most pa-
tients achieve an excellent response even prior to the first ASCT
with novel regimens such as VRD, the role of double ASCT in mye-
loma has diminished greatly.
Allogeneic bone marrow transplantation
Unfortunately, allogeneic bone marrow transplantation has not
consistently shown a benefit in myeloma, and it is associated with
high mortality and morbidity rates. Hence the procedure is mainly
investigational, and outside of a trial setting reserved for selected
young patients with high-​risk disease that has relapsed after initial
therapy and ASCT.
Maintenance therapy
The role of consolidation and maintenance after ASCT has
been studied extensively in myeloma. Most initial trials were
Table 22.4.6.5  Major treatment regimens in plasma cell myeloma
Regimen
Usual dosing schedulea
Doublets
Lenalidomide–​dexamethasone (Rd)
Lenalidomide 25 mg orally days 1–​21 every 28 days
Dexamethasone 40 mg orally days 1, 8, 15, 22 every 28 days
Repeated every 4 weeks
Pomalidomide–​dexamethasone
(Pom/​Dex)
Pomalidomide 4 mg days 1–​21
Dexamethasone 40 mg orally on days on days 1, 8, 15, 22
Repeated every 4 weeks
Triplets
Bortezomib–​thalidomide–​
dexamethasone (VTD)b
Bortezomib 1.3 mg/​m2 intravenous days 1, 8, 15, 22
Thalidomide 100–​200 mg orally days 1–​21
Dexamethasone 20 mg on day of and day after bortezomib (or 40 mg days 1, 8, 15, 22)
Repeated every 4 weeks × 4 cycles as pretransplant induction therapy
Bortezomib–​cyclophosphamide–​
dexamethasoneb (VCD or CyBorD)
Cyclophosphamide 300 mg/​m2 orally on days 1, 8, 15, and 22
Bortezomib 1.3 mg/​m2 intravenously on days 1, 8, 15, 22
Dexamethasone 40 mg orally on days on days 1, 8, 15, 22
Repeated every 4 weeksc
Bortezomib–​lenalidomide–​
dexamethasone (VRD)b
Bortezomib 1.3 mg/​m2 intravenously days 1, 8, 15
Lenalidomide 25 mg orally days 1–​14
Dexamethasone 20 mg on day of and day after bortezomib (or 40 mg days 1, 8, 15, 22)
Repeated every 3 weeksd
Carfilzomib–​ cyclophosphamide–​
dexamethasone (CCyD)e
Carfilzomib 20 mg/​m2 (cycle 1) and 36 mg/​m2 (subsequent cycles) intravenously on days 1, 2, 8, 9, 15, 16
Cyclophosphamide 300 mg/​m2 orally on days 1, 8, 15
Dexamethasone 40 mg orally on days on days 1, 8, 15
Repeated every 4 weeksc
Carfilzomib–​lenalidomide–​
dexamethasone (KRD)
Carfilzomib 27 mg/​m2 intravenously on days 1, 2, 8, 9, 15, 16 (note: cycle 1, day 1 and 2 carfilzomib dose is 20 mg/​m2)
Lenalidomide 25 mg orally days 1–​21
Dexamethasone 20 mg on day of and day after bortezomib (or 40 mg days 1, 8, 15, 22)
Repeated every 4 weeks
Carfilzomib–​pomalidomide–​
dexamethasone
Carfilzomib 27 mg/​m2 intravenously on days 1, 2, 8, 9, 15, 16 (note: cycle 1, day 1 and 2 carfilzomib dose is 20 mg/​m2)
Pomalidomide 4 mg orally days 1–​21
Dexamethasone 40 mg days 1, 8, 15, 22
Repeated every 4 weeks
a All doses need to be adjusted for performance status, renal function, blood counts, and other toxicities.
b Doses of dexamethasone and/​or bortezomib reduced based on subsequent data showing lower toxicity and similar efficacy with reduced doses.
c The day 22 dose of all three drugs is omitted if counts are low, or after initial response to improve tolerability, or when the regimen is used as maintenance therapy; when used as
maintenance therapy for high-​risk patients, further delays can be instituted between cycles.
d Omit day 15 dose if counts are low or when the regimen is used as maintenance therapy; when used as maintenance therapy for high-​risk patients, lenalidomide dose may be
decreased to 10–​15 mg per day, and delays can be instituted between cycles as done in total therapy protocols.
e Dosing based on trial in newly diagnosed patients; in relapsed patients cycle 2 carfilzomib dose is 27 mg/​m2 consistent with approval summary.
Modified from Rajkumar SV (2014). Multiple myeloma: 2014 update on diagnosis, risk-​stratification, and management. Am J Hematol, 89, 998–​1009. Copyright © 2014, John
Wiley and Sons.
section 22  Haematological disorders
5318
disappointing. More recent studies with lenalidomide and with
bortezomib have however shown promise. McCarthy et al., found
superior PFS and OS with lenalidomide maintenance compared with
placebo; the OS survival at 3 years was 88% with lenalidomide and
80% in the placebo regimen (hazard ratio 0.62). However, Attal and
colleagues in a similar trial found that PFS was longer, 41 months
with maintenance lenalidomide and 23  months for placebo, but
OS was virtually the same. These data are conflicting, and the un-
answered question is whether patients who start lenalidomide at the
first sign of relapse may do as well (with fewer adverse effects) com-
pared with patients who start it at the outset following ASCT. One
concern about routine post-​ASCT maintenance is that there was an
excess risk of secondary cancers in both trials in patients receiving
lenalidomide maintenance compared with placebo. At this point,
we are hesitant to recommend lenalidomide maintenance to all pa-
tients. Lenalidomide maintenance can be considered post ASCT, es-
pecially in standard-​risk myeloma patients who have not achieved a
very good partial response following ASCT. In high-​risk myeloma,
bortezomib maintenance appears promising. In a randomized trial,
patients receiving 2 years of bortezomib given every other week as
post-​transplant maintenance had superior outcomes compared with
those receiving thalidomide maintenance.
Following ASCT, and during the observation/​maintenance phase,
patients should be monitored closely every 3 to 4 months. The me-
dian time at which relapse occurs is approximately 2 years in pa-
tients not receiving maintenance, and 4 years in patients receiving
maintenance.
Patients not eligible for stem cell transplantation
Initial therapy for patients who are considered to be not eligible for
ASCT is usually given for approximately 12 months. We prefer VRD
or daratumumab, lenalidomide, dexamethasone (DRD) in fit pa-
tients, and Rd in patients who are frail. Following initial therapy,
maintenance is usually administered. In patients who are treated
with VRD, treatment is continued for 9–12 months and is then fol-
lowed by lenalidomide maintenance (standard risk myeloma) or
bortezomib maintenance (high risk myeloma). In patients treated
with DRD, the same regimen is continued with a lower intensity
after the first 6 months.
Treatment of relapsed refractory myeloma
Patients with myeloma undergo multiple remissions and relapses.
Each remission is typically shorter than the previous one. As the
number of new drugs increases, so does the number of available
combinations to treat relapse. Some key principles are important
to keep in mind when treating relapse. First, in general, treatment
started at relapse is continued until progression. However, in some
regimens such as those using carfilzomib or alkylators it may be
reasonable to stop therapy with these drugs once a stable plateau
has been reached in order to minimize risks of serious toxicity.
Second, the choice of the regimen is based on the aggressiveness
of the relapse. Thus patients with more aggressive relapse may
require a multidrug regimen, more indolent relapses can some-
times be managed with doublets or triplets. Third, in eligible pa-
tients, ASCT should be included in the consideration if the patient
has never had an ASCT, or if the remission duration with a prior
ASCT exceeds 18  months (unmaintained) or 36  months (with
maintenance).
New agents approved for the treatment of relapsed myeloma
include carfilzomib, pomalidomide, panobinostat, daratumumab,
elotuzumab, and ixazomib. The most common regimens and new
drugs used in the treatment of relapsed refractory myeloma are
discussed in the following sections. These drugs used alone and in
combination are the principal options for the treatment of relapse.
Regimens used in initial therapy
In patients who relapse, the regimens listed as options for initial
therapy such as VRD, DRD, VTD, VCD, etc. can all be considered.
If patients had responded well to a given regimen, and then re-
lapsed months to years after stopping therapy, the same regimen
can be reinstituted. Alternatively, a regimen substantially different
than the one used for initial therapy can be tried; for example, DRD
in a patient who is relapsing after initial therapy with VRD. Triplet
regimens such as VRD, DRD, and VCD are well tolerated when
low-​dose dexamethasone and weekly subcutaneous bortezomib
schedules are used.
Carfilzomib
Carfilzomib is a keto-​epoxide tetrapeptide proteasome inhibitor
that has shown efficacy in relapsed myeloma. In a phase III trial of
792 patients, carfilzomib, lenalidomide, and dexamethasone (KRd)
was associated with better response rates, PFS, and OS compared
with Rd. PFS was 26.3 months with KRD versus 17.6 months in the
control group (P = 0.0001). Support for carfilzomib as a more potent
proteasome inhibitor than bortezomib comes from a randomized
trial in which carfilzomib/​dexamethasone demonstrated a doub-
ling of PFS compared with bortezomib/​dexamethasone in relapsed
myeloma; PFS of 18.7 months versus 9.4 months, respectively (P
<0.001). However, the dose of carfilzomib used in this trial (56 mg/​
m2) is much higher, and carries a much higher monetary cost com-
pared with bortezomib. Carfilzomib has lower neurotoxicity than
bortezomib, but a small proportion (5%) of patients may experience
serious cardiac side effects.
Pomalidomide
Pomalidomide is an analogue of lenalidomide and thalidomide with
significant activity in relapsed refractory myeloma, even in patients
failing lenalidomide and bortezomib. In a randomized trial of 302
patients with refractory myeloma, pomalidomide plus low-​dose
dexamethasone was found superior to high-​dose dexamethasone.
The doublet regimen of Pd is a reasonable option for patients with
indolent relapse. Pomalidomide can also be administered in com-
binations such as carfilzomib, pomalidomide, and dexamethasone
(car/​pom/​dex).
Panobinostat
Panobinostat is a pan-​deacetylase inhibitor that blocks the
aggresome pathway, an alternative route for cells to bypass the le-
thal effects of proteasome inhibition. By combining bortezomib and
panobinostat, there is simultaneous blockade of both proteasome
and aggresome pathways. In a randomized trial of 768 patients,
bortezomib/​dexamethasone plus panobinostat was associated with
superior PFS compared with bortezomib/​dexamethasone plus pla-
cebo. However, panobinostat was associated with grade 3 diarrhoea
in approximately 25% of patients, and care should be exercised when
using this drug.
22.4.6  Plasma cell myeloma and related monoclonal gammopathies
5319
Elotuzumab
Elotuzumab, a monoclonal antibody targeting the signalling
lymphocytic activation molecule F7 (SLAMF7), does not have any
single-​agent activity, but synergizes with Rd. In a phase III trial of
646 patients, elotuzumab plus Rd was superior to Rd in terms of
PFS, median PFS 19.4 months versus 14.9 months, respectively (P
<0.001). Elotuzumab is well tolerated.
Daratumumab
Daratumumab, a monoclonal antibody targeting CD38, has
shown promise in relapsed, refractory myeloma. In a phase II trial,
daratumumab as a single agent produced a response rate of approxi-
mately 30% in heavily pretreated patients. In randomized trials,
DRD and daratumumab, bortezomib, dexamethasone (DVD) have
shown superiority over doublet regimens.
Ixazomib
Ixazomib is an oral proteasome inhibitor that is active in relapsed
myeloma. In a randomized controlled trial in relapsed myeloma,
ixazomib, lenalidomide, and dexamethasone (IRd) was found to im-
prove PFS compared with Rd. Ixazomib needs to be administered
only once weekly, a major advantage when considering long-​term
therapy. It has lower risk of neurotoxicity compared with bortezomib.
Emerging options
Promising agents in development for myeloma include marizomib,
a new proteasome inhibitor, oprozmib, an oral proteasome inhibitor
related to carfilzomib; filanesib, a kinesin spindle protein inhibitor;
dinaciclib, a cyclin-​dependent kinase inhibitor; venetoclax, a se-
lective BCL-​2 inhibitor, and LGH-​447, a pan PIM kinase inhibitor.
Each of these drugs has single-​agent activity in relapsed myeloma.
Supportive care
Skeletal complications
Bisphosphonates are important for the treatment of patients with
plasma cell myeloma and associated bony disease. Pamidronate
90 mg intravenously over at least 2 h or zoledronic acid 4 mg over
at least 15 min every 3 to 4 weeks are recommended. Intravenous
bisphosphonates should be continued for 1 to 2 years and if
the patient has stable disease, bisphosphonates may be discon-
tinued. Bisphosphonates should be resumed upon relapse with
new skeletal-​related events. Other potential side effects of intra-
venous bisphosphonates include the development of proteinuria or,
rarely, acute kidney injury. Osteonecrosis of the jaw has also been
­described. Denosumab, is an alternative to bisphosphonates in pa-
tients with renal dysfunction.
Both vertebroplasty (injection of methyl methacrylate into a
collapsed vertebral body) and kyphoplasty (introduction of an in-
flatable bone tamp into the vertebral body and after inflation, the
injection of methyl methacrylate into the cavity) have been used to
decrease pain and help restore vertebral height. Pain relief is gener-
ally rapid and may be long-​lasting.
Patients should be encouraged to be as active as possible, but they
must avoid undue trauma. Fixation of fractures or pending fractures
with an intramedullary rod and methyl methacrylate has produced
good results. Bone pain should be treated with analgesics or nar-
cotics as necessary.
Hypercalcaemia
This is present in 15% of patients at diagnosis and should be sus-
pected in the presence of anorexia, nausea, vomiting, polyuria,
polydipsia, constipation, weakness, confusion, or stupor. If un-
treated, renal insufficiency develops. Hydration with saline, cor-
ticosteroids, and intravenous bisphosphonates are the mainstay
of therapy. Calcitonin may be helpful if the patient is refractory to
bisphosphonates. Haemodialysis may be useful in some patients
with severe, resistant hypercalcaemia. The dose of zoledronic acid
must be reduced in patients with renal failure.
Light-​chain cast nephropathy
Maintenance of a high urine output (3 litres/​day) is important.
Nonsteroidal anti-​inflammatory agents, dehydration, infections, or
radiographic contrast media may contribute to acute kidney injury
and should be avoided. Patients with acute or subacute kidney injury
should be treated with a regimen such as VCD or VTD to reduce the
tumour mass as quickly as possible. Plasmapheresis may be useful
in acute kidney injury, but patients with irreversible changes are un-
likely to benefit. Haemodialysis or peritoneal dialysis is necessary in
the event of symptomatic azotaemia.
Infection
Appropriate antibiotic therapy for bacterial infections is essential.
Patients should receive pneumococcal and influenza vaccination
despite their suboptimal antibody response. Prophylaxis against
pneumocystis pneumonia should be considered in patients re-
ceiving high-​dose corticosteroids. Intravenously administered
gammaglobulin can be used for severe recurrent infections.
Neurological
Spinal cord compression should be suspected in patients with back
pain who develop weakness or paraesthesiae of the lower extrem-
ities or bladder or bowel dysfunction. Imaging by MRI or CT must
be performed immediately. Radiation therapy and dexametha-
sone are usually effective, and surgical decompression is rarely
necessary.
Hyperviscosity
This is characterized by oral or nasal bleeding, blurred vision, paraes-
thesiae, headache, reduced cerebration, or congestive heart failure.
Serum viscosity levels do not correlate well with the symptoms or
clinical findings. Plasmapheresis promptly relieves the symptoms
and should be done regardless of the viscosity level if the patient has
signs or symptoms of hyperviscosity.
Anaemia
Anaemia occurs in almost all patients during the course of
­myeloma. Erythropoietin (40 000 U subcutaneously weekly) or
darbepoetin (200 µg subcutaneously every 2 weeks) is beneficial.
Blood transfusions are indicated for patients with symptomatic
anaemia who do not benefit from other therapy. Iron, folate, or
vitamin B12 deficiency may be responsible for anaemia and must be
recognized and treated.
Emotional support
All patients with myeloma need substantial and continuing emo-
tional support. The physician’s approach must be positive and
section 22  Haematological disorders
5320
emphasize the potential benefits of therapy. It is reassuring for pa-
tients to know that many survive for 10 years or more. It is vital that
the physician caring for patients with myeloma has the interest and
capacity to deal with an incurable disease over the space of years
with assurance, sympathy, and resourcefulness.
Variant forms of plasma cell myeloma
Plasma cell leukaemia
Plasma cell leukaemia is defined as the presence of more than 5%
plasma cells in the peripheral blood on regular white blood cell dif-
ferential count examination. It is classified as primary when it pre-
sents de novo (60% of cases) and as secondary when it is a leukaemia
transformation of previously recognized plasma cell myeloma
(40%). Patients with primary plasma cell leukaemia are younger and
have a higher platelet count, fewer bone lesions, a smaller serum
paraprotein, a greater incidence of hepatosplenomegaly and lymph-
adenopathy, and a longer duration of survival than patients with sec-
ondary plasma cell leukaemia. Cytogenetic abnormalities are more
common than in patients with plasma cell myeloma. Treatment is
with multidrug initial therapy followed by ASCT and bortezomib-​
based maintenance. Those with secondary plasma cell leukaemia
rarely respond to chemotherapy because they already have received
treatment and are refractory.
POEMS syndrome (osteosclerotic myeloma)
This is characterized by polyneuropathy (P), organomegaly (O),
endocrinopathy (E), M-​protein (M), and skin changes (S) (Table
22.4.6.1). The major clinical finding is a chronic inflammatory–​
demyelinating neuropathy with predominantly motor disability.
Sclerotic bone lesions are found in almost all patients. The cranial
nerves are not involved except for the presence of papilloedema,
which is seen in 30 to 40%. Hepatomegaly occurs in almost half of
patients, but splenomegaly and lymphadenopathy occur in a mi-
nority. Hyperpigmentation and hypertrichosis are frequent but may
be easily overlooked. Gynaecomastia and atrophic testes as well as
clubbing of the fingers and toes may be present. Angiomatous le-
sions of the trunk are often prominent. Pulmonary hypertension
has been recognized in several instances. Ascites, pleural effusion,
and peripheral oedema may be present. In contrast to plasma cell
myeloma, the haemoglobin level is usually normal or increased,
and thrombocytosis is common. The bone marrow usually contains
fewer than 5% monoclonal plasma cells, and hypercalcaemia and
renal insufficiency rarely occur. Almost all patients have a λ light
chain, and IgA is the most common heavy-​chain type. Castleman’s
disease may be present. Vascular endothelial growth factor levels
may be 5-​ to 10-​fold higher in POEMS syndrome compared with
normal controls. If the skeletal lesions are in a limited area, radiation
almost always produces a substantial improvement of the neur-
opathy. If widespread osteosclerotic lesions exist, an ASCT should be
considered for therapy. Chemotherapy similar to that used in mye-
loma may be helpful. The median survival is approximately 15 years.
Solitary plasmacytoma of bone
The diagnosis depends on histological evidence of a plasma cell
tumour but no evidence of multiple myeloma (Table 22.4.6.1).
A  small M-​protein may be found in the serum or urine, but it
usually disappears after radiation of a solitary lesion. In a study
of 116 patients with solitary plasmacytoma of bone, the persist-
ence of a serum M-​protein level of 5 g/​litre or more 1 to 2 years
after diagnosis and an abnormal FLC ratio at the time of diagnosis
are predictive of disease progression. Some patients may have up
to 10% clonal plasma cells, and are considered to have both soli-
tary plasmacytoma with minimal marrow involvement. Treatment
consists of tumouricidal radiation (40–​50 Gy). Progression to overt
myeloma develops in approximately 10% of patients with solitary
plasmacytoma, and in 60% of patients with solitary plasmacytoma
with minimal marrow involvement over the next 3 years. Baseline
PET-​CT and/​or MRI scans are helping in making the accurate
diagnosis at the outset.
Solitary extramedullary plasmacytoma
This is a plasma cell tumour that arises outside the bone marrow.
It is located in the upper respiratory tract in approximately 80% of
cases, and the nasal cavity and sinuses, nasopharynx, and larynx
are most often involved. The gastrointestinal tract, central ner-
vous system, urinary bladder, thyroid, breast, testes, parotid gland,
and lymph nodes have all been reported as the initial site of an
extramedullary plasmacytoma. There is a predominance of IgA
M-​protein in extramedullary plasmacytomas. The diagnosis de-
pends on the finding of a plasma cell tumour in an extramedullary
location and the absence of myeloma on bone marrow examin-
ation, radiography, and appropriate studies of serum and urine.
Treatment consists of tumouricidal radiation (40–​50 Gy). Regional
occurrences develop in approximately 10% of patients, while devel-
opment of typical plasma cell myeloma occurs in 20%.
Waldenström’s macroglobulinaemia
This malignant lymphoplasmacytic proliferative disorder produces
a high concentration of immunoglobulin M (IgM) paraprotein. It
bears similarities to myeloma, lymphoma, and chronic lympho-
cytic leukaemia. The incidence rate is 0.5/​100 000. The median age
is approximately 65 years, and 60% of patients are male.
Clinical findings
Weakness and fatigue are the most common symptoms of WM.
Chronic nasal bleeding or oozing from the gums is character-
istic, but postsurgical or gastrointestinal bleeding may also occur.
Blurring or loss of vision may be prominent. Dyspnoea and con-
gestive heart failure may develop. Dizziness, headaches, vertigo,
nystagmus, ataxia, and diplopia have been seen. Constitutional
symptoms including fever, night sweats, and loss of weight may
be present. Bone pain is rare. Hepatomegaly occurs in about 25%
of patients at diagnosis, and splenomegaly and lymphadenopathy
are slightly less common. Retinal vein engorgement and flame-​
shaped haemorrhages are common and are a better measure of
symptomatic hyperviscosity syndrome than is the measurement
of serum viscosity.
Pulmonary involvement may be manifested by diffuse pulmonary
infiltrates, isolated masses, or pleural effusion. Retroperitoneal and
mesenteric lymphadenopathy are common, but they are usually
asymptomatic. The most common neurological manifestation is
sensorimotor peripheral neuropathy. It is related to amyloid depos-
ition in some instances.
22.4.6  Plasma cell myeloma and related monoclonal gammopathies
5321
Laboratory findings
Anaemia is found in most patients with symptomatic WM.
Spuriously low haemoglobin and haematocrit levels may re-
sult from an increased plasma volume due to the large amount of
paraprotein.
Serum protein electrophoresis reveals a tall, narrow spike or
dense band usually migrating in the γ area. About 75% of the IgM
paraproteins are κ. The IgM level obtained by nephelometry is often
10 to 30 g/​litre more than that found with serum protein electrophor-
esis. A reduction of uninvolved IgG and IgA immunoglobulins is less
striking than in multiple myeloma. About 10% of macroglobulins
precipitate in the cold (cryoglobulin) but are almost always asymp-
tomatic. A monoclonal light chain detected by immunofixation is
present in the urine in 75% of patients, but it is usually small.
The bone marrow aspirate is often hypocellular, but the biopsy
specimen is usually hypercellular and extensively infiltrated with
lymphoid or plasmacytoid cells (lymphoplasmacytic lymphoma).
Increased mast cells are frequently present.
Diagnosis
WM arises from CD25+CD22+low activated B lymphocytes. The
diagnosis of WM depends on the presence of an IgM paraprotein
and a 10% or greater monoclonal lymphoplasmacytic infiltration
of the bone marrow producing symptoms and physical findings
consistent with WM. The lymphoplasmacytic cells express CD19,
CD20, and CD22. Most patients with WM have a recurrent somatic
mutation, L265P, involving the MYD88 gene. The differential diag-
nosis includes plasma cell myeloma, MGUS of the IgM type, smoul-
dering WM, chronic lymphocytic leukaemia, lymphoma, and other
undifferentiated lymphoplasmacytic proliferative disorders. A var-
iety of prognostic models have been reported.
Treatment
Patients with WM should not be treated unless they are symptom-
atic. Even in the presence of 10% or more clonal infiltration of the
bone marrow, patients without end-​organ damage can remain stable
without therapy for extended periods of time. Such patients, some-
times termed smouldering WM, occupy a clinical stage in between
IgM MGUS and WM.
In symptomatic patients, therapy involved rituximab-​based
combinations. Rituximab, a monoclonal antibody directed against
the CD20 antigenic determinant, produces a response in approxi-
mately one-​half of patients. It is often given with dexamethasone
and cyclophosphamide, or with bendamustine. The IgM levels
may temporarily increase following therapy (flare). The response
to rituximab may be delayed and maximum response may not
occur until several months after therapy. Symptoms and findings
of hyperviscosity are quickly controlled by plasmapheresis with a
cell separator.
Options for therapy of relapsed WM besides regimens used in
the frontline setting include ibrutinib, bortezomib plus rituximab
and dexamethasone, purine nucleoside analogues (cladribine,
fludarabine), carfilzomib, and immunomodulatory agents (thal-
idomide, lenalidomide). Ibrutinib is an irreversible and selective
inhibitor of Bruton tyrosine kinase, a signalling molecule in the
B-​cell antigen receptor cascade. In a phase II study including 63
patients with symptomatic relapsed/​refractory WM, the overall
response rate was 81%. Chlorambucil, given continuously or
intermittently, has been a standard treatment for more than
50 years, but has now fallen out of favour. Everolimus (RAD-​001),
perifosine, or alemtuzumab, an anti-​CD52 antibody, have all
shown activity. The median duration of survival for patients with
WM is more than 5 years.
Heavy-​chain diseases
Heavy-​chain diseases are clonal plasma cell disorders where intact
IgH is secreted without any companion light chain. These disorders
are rare, and very few data are available on optimal therapy.
γ-​Heavy-​chain diseases
The paraprotein consists of a monoclonal γ chain with significant
amino acid deletions. The initial presentation is often a lymphoma-​
like illness, but the symptoms and clinical findings are diverse and
range from an aggressive lymphoproliferative process to an asymp-
tomatic state. Weakness, fatigue, and fever are the most common
presenting symptoms. The serum protein electrophoretic pattern
usually shows a broad-​based band more suggestive of a polyclonal
than an M-​protein. Symptomatic patients should be treated with
chemotherapy regimens similar to ones used on non-​Hodgkin’s
lymphoma. The median duration of survival in a series of 23 patients
was 7.4 years (range 1 month to 21 years).
α-​Heavy-​chain diseases
α-​Heavy-​chain disease is the most common type of heavy-​chain
disease, with more than 400 reported patients since its recogni-
tion. Most patients have been from relatively poor countries in
the Mediterranean region and Middle East. Gastrointestinal tract
involvement is most common and is manifested by malabsorp-
tion with loss of weight, diarrhoea, and steatorrhoea. It is similar
to ‘immunoproliferative small intestinal disease’ (IPSID). The
serum protein electrophoretic pattern shows no spike. The diag-
nosis depends on recognition of a monoclonal α heavy chain in
the serum or jejunal fluid. α-​Heavy-​chain disease is progressive
and fatal without therapy. Antibiotics may produce remission,
particularly if given early in the course of the disease. Patients
who have advanced disease or who do not respond to antibiotics
should be treated with a combination of chemotherapy consisting
of chemotherapy regimens similar to ones used on non-​Hodgkin
lymphoma.
μ-​Heavy-​chain diseases
µ-​Heavy-​chain disease is characterized by the presence of a mono-
clonal µ-​chain fragment in the serum. Most patients have a chronic
lymphoproliferative process resembling chronic lymphocytic leu-
kaemia or lymphoma. Fewer than 40 cases have been reported. The
serum protein electrophoretic pattern contains a spike or localized
band in about 40% of patients. Vacuolization of the plasma cells in
the marrow is an important clue for the diagnosis of µ-​heavy-​chain
disease. The course of µ-​heavy-​chain disease is variable, with a me-
dian survival of approximately 2 years. Patients should be treated
with chemotherapy regimens similar to ones used on non-​Hodgkin
lymphoma.
section 22  Haematological disorders
5322
Immunoglobulin light-​chain amyloidosis
Amyloid is a substance consisting of fibrils that appear homoge-
neous and amorphous under the light microscope and stain pink
with haematoxylin–​eosin. With polarized light, amyloid stained
with Congo red produces an apple-​green birefringence. Linear,
nonbranching, aggregated fibrils 7.5 to 10 nm wide and of indef-
inite length are seen with electron microscopy. In AL amyloidosis,
these fibrils consist of monoclonal κ or λ light chains. Other forms
of amyloidosis exist, and are caused by a variety of different pro-
teins such as protein A  in secondary amyloidosis, transthyretin
(prealbumin) in familial or senile systemic amyloidosis, and β2-​
microglobulin in dialysis-​associated amyloidosis. More than 25 dif-
ferent proteins may form amyloid fibrils. Only AL amyloidosis is a
monoclonal gammopathy; the other forms are not related to plasma
cell disorders and are not discussed here.
Aetiology and epidemiology
The annual incidence of AL amyloidosis is 0.9/​100 000. The median
age at diagnosis is 65 years, and only 1% of patients are younger than
40 years. The cause of AL amyloidosis is unknown.
Clinical features
The clinical features depend on the specific organ involved, and
the number of organs that are involved in AL. Weakness, fatigue,
and weight loss are the most common initial symptoms. Light-​
headedness, syncope, change in the tongue or voice, jaw or hip
claudication, paraesthesiae, dyspnoea, and oedema are the most
frequent symptoms. Macroglossia is present in 10% of patients,
and purpura, particularly in the periorbital and facial areas, is
found in 15%. The liver is palpable in 25% of patients, but spleno-
megaly occurs in only 5%. Nephrotic syndrome or renal failure is
found in more than 25% of patients at diagnosis (Fig. 22.4.6.5).
Congestive heart failure, carpal tunnel syndrome, sensorimotor
peripheral neuropathy, and orthostatic hypotension are other
important features. The presence of one of these syndromes and
concomitant presence of an M-​protein are strong indications of
AL, for which appropriate biopsy specimens must be taken for
diagnosis.
Laboratory findings
The serum protein electrophoretic pattern shows a modest-​sized lo-
calized band or spike in about half of the patients (median 14 g/​litre).
An M-​protein is found in the serum or urine in 90% of patients,
and λ light chains are twice as common as κ. The bone marrow con-
tains 5% or less monoclonal plasma cells in almost half of patients.
A paraprotein in the serum or urine or a monoclonal proliferation of
plasma cells in the bone marrow occurs in 98% of patients with AL
amyloidosis. Only one-​fifth of patients have more than 20% plasma
cells in the bone marrow, but they usually do not have the other fea-
tures of plasma cell myeloma.
An increased serum alkaline phosphatase level is not un-
common. Hyperbilirubinaemia is infrequent, but when present
it is an ominous sign. The coagulation factor X concentration is
decreased in more than 10% of patients but is rarely the cause of
bleeding.
Congestive heart failure is present in about 20% of patients at
diagnosis. Electrocardiography frequently reveals low voltage in the
limb leads or characteristics consistent with anteroseptal infarction
(loss of anterior forces). Arrhythmias, including atrial fibrillation
or heart block, are common. Almost two-​thirds of patients have an
abnormal echocardiogram at diagnosis. Early cardiac involvement
is characterized by abnormal relaxation followed by the features of
constrictive cardiomyopathy. Amyloid heart disease may closely
resemble constrictive pericarditis or hypertrophic obstructive car-
diomyopathy. A sensorimotor peripheral neuropathy is present in
about 15% of patients at diagnosis. Autonomic dysfunction may be
a prominent feature and is often manifested by orthostatic hypoten-
sion, diarrhoea, and impotence.
Diagnosis
The diagnosis depends on appropriate systemic syndrome, evi-
dence of a monoclonal plasma cell disorder, the demonstration
of amyloid deposits, and mass spectroscopy showing that the
fibrils are composed of immunoglobulin light chains. The pos-
sibility of AL amyloidosis must be considered in patients who
have an M-​protein in the serum or urine and who present with
unexplained nephrotic syndrome, congestive heart failure, sen-
sorimotor peripheral neuropathy, carpal tunnel syndrome, hep-
atomegaly, or malabsorption syndrome. The initial diagnostic
procedure should be an abdominal fat aspiration, which is posi-
tive in about 75% of patients (Fig. 22.4.6.6). A bone marrow as-
piration and biopsy should be done to determine the degree of
plasmacytosis, and amyloid stains of the biopsy specimen will be
positive in slightly more than half of patients. The abdominal fat
aspirate and/​or bone marrow biopsy is positive in 90% of cases;
if negative, a biopsy of a suspected involved organ such as the
kidney, liver, heart, or sural nerve is indicated. We perform laser
microdissection of Congo Red staining material from the biopsy
specimens embedded in paraffin and then subject the specimen to
tandem mass spectrometry-​based proteomic analysis. Iodine-​123
labelled serum amyloid-​P component scintigraphy can be used
for identifying and monitoring the extent of systemic amyloidosis
but it is not readily available.
40
20
0
Positive (%)
10
30
2
5
0.5
0.5
17
21
17
28
Nephrotic/
renal failure
n = 142
CHF
n = 104
Carpal
tunnel
n = 102
Peripheral
neuropathy
n = 81
Ortho
hypo
n = 58
11
1.5
At diagnosis
During follow-up
n = 474
Fig. 22.4.6.5  Frequency of amyloid syndromes at diagnosis of
immunoglobulin light chain amyloidosis. CHF, congestive heart failure;
Ortho hypo, orthostatic hypotension.
From Kyle RA, Gertz MA (1995). Primary systemic amyloidosis: clinical and
laboratory features in 474 cases. Semin Hematol, 32, 45–​59. By permission of
W B Saunders Company.
22.4.6  Plasma cell myeloma and related monoclonal gammopathies
5323
Prognosis
The median duration of survival for 474 patients with AL within
1 month of diagnosis was 13 months. Those presenting with con-
gestive heart failure had a median survival of 4 months. Elevated
levels of N-​terminal pro-​brain natriuretic peptide (NT-​proBNP),
as well as cardiac troponins are important prognostic features.
In 261 patients with newly diagnosed AL amyloidosis, the sur-
vival was 6 months in patients with a detectable level of cardiac
troponin T compared with 22  months for those who did not.
A prognostic model consisting of NT-​proBNP greater than 332
mg/​litre and cardiac troponin T greater than 0.035 mcg/​litre were
classified as stages I, II, or III, depending on whether none, one, or
both NT-​proBNP and cardiac troponins were above these levels.
The survivals for stages I, II, and III were 26, 11, and 4 months,
respectively.
Treatment
The goal of therapy in AL amyloidosis is to eradicate the plasma cell
clone responsible for the monoclonal light chain production. This is
achieved with therapy similar to that used in plasma cell myeloma.
Specifically VCD is used as initial therapy in most patients. In eli-
gible patients without cardiac or multiorgan involvement, ASCT is
an option. But even with careful selection, the 100-​day treatment
related mortality is approximately 5%, in contrast to 1 to 2% for mul-
tiple myeloma. There are emerging data that doxycycline may pro-
long survival in AL amyloidosis by inhibiting deposition of amyloid
fibrils. Trials are also ongoing with other agents in an attempt to pre-
vent amyloid fibril deposition.
Future directions
Despite the introduction of several new agents, plasma cell myeloma
and related disorders remain incurable in most patients. A better
understanding of disease biology and identification of new drugs in
the last 10 years has dramatically improved OS. Additional prom-
ising treatment options are certain to emerge. At present, there are
also trials going on to determine if early therapy at the SM stage
can lead to a cure. Some of the problem areas requiring additional
attention include treatment of high-​risk myeloma, plasma cell leu-
kaemia, extramedullary relapse, and AL amyloidosis. Prediction of
response to therapy may lead to more personalized care and is a
major goal for the future.
FURTHER READING
Attal M, et al. (1996). A prospective, randomized trial of autologous
bone marrow transplantation and chemotherapy in multiple mye-
loma. Intergroupe Francais du Myelome. N Engl J Med, 335, 91–​7.
Attal M, et al. (2012). Lenalidomide maintenance after stem-​cell trans-
plantation for multiple myeloma. N Engl J Med, 366, 1782–​91.
Benboubker L, et  al. (2014). Lenalidomide and dexamethasone in
transplant-​ineligible patients with myeloma. N Engl J Med, 371,
906–​17.
Dispenzieri A, et al. (2013). Prevalence and risk of progression of light-​
chain monoclonal gammopathy of undetermined significance:  a
retrospective population-​based cohort study. Lancet, 375, 1721–​8.
Falk RH, Comenzo RL, Skinner M (1997). The systemic amyloidoses.
N Engl J Med, 337, 898–​909.
Krishnan A, et al. (2011). Autologous haemopoietic stem-​cell trans-
plantation followed by allogeneic or autologous haemopoietic
stem-​cell transplantation in patients with multiple myeloma (BMT
CTN 0102): a phase 3 biological assignment trial. Lancet Oncol, 12,
1195–​203.
Kumar S, et al. (2012). Trisomies in multiple myeloma: impact on sur-
vival in patients with high-​risk cytogenetics. Blood, 119, 2100–​5.
Kumar SK, et al. (2014). Continued improvement in survival in mul-
tiple myeloma: changes in early mortality and outcomes in older pa-
tients. Leukemia, 28, 1122–​8.
Kyle RA (2000). Multiple myeloma:  an odyssey of discovery. Br J
Haematol, 111, 1035–​44.
Kyle RA, Gertz MA (1995). Primary systemic amyloidosis: clinical and
laboratory features in 474 cases. Semin Hematol, 32, 45–​59.
Kyle RA, Rajkumar SV (2004). Multiple myeloma: drug therapy (re-
view article). N Engl J Med, 351, 1860–​73.
Kyle RA, Rajkumar SV (2006). Monoclonal gammopathy of undeter-
mined significance. Br J Haematol, 134, 573–​89.
Abdominal
fat
n = 212
100
80
60
40
20
0
Positive (%)
80
56
75
94
82
97
83
90
86
100
Bone
marrow
394
Rectum
194
Kidney
81
Carpal
ligament
20
Liver
32
Small
intestine
23
Skin
19
Sural
nerve
21
Heart
16
Fig. 22.4.6.6  Diagnosis of amyloidosis on the basis of deposits in tissues.
From Kyle RA, Gertz MA (1995). Primary systemic amyloidosis: clinical and laboratory features in 474 cases. Semin
Hematol, 32, 45–​59. By permission of W B Saunders Company.
section 22  Haematological disorders
5324
Kyle RA, et al. (2003). Review of 1027 patients with newly diagnosed
multiple myeloma. Mayo Clin Proc, 78, 21–​33.
Kyle RA, et al. (2004). Long-​term follow-​up of 241 patients with mono-
clonal gammopathy of undetermined significance:  the original
Mayo Clinic series 25 years later. Mayo Clin Proc, 79, 859–​66.
Kyle RA, et al. (2006). Prevalence of monoclonal gammopathy of un-
determined significance. N Engl J Med, 354, 1362–​9.
Kyle RA, et al. (2007). Clinical course and prognosis of smoldering
(asymptomatic) multiple myeloma. N Engl J Med, 356, 2582–​90.
Kyle RA, et  al. (2012). Progression in smoldering Waldenstrom
macroglobulinemia: long-​term results. Blood, 119, 4462–​6.
Landgren O, et al. (2014). Racial disparities in the prevalence of mono-
clonal gammopathies:  a population-​based study of 12 482 per-
sons from the national health and nutritional examination survey.
Leukemia, 28, 1537–​42.
Lonial S, et al. (2015). Elotuzumab therapy for relapsed or refractory
multiple myeloma. N Engl J Med, 373, 621–​31.
McCarthy PL, et al. (2012). Lenalidomide after stem-​cell transplant-
ation for multiple myeloma. N Engl J Med, 366, 1770–​81.
Palumbo A, et  al. (2012). Continuous lenalidomide treatment for
newly diagnosed multiple myeloma. N Engl J Med, 366, 1759–​69.
Palumbo A, et al. (2015). Revised International Staging System for
Multiple Myeloma: A Report From International Myeloma Working
Group. J Clin Oncol, 33, 2863–​9.
Rajan AM, et al. (2012). Interpretation of cytogenetic results in mul-
tiple myeloma for clinical practice. Blood Cancer J, 5, e365.
Rajkumar SV (2014). Multiple myeloma: 2014 Update on diagnosis,
risk-​stratification, and management. Am J Hematol, 89, 998–​1009.
Rajkumar SV, et al. (2005). Serum free light chain ratio is an inde-
pendent risk factor for progression in monoclonal gammopathy of
undetermined significance. Blood, 106, 812–​7.
Rajkumar SV, et al. (2010). Lenalidomide plus high-​dose dexametha-
sone versus lenalidomide plus low-​dose dexamethasone as initial
therapy for newly diagnosed multiple myeloma: an open-​label ran-
domised controlled trial. Lancet Oncol, 11, 29–​37.
Rajkumar SV, et al. (2011). Consensus recommendations for the uni-
form reporting of clinical trials: report of the International Myeloma
Workshop Consensus Panel 1. Blood, 117, 4691–​5.
Rajkumar SV, et al. (2014). International Myeloma Working Group
Updated Criteria for the Diagnosis of Multiple Myeloma. Lancet
Oncol, 15, e538–​48.
Rajkumar SV, et al. (2015). Smoldering multiple myeloma. Blood, 125,
3069–​75.
Richardson PG, et al. (2003). A phase 2 study of bortezomib in re-
lapsed, refractory myeloma. N Engl J Med, 348, 2609–​17.
Roodman GD (2009). Pathogenesis of myeloma bone disease.
Leukemia, 23, 435–​41.
San Miguel JF, et al. (2013). Pomalidomide plus low-​dose dexametha-
sone versus high-​dose dexamethasone alone for patients with re-
lapsed and refractory multiple myeloma (MM-​003): a randomised,
open-​label, phase 3 trial. Lancet Oncol, 14, 1055–​66.
San-​Miguel JF, et al. (2014). Randomized phase 3 trial of the deacetylase
inhibitor panobinostat plus bortezomib and dexamethasone versus
placebo plus bortezomib and dexamethasone in relapsed or relapsed
and refractory multiple myeloma. Lancet Oncol, 15, 1195–​206.
Singhal S, et al. (1999). Antitumor activity of thalidomide in refractory
multiple myeloma. N Engl J Med, 341, 1565–​71 (erratum in N Engl J
Med 2000, 342, 364).
Sonneveld P, et  al. (2012). Bortezomib induction and maintenance
treatment in patients with newly diagnosed multiple myeloma: re-
sults of the randomized phase III HOVON-​65/​GMMG-​HD4 Trial. J
Clin Oncol, 30, 2946–​55.

22.5 Bone marrow failure 5325 22.5.1 Inherited bon
22.5 Bone marrow failure 5325 22.5.1 Inherited bone marrow failure syndromes 5325 Irene Roberts and Inderjeet S. Dokal
CONTENTS
22.5.1	 Inherited bone marrow failure syndromes  5325
Irene Roberts and Inderjeet S. Dokal
22.5.2	 Acquired aplastic anaemia and pure red cell
aplasia  5336
Judith C.W. Marsh, Shreyans Gandhi, and Ghulam J. Mufti
22.5.3	 Paroxysmal nocturnal haemoglobinuria  5348
Lucio Luzzatto
22.5.1  Inherited bone marrow
failure syndromes
Irene Roberts and Inderjeet S. Dokal
ESSENTIALS
Inherited forms of bone marrow failure may involve all haematopoi-
etic lineages or a single lineage. They are rare, but collectively account
for 20 to 30% of patients presenting with aplastic anaemia. They may
present at birth or in infancy or childhood, but also sometimes in
adults. Associated somatic abnormalities may be helpful in diagnosis.
Two of the best characterized syndromes are Fanconi anaemia and
dyskeratosis congenita, both frequently associated with gener­alized
bone marrow failure. Other well-​recognized disorders lead to much
more specific abnormalities affecting a single cell type, e.g. impaired
red cell production in Diamond–​Blackfan anaemia and impaired neu-
trophil production in Shwachman–​Diamond syndrome, and reduced
platelet production in thrombocytopenia with absent radii syndrome.
Advances in understanding the genetics of inherited bone marrow
failure syndromes have provided valuable insight into their patho-
physiology, and also into normal haematopoiesis.
Introduction
Bone marrow failure (BMF) is the commonly used term to describe
impaired production of normal peripheral blood cells as a result of
reduced numbers of haematopoietic stem and progenitor cells in the
bone marrow (BM). Inherited forms of BMF may involve all haem-
atopoietic lineages or a single lineage. In some cases, patients may
present with a single cytopenia (e.g. thrombocytopenia or anaemia)
before progressing to pancytopenia later in the course of the disease.
Where all haematopoietic lineages are involved, the term ‘aplastic
anaemia’ (AA) is sometimes used since BM biopsies from affected
patients show striking hypocellularity, although this term is rather
misleading since such patients also have thrombocytopenia and
neutropenia. The inherited BMF syndromes are rare. Although their
precise incidence and prevalence is not known, collectively they rep-
resent approximately 20 to 30% of patients presenting with AA and
constitute a significant clinical burden, as many are associated with
premature mortality.
The main types of inherited BMF are listed in Box 22.5.1.1. Many
have associated somatic abnormalities which in some cases are
very helpful in making a presumptive diagnosis at an early stage
(Table 22.5.1.1). Inherited BMF syndromes may present at birth or
at a variable time thereafter, including in adulthood in some cases.
The genetic basis for an increasing proportion of cases of inherited
22.5
Bone marrow failure
Box 22.5.1.1  Inherited BMF syndromes
Pancytopenia (usually associated with a global
haematopoietic defect)
•	 Fanconi anaemia
•	 Dyskeratosis congenita
•	 Shwachman–​Diamond syndrome
•	 Reticular dysgenesis
•	 Pearson syndrome
•	 Familial aplastic anaemia (autosomal and X-​linked forms)
•	 Myelodysplasia (MDS)
•	 Nonhaematological syndromes (Down syndrome, Dubowitz’s
syndrome)
Single cytopenia (usually)
•	 Anaemia:
—​	 Diamond–​Blackfan anaemia
—	 ​Congenital dyserythropoietic anaemia
•	 Neutropenia:
—​	 Severe congenital neutropenia
—​	 Cyclic neutropenia
•	 Thrombocytopenia:
—​	 Thrombocytopenia with absent radii (TAR) syndrome
—​	 Congenital amegakaryocytic thrombocytopenia (CAMT)
section 22  Haematological disorders
5326
BMF is now known (Table 22.5.1.1). Advances in understanding
the genetics have provided valuable insight not only into the patho-
physiology of these syndromes, but also into normal haematopoi-
esis. Two of the best characterized syndromes are Fanconi’s anaemia
(FA) and dyskeratosis congenita (DC). These two syndromes are
frequently associated with generalized BMF/​AA and are discussed
in some detail in the following sections. By contrast, there are sev-
eral well-​recognized disorders where the majority of patients have
much more specific abnormalities affecting a single cell type, such
as impaired red cell production in Diamond–​Blackfan anaemia
(DBA) and congenital dyserythropoietic anaemia (CDA); impaired
neutrophil production in Shwachman–​Diamond syndrome (SDS);
and reduced platelet production in thrombocytopenia with absent
radii (TAR) syndrome and congenital amegakaryocytic thrombo-
cytopenia (CAMT). Each of these conditions is reviewed in this
chapter.
Fanconi anaemia
Aetiology
The clinical syndrome of FA (Fig. 22.5.1.1) was first described by
Guido Fanconi in 1927. The vast majority of cases are inherited in
an autosomal recessive manner although occasional families dem-
onstrate an X-​linked mode of inheritance. FA is characterized by
progressive BMF and an increased predisposition to malignancy, es-
pecially acute myeloid leukaemia (AML).
Pathogenesis and pathology
The hallmark of the cellular abnormality in FA is genomic in-
stability. This is manifest as a high frequency of spontaneous
chromosomal breakage and hypersensitivity to DNA cross-​linking
agents such as diepoxybutane and mitomycin C (MMC) (Fig.
22.5.1.2). Indeed, the most useful screening test for FA remains
the demonstration of increased chromosomal breakage in FA cells
compared with normal controls after exposure to diepoxybutane/​
mitomycin C. This ‘FA cell phenotype’ has facilitated the identifica-
tion of the complex genetics of this disease. There are currently 18
subtypes (or complementation groups) each caused by a mutation
in one of the following genes: FANCA, FANCB, FANCC, FANCD1,
FANCD2, FANCE, FANCF, FANCG, FANCI, FANCJ, FANCL,
FANCM, FANCN, FANCO, FANCP, FANCQ, FANCS, and FANCT.
In vitro gene transfer studies confirm the important role played by
these genes: introduction of the relevant wild-​type FA gene into
FA human lymphoid and haematopoietic cells normalizes their re-
sponse to mitomycin C and chromosomal breakage and restores
their growth to normal.
The proteins encoded by the FA genes participate in a complex
network essential for normal DNA repair. Nine of the proteins
(FANCA, FANCB, FANCC, FANCE, FANCF, FANCG, FANCL,
FANCM, and FANCT) form the FA core complex which is necessary
for the activation of a second complex, the FANCI–​FANCD2 pro-
tein complex, to a monoubiquitinated form (FANCI–​FANCD2-​Ub)
(Fig. 22.5.1.3). This FANCI–​FANCD2-​Ub complex then interacts
with DNA repair proteins (including BRCA2, BRCA1, and RAD51)
to bring about the repair of the DNA damage. These observations
have linked the FA proteins with BRCA1 and BRCA2 (FANCD1) in
a DNA damage response pathway ‘The FA/​BRCA pathway’. Indeed,
FA-​D1 patients have biallelic mutations in BRCA2. In cells lacking
BRCA2, repair of DNA damage by homologous recombination is
inaccurate and such cells are hypersensitive to DNA cross-​linking
agents. More recently, it has been shown that two other FA genes
(FANCJ and FANCN) encode BRCA1 and BRCA2 partners, BRIP1
and PALB2, respectively. Mutations in these two FA genes have a
specific link to cancer predisposition. Biallelic BRCA2 (FA-​D1) and
PALB2 (FA-​N) mutations, in contrast to mutations in the other FA
genes, are associated with high risks of solid childhood malignancies
(e.g. Wilms’ tumour and medulloblastoma) while heterozygous mu-
tations specifically in BRCA1 (FA-​S), (FA-​D1), PALB2 (FA-​N), and
BRIP1 (FA-​J) confer an increased risk of breast cancer. These differ-
ences highlight the complex relationship between the FA proteins
and their interactions with other molecules both at the molecular
and clinical level.
It is now clear that the clinical and biological defects in FA overlap
with those of several other chromosomal breakage syndromes,
including Seckel’s syndrome and the Nijmegen breakage syndrome
(NBS). Consistent with this, considerable evidence indicates that
the causative genes are involved in some of the same pathways. For
example, activation of the FA–​BRCA pathway in response to DNA
damage (e.g. replication fork arrest) involves ATR (ataxia telangi-
ectasia), the gene for which is mutated in a subset of patients with
Seckel’s syndrome (Fig. 22.5.1.3). ATR appears to directly regulate the
Table 22.5.1.1  Characteristics of the BMF syndromes
FA
DC
SDS
DBA
CDA
TAR
SCN
IAA
Other#
Inheritance pattern
AR, XLR
XLR, AR, AD
AR
AD, XLR
AR, AD
AR
AD, AR
?
AD and AR
Somatic abnormalities
Yes
Yes
Yes
Yes
Rare
Yes
Rare
?None
Yes
Bone marrow failure
AA (>90%)
AA (c.80%)
AA (20%)
RCA
Eryth
Megs
Neutropenia
Yes (100%)
Yes
Short telomeres
Yes
Yes
Yes
?
?
?
?
Yes
No
Malignancy
Yes
Yes
Yes
Yes
?No
?No
Yes
Yes
?
Chromosome instability
Yes
Yes
Yes
?
?
?
?
Yes
?
Genes identified
18
12
1
13
4
1
5+
a
3
AA, aplastic anaemia; AD, autosomal dominant; AR, autosomal recessive; CDA, congenital dyserythropoietic anaemia; DBA, Diamond–​Blackfan anaemia; DC, dyskeratosis congenita;
FA, Fanconi’s anaemia; IAA, idiopathic aplastic anaemia; Eryth, ineffective erythropoiesis; Megs, low megakaryocytes; RCA, red cell aplasia; SCN, severe congenital neutropenia; SDS,
Shwachman–​Diamond syndrome; TAR, thrombocytopenia with absent radii; XLR, X-​linked recessive. Other#, patients in whom genetic defects have been identified (e g. in SRP72,
ERCC6L2) but who do not fit into previously recognized categories.
a Heterozygous mutations in TERC and TERT are risk factors for some cases of AA.
22.5.1  Inherited bone marrow failure syndromes
5327
FA pathway as it is required for the monoubiquitination of FANCD2
and FANCI. It is interesting to note that ATR-​Seckel cells also exhibit
defects in FANCD2 monoubiquitination and that NBS cells with
(mutations in NBN) show defects in FANCD2 monoubiquitination.
Although there has been considerable progress in elucidating the
molecular pathogenesis of FA in recent decades, our understanding
remains incomplete. Studies in animal models and human samples
show that FA cells display other abnormalities as well as DNA repair
defects, including hypersensitivity to oxygen, accelerated telomere
shortening, abnormal cell cycle kinetics, upregulation of p53, and
overactivation of the mitogen-​activated protein kinase (MAPK)
pathways leading to overproduction of tumour necrosis factor-​
α (TNFα). Mouse models of FA have shown that haematopoietic
progenitors are hypersensitive to TNFα and interferon-​γ (IFNγ).
There is some evidence that this effect is mediated by fas-​induced
apoptosis and this, perhaps together with aldehyde-​induced
genotoxicity affecting haematopoietic stem cells, may explain the
development of progressive BMF in FA. There is also evidence that
some of the features of the pathophysiology of BMF are common
to both FA and idiopathic AA. Patients with idiopathic AA often
have increased IFNγ levels and short telomeres are a feature in both
conditions and in DC.
(a1)
(b1)
(c1)
(a2)
(b2)
(c2)
(a3)
(b3)
(c3)
Fig. 22.5.1.1  Fanconi anaemia (FA). (a) Photographs of patients with FA (a1–​a3) with small mouth and chin (‘Fanconi facies’).
(b) Abnormalities of pigmentation (hyper-​ and hypopigmentation) on the abdomen (b1) with a close up (b2) of a café-​au-​lait
spot and a hypopigmented patch. The bottom photograph (b3) shows the back of an FA patient demonstrating lumbar scoliosis.
(c) Hands/​forearms of FA children showing hypoplastic thumbs (c1), rudimentary (‘dangling’) thumbs (c2), and a radiograph (c3)
showing rudimentary thumb (skeletal) development.
section 22  Haematological disorders
5328
ctb
ctb
(a)
(b)
mci
ctg
ctb
ctb
ring/tri
qr
tri
tri
Fig. 22.5.1.2  (a, b) Chromosomal abnormalities seen in FA lymphocytes following incubation with diepoxybutane. ctb, chromatid break; ctg,
chromatid gap; mci, multiple chromatid interchanges (complex rearrangement); tri, triradial; qr, quadriradial.
Courtesy of Nicola Foot, Hammersmith Hospital.
Constitutional biallelic mutations in Fanconi anaemia genes
MAPKs
Altered
oxidative
stress
response
Other
aberrations
e.g. defective
telomere
maintenance
Enviromental
factors
Sunlight
Smoking
Infections
Alcohol
FA phenotype
Genomic instability
Altered cell checkpoints and survival
DNA repair foci
FA core
Complex
ABCE
FGLM
I
I
UB
BRCA2
(D1)
BRCA1
ERCC4
(Q)
RAD51
(0)
Altered DNA damage response
‘FA-BRCA pathway’
ATR
DNA damage
TNF-α
D2
D2
BRIP1
(J)
PALB2
(N)
SLX4
(P)
Fig. 22.5.1.3  Schematic representation of the FA–​BRCA pathway and related networks. The diagram shows that
the constitutional mutations in FA cells lead to aberration of the FA–​BRCA pathway, abnormal handling of oxidative
stress, aberrant activation of mitogen-​activated protein kinases (MAPKs), defective telomere maintenance, as well
as other biological aberrations. The net impact of these is increased genomic instability and altered cell survival/​
checkpoints. The diagram also highlights the potential role of environmental factors such as smoking in adding
to the effect of the FA mutations. Within the FA–​BRCA pathway, the proteins shown in yellow are those mutated
in different FA patients. The FA core complex consists of nine FA-​proteins (A, B, C, E, F, G, L, M, and T) and this,
together with ATR (ataxia telangiectasia and RAD3-​related protein), is essential for activation (ubiquitination) of the
I–​D2 complex after DNA damage. Activated I–​D2-​Ub translocates to DNA repair foci where it associates with other
DNA damage response proteins including BRCA2, RAD51, and SLX4 and participates in DNA repair. TNF, tumour
necrosis factor; Ub, ubiquitination.
22.5.1  Inherited bone marrow failure syndromes
5329
Clinical features
As well as progressive BMF, most patients with FA have at least one
somatic abnormality (Table 22.5.1.2). The most common clinical
signs are skeletal abnormalities, particularly absent thumbs and/​or
radial hypoplasia; skin lesions, particularly café-​au-​lait spots; short
stature; and microphthalmia (Fig. 22.5.1.1). However the devel-
opment of almost all organs and tissues may be affected, including
genitourinary abnormalities, such as underdeveloped gonads and
horseshoe kidneys; and gastrointestinal, cardiac, and neurological
anomalies (Table 22.5.1.2). There is marked variation between pa-
tients both in the pattern of somatic abnormalities and in the course
of the disease. It is important to note that almost one-​third of pa-
tients have no physical abnormalities, making the diagnosis of FA on
clinical grounds difficult and unreliable.
Although the somatic abnormalities of FA will be evident at
birth, the blood count in neonates is usually normal. Pancytopenia
develops insidiously as BM cellularity falls and presents in most
cases between the ages of 5 and 10 years (median age 7 years), al-
though late presentation in adolescents or adults sometimes occurs.
Typically the first haematological signs of BMF in FA are thrombo-
cytopenia and anaemia and the neutrophil count is often maintained
in the early stages. As BMF progresses, the main causes of death are
fatal haemorrhage or infection due to progressive BMF.
As patients survive longer, the increased risk of leukaemia and
other malignancies in FA is becoming more evident. By the age of
40 years, the cumulative incidence of haematological malignancy
is 33% and of nonhaematological malignancies is 28%. The most
frequent haematological malignancy is acute myeloid leukaemia.
Amongst the nonhaematological cancers the most frequent are hep-
atic tumours and squamous cell carcinoma of the vulva, oesophagus,
head, and neck. Anecdotal evidence, supported by recent data from
the German FA registry, suggests that malignancies occur mainly
in patients with late-​onset BMF and longer survival, with a me-
dian age of 13 years for leukaemia and 25 years for solid tumours.
Long-​term follow-​up of FA patients treated by haematopoietic
stem cell transplantation (HSCT) also shows a higher incidence of
nonhaematological malignancies in patients with FA than in pa-
tients with other types of BMF who have undergone HSCT.
An additional feature of FA is the occurrence, in a small pro-
portion of patients of somatic mosaicism. This occurs when the
‘pathogenic’ allele reverts to ‘wild type’ in a single haematopoietic
(somatic) cell, that is, the reverted cell effectively becomes a ‘het-
erozygous cell’, which would be expected to have a growth/​survival
advantage in the background of FA cells. Such FA mosaic patients
may then experience an improvement in their blood count and
a return of the diepoxybutane/​mitomycin C test to normal indi-
cating that a single haematopoietic stem cell may be sufficient to
restore adequate haematopoiesis. This process can be viewed as a
natural form of haematopoietic stem cell gene therapy suggesting
that therapeutic gene therapy might be a valuable approach for
BMF in FA.
Treatment
The most common indication for treatment in FA is BMF. Although
HSCT is the only curative treatment, anabolic steroids, such as
oxymetholone and danazol, achieve useful haematological responses
in 50 to 70% of patients by delaying the need for red cell or platelet
transfusions or HSCT. Unfortunately, anabolic steroids often have
serious side effects, such as liver dysfunction, and so patients need to
be monitored closely and many patients become refractory. HSCT
protocols have to be adapted specifically for FA patients because of
their hypersensitivity to irradiation and alkylating agents. Reduced
doses of cyclophosphamide (20–​40 mg/​kg) are typically used as well
as protocols that avoid the need for irradiation or use lower doses
than usual for HSCT by using immunosuppressive drugs such as
fludarabine. Gene therapy protocols are also being investigated in
FA as a means of ameliorating the BMF.
Prognosis
Despite improvements in supportive care and HSCT and a dramatic
increase in our understanding of the pathogenesis of FA, the prog-
nosis remains poor. By the age of 40 years, patients with FA have
a cumulative incidence of BMF of approximately 90% as well as a
cumulative incidence of haematological malignancy of 33% and
of nonhaematological malignancies of 28%. Median survival in
the largest reported series to date (International Fanconi Anaemia
Registry) was 24  years. Major challenges in the future include
developing better treatment of malignancies and the management
of complications (e.g. pulmonary disease) in adulthood.
Dyskeratosis congenita
Aetiology
The classical DC clinical triad of abnormal skin pigmentation, nail
dystrophy, and mucosal leucoplakia (Fig. 22.5.1.4) was first de-
scribed by Jacobi in 1906 and Zinsser in 1910. DC is now known to
affect almost all tissues but the principal cause of early mortality is
BMF. Like FA, patients with DC also have a predisposition to ma-
lignancy. DC is now regarded as one of a group of disorders known
as ‘telomeropathies’, characterized by abnormal telomere function.
It is a very heterogeneous disorder, both clinically and genetically;
Table 22.5.1.2  Somatic abnormalities in FA
Abnormality
Percentage of patients
Skeletal (radial ray, vertebral, scoliosis, rib)
71
Skin pigmentation (café-​au-​lait, hyper-​ and
hypopigmentation)
64
Short stature
63
Eyes (microphthalmia)
38
Renal and urinary tract
34
Male genital
20
Mental retardation
16
Gastrointestinal (e.g. anorectal, duodenal atresia)
14
Heart
13
Hearing
11
Central nervous system (e.g. hydrocephalus,
septum pellucidum)
  8
No abnormalities
30
Source data from Auerbach, A., Buchwald, M. & Joenie, H. (2001) In: The metabolic and
molecular bases of inherited diseases (eds. C. Scriver, W. Sly, B. Childs, A. Beaudet, D. Valle,
K. Kinzler and B. Vogelstein). McGraw Hill, New York.
section 22  Haematological disorders
5330
X-​linked recessive, autosomal dominant, and autosomal recessive
forms of the disease are recognized.
Pathogenesis and pathology
The genes responsible for approximately 70% of cases have now been
identified. The DKC1 gene, which is responsible for all cases of X-​
linked DC, was the first to be identified, in 1998. This gene is highly
conserved and encodes the protein dyskerin which is a core compo-
nent of telomerase. Mutations in DKC1 also underlie some cases of
a severe, multisystem form of DC, known as Hoyeraal–​Hreidarsson
syndrome (HH).
Autosomal dominant DC is heterogeneous and to date, heterozy-
gous mutations in three genes (TERC, TERT, and TIN2) have been
characterized. Like dyskerin, TERC (telomerase RNA component)
and TERT (telomerase reverse transcriptase) are key core compo-
nents of telomerase, a ribonucleoprotein essential for maintaining
telomere length in rapidly dividing cells such as haematopoietic
cells (Fig. 22.5.1.5). TERT catalyses the addition of repeats and
TERC acts as the template. In cells which lack telomerase, the telo-
meres shorten with each successive round of replication until they
reach a critical length and the cells enter senescence. Accordingly,
patients with DKC1 and TERC gene mutations have very short
telomeres compared to age-​matched controls. Families with het-
erozygous TERC and TERT gene mutations frequently exhibit the
phenomenon of genetic anticipation whereby the disease mani-
fests at a much younger age and with greater severity in the off-
spring of an affected parent despite the same pathogenic telomerase
mutation.
Autosomal dominant DC can also be caused by mutations in the
TIN2 protein, a component of the shelterin complex which has at
least three effects on telomeres. Firstly, the complex determines the
structure of the telomeric terminus; secondly, it is implicated in the
generation of t-​loops; and thirdly, it controls the synthesis of telo-
meric DNA by telomerase. Heterozygous TIN2 mutations have now
been identified in a subset of patients with DC, HH, AA, and Revesz’s
syndrome (bilateral exudative retinopathy, BMF, nail dystrophy, fine
hair, cerebellar hypoplasia, and growth retardation). Patients with
TIN2 mutations tend to have very short telomeres and severe dis-
ease. Almost all affected patients have de novo TIN2 mutations in
contrast to those with TERC or TERT mutations.
Autosomal recessive DC is also a heterogeneous disease. To
date, mutations in eight genes have been identified: NOP10, TERT,
NHP2, TCAB1, RTEL1, CTC1, PARN, and USB1, all of which re-
sult in reduced telomere length apart from USB1. Homozygous
NOP10 mutations are associated with reduced telomere length
and reduced TERC levels. Biallelic TERT mutations, in contrast to
heterozygous TERT mutations, also have greatly reduced telomere
length and telomerase activity, as do patients with biallelic mu-
tations in NHP2. Both NOP10 and NHP2 are components of H/​
ACA ribonucleoprotein complex (H/​ACA RNP). This complex is
comprised of a RNA molecule and four highly conserved proteins,
dyskerin, GAR1, NOP10, and NHP2, involved in ribosome biogen-
esis, pre-​mRNA splicing, and telomere maintenance. To date, mu-
tations have been identified in all components of this H/​ACA RNP
complex in patients with DC except for GAR1. Compound hetero-
zygous mutations in the TCAB1 gene, which encodes a telomerase
holoenzyme, cause short telomeres by disrupting telomere localiza-
tion to Cajal bodies, resulting in misdirection of telomerase RNA
to nucleoli. Biallelic mutations in the RTEL1 (regulator of telomere
length 1) gene result in telomeres which are not only very short,
but also have qualitative defects and are associated with severe DC
presumably because RTEL1 is a helicase with an important role in
homologous recombination and telomere maintenance. Biallelic
mutations in the CTC1 (conserved telomere maintenance compo-
nent 1) gene were initially identified in the pleotropic syndrome
Coats plus (characterized by retinopathy, intracranial calcifications
and cysts, osteopenia, and gastrointestinal abnormalities) but are
now known to also be a rare cause of DC. Finally, patients with USB1
mutations appear to be a different biological subtype of DC; clinic-
ally, affected individuals have poikiloderma with neutropenia and
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 22.5.1.4  Photographs of patients with DC showing abnormal skin pigmentation (a, b, c), nail dystrophy (d, e), and leucoplakia of the tongue (f).
22.5.1  Inherited bone marrow failure syndromes
5331
Rothmund–​Thomson syndrome and, unlike other subtypes of DC,
telomere length is normal.
Clinical features
In addition to the typical mucocutaneous triad of abnormal skin
pigmentation, nail dystrophy, and leucoplakia, patients with DC
may also have a variety of noncutaneous (dental, gastrointestinal,
genitourinary, neurological, ophthalmic, pulmonary, and skeletal)
abnormalities (Fig. 22.5.1.4 and Table 22.5.1.3). Changes in skin
pigmentation and in the nails generally appear first, usually by
the age of 10 years. In most cases, these changes are followed by
BMF which typically develops below the age of 20  years and
80 to 90% of patients will have BM abnormalities by the age of
30 years. The minimal clinical criteria for diagnosis of DC are the
presence of at least two out of the four major features (abnormal
skin pigmentation, nail dystrophy, leucoplakia, and BMF) and
two or more of the other somatic features known to occur in DC.
The main causes of death in DC are BMF/​immunodeficiency
(c.60–​70%), pulmonary complications (c.10–​15%), and malig-
nancy (c.10%).
It is important to note that the diagnosis of DC on clinical grounds
can be very difficult because of the heterogeneity of the condition.
BM abnormalities may appear before the mucocutaneous manifest-
ations, leading to patients being misdiagnosed as having ‘idiopathic
aplastic anaemia’. The clinical presentation in patients with TERT
mutations is highly variable ranging from near DC phenotype to just
AA. Heterozygous TERC mutations may occur in patients with AA
and in patients with myelodysplastic syndrome (MDS) as well as in
DC. Furthermore, heterozygous mutations in TERT and TERC have
been identified in some patients with idiopathic pulmonary fibrosis,
liver disease, and leukaemia.
Treatment and prognosis
As with FA, anabolic steroids (oxymetholone and danazol) produce
a valuable haematological response in around two-​thirds of patients
by delaying the need for red cell or platelet transfusions or HSCT
for several years. Patients with DC may respond to a dose as low as
0.25 mg oxymetholone/​kg per day and this can be increased, if neces-
sary to 2 to 5 mg/​kg per day, although close monitoring is essential in
view of the side effects, particularly liver toxicity. However, the only
long-​term cure for the BMF is HSCT. Transplant-​related mortality is
higher in DC than in other BMF syndromes, mainly because of pul-
monary and vascular complications, probably due to the underlying
telomere defect. HSCT protocols have been adapted to try to re-
duce HSCT-​related toxicity, for example, by using nonmyeloablative
fludarabine-​based protocols as for FA, although the long-​term bene-
fits of this approach are not yet clear. The main causes of death in DC
are BMF and HSCT-​related toxicity.
Shelterin
Tankyrase
TRF2
RAP1
USB1
snRNA processing
TCAB1
Cajal body
Telomerase
Genes mutated in dyskeratosis congenita and related
bone marrow failure syndromes—“the telomereopathies”
Helicases
RTEL1
POT1
TPP1
TERT
TERC
Dyskerin
NOP10
GAR1
NHP2
1
Capping complex
TEN1
CTC1
STN1
TIN2
TRF1
Fig. 22.5.1.5  Schematic representation of complexes involved in telomere maintenance. The telomerase complex includes TERC, TERT,
dyskerin, NOP10, NHP2, and GAR1. The shelterin complex includes the six proteins TRF1, TRF2, TPP1, POT1, RAP1, and TIN2. The telomere
capping (CST) complex is composed of CTC1, STN1, and TEN1. Protein/​RNA names indicated by red arrows are mutated in DC and related
disorders: Hemizygous DKC1 (dyskerin) mutations are observed in X-​linked DC and HH. Heterozygous TERC mutations are associated with DC,
AA, MDS, AML, and pulmonary fibrosis. Heterozygous TERT mutations are responsible for some cases of AA, DC, MDS, AML, and pulmonary/​liver
fibrosis. Biallelic mutations in TERT can cause classic DC and HH. Heterozygous TIN2 mutations have been observed in DC, AA, HH, and Revesz’s
syndrome. Biallelic NOP10, NHP2, TCAB1, USB1, and CTC1 mutations have been seen in autosomal recessive DC. Biallelic RTEL1 and PARN
mutations are observed in autosomal recessive HH.
section 22  Haematological disorders
5332
Shwachman–​Diamond syndrome
Aetiology
SDS is an autosomal recessive disorder characterized by exocrine
pancreatic insufficiency, BMF and, in some patients, a variety of
other somatic abnormalities, particularly involving the skeletal
system. SDS, which was first reported independently by Shwachman
and colleagues and Bodian and colleagues in 1964, is an example of a
group of ribosome biogenesis disorders known as ‘ribosomopathies’.
The risk of leukaemia, as in FA and DC, is increased.
Pathology and pathogenesis
More than 90% of SDS patients have biallelic mutations in the SBDS
gene which encodes for a protein which plays an important role in
the maturation of the large (60S) ribosomal protein (RP) subunit
(Fig. 22.5.1.6).
Clinical features
Exocrine pancreatic insufficiency and BMF are the hallmarks of
SDS, occurring in all patients. Signs of pancreatic insufficiency, typ-
ically failure to thrive and malabsorption, are usually apparent early
in infancy, although improvement in pancreatic function with age
has been reported in a subset of SDS patients. Associated skeletal
abnormalities are common in SDS, including short stature (c.70%),
metaphyseal dysostosis (75%), rib and thoracic cage abnormalities,
hypertelorism, syndactyly, cleft palate, and dental dysplasia. Other
abnormalities can include an ichthyotic skin rash (c.60%) or skin
Table 22.5.1.3  Somatic abnormalities in dyskeratosis congenita
Abnormality
Percentage of patients
Abnormal skin pigmentation
89
Nail dystrophy
88
BMF
85.5
Leucoplakia
78
Epiphora
30.5
Learning difficulties/​developmental delay/​mental
retardation
25.4
Pulmonary disease
20.3
Short stature
19.5
Extensive dental caries/​loss
16.9
Oesophageal stricture
16.9
Premature hair loss/​greying/​sparse eyelashes
16.1
Hyperhidrosis
15.3
Malignancy
9.8
Intrauterine growth retardation
7.6
Liver disease/​peptic ulceration/​enteropathy
7.3
Ataxia/​cerebellar hypoplasia
6.8
Hypogonadism/​undescended testes
5.9
Microcephaly
5.9
Urethral stricture/​phimosis
5.1
Osteoporosis/​aseptic necrosis/​scoliosis
5.1
Deafness
0.8
Ribosomal DNA
45S rRNA
30S
18S
Nucleus
Cytoplasm
40S subunit
80S ribosome
60S subunit
DBA RPS7,RPS10,
RPS17, RPS19, RPS24,
RPS26, RPS29
DBA RPL5,RPL11,
RPL15, RPL26, RPL35a
5q- syndrome
RPS14
SDS
SBDS
5.8S
28S
5S
32S
Fig. 22.5.1.6  Schematic diagram showing scheme of ribosomal (r)RNA processing in human cells
and the points at which this is possibly disrupted in the different BMF syndromes. The rRNAs are
transcribed by RNA polymerase I as a single precursor transcript (45S rRNA). The 45S rRNA is then
processed to 18S, 5.8S, and 28S rRNAs. The 18S is a component of the 40S ribosomal subunit. The
5.8S and 28S together with 5S (synthesized independently) are components of the 60S ribosomal
subunit. The 40S and 60S subunits are assembled to form the 80S ribosome. The processing
steps affected in DBA (heterozygous mutations in RPS7, RPS10, RPS17, RPS24, RPS26, RPS29,
RPL5, RPL11, RPL26, and RPL35a), 5q–​ syndrome (haploinsufficiency of RPS14) and SDS (biallelic
mutations in SBDS) are indicated by the different coloured stars.
22.5.1  Inherited bone marrow failure syndromes
5333
pigmentation, hepatomegaly/​protuberant abdomen, and ptosis.
BMF usually presents as an isolated neutropenia but approximately
20% of patients have pancytopenia. AML or MDS develops in ap-
proximately 25% of patients and appears to be more common in
affected males. The age at which leukaemia develops varies widely,
from 1 to 43 years.
The diagnosis of SDS is made by a combination of clinical features,
family history, the clinical combination of exocrine pancreatic insuf-
ficiency with BMF, and, in recent years, mutational analysis of the
SBDS gene. The main differential diagnosis is Pearson syndrome
which also presents with exocrine pancreatic insufficiency and BMF
but is distinguished by the presence of a mitochondrial DNA dele-
tion and characteristic erythroid abnormalities in the BM.
Treatment and prognosis
The malabsorption in patients with SDS responds to treatment
with oral pancreatic enzymes. Isolated neutropenia is usually man-
aged by measures to prevent, or promptly treat, infection together
with injections of granulocyte colony-​stimulating factor (G-​CSF)
to boost neutrophil production if required. As for FA and DC,
oxymetholone may be useful for delaying the need for red cell and/​
or platelet-​transfusions or HSCT for patients with severe, symptom-
atic pancytopenia. The development of leukaemia usually has a poor
prognosis as the response to standard AML chemotherapy is poor.
HSCT is the only curative treatment for BMF and often the best op-
tion for AML, using conditioning regimens that include fludarabine.
It is noteworthy that an increased frequency of nonhaematological
malignancies has not been reported in SDS.
Diamond–​Blackfan anaemia
Aetiology
DBA is a rare inherited form of red cell aplasia (five cases/​million
livebirths) which usually presents in early infancy. The vast ma-
jority of cases exhibit an autosomal dominant pattern of inherit-
ance and at least half of these cases are associated with heterozygous
mutations of one of the RP genes. Like SDS, DBA is considered a
ribosomopathy and many patients have associated somatic anom-
alies, particularly affecting the skeleton. The cause of the remaining
autosomal dominant cases remains to be determined. Occasional
families with a DBA-​like pattern of disease and X-​linked inheritance
have been found to be due to mutations in the haematopoietic tran-
scription factor gene GATA1. Occasional cases of AML, MDS, and
progressive BMF have been reported.
Pathology and pathogenesis
Mutations in at least 12 different RP genes have now been reported
in patients with DBA and provide the genetic basis for around two-​
thirds of cases. The gene mutations affect either the small RP unit
(RPS7, RPS10, RPS17, RPS19, RPS24, RPS26, and RPS29) or the
large ­ribosomal subunit (RPL5, RPL11, RPL15, RPL26, and RPL35a) ­
(Fig. 22.5.1.6). This suggests that the primary defect in DBA is
­defective ribosome biogenesis, which then leads to other bio-
logical defects including increased apoptosis, upregulation of p53,
and ­defective erythroid progenitor development. The exact mech-
anism by which RP protein mutations cause selective impairment
of red cell production is unclear. This question is further com-
plicated by evidence from some studies that defects in other
haematopoietic lineages may be present in some patients with re-
fractory DBA. Recently, constitutional hemizygous GATA1 muta-
tions have been identified in rare patients with ‘DBA-​like’ disease;
in these cases the mechanism of disease is likely to be completely
different.
Clinical features
DBA usually presents in early infancy, with features of anaemia such
as pallor or failure to thrive. Data from the DBA Registry of North
America found a median age at presentation of 8 weeks with 93%
of patients presenting in the first year of life. Occasional cases pre-
sent before birth (as fetal anaemia) or in adulthood. The hallmark of
classical DBA is reduced numbers of BM erythroid precursors with
resultant normochromic macrocytic anaemia. Around 50% of pa-
tients have associated somatic abnormalities typically affecting the
craniofacial bones (high-​arched palate, cleft lip, hypertelorism, and
flat nasal bridge) and upper limb, especially the thumb. Around 30%
of patients with DBA are below the third centile for height. There
is also an increased frequency of cardiac and urogenital malforma-
tions. The anaemia at presentation is often severe (haemoglobin <50
g/​litre). The presence of severe anaemia together with a low reticu-
locyte count and normal platelet and leucocyte counts is typical.
The diagnosis is made when these features are accompanied by a
normocellular BM with a selective loss or severe reduction in im-
mature erythroid cells while other cell types are normal. Red cell
adenosine deaminase is elevated in most cases and may provide
useful supportive evidence for a diagnosis of DBA but is not specific.
Where available, mutational analysis of ribosomal protein genes is
useful to confirm the diagnosis and establish which family members
are affected.
The genotype–​phenotype relationships between the different
mutations and disease manifestations remain to be clarified al-
though recent studies suggest that patients with RPL5 mutations
tend to have multiple physical abnormalities, including craniofacial,
thumb, and heart anomalies, whereas isolated thumb malforma-
tions are predominantly seen in patients with heterozygous RPL11
mutations. Further collaborative studies will be needed to validate
these findings and identify other associations that are likely to be
important both for genetic counselling and providing mechanistic
insight into the pathogenesis of DBA. As for the other inherited
BMF syndromes, patients with DBA appear to be at increased risk of
developing AML and MDS, although relatively few cases have been
reported to date. DBA has also been reported to evolve to AA in
some patients. Even in patients without AA, BM cellularity is often
reduced and such patients are more likely to develop neutropenia
and/​or thrombocytopenia.
Treatment and prognosis
Most patients with DBA require treatment in order to maintain a
satisfactory haemoglobin concentration compatible with normal
growth and development. The mainstay of treatment is oral cortico-
steroids to which approximately 80% of patients will respond ini-
tially. Steroids should be used in as low a dose as possible, or given on
alternate days, due to the toxicity of chronic steroid therapy. Patients
with DBA who are steroid refractory, or require very high doses
to maintain a satisfactory haemoglobin concentration, are treated
section 22  Haematological disorders
5334
with regular red cell transfusions, usually at 4-​ to 5-​weekly inter-
vals. Infants under the age of 12 months are also treated with regular
red cell transfusions to avoid steroid toxicity in this age group. The
main complication of chronic red cell transfusion therapy is iron
overload and all regularly transfused patients with DBA also require
treatment with an iron-​chelating agent. The only curative option for
DBA is HSCT. Patients and families need to weigh up the relative
benefits and risks of HSCT versus life-​long transfusion/​chelation.
Data from the DBA of North America Registry show that the actu-
arial survival rates at ages over 40 years were 100% for those in sus-
tained remission, 87% for steroid-​maintained patients, and 57% for
transfusion-​dependent patients. Of the 36 deaths they reported, 25
were treatment-​related, including 14 patients who died as a result of
HSCT-​related complications.
Congenital dyserythropoietic anaemias
Aetiology
The CDAs represent a heterogeneous group of disorders of erythro-
poiesis characterized by anaemia and ineffective erythropoiesis.
First by Crookston and colleagues in 1966, these disorders are clas-
sified into three main types (I, II, and III) of which type II is the
commonest. Type I and type II CDA are inherited in an autosomal
recessive fashion while type III, which is rare, is autosomal dom-
inant. The genes responsible for most cases of type II and some cases
of type I and type III CDA have been identified. Additional vari-
ants of CDA are described and this classification system is likely
to be further refined as the genetic basis of these disorders is un-
covered. Mutations in the transcription factor genes GATA1 and
KLF1 have recently been identified as the molecular basis for some
of these cases.
CDA type I
The majority of cases of CDA type I are due to mutations in the
CDAN1 gene which encodes a ubiquitous protein codanin-​1 likely
to be involved in intracellular transport. Occasional cases due to
mutations in the C15ORF41 gene which encodes a protein of un-
known function have also been described. Patients usually present
with splenomegaly and mild to moderate macrocytic anaemia
(c.70–​120 g/​litre). Some patients also exhibit nonhaematological
features such as skeletal abnormalities and/​or abnormal skin pig-
mentation. The diagnosis is made after expert review of the blood
film and BM, including electron microscopy. The presence of inter-
nuclear chromatin bridging and binuclearity of the erythroblasts
is characteristic but not specific and the defining feature on elec-
tron microscopy is the presence of a spongy (‘Swiss cheese’) ap-
pearance of the heterochromatin in the majority of erythroblasts.
Where possible, the diagnosis should be confirmed by molecular
genetic analysis because the morphological abnormalities may be
confused with other inherited red cell disorders, such as pyruvate
kinase deficiency. Markers of haemolysis may also be present (ele-
vated lactate dehydrogenase and bilirubin). Patients with mild
anaemia do not require treatment. For those with more severe an-
aemia, the options are regular red cell transfusion with chelation or
HSCT. A proportion of patients with CDA type I respond to treat-
ment with α-​interferon, becoming transfusion independent by an
unknown mechanism.
CDA type II
Originally known as HEMPAS (hereditary erythroblastic multi­
nuclearity with a positive acidified serum lysis test), CDA type II is
now known to be caused in the vast majority of patients by mutations
in the SEC23B gene. This gene encodes a secretory pathway protein
and more than 60 different mutations have now been reported in
affected individuals. Most patients present with anaemia with or
without a variable degree of jaundice, hepatomegaly, splenomegaly,
and cirrhosis. Typically the anaemia is moderate (haemoglobin
80–​110 g/​litre) but approximately 10% of cases are transfusion de-
pendent and some cases present with anaemia at or before birth.
Peripheral blood red cell morphology is usually unremarkable but
the BM findings are characteristic with more than 10% binucleate or
multinucleate erythroblasts and peripheral cisternae (‘double mem-
brane’) beneath the plasma membrane of erythroid cells on electron
microscopy. The diagnosis is now made by genetic analysis of cases
with typical clinical and BM findings; the acidified lysis test (Ham’s
test) is no longer used. Most patients with CDA type II are not trans-
fusion dependent. For those that are splenectomy may lead to trans-
fusion independence. Patients with CDA type II, whether transfused
or not, have an increased risk of iron overload and should be moni-
tored from the age of 10 years as iron chelation therapy may be re-
quired. HSCT is the only cure and may be valuable in severe cases.
CDA type III
This is extremely rare. Although usually autosomal dominant, oc-
casional sporadic cases are reported. The autosomal dominant form
of CDA type III is caused by mutations in the KIF23 gene which en-
codes a protein essential for cytokinesis. The molecular basis for the
sporadic form of CDA type III is unknown. The diagnosis is usually
made as a result of the characteristic giant multinucleate erythro-
blasts seen by in the BM by light microscopy. Treatment is with red
cell transfusion if required.
Congenital and cyclic neutropenias
Aetiology
Severe congenital neutropenia (SCN) is clinically and genetically
heterogeneous. The majority of cases are autosomal dominant and
are due to mutations in the ELANE gene which may also present
as cyclic neutropenia. Occasional autosomal dominant cases are
due to mutations in the GFI1 gene while the genetic basis of other
cases remains to be determined. Around 30 to 40% of cases of SCN,
including SDS as described earlier, are autosomal recessive. To date,
mutations in the HAX1, G6PC3, VPS45, JAGN1, and CLPB genes
have been described in these affected families. Kostmann’s syn-
drome, the first reported form of SCN, is now known to be caused
by mutations in the HAX1 gene.
Pathology and pathogenesis
The most commonly mutated gene in SCN and cyclic neutropenia is
the ELANE gene which encodes neutrophil elastase, a serine protease
synthesized mainly at the promyelocyte stage of neutrophil differen-
tiation. Mutations in ELANE lead to accumulation of nonfunctional
neutrophil elastase within the cell which triggers an unfolded pro-
tein response leading to maturational arrest. In cyclic neutropenia,
the ELANE mutations are usually clustered around the active site in
22.5.1  Inherited bone marrow failure syndromes
5335
contrast to those found in SCN although the mechanism by which
this leads to the different manifestations of SCN and cyclic neutro-
penia is unclear. The second most commonly mutated gene in SCN
is HAX1. Biallelic HAX1 mutations lead to premature cell death
suggesting that HAX1 protein, which is a critical regulator of mito-
chondrial membrane potential and cellular viability, plays a role in
apoptosis.
Clinical features
SCN usually presents at birth or the first few weeks of life due to
the combination of severe neutropenia (typically <0.2 × 109/​litre)
and bacterial infection. Affected individuals have severe, recurrent
bacterial infections throughout life leading to premature death in
childhood unless diagnosed and treated promptly. The diagnosis is
made by a combination of the clinical history, blood count, and BM
abnormalities and should be confirmed by genetic analysis as this is
essential for genetic counselling. Despite the low neutrophil count,
the haemoglobin and platelet counts are usually normal. The BM in
SCN shows maturation arrest at the promyelocyte/​myelocyte stage.
Cyclical neutropenia is characterized by a neutrophil count that usu-
ally reaches a nadir with a 21-​day periodicity. Around the nadir, pa-
tients may develop fever and mouth ulcers.
Treatment and prognosis
The mainstay of treatment for SCN is G-​CSF. The vast majority of
patients respond and will remain on life-​long treatment. It is now
clear that patients with SCN have a markedly increased risk of
developing leukaemia (AML) and MDS which increases with age,
approaching 25% by age 25 years. The majority of SCN patients who
develop AML have a mutation in the G-​CSF receptor gene (CSF3R);
however, the precise contribution of G-​CSF therapy to the develop-
ment of CSF3R mutations remains unclear. The risk of AML appears
to be lower in cyclic neutropenia. The only curative treatment for
SCN is HSCT which is a valuable option especially for patients who
become refractory to G-​CSF therapy.
Congenital thrombocytopenias
Aetiology
A number of rare BMF disorders may present with isolated
thrombocytopenia due to reduced platelet production, the best
recognized being TAR syndrome and CAMT. The inheritance of
TAR is interesting as recent data indicate that it is caused by com-
pound inheritance of a low-​frequency regulatory single nucleotide
polymorphism together with a rare null mutation in the RBM8A
gene (which encodes a subunit of the exon-​junction complex).
CAMT is genetically heterogeneous. The majority of reported cases
are autosomal recessive and caused by mutations in the MPL gene
which encodes the thrombopoietin receptor. A  subgroup of pa-
tients with CAMT exhibit an X-​linked inheritance pattern. An add-
itional syndrome with similarities to TAR has also been described,
amegakaryocytic thrombocytopenia with radio-​ulnar synostosis
(ATRUS or RUSAT syndrome). Mutations in the HOXA11 gene
and in the MECOM gene have been identified in different families
with ATRUS/​RUSAT syndrome. Some patients with mutations in
the MECOM gene do progress to BMF or AML.
TAR syndrome
This syndrome presents in the neonatal period with distinctive
skeletal abnormalities, particularly bilateral radial aplasia, to-
gether with thrombocytopenia often with associated bleeding.
Some affected individuals have additional skeletal abnormalities
(absent ulnae, absent humeri, and/​or clinodactyly) or somatic
abnormalities, such as microcephaly, hypertelorism, strabismus,
and/​or heart defects, and approximately 50% of patients have
cow’s milk intolerance. The platelet count is usually less than
50 × 109/​litre and the white cell count is elevated in more than 90%
of patients, sometimes exceeding 100 × 109/​litre and mimicking
congenital leukaemia. BM examination shows reduced or absent
megakaryocytes. Thrombopoietin receptor and thrombopoietin
levels are both normal. The mainstay of treatment for TAR is platelet
transfusion. However, although death in the neonatal period due
to intracranial haemorrhage may occur, babies that survive the
first year of life generally do well since the platelet count spon-
taneously improves and is usually maintained at low normal levels
thereafter. In contrast to CAMT (see next paragraph), BMF is not a
feature but AML has been reported suggesting that TAR syndrome
may be a preleukaemic syndrome.
CAMT
This rare disorder typically presents in infancy and is characterized by
isolated thrombocytopenia and absent or reduced megakaryocytes
in the BM. Most patients have no associated somatic abnormal-
ities, although some patients with MPL gene mutations have cen-
tral nervous system abnormalities, such as cerebral and cerebellar
hypoplasia. In contrast to TAR syndrome, approximately 50% of pa-
tients develop BMF, usually by the age of 5 years, with a hypocellular
BM, reduced haemoglobin, and reduced leucocyte count in addition
to thrombocytopenia. The treatment of choice for patients with se-
vere thrombocytopenia and/​or BMF is HSCT.
FURTHER READING
Alter BP (2014). Fanconi anaemia and the development of leukemia.
Best Pract Res Clin Haematol, 27, 214–​21.
Bessler M, et al. (2015). Inherited bone marrow failure syndromes.
In: Orkin SH, et al. (eds) Hematology of infancy and childhood, 8th
edition, pp. 182–​253. WB Saunders, Philadelphia.
Bluteau O, Sebert M, Le Blanc T, et al. (2018). A landscape of germ
line mutations in a cohort of inherited bone marrow failure patients.
Blood, 131, 717–32.
Dokal I (2014). Inherited bone marrow failure syndromes. EHA 19
Education Book, 8, 299–​308.
Dror Y, et  al. (2011). Draft consensus guidelines for diagnosis and
treatment of Shwachman–​Diamond syndrome. Ann NY Acad Sci,
1242, 40–​55.
Gerrad G, et  al. (2013). Target enrichment and high-​throughput
sequencing of 80 ribosomal protein genes to identify mutations asso-
ciated with Diamond-​Blackfan anaemia. Br J Haematol, 162, 530–​6.
Gramatges MM, Bertuch AA (2013). Short telomeres:  from
dyskeratosis congenita to sporadic aplastic anemia and malignancy.
Transl Res, 162, 353–​63.
Hauck F, Klein C (2013). Pathogenic mechanisms and clinical implica-
tions of congenital neutropenia syndromes. Curr Opin Allergy Clin
Immunol, 13, 596–​606.

22.5.2 Acquired aplastic anaemia and pure red cell
22.5.2 Acquired aplastic anaemia and pure red cell aplasia 5336 Judith C.W. Marsh, Shreyans Gandhi, and Ghulam J. Mufti
section 22  Haematological disorders
5336
Iolascon A, et al. (2013). Congenital dyserythropoietic anemias: mo-
lecular insights and diagnostic approach. Blood, 122, 2162–​6.
Kottemann MC, Smogorzewska A (2013). Fanconi anaemia and the
repair of Watson and Crick DNA crosslinks. Nature, 493, 356–​63.
Ruggero D, Shimamura A (2014). Marrow failure: a window into ribo-
some biology. Blood, 124, 2784–​92.
Russo R, et al. (2014). Retrospective cohort study of 205 cases with
congenital dyserythropoietic anemia type II: definition of clinical
and molecular spectrum and identification of new diagnostic scores.
Am J Hematol, 89, E169–​75.
Savoia A (2016). Molecular basis of inherited thrombocytopenias.
Clin Genet, 89,154–​62.
Touw I (2015). Game of clones: the genomic evolution of severe con-
genital neutropenia. Hematology Am Soc Hematol Educ Program,
2015, 1–​7.
Vlachos A, Blanc L, Lipton JM (2014). Diamond Blackfan anemia: a
model for the translational approach to understanding human dis-
ease. Expert Rev Hematol, 7, 359–​72.
22.5.2  Acquired aplastic anaemia
and pure red cell aplasia
 Judith C.W. Marsh, Shreyans Gandhi, and
Ghulam J. Mufti
ESSENTIALS
Aplastic anaemia
Aplastic anaemia (AA) is a rare bone marrow failure (BMF) disorder
characterized by pancytopenia and a hypocellular bone marrow. AA
is commonly acquired, immune mediated, and idiopathic in nature.
Activated autoreactive, cytotoxic CD8+ T cells are present but re-
cent work has shown that CD4+ T cells appear to be more important
in the pathogenesis of acquired AA. The immune nature of acquired
AA provides the rationale for one of the treatment options, namely
immunosuppressive therapy.
First-​line treatment of acquired AA is either immunosuppressive
therapy with antithymocyte globulin and ciclosporin or allogeneic
haematopoietic stem cell transplantation (HSCT). Both modal-
ities offer excellent survival. Patients treated with immunosuppres-
sive therapy (IST) are at later risk of relapse and clonal evolution to
myelodysplastic syndrome and acute myeloid leukaemia, so require
long-​term follow-​up. HSCT, if successful, is curative, but risks include
graft rejection, infections, and graft-​versus-​host disease (GVHD); re-
cent changes to the transplant conditioning regimen have reduced
the GVHD risk. Eltrombopag is now licensed to treat severe AA that
is refractory to IST with response seen in 40–50% of patients, and its
use with upfront IST is being explored further.
There is increasing awareness of inherited AA which may present
not only in childhood with somatic anomalies, but also in adulthood.
Adults with later-​onset inherited AA often lack the somatic anomalies
seen in children resulting frequently in delayed diagnosis and/​or mis-
diagnosis as acquired AA.
Pure red cell aplasia
Pure red cell aplasia (PRCA) is a form of BMF characterized by severe
anaemia with reticulocytopenia and reduced erythroid progenitors
in the bone marrow. PRCA most commonly is an acquired disorder
and immune mediated, and often occurs in association with a wide
range of conditions.
Diamond–​Blackfan anaemia, an inherited form of PRCA, is an-
other example of a ribosomopathy, and is caused by mutations in
one of many ribosomal protein genes, resulting in haploinsufficiency
(see also Chapter 22.5.1).
Introduction—​the bone marrow failure disorders
Aplastic anaemia (AA) and pure red cell aplasia (PRCA) represent
two examples of the bone marrow failure (BMF) disorders illustrated
in Fig. 22.5.2.1. BMF in this context implies a reduction in one or
more circulating blood cell types (red cells, white cells, and platelets)
due to failure of production by the bone marrow (BM), resulting in
single cytopenias such as anaemia in PRCA, or pancytopenia as in
AA. In AA, there is BMF within the haematopoietic stem cell com-
partment and in PRCA either production failure or defective matur-
ation of erythroid progenitors.
Most cases of AA and PRCA are acquired and have an immune
basis, but recently there has been increasing awareness of constitu-
tional and often inherited forms of BMF, presenting not only in chil-
dren but also later in adulthood. Often these arise in the absence of
classical somatic anomalies but instead have atypical clinical features
such as pulmonary fibrosis and cirrhosis. Examples of congenital
AA include Fanconi’s anaemia and dyskeratosis congenita (DC),
and Diamond–​Blackfan anaemia—​the inherited form of PRCA.
There is a marked overlap between AA and paroxysmal nocturnal
haemoglobinuria (PNH), which arises from an acquired somatic mu-
tation in the PIGA gene in the haematopoietic stem cell. PNH may
arise from, or evolve to, AA, and small PNH clones are found in about
40% of AA patients. The BMF disorders, whether acquired or in-
herited, but particularly the latter, may undergo later clonal evolution
and malignant transformation to myelodysplastic syndrome (MDS)
and acute myeloid leukaemia (AML) due to genomic instability.
With the development of high-​throughput sequencing technolo-
gies, it has become apparent that somatic mutations are present in
approximately 75% of patients with AA. In addition to the expected
PIGA mutations associated with PNH clones and STAT3/​STAT5b
mutations which likely characterize subclinical T-​cell large granular
lymphocyte leukaemia (T-​LGL) clones within AA, a number of
mutations commonly associated with MDS/​AML have been de-
scribed in the absence of any morphological features of these dis-
orders. Approximately 20% of patients with AA have been shown
to have such mutations with commonly mutated genes including
DNMT3A, ASXL1, and BCOR/​BCORL. Similar mutations have
also been detected in the peripheral blood of healthy individuals,
albeit occurring less frequently and at an older age. This phenom-
enon, termed clonal haematopoiesis of indeterminate potential
(CHIP), is associated with a higher risk of developing a haemato-
logical malignancy (e.g. MDS/​AML). The increased prevalence of
CHIP in AA, particularly at younger ages, may reflect the inability of
the damaged immune system to either eliminate or at least control
22.5.2  Acquired aplastic anaemia and pure red cell aplasia
5337
the abnormal haematopoietic clones which might normally emerge
with increasing years. Assessment of mutational status in AA is of
clinical relevance due to evidence of an association between gene
disruption and clinical outcomes: mutations in PIGA and BCOR/​
BCORL are associated with a better prognosis whereas mutations in
ASXL1 and DNMT3A are associated with a poorer response to im-
munosuppressive treatment, potentially indicating patients suitable
for early allogeneic stem cell transplant.
Aplastic anaemia
Defining aplastic anaemia
The traditional definition of AA is a pancytopenia with a hypocellular
BM in the absence of an abnormal infiltrate and with no increase in
reticulin. Both a BM aspirate and trephine biopsy are required for
the diagnosis. There must be at least two of the following findings:
(1) haemoglobin concentration less than 100 g/​litre; (2) platelet count
less than 50 × 109/​litre; (3) neutrophil count less than 1.5 × 109/​litre.
However, additional aspects of the disease should be evaluated and
included in the definition, as summarized in Table 22.5.2.1.
A constitutional form of AA should be excluded by taking a de-
tailed medical and family history, clinical examination, and labora-
tory investigation, as this has important implications for clinical
management and screening of family members (Table 22.5.2.2, and
Chapter 22.5.1).
The severity of the disease is graded as for acquired AA
(Table 22.5.2.1).
Most cases (80%) of acquired AA are idiopathic, but a careful drug
and exposure history should be taken to attempt to exclude a drug-​
or chemical-​induced cause. However, there are often confounding
factors and no tests are available to prove causality. Nevertheless, any
putative drug should be discontinued at presentation of the AA. See
Table 22.5.2.2 for a list of licensed drugs reported as a possible cause
of idiopathic AA.
All patients should be assessed for the presence of abnormal clones,
namely PNH clones and cytogenetic or molecular clonal changes as-
sociated with MDS. Occasionally, T-​LGL/​T-​lymphocytosis clones
may be present (see ‘Aetiology and incidence’). Using conventional
BM karyotyping, abnormal cytogenetic clones are detected in up to
12% of AA patients in the absence of morphological features of MDS/​
AML. The most common clones found are trisomy 8, trisomy 6, and
monosomy 7, among others. Their prognostic significance varies;
trisomy 8 is associated with a good response to immunosuppressive
treatment whereas monosomy 7 is associated with a poor prognosis
and a high risk of MDS/​AML. Because it is often difficult to obtain
sufficient metaphases due to the BM hypocellularity, cytology/​fluor-
escence in situ hybridization (FISH) or targeted next-​generation
sequencing should be performed in such situations. Flow cytometry
is used to detect PNH clones in the blood, with around 40% of pa-
tients showing deficient expression of glycosylphosphatidylinositol
(GPI) proteins on the cell surface of all blood cells. Small PNH clones
(in general affecting <10% blood cells) is termed subclinical PNH as
there is no clinical or laboratory evidence of haemolysis, but in 10%
of patients the size of the PNH clone is larger, resulting in ‘classical’
haemolytic PNH. The deficient expression of GPI-​anchored proteins
is due to an acquired mutation in the PIGA gene located on the X
chromosome and arising from the haematopoietic stem cells, but
genetic mutation analysis of PIGA does not form part of the routine
diagnostic investigation.
Finally, because of the rarity of AA, it is important to consider pos-
sible alternative causes of the pancytopenia and hypocellular BM.
RBDS
VSAA
SAA
NSAA
PNH
Hypocellular MDS
AML
Low
risk
High
risk
MDS
TLGL
(clonal or polyclonal)
GATA2
defic
DC
AA/PNH
FA
PIGA
STAT3
MDS/AML-associated somatic mutations
SBDS, RPS5,
RPS11, RPS 19…
DKC1, TERC, TERT,
TINF2, RTEL1….
FANC
GATA2
HLA
Hypo
AML
Germline mutations
Somatic (acquired) mutations
Fig. 22.5.2.1  The association of aplastic anaemia with acquired and constitutional disorders of
clonal haematopoiesis. AA, aplastic anaemia; AML, acute myeloid leukaemia; DC, Dyskeratosis
congenita; FA, Fanconi anaemia; MDS, myelodysplastic syndrome; NSAA, non-severe AA; PNH,
paroxysmal nocturnal haemoglobinuria; RBDS, Ribosomal dysgenesis syndromes; SAA, severe AA;
TLGL, T-large granular lymphocytosis; VSAA, very severe AA.
section 22  Haematological disorders
5338
The most difficult condition to distinguish from AA is hypocellular
MDS as the latter may share many features of AA. In most cases
of MDS the BM is hypercellular, but 10 to 15% of patients have a
hypocellular marrow. The findings of dysplastic neutrophils, dys-
plastic megakaryocytes, increased BM reticulin, ringed sideroblasts,
and the presence of blasts/increased CD34+ cells in the blood or
BM, all suggest hypocellular MDS. Even then, the distinction be-
tween AA and hypocellular MDS may not always be certain and
some patients appear to have an ‘AA/​hypocellular MDS’ overlap
syndrome. However, a recent integrated scoring system that incorp-
orates cytohistological, and genetic mutations highly specific for
MDS, enables clear separation of hypocellular MDS cases into two
distinct groups, one with clinical and genetic features highly con-
sistent with a clonal disease with high risk of leukemic evolution,
and the other with features more consistent with a nonmalignant
bone marrow failure, very low risk of transformation and better
survival. Other conditions that are to be considered in the differ-
ential diagnosis of AA include hypocellular AML, or especially in
children, acute lymphoblastic leukaemia, hairy cell leukaemia,
myelofibrosis, lymphoma, atypical mycobacterial infection, and an-
orexia nervosa, which can all sometimes present with pancytopenia
and a hypocellular BM.
Acquired aplastic anaemia
Aetiology and incidence
Most cases (around 80%) of acquired AA are considered to be idio-
pathic. There is a biphasic age distribution with peaks from 10 to
25 years and over 60 years. There is no significant difference in inci-
dence between males and females. Because AA is a rare disease, only
large national and international prospective studies will provide
meaningful data on the aetiology of this condition. The incidence in
Table 22.5.2.1  Defining aplastic anaemia
Confirmation of diagnosis
1 Traditional definition
Pancytopenia with hypocellular BM, haematopoietic tissue replaced by fat cells,
in absence of abnormal infiltrate or increase in reticulin At least 2 of the following
required: Hb <100 g/​litre; platelet count <50 × 109/​litre; and neutrophil count
<1.5 × 109/​litre
FBC, reticulocyte count, blood film examination, BM aspirate and trephine
2 Is the diagnosis really AA?
Or is there another cause for pancytopenia and hypocellular BM?
Exclude hypocellular MDS/​AML, hypocellular ALL especially in children, hairy
cell leukaemia, myelofibrosis, lymphoma, atypical mycobacterial infection,
anorexia nervosa
3 Is the disease an inherited bone marrow failure syndrome?
Clues in medical history, extended family history, and clinical examination
Fanconi anaemia
Presence of café-​au-​lait spots, short stature, anomalies of upper extremities, etc.
Increased chromosome breakages of peripheral blood lymphocytes with DEB/​MMC
Inherited telomeropathy/Dyskeratosis congenita
Nail dystrophy, leukoplakia and skin pigmentation, pulmonary fibrosis, cirrhosis,
premature hair greying, avascular necrosis
Shwachman–​Diamond syndrome
History of pancreatic exocrine insufficiency, neutropenia prior to AA, short stature
GATA2 deficiency
Monocytopenia, B, NK and dendritic cell deficiency, warts, non-tuberculous
mycobacterial infections, lymphoedema
4 What is the aetiology?
Idiopathic
Around 80% of cases
Post-​hepatitic
Liver function tests, viral studies (hepatitis A, B, C, G, usually negative)
Viral infection
Screen for HIV, hepatitis A, B, C, EBV
Drugs and chemicals; environmental/​occupational exposures
Careful drug and exposure history, but no tests available to prove association
PNH
Flow cytometry of GPI-​anchored proteins on red cells and granulocytes
Rarely: pregnancy, systemic lupus erythematosus, thymoma, eosinophilic fasciitis
5 Are there abnormal clones present?
PNH
Flow cytometry as above
MDS/​CHIP
BM metaphase cytogenetics ± FISH/SNP-A karyotyping
T-​LGL clone
Morphology, flow cytometry and TCR gene rearrangement studies
6 How severe is the disease?
Severe AA
Criteria: BM cellularity <25% or 25–​50% with <30% residual haematopoietic
cells, with 2 out of 3 of the following:
neutrophils <0.5 × 109/​litre, platelets <20 × 109/​litre, reticulocytes <20 × 109/​
litre, reticulocytes <60 × 109/​litre
Very severe AA
Same as for severe AA, except neutrophil count <0.2 × 109/​litre
AA, ALL, acute lymphoblastic leukaemia; aplastic anaemia; BM, bone marrow; CHIP, clonal haematopoiesis of indeterminate potential; DEB, diepoxybutane; EBV, Epstein–​Barr
virus; FBC, full blood count; FISH, fluorescence in situ hybridization Hb, haemoglobin; MDS/​AML, myelodysplastic syndrome/​acute myeloid leukaemia; MMC, mitomycin C; PNH,
paroxysmal nocturnal haemoglobinuria; SNP-A, single nucleotide polymorphism array-based; TCR, T-​cell receptor; T-​LGL, T-​large granular lymphocyte.
22.5.2  Acquired aplastic anaemia and pure red cell aplasia
5339
the West is about one to two per million per year, but it occurs more
commonly in eastern Asia, with a two-​ to fourfold higher incidence.
Reasons for this difference in incidence are not known, but may in-
clude infections and genetic factors. In rural areas of Thailand, the
use of nonbottled water, agricultural pesticides, nonmedical needle
exposures, and exposure of farmers to ducks and geese are signifi-
cant environmental risk factors for developing AA.
Many drugs and chemicals have been implicated in the aeti-
ology of AA, but for only a few is there strong evidence for an
association, and even then it is usually impossible to prove caus-
ality (Table 22.5.2.3). A careful drug history must be obtained.
Drug exposure in the year preceding presentation should be de-
tailed. Earlier exposures should be recorded but are not likely to
be relevant unless the particular drug or drug group has been
taken again during the presumed critical period. If the patient is
taking several drugs which may have been implicated in AA, even
if the evidence is based on case report(s) alone, then all the pu-
tative drugs should be discontinued and the patient should not
be rechallenged with the drugs at a later stage after recovery of
the blood count. Drugs most commonly implicated include anti-
biotics and nonsteroidal anti-​inflammatory drugs. Post-​hepatitic
AA accounts for about 5% of all cases, in most cases being non-​A,
Table 22.5.2.2  Indicators of possible inherited AA from the clinical and family history and clinical examination
Clinical history
Family history (first-​degree relatives and
extended family)
Clinical examination
Birth history
Unexplained anaemia, thrombocytopenia or
neutropenia, or pancytopenia
Short stature
Failure to thrive as child
Thrombocytopenia
Skeletal anomalies
Malabsorption
Neutropenia
Abnormal facies
Developmental delay, short stature
Aplastic anaemia
Nail dystrophy, examine hands and feet
Skin or nail problems
MDS
Leukoplakia
Premature greying of hair and use of hair dyes
AML
High-​arched palate
Eye or ear problems
Cancers
Abnormal dentition
Liver problems (cirrhosis, portal hypertension,
splenomegaly)
Liver issues, including those attributed to
alcohol excess
Skin abnormalities: reticulate pigmentation,
hyperpigmentation (café-​au-​lait spots), hypopigmentation
Lung fibrosis
Cirrhosis
Cardiac murmurs and abnormal heart sounds
Osteoporosis
Lung fibrosis
Pulmonary disease
Avascular necrosis of bone
Early childhood deaths
Genitourinary anomalies
Joint or bone abnormalities; corrective
orthopaedic surgery
Osteoporosis
Lymphoedema
Learning disabilities
Learning disabilities
Table 22.5.2.3  Currently licensed drugs and occupational exposures reported as a probable cause of aplastic anaemia
(a) Currently licensed drugs
Antibiotics
Chloramphenicola, sulfonamides, co-​trimoxazole, linezolid
Anti-​inflammatories
Phenylbutazone, indomethacin, diclofenac, naproxen, piroxicam, gold,
penicillamine
Anticonvulsants
Phenytoin, carbamazepine
Antithyroid
Carbimazoleb, thiouracil
Antidepressants
Dothiepin, phenothiazides
Antidiabetic
Chlorpropamide, tolbutamide
Antimalarial
Chloroquine
Others
Mebendazole, thiazidesc, allopurinol
(b) Occupational and environmental exposures
Agent
Evidence base
Benzene
Large industrial studies, case–​control study from Thailand
Pesticides: organochlorines, e.g. lindane, organophosphates, pentachlorophenol
Literature review of case reports and UK case–​control study
Cutting oils and lubricating agents
UK case–​control study
Recreational drugs, e.g. methylenedioxymethamphetamine (MDMA, ecstasy)
Case reports
a There is no evidence for an association between chloramphenicol eye drops and aplastic anaemia.
b More likely to cause neutropenia alone.
c From a case–​control study in Thailand.
section 22  Haematological disorders
5340
non-​B, non-​C, non-​G. Patients usually present with jaundice and
hepatic symptoms, then, on average 6 weeks later, develop pan-
cytopenia when the liver function has usually improved. Rarely
AA follows Epstein–​Barr virus (EBV) infection. A careful occupa-
tional history may reveal exposure to chemicals or pesticides that
have been associated with AA.
The association of AA with PNH was discussed earlier. T-​LGL is
characterized by chronic proliferation of cytotoxic T-​lymphocytes
(CTLs) and is commonly associated with single cytopenias such as
PRCA and autoimmune neutropenia. T-​LGL clones have also been
reported in AA and PNH as well as MDS, as shown by T-​cell re-
ceptor (TCR) gene rearrangement, and more recently using deep
sequencing of the TCR. The haematological target of these expanded
T-​LGL clones is as yet unknown, but the finding of STAT3 mutations
in CTLs in some patients with AA (and MDS and PNH) provides
further insight into the mechanism of their proliferation and their
potential role in immune-​mediated AA.
Pathogenesis
AA is characterized by a severe quantitative defect in the haemato-
poietic stem progenitor cell compartment. The primitive long-​term
culture-​initiating cells and more mature haematopoietic progenitors
in the BM (colony-​forming cells) of all cell lineages are reduced or
absent. There is a reduction in the percentage of CD34+ BM cells,
and they are more apoptotic than normal CD34+ cells.
There is strong evidence for an immune-mediated pathogenesis
for idiopathic AA. Most patients (approximately 70%) will respond
to immunosuppressive therapy, and there are multiple data from
experimental studies. Following an initial BM insult, likely virus
infection, there is immune recognition of a neoantigen or aber-
rantly expressed autoantigen that is presented in the context of class
I HLA molecules on HSC. The immune response is characterised
by marked expansion of CD4+ helper Th1 (clonal), Th2 and Th17
cells, and a decrease in T regulatory cells (Tregs) (Fig. 22.5.2.2). AA
Tregs, as defined by CD4+, FOXP3, CD127, are also dysfunctional
and are unable to suppress autoreactive CD8+ CTLs, resulting in
oligoclonal CTL expansion. The Treg immune signature of AA has
been defined, with two distinct Treg subsets, Treg subpopulations
(Treg A and Treg B) that differ in number and immune-phenotype
among AA responders and non-responders to IST. Treg B popula-
tion predominates in responders to IST and they have a memory/
activated phenotype (as shown by high expression of CD95, CCR4
and CD45RO). The CD8+ CTL expansion, along with an increase
in proinflammatory cytokines interferon-γ (IFN-γ) and tumour ne-
crosis factor-α (TNF-α), leads to death of haematopoietic stem cells
(HSC) by apoptosis. Some clones, however, are able to escape this
Fig. 22.5.2.2  Pathogenesis of acquired aplastic anaemia: immune-​mediated BM apoptosis. FasL, Fas ligand; FasR, Fas receptor; NK, NK cells; TCR,
T-​cell receptor.
22.5.2  Acquired aplastic anaemia and pure red cell aplasia
5341
immune attack and preferentially expand, for example, PNH clones,
+8, del13q clones and somatic loss of class I HLA alleles through
CN-LOH 6p or loss of function somatic mutations. The presence
of such ‘immune escape clones’ is associated with good response to
IST and low risk of clonal evolution to MDS/AML in adult patients.
In AA, the increased proliferative pressure on a reduced HSC
pool, results in telomere loss, leading to genomic instability, with ac-
quisition of somatic mutations and risk of transformation to MDS/
AML. Short telomeres, in the absence of an inherited telomeropathy,
are detected in 30% of patients with acquired AA. In contrast to
the immune response in AA, in low risk MDS there are features of
early, low grade/smouldering inflammation with increased Th17
and normal Tregs, followed by subsequent reduction in Th17 and
expansion of Tregs in high risk MDS. Treg expansion is associated
with expansion of the innate immune effector cells, namely myeloid
derived suppressor cells (MDSCs) that have a potent immunosup-
pressive effect, so there is suppression of immune mediated tumour
surveillance, immune subversion and immune escape that facilitates
progression towards AML.
Clinical features
Taking a detailed and extended family history and careful clinical
examination will increase the chance of early detection of constitu-
tional AA (Table 22.5.2.2). Most cases of AA are acquired and have an
immune basis, but in recent years, there has been increasing aware-
ness of constitutional and often inherited forms of BMF, presenting
not only in children, but also later in adulthood. Furthermore, with
the advent of next generation sequencing, an increasing number of
genetically defined BMF syndromes have been described, aside from
the previously known examples such as Fanconi anaemia, classical
X-linked dyskeratosis congenita (DC) and Schwachman-Diamond
syndrome. These syndromes presenting in adulthood often lack the
classical clinical features, and instead have atypical features such as
pulmonary fibrosis and cirrhosis, premature hair greying, avascular
necrosis, as seen in autosomal dominant DC, now termed ‘inherited
telomeropathies’, due to germline mutations in the telomere gene
complex, of which 12 have so far been described, including TERT,
TERC, RTEL1, among others. GATA2 deficiency arises as a de novo
mutation and then is subsequently inherited as autosomal dominant
disorder. It may present with BMF and progress to MDS/AML. It
is characterised by monocytopenia, B, MK and dendritic cell defi-
ciency, and other features include lymphoedema, non-tuberculous
mycobacterial disease, viral warts, pulmonary alveolar proteinosis.
These and other types of inherited BMF disorders are discussed in
Chapter 22.5.1.
Missing a diagnosis of constitutional AA will result in wrong
treatment, potential risk of fatal outcome after haemopoietic stem
cell transplantation (HSCT) since the HSCT conditioning regimen
is different, inappropriate use of an asymptomatic undiagnosed sib-
ling donor with the same genetic mutation, and failure to refer for
genetic counselling and for long term cancer surveillance.
Patients with AA most commonly present with symptoms of an-
aemia and skin or mucosal haemorrhage (ecchymoses or petechiae),
or visual disturbance due to retinal haemorrhage. Infection may be
a presenting feature, but is less common. There is no lymphaden-
opathy or hepatosplenomegaly (in the absence of infection) and
these findings strongly suggest another diagnosis. A preceding
history of jaundice, usually 2 to 3 months before, may also indi-
cate a post-hepatitic AA. There may be specific clinical features to
indicate a possible inherited bone marrow failure disorder, but some
affected patients may have none of these clinical features. Adults
with late onset inherited AA often lack the classical features of in-
herited disorders such as Fanconi anaemia or DC. The autosomal
dominant forms of DC usually lack the classical mucocutaneous
features of DC but instead may present with features that include
premature greying of the hair, pulmonary fibrosis, and cirrhosis or
non-cirrhotic portal hypertension. A detailed family history should
be taken as this may reveal similar findings in family members
and/or a history of low blood counts or macrocytosis and cancer
(Table 22.5.2.2).
Clinical investigations
The following investigations are required in the evaluation of pa-
tients with suspected AA.
Full blood count
This typically shows pancytopenia. In the early stages, isolated
cytopenia, particularly thrombocytopenia, may occur. Anaemia is
accompanied by reticulocytopenia, and macrocytosis is common.
Careful examination of the blood film is essential to exclude the
presence of dysplastic neutrophils and platelets, blasts, and other
abnormal cells such as hairy cells.
Bone marrow aspirate and trephine biopsy
The BM is hypocellular with prominent fat spaces and variable
amounts of residual haematopoietic cells. Erythroid precursors,
megakaryocytes, and granulocytic precursors are reduced or ab-
sent. Lymphocytes, macrophages, plasma cells, and mast cells ap-
pear prominent. A trephine is essential to assess overall cellularity,
to exclude an abnormal infiltrate and other disorders. Sometimes
the BM is patchy, with hypocellular and cellular areas (Fig. 22.5.2.3).
Immunostaining on the trephine may be particularly helpful, for ex-
ample, CD34 positive cells to identify blasts in AML and CD61 to
identify dysmegakaryopoiesis in MDS. A reticulin stain should be
routinely performed because an increase in BM reticulin is not in
keeping with a diagnosis of AA and may instead indicate a diagnosis
of MDS or hairy cell leukaemia, for example.
Liver function tests and virology
In post-​hepatitic AA, the serology is usually negative for all the
known hepatitis viruses. Abnormal liver function tests may occur
in DC secondary to cirrhosis or noncirrhotic portal hypertension.
Blood should be sent to test for hepatitis A antibody, hepatitis B sur-
face antigen, hepatitis C antibody, HIV antibody I and II, and EBV,
as these are rare causes of AA; parvovirus may causes red cell aplasia,
but not typically AA.
Vitamin B12 and folate levels
These should be tested to exclude megaloblastic anaemia, which
when severe may present with pancytopenia.
Antinuclear antibody and anti-​DNA antibody
These should be tested to exclude systemic lupus erythematosus
which is a rare cause of AA.
PNH screen
Multiparametric flow cytometry is used to identify clones of cells
that lack GPI-​anchored proteins, such as CD55 and CD59. This is a
section 22  Haematological disorders
5342
sensitive and quantitative test for analysis of PNH populations. The
old test for PNH was Ham’s test, but this was relatively insensitive
and only assessed red blood cells for lysis in the presence of added
complement. Additionally, recent blood transfusion may mask and/​
or underestimate the size of the red cell PNH clone. In contrast,
flow cytometry analyses expression of GPI-​anchored proteins not
only on the surface of red cells but also granulocytes and mono-
cytes, and can detect much smaller PNH clones than Ham’s test.
Bacterial aerolysin selectively binds to the GPI anchor and causes
lysis of normal cells but not PNH cells that lack the GPI anchor.
FLAER (fluorescein-​labelled aerolysin) is a labelled, inactive variant
of aerolysin that does not cause lysis of cells, but provides a more ac-
curate and more sensitive assay than flow cytometry alone to detect
PNH cells as it binds to normal but not PNH cells. FLAER is often
combined with flow cytometry. Flow cytometry is more sensitive
at diagnosing small PNH clones than molecular testing for PIGA
gene mutations, hence detection of PIGA gene mutations does not
form part of routine diagnostic investigations. Small PNH clones (in
general affecting < 10% blood cells) is termed sub-clinical PNH as
there is no clinical or laboratory evidence of haemolysis, but in 10%
of patients the size of the PNH clone is larger resulting in ‘classical’
haemolytic PNH.
Marrow cytogenetics and molecular genetics
Abnormal cytogenetic clones may be present in up to 12% of pa-
tients with otherwise typical AA, and do not necessarily indicate
MDS or leukaemia, but do require following carefully. If insuffi-
cient metaphases are obtained, cytology/​FISH should be performed,
particularly for deletions of 5 and 7 and for trisomy 8, and/or single
­nucleotide polymorphism array-based (SNP-A) karyotyping.
Somatic mutations
Many major centres offer testing for acquired somatic mutations in
myeloid specific genes, such as ASXL1, DNMT3A using customised
diagnostic gene panels. Targeted next-​generation sequencing of mu-
tation hotspots in key genes such as STAT3, DNMT3a, and ASXL1
may also be of utility.
Screen for inherited disorders
For all new patients up to middle age and older patients who
are potential candidates for BM transplantation, or in whom in-
herited AA is suspected, peripheral blood lymphocytes should be
tested for chromosomal breakages to exclude Fanconi’s anaemia.
Certain specialist centres provide testing for telomere length
and NGS constitutional BMF gene panels. Following the 100 K
Genome Project, future testing is planned for all patients by NHS
England.
Chest radiograph and abdominal ultrasonography
Radiological investigations, including chest radiograph and an
abdominal ultrasound scan should be performed to exclude in-
fection, and an enlarged spleen and/​or enlarged lymph nodes,
respectively. In younger patients, abnormal or anatomically dis-
placed kidneys are features of Fanconi’s anaemia. Radiography of
the hands and forearms may be indicated for specific inherited
forms of BMF.
(b)
(a)
(d)
(c)
Fig. 22.5.2.3  Bone marrow trephine biopsy sections. (a) Normal bone, showing normal
distribution of haematopoietic cells and fat cells within the bone trabeculum; (b) severe AA
showing replacement of haematopoietic cells by fat cells; (c) nonsevere AA showing patchy
distribution of remaining haematopoietic cells; (d) high-​power view of severe AA showing fat
cells interspersed by lymphocytes and macrophages.
22.5.2  Acquired aplastic anaemia and pure red cell aplasia
5343
Treatment and prognosis
Once the diagnosis is firmly established and the disease has been
carefully defined as described previously, the treatment plan for the
patient should be mapped out and discussed. Treatment is influ-
enced by the age of the patient, disease severity, and the availability
of a suitable stem cell donor.
Supportive care
Transfusions  Red cell and platelet transfusions are given to help
maintain safe blood counts. Current national and European guide-
lines recommend to give prophylactic platelet transfusions when
the platelet count is below 10 × 109/​litre (or higher in the presence
of fever or bleeding). Purpura, menorrhagia, and spontaneous
bleeding, mostly from the gums and buccal mucosa, usually develop
below this level, but there is also a risk of life-​threatening haemor-
rhage, particularly in the presence of infection.
Transfusions may induce alloimmunization to leucocytes present
in red cell and platelet transfusions by generating human leucocyte
antigen (HLA) or non-​HLA (minor histocompatibility) antibodies
or platelet-​specific antibodies, which cause refractoriness to platelet
transfusions as well as increasingly the risk of graft rejection fol-
lowing allogeneic haematopoietic stem cell transplantation (HSCT).
The risk of alloimmunization to leucocytes has diminished since the
widespread introduction of leucodepletion and irradiation of blood
products, but still remains a problem. If patients develop HLA anti-
bodies, HLA-​compatible platelets may be needed. Other important
practical and are refractory to random donor platelets, measures to
help prevent bleeding include good dental hygiene, the use of oral
tranexamic acid, and control of menorrhagia with appropriate hor-
mone therapy.
Transfusional haemosiderosis occurs with prolonged red cell
transfusions. Iron chelation therapy with desferrioxamine or
deferasirox may be indicated, especially in patients who are trans-
plant candidates. For patients with haematological disorders under-
going HSCT, transplant-​related mortality is increased if the serum
ferritin is greater than 2000 μg/​litre, and many centres commence
iron chelation when the serum ferritin is greater than 1000 μg/​litre.
Infections  Patients with AA are at risk of bacterial and fungal infec-
tions, with the level of risk depending on the degree of neutropenia.
Severe neutropenia (neutrophils <0.5 × 109/​litre, and especially
<0.2 × 109/​litre) carries a high risk of systemic infection arising from
endogenous organisms, especially Gram-​negative but also Gram-​
positive bacteria. Fungal infections, particularly aspergillus, occur
in severe neutropenia. Patients with neutrophil count <0.5 × 109/litre
should receive antibiotic and antifungal prophylaxis as recom-
mended by NICE and national AA guidelines. Oral hygiene is
important. Entry sites for venous access are potential sources of
systemic infection. Fever should be treated with broad-​spectrum
antibiotics, without waiting for laboratory identification of or-
ganisms. Systemic antifungal drugs should be commenced early
if fevers fail to resolve with intravenous antibiotics. Granulocyte
colony-​stimulating factor (G-​CSF) is usually ineffective in severe
AA because of a severe reduction or absence of myeloid progenitor
cells. Granulocyte transfusions are sometimes administered in life-​
threatening infections.
Psychological support  Psychological support for the patient,
family, and close friends is of great importance. AA is a rare dis-
ease and requires careful explanation of its nature and prognosis.
Patients should be given the opportunity to be referred to a centre
that specializes in the management of AA. The chronic nature and
slow response to treatment should be stressed early in the disease.
There is an excellent patient support group based in the United
Kingdom (http://​www.theaat.org.uk).
Specific treatment
The treatment choice is essentially either allogeneic HSCT or im-
munosuppressive therapy using horse ATG and ciclosporin. Horse
ATG has been found to be superior to rabbit ATG and is therefore
the initial treatment of choice if available (see ‘Immunosuppressive
Age of patient
Co-morbidities
≤ 35 y
50–60 y
HLA identical
sibling donor
No
Yes
Unrelated
donor HSCT
Children
35–50 or 60y?
Or
Or
Horse ATG with
ciclosporin
HLA matched
sibling HSCT
If no response
2nd ATG,
danazol,
haplo/cord HCT
Eltrombopag
Fig. 22.5.2.4  Algorithm for treatment of severe aplastic anaemia. HSCT, haematopoietic stem cell
transplantation.
section 22  Haematological disorders
5344
therapy (ATG and ciclosporin)’). An algorithm summarizing treat-
ment options for severe AA is shown in Fig. 22.5.2.4.
Haematopoietic stem cell transplantation  First-​line treatment for
younger severe AA patients is allogeneic HSCT from an HLA-​identical
sibling donor, with 70 to 90% long-​term survival (Fig. 22.5.2.5).
Results of HSCT in older patients are less successful so older pa-
tients should receive immunosuppressive therapy with ATG and
ciclosporin as first-​line treatment. One possible immunosuppres-
sive, nonmyeloablative conditioning regimen is employed for
HSCT in AA. Cyclophosphamide (CY) 200 mg/kg with ATG or
alemtuzumab is used for patients aged <30 years, and fludarabine,
low dose CY with ATG or alemtuzumab (‘FCC’) for patients aged
30 years. Alemtuzumab is associated with a lower incidence of
chronic graft-​versus-​host disease (GVHD). Ciclosporin (usu-
ally with methotrexate, unless using alemtuzumab) is given to
suppress GVHD and to aid engraftment. Irradiation should not
be used in AA patients receiving transplants from matched sib-
ling donors. Children grow and develop normally and fertility is
well preserved post transplantation. Chronic GVHD and infection
are the main causes of transplant-​related morbidity and mortality.
Graft rejection may be early, with failure of engraftment of host
cells, or late, after initial engraftment. It occurs in 10 to 15% of pa-
tients. Analysis of chimerism on myeloid cells (CD15) and T cells
(CD3) in patients receiving alemtuzumab-​based conditioning
demonstrates that stable mixed T-​cell chimerism in the presence
of full donor myeloid chimerism is associated with excellent sur-
vival and a low incidence of chronic GVHD. Using an ATG-​based
conditioning regimen, BM stem cells are used instead of periph-
eral blood stem cells, but either can be used with alemtuzumab-
based conditioning. It is important to give at least 3 × 108 nucleated
marrow cells/​kg body weight (or >2 × 106 CD34+ cells/​kg) to re-
duce the risk of graft rejection.
Matched unrelated donor (MUD) HSCT is considered for patients
who have no HLA-​compatible donor and who have failed treatment
with one course of immunosuppressive therapy. However, for chil-
dren who lack a matched sibling donor, first line MUD HSCT may be
considered. Because of recent improvements in outcomes after MUD
HSCT in adults with SAA, there is currently active debate for consid-
ering first line MUD HSCT as an option for adults  who need urgent
HSCT. Patients undergoing MUD HSCT should receive a fludarabine,
low dose CY based conditioning with ATG or alemtuzumab. Low dose
total body irradiation (2Gy) is also needed when using ATG but not
with alemtuzumab conditioning.
Survival
Days following Transplantation (OS)
0,00
0.0
0.2
0.4
0.6
0.8
1.0
1000,00
2000,00
3000,00
Age >50yrs (n = 12)
72.9%
86.5%
Effect of age < 50 or >50 years
(d)
Age <50yrs (n = 33)
4000,00
P = 0.02
MSD (n = 21)
UD (n = 29)
P = 0.34
83%
Matched sibling and UD HSCT
(c)
Survival
Days following Transplantation (OS)
0,00
0.0
0.2
0.4
0.6
0.8
1.0
1000,00
2000,00
3000,00
4000,00
95%
100
75
50
25
0
0
1000
2000
3000
4000
Days from transplant
FCA; n–52
FCA- TBI; n = 48
79%
73%
(b)
Unrelated donor HSCT
Survival (%)
1,000
0,750
0,500
0,250
0,000
0,0
1000,0
Matched sibling donor HSCT
(a)
2000,0
3000,0
4000,0
Age ≥ 20, n = 890
P<0.0001
70%
87%
Age ≥ 20, n = 996
Fig. 22.5.2.5  Overall survival after matched sibling donor and unrelated donor HSCT for severe AA. (a, b) Data from the
European Group for Blood and Marrow Transplantation (EBMT) Registry. Conditioning regimens used for matched sibling
transplants comprised high-dose cyclophosphamide with ATG, and for unrelated donor transplants fludarabine, low-dose
cyclophosphamide and ATG, with or without low-dose total body irradiation. (c, d) Results from a multicentre study from
the United Kingdom and Toronto, Canada, using alemtuzumab-based conditioning instead of ATG.
(a, b) Bacigalupo A et al., 2012, with permission.
22.5.2  Acquired aplastic anaemia and pure red cell aplasia
5345
For patients who lack both a suitable matched sibling and unrelated
donor, alternative transplant approaches currently being explored
are cord blood HSCT and haploidentical HSCT. A haploidentical
donor, whether a parent or sibling or child, is usually readily avail-
able for most patients. In contrast, cord blood transplant often
requires 2 cord units for adult HSCT and is more expensive than
haploidentical HSCT. Haploidentical HSCT performed during
the 1980s was usually unsuccessful due to a very high incidence of
GVHD and graft rejection, but the more recent use of high-​dose
cyclophosphamide given in the immediate post-​transplant phase to
eliminate alloreactive donor T cells is showing promising potential.
Immunosuppressive therapy (ATG and ciclosporin)  ATG is
a polyclonal IgG antibody preparation prepared by immunizing
horses, rabbits, or pigs with human thymocytes. Its mechanism of
action in AA is not entirely clear, but may be partly due to depletion
of autoreactive, cytotoxic T cells, in addition to direct stimulation of
residual BM CD34+ cells or a reduction in their degree of apoptosis,
or other mechanisms. The response rate to ATG in AA is increased
when it is combined with ciclosporin. Horse ATG is preferred to
rabbit ATG as it results in a higher response rate (68% at 6 months
compared to 37% for rabbit ATG) and better survival, despite rabbit
ATG being more immunosuppressive than horse ATG in terms of
the degree and duration of lymphodepletion, indicating that different
mechanisms are important in the mode of action of the two agents.
ATG and ciclosporin are indicated for patients who are ineligible for
HSCT or whose disease is not severe enough to warrant a transplant.
This includes older patients with nonsevere AA, and younger patients
who have severe disease but who lack an HLA-​identical sibling donor.
ATG is highly immunosuppressive and must be given as an in-
patient treatment, preferably with patients nursed in isolation fa-
cilities. Response is delayed, rarely occurring before 3 to 4 months.
Around 70% will respond and achieve normal or near-​normal
blood counts, with a 5-​year overall survival of 80%. Survival is
better in younger patients compared to older patients, and better
in patients with severe AA compared to those with very severe AA
(Fig. 22.5.2.6). Patients require long-​term follow-​up for later events,
such as relapse of AA which occurs in up to 30%, necessitating fur-
ther treatment with ATG or consideration for HSCT. Later clonal
evolution to MDS/​AML occurs in 15% of patients. Retreatment with
a second course of ATG results in response rates of about 35% for
nonresponse and 50 to 60% for relapse after a first course of ATG.
Patients with nonsevere AA who are not dependent on red cell or
platelet transfusions may remain stable for months or years without
definitive treatment. They should have their blood count monitored
regularly, and if it worsens such that they require transfusions, they
may then be treated with ATG and ciclosporin. Oral ciclosporin may
be used on its own, although the response rate is lower than with
the combination of ATG and ciclosporin. Patients with inherited
AA more commonly present with nonsevere AA, so a high degree
of suspicion should be maintained and such patients investigated
as thoroughly as possible for a potential inherited BMF disorder.
In addition, among the nonresponders to ATG, there may be some
who have an unsuspected, underlying inherited BMF syndrome.
Other immunosuppressive agents  In an attempt to improve
the response to ATG and ciclosporin, the additional use of either
mycophenolate or sirolimus has been assessed in clinical trials, but nei-
ther agent improves the response rate or survival. Alemtuzumab has
been used to treat AA, and the response rate is comparable to ATG for
relapsed AA (56%) and refractory AA (37%) but lower for untreated
AA (19% response), so its use is not recommended as first-​line therapy.
High-​dose cyclophosphamide (200 mg/​kg body weight) without
haematopoietic stem cell support had been proposed as an alter-
native treatment for refractory AA. However, a prospective ran-
domized study comparing cyclophosphamide and ciclosporin
with ATG and ciclosporin was terminated prematurely because
of a high incidence of systemic fungal infections and early deaths
Fig. 22.5.2.6  Overall survival (OS) of patients with severe aplastic anaemia transplanted with ‘FCC’ conditioning regimen. (a) OS for patients
transplanted from matched sibling donors (MSD) and matched unrelated donors (MUD). (b) OS according to co-morbidity index. (c) OS comparing
patients aged <50 years with those aged ≥50 years. GVHD, graft versus host disease; CI, cumulative incidence.
Reproduced with permission from Marsh JCW, et al. (2011). Alemtuzumab with fludarabine and cyclo­phosphamide reduces chronic graft versus host disease after allogeneic
stem cell transplantation for acquired aplastic anemia. Blood, 118, 2351–7.
section 22  Haematological disorders
5346
after cyclophosphamide. Furthermore, when the dose was reduced
to 120 mg/​kg there were still unacceptable toxicities and mortality
on account of the predictably prolonged period of neutropenia and
thrombocytopenia. Consequently, the use of cyclophosphamide
alone to treat AA is not now generally recommended.
Oxymetholone and corticosteroids  Oxymetholone is sometimes
used in AA, most often in inherited AA such as FA and DC when
HSCT is not possible. Up to 25% of patients with severe acquired,
refractory AA will respond to this drug. A response to androgens
should raise the possibility of an inherited BMF disorder. The hepato-
toxicity and virilizing side effects restrict their use. Corticosteroids
have no role in the treatment of AA, other than helping to prevent
the complication of serum sickness following treatment with ATG.
Corticosteroids are not effective in treating AA and increase the risk
of infections and later complications such as avascular necrosis.
Haematopoietic growth factors  As serum levels of endogenous
haematopoietic growth factors such as G-​CSF, IL-​3, erythropoietin
(EPO), thrombopoietin, and stem cell factor are invariably markedly
elevated in AA, clinical treatment with any of these agents is inef-
fective in most cases. There is no indication for using G-​CSF or EPO
to treat the underlying disease.
Eltrombopag  The thrombopoietin receptor agonist, eltrombopag,
has been licensed for treatment of refractory severe AA patients who
have failed to respond to a course of IST. Response rates are 40–50%,
but there is a 20% risk of early cytogenetic clones including monosomy
7, so patients must be monitored with regular BM cytogenetic analysis.
More recently, a phase II study of horse ATG, ciclosporin and
eltrombopag for first line treatment of severe AA reported mark-
edly high response rate compared to historical data. A European
prospective randomised study (‘RACE’ study) comparing ATG
and ciclosporin with or without eltrombopag as first line therapy,
has recently been completed and data analysis is in progress.
Inherited aplastic anaemia
(See Chapter 22.5.1.)
Early diagnosis is important to (1) initiate a management plan and
to avoid inappropriate treatment that would be given for acquired AA,
(2) help prevent multiorgan complications from treatment, (3) institute
genetic counselling, (4) ensure selection of appropriate donors for fu-
ture HSCT, and (5) to ensure long term cancer surveillance programme.
Bone marrow failure affecting the erythroid cell
lineage: pure red-​cell aplasia
PRCA is an acquired or inherited disorder, characterized haemato­
logically by severe normocytic, normochromic anaemia (the an-
aemia in inherited PRCA is macrocytic), reticulocytopenia, and
reduced or absent erythroid progenitors in the BM. The white cell,
neutrophil, and platelet counts are normal. The aetiology and patho-
physiology are heterogeneous.
Acquired PRCA
Aetiology
Acquired PRCA can present in different forms. It may present in
children with transient anaemia following a viral infection (tran-
sient erythroblastopenia of childhood (TEC)). In a proportion of
cases of TEC, there may also be neutropenia. Occasionally, cases
occur in clusters, suggesting exposure to a virus or toxin, but a
common virus has not been proven. The natural history is of spon-
taneous recovery over a few weeks. The most important differen-
tial diagnosis is inherited PRCA, Diamond–​Blackfan anaemia, but
in contrast to Diamond–​Blackfan anaemia, the presentation of
TEC is later (most cases occur after 1 year of age), the anaemia is
normocytic and normochromic, levels of fetal haemoglobin and red
cell adenine deaminase are normal and patients lack somatic anom-
alies of Diamond–​Blackfan anaemia. Another acquired condition
to exclude in children is Pearson’s syndrome, a rare sporadic mito-
chondrial cytopathy due to mutations in mitochondrial DNA and
characterized by sideroblastic anaemia with vacuolization of haem-
atological precursors, neutropenia and thrombocytopenia, exocrine
pancreatic insufficiency, and metabolic acidosis.
Acquired PRCA may be immune mediated and idiopathic, but
all cases of acquired PRCA should be investigated thoroughly for a
possible secondary cause, as PRCA may occur in association with a
wide range of disorders or conditions (Table 22.5.2.4).
A thymoma occurs in up to 10% of patients with PRCA and up
to 5% of patients with thymoma have PRCA. PRCA may precede,
Table 22.5.2.4  Secondary causes of pure red cell aplasia
Associations
Examples/​comments
Thymoma
May be associated with hypogammaglobulinaemia (Good’s syndrome) and myasthenia gravis
Systemic autoimmune disorders
Rheumatoid arthritis, systemic lupus erythematosus, Sjögren’s syndrome, mixed
connective tissue disease
B-​lymphoproliferative and plasma cell disorders
Chronic lymphocytic leukaemia, lymphoma, Waldenstroms macroglobulinaemia, MGUS
T-​LGL
Myeloid blood disorders
MDS, myeloproliferative disorder, e.g. myelofibrosis
Solid tumours
Gastric carcinoma, renal cell carcinoma
Infections
B19 parvovirus, HIV, EBV, CMV, hepatitis B, C
Drugs
Phenytoin, isoniazid, rifampicin, azathioprine, procainamide, EPO in CKD (anti-​EPO antibodies)
ABO incompatible haematopoietic stem cell transplantation
Recipient antibody against incompatible donor ABO blood group. Recent reports of
response to daratumumab (anti CD38 monoclonal antibody)
Pregnancy
CKD, chronic kidney disease; CMV, cytomegalovirus; EPO, erythropoietin; MDS, myelodysplastic syndrome; T-​LGL, T-​cell large granular lymphocytic leukaemia.
22.5.2  Acquired aplastic anaemia and pure red cell aplasia
5347
accompany, or follow the development of a thymoma. Myasthenia
gravis is a frequently reported autoimmune association with thymoma.
Good’s syndrome describes the occurrence of thymoma with im-
munodeficiency, characterized by hypogammaglobulinaemia, and
B-​ and T-​cell dysfunction. Because thymoma and PRCA is a rare as-
sociation, optimal treatment is not known. Historical data report that
surgical resection of the thymoma induces remission of the PRCA in
about 25% of cases; the PRCA may also relapse later after thymec-
tomy. A study from the Mayo Clinic reported on the outcome of 13
patients seen with PRCA and thymoma over a 50-​year period. Out of
12 patients who underwent thymectomy, only one achieved a normal
haemoglobin level and all other patients required additional treat-
ment for the PRCA. At a median follow-​up of 26 months, four were in
complete remission and eight required continued blood transfusions.
Nevertheless, it is proposed that thymectomy should be considered if
the patient is clinically fit enough for surgery.
PRCA may occur in association with a wide range of systemic auto-
immune disorders, including rheumatoid arthritis, systemic lupus
erythematosus, and Sjögren’s syndrome. The most common viral in-
fection associated with PRCA is B19 parvovirus, and rarely following
EBV, cytomegalovirus (CMV), or hepatitis B and C.  Parvovirus-​
associated PRCA may occur in (1) patients with haemolytic anaemia
such as hereditary spherocytosis, sickle cell disease, pyruvate kinase
deficiency, and glucose-​6-​phosphate dehydrogenase deficiency (pro-
ducing a so-​called aplastic crisis); and (2) immunocompromised pa-
tients, for example, transplant recipients, patients with HIV, or during
chemotherapy or monoclonal antibody therapy for lymphoma. The re-
ceptor for B19 parvovirus is the blood group P antigen expressed on the
CFU-​E erythroid progenitors and mature red cells. The virus invades
and destroys the cells, resulting in sudden and severe anaemia with
reticulocytopenia. PRCA secondary to parvovirus infection responds
well to intravenous immunoglobulin. Lymphoproliferative disorders,
such as chronic lymphocytic leukaemia and Hodgkin’s or non-​
Hodgkin’s lymphoma, may sometimes be complicated by PRCA. In con-
trast, PRCA occurs more commonly with T-​lymphocytosis (T-​LGL).
T-​LGL may also be associated with other cytopenias, such as auto-
immune neutropenia or immune thrombocytopenic purpura. It is a
clonal disorder as demonstrated by rearrangement of the TCR. The
immunophenotype is typically CD3, CD8, and TCRαβ positive, more
rarely CD4 TCRαβ or TCRγδ positive.
A long list of drugs has been incriminated in the aetiology of red
cell aplasia, but the association is very rare and mostly there is only a
single case report for each drug. Exceptions are azathioprine, pheny-
toin, procainamide, and isoniazid, for which more cases have been re-
ported. Severe and sudden onset of PRCA due to EPO may occur. This
is due to anti-​EPO antibodies against endogenous EPO, and in most
cases it was associated with the use of epoetin-​α (Eprex) when given
subcutaneously in chronic kidney disease. EPO-​induced PRCA had a
peak incidence in 2001 to 2002 and was most likely due to changes in
the formulation of the drug and the use of uncoated rubber stoppers.
Clinical investigations
1.	 Full blood count:  anaemia, reticulocytopenia, normal white
cell, neutrophil, and platelet counts. Normal mean cell volume
and mean cell haemoglobin in acquired PRCA; macrocytosis in
Diamond–​Blackfan anaemia. Examine blood film for T-​LGLs,
smear cells, and small mature lymphocytes of B-​CLL.
2.	 BM aspirate and trephine: absence or reduction in erythroid pro-
genitors; sometimes maturation arrest at proerythroblast stage;
giant pronormoblasts in parvovirus-​induced PRCA. Examine
BM for secondary MDS, myeloproliferative disorder and CLL,
lymphoma, and T-​LGL. Parvovirus B19 DNA in situ hybridiza-
tion study on BM slides.
3.	 Autoimmune profile to exclude rheumatoid arthritis, systemic
lupus erythematosus, and other autoimmune disorders.
4.	 Anticholinesterase receptor antibodies to exclude myasthenia
gravis.
5.	 Serum immunoglobulins to exclude hypogammaglobulinaemia.
6.	 Viral screen:  parvovirus serology and/​or B19 DNA in serum,
HIV, EBV, CMV, and hepatitis B and C.
7.	 Immunophenotyping and gene rearrangement studies for heavy-​
chain and TCR monoclonal expansions.
8.	 CT scan of chest to exclude thymoma; CT scan of chest, ab-
domen, and pelvis to exclude lymphoma.
9.	 For Diamond–​Blackfan anaemia, elevated erythrocyte adenine
deaminase in 80 to 85% of patients; fetal haemoglobin often elevated
and i red blood cell antigen expressed. Research testing for mutation
in one of the known ribosomal protein genes (see Chapter 22.5.1).
Treatment
Blood transfusions are required until recovery occurs. Any drugs that
have been implicated in the disease should be discontinued. Thymectomy
should be considered in patients with a thymoma. Intravenous im-
munoglobulin is indicated for parvovirus-​induced PRCA. Secondary
causes of PRCA may respond to treatment of the underlying condition
but specific treatment of the PRCA is often required.
Immunosuppression is the mainstay of treatment for idiopathic
acquired and secondary PRCA not responding to treatment of the
underling disorder, based on the findings of a wide range of im-
munological abnormalities involving serum autoantibodies, and
T-​cell and NK cell-​mediated effects on erythropoiesis. The op-
timal treatment of PRCA is restricted by the rarity of the condition.
Historically, the first-​line treatment used to be prednisolone at a
dose of 1 mg/​kg per day, to which 30 to 60% of patients responded.
Issues relating to corticosteroids are (1) relapse which is common
and occurs in around 80% of patients; and (2) toxicity, for example,
infection, diabetes, weight gain, avascular necrosis and osteoporosis,
especially with long-​term use.
For nonresponders, remission may be induced by cytotoxic im-
munosuppressants such as azathioprine and cyclophosphamide,
but there are concerns about the long-​term risk of malignancies and
gonadal toxicity. ATG has sometimes been used, with anecdotal re-
ports of response in PRCA, but this requires inpatient treatment and
the main side effects include infection and severe allergic reaction.
For the above-​mentioned reasons, there has been a move away
from corticosteroids and cytotoxic immunosuppressants in the
treatment of PRCA. Instead, oral ciclosporin (CSA) has become
the first-​line treatment. CSA was first proposed as a first-​line treat-
ment for PRCA back in 1990 when review of the literature showed
an overall response rate of 65%. Subsequent clinical studies, most
from Japanese national surveys, report excellent results with CSA.
For primary acquired PRCA, the response rate is 74 to 87% and 60 to
73% for secondary PRCA. On withdrawal of CSA, the risk of relapse
is high. It appears that continued CSA, at a lower dose, is required
to maintain remission in the long term. It is important to monitor
patients’ renal function, blood pressure, and CSA blood levels for
potential long-​term toxicity of the drug.

22.5.3 Paroxysmal nocturnal haemoglobinuria 5348 L
22.5.3 Paroxysmal nocturnal haemoglobinuria 5348 Lucio Luzzatto
section 22  Haematological disorders
5348
There are anecdotal reports on the use of monoclonal antibody
therapy with alemtuzumab (anti-​CD52) and rituximab (anti-​
CD20). Alemtuzumab has been used in combination with CSA in
a small number of patients. However, relapse may occur and con-
tinued ‘maintenance’ with CSA may be required. Patients need to be
carefully monitored for infections. Anecdotal responses have been
reported with rituximab notably in patients with underlying B-​cell
lymphoproliferative disorders.
Likely future developments
With the advent of next-​generation, high-​throughput DNA sequen-
cing, and the success of the 100K Genome project, increasing num-
bers of genes involved in inherited BMF disorders will be identified.
This will form the basis of a routine clinical screening test for all
newly presenting patients, and is likely to have a major impact on
clinical management decisions. It is predicted that more patients
with apparent acquired BMF will have a mutation in one or more
genes that are involved in the pathogenesis of BMF.
Further understanding of the immunological changes that occur
in acquired AA is likely to result in the availability of predictive tools
for response to immunosuppressive therapy and for later relapse.
Alternative approaches to HSCT for those patients lacking a suit-
ably matched sibling or volunteer unrelated donor may include
further evaluation of haploidentical HSCT with the use of post-​
transplantation high-​dose cyclophosphamide. It is almost always
possible to find a potential haploidentical donor. Novel cellular ther-
apies using, for example, ex vivo expanded autologous Tregs are in
development following the in vitro expansion of AA Tregs that are
functional, stable and polyclonal.
As the success of HSCT continues to improve, it is likely that this
treatment is already being extended to older patients.
FURTHER READING
Alter BP (2005). Bone marrow failure: a child is not just a small adult
(but an adult can have a childhood disease). Hematology, 2005,
96–​103.
Bacigalupo A, et al. (2015). For the Aplastic Anemia Working Party
of the European Group for Blood and Marrow Transplantation
(WPSAA-EBMT). Current outcome of HLA identical sibling vs. un-
related donor transplants in severe aplastic anemia: an EBMT ana-
lysis. Haematologica, 100, 696–​702.
Bono E, et al. (2019). Clinical, histopathological and molecular char-
acterization of hypoplastic myelodysplastic syndrome. Leukemia,
doi: 10.1038/s41375-019-0457-1.
Calado RT, Young NS (2009). Telomere diseases. N Engl J Med, 361,
2353–​65.
Casadevall N, et al. (2002). Pure red-​cell aplasia and antierythropoietin
antibodies in patients treated with recombinant erythropoietin.
N Engl J Med, 346, 469–​75.
Killick SB, et al. (2016). British Society for Standards in Haematology.
Guidelines for the diagnosis and management of adult aplastic an-
aemia. Br J Haematol, 172, 187–207.
Kordasti S, et al. (2016). Deep phenotyping of Tregs identifies an im-
mune signature for idiopathic aplastic anemia and predicts response
to treatment. Blood, 128, 1193–205.
Kulasekararaj AG, et al. (2014). Somatic mutations identify a subgroup
of aplastic anemia patients who progress to myelodysplastic syn-
drome. Blood, 124, 2698–704.
Luzzatto L, Risitano AM (2018). Advances in understanding the
pathogenesis of acquired aplastic anaemia. Br J Haematol, 182,
758–76.
Macdougall IC, et al. (2009). A peptide-​based erythropoietin-​receptor
agonist for pure red-​cell aplasia. N Engl J Med, 361, 1848–​55.
Marsh JCW, Mufti GJ (2018). Somatic mutations in aplastic anaemia.
Hematol Oncol Clin N Am, 32, 595–​607.
Marsh JCW, Risitano AM, Mufti GJ (2019). The case for upfront HLA-
matched unrelated donor HCT as a curative option for adult ac-
quired severe aplastic anemia. Biol Blood Marrow Transplant, pii:
S1083-8791(19)30323-4. doi: 10.1016/j.bbmt.2019.05.012.
Marsh JCW, et al. (2011). Alemtuzumab with fludarabine and cyclo-
phosphamide reduces chronic graft versus host disease after allo-
geneic stem cell transplantation for acquired aplastic anemia. Blood,
118, 2351–​7.
Means RT (2016). Pure red cell aplasia. Blood, 128, 2504–9.
Rossert J, Casadevall N, Eckardt KU (2004). Anti-​erythropoietin anti-
bodies and pure red cell aplasia. J Am Soc Nephrol, 15, 398–​406.
Soulier J (2011). Understanding and management of inherited bone
marrow failure syndromes: Fanconi anemia. Hematology, 2011, 492–​7.
Tichelli A, et  al. (2011). A randomized controlled study in
newly-​diagnosed severe aplastic anemia patients receiving
antithymocyte globulin, ciclosporin, with or without G-​CSF.
Blood, 118, 2351–​7.
Winkler T, et al. (2019). Treatment optimization and genomic out-
comes in refractory severe aplastic anemia treated with eltrombopag.
Blood, 133, 2575–85.
Yoshizato T, et al. (2015). Somatic mutations and clonal hematopoiesis
in aplastic anemia. N Engl J Med, 373, 35–​47.
Young NS (2018). Aplastic anemia. N Engl J Med, 379, 1643–​56.
22.5.3  Paroxysmal nocturnal
haemoglobinuria
Lucio Luzzatto
ESSENTIALS
Paroxysmal nocturnal haemoglobinuria (PNH) is a unique disorder
in which many of the patient’s red cells have an abnormal suscep-
tibility to activated complement. This results from the presence of
a clone that originates from a haematopoietic stem cell bearing an
acquired somatic mutation in the X-​linked gene PIGA, required for
the biosynthesis of the glycosylphosphatidylinositol molecule which
anchors many proteins to the cell membrane, including the comple-
ment regulators CD59 and CD55.
The ‘classical’ presentation is with ‘passing blood instead of urine’
(haemoglobinuria). Sometimes the patient presents with the full triad
of (1) haemolytic anaemia, (2) pancytopenia, and (3) thrombosis—​
most commonly of intra-​abdominal veins. An element of bone
22.5.3  Paroxysmal nocturnal haemoglobinuria
5349
marrow failure is always present; and sometimes the disease may be
preceded by or may evolve to bone marrow aplasia indistinguish-
able from acquired aplastic anaemia. Definitive diagnosis is based
on demonstrating the presence of a discrete population of ‘PNH red
blood cells’ by flow cytometry using anti-​CD59. In most cases, espe-
cially when the patient is transfusion dependent and/​or has severe
signs and symptoms, there is an indication for long-​term treatment
with the complement inhibitor eculizumab.
Definition
Paroxysmal nocturnal haemoglobinuria (PNH) is an acquired
chronic disorder characterized by persistent intravascular haem-
olysis, subject to recurrent exacerbations, often associated with
cytopenias, and with a distinct tendency to venous thrombosis. The
triad of haemolytic anaemia, pancytopenia, and thrombosis makes
PNH a truly unique clinical condition: however, even in the absence
of one or more of these manifestations a conclusive diagnosis can be
made by appropriate laboratory investigations.
Epidemiology
PNH is encountered in all populations throughout the world, and
it can affect people of all socioeconomic groups. The prevalence
of PNH is not accurately known: however, it is rarer than the re-
lated disorder, acquired aplastic anaemia (AAA). An estimate of
the prevalence of PNH is between 1 in 100 000 and 1 in 1 million.
Like AAA, PNH may be somewhat less rare in South-​East and East
Asia. Most patients present as young adults, but we have seen PNH
in a 2-​year-​old child and in people in their seventies. PNH has
never been reported as a congenital disease, and there is only one
isolated case with inherited susceptibility. The sex ratio is not far
from even.
Clinical features
The patient may seek medical attention because, one morning,
he or she has ‘passed blood instead of urine’. This distressing or
frightening event—​the direct evidence of haemoglobinuria—​may
be regarded as the classic presentation; however, more frequently
the patient presents as a problem in the differential diagnosis of
anaemia, whether symptomatic or discovered incidentally. The
anaemia may be associated with jaundice, with neutropenia, with
thrombocytopenia, or any combination of these. Recurrent attacks
of abdominal pain are not uncommon; dysphagia and erectile dys-
function less common. In some patients, venous thrombosis may
be the first clinical manifestation. Although any vein may be af-
fected, the most common localization is intra-​abdominal, and
when thrombosis affects the hepatic veins, it may produce acute
hepatomegaly and ascites—​that is, Budd–​Chiari syndrome.
The natural history of PNH can extend over decades. In the past,
with only supportive treatment, the median survival was estimated
to be about 10 to 20 years (Fig. 22.5.3.1), with the most common
causes of death being thrombosis, infection associated with severe
neutropenia, and haemorrhage associated with severe thrombo-
cytopenia. With contemporary treatment including eculizumab (see
‘Eculizumab’), the lifespan may be nearing normal.
A patient with PNH may have a history of previous AAA. In
fact, with improved laboratory diagnosis the transition from AAA
to PNH is being documented increasingly, to the extent that PNH
evolving on a background of bone marrow failure may be the rule
rather than the exception. Conversely, a picture of overt AAA may
develop after years of PNH. Rarely (estimated at 1–​2% of all cases),
PNH may terminate in acute myeloid leukaemia. By contrast, full
spontaneous recovery from PNH has also been well documented.
Laboratory investigations and diagnosis
The most consistent finding in the blood count is anaemia, which may
range from mild to moderate to very severe. The anaemia is usually
normo-​macrocytic; a high mean cell volume is usually largely ac-
counted for by reticulocytosis, which may be quite marked—​up to
20%. The anaemia may become microcytic if the patient is allowed
to become iron deficient as a result of chronic urinary blood loss
through haemoglobinuria. The red cell morphology is otherwise usu-
ally normal. There may be neutropenia and/​or thrombocytopenia.
Unconjugated bilirubin is mildly or moderately elevated, lactate
dehydrogenase is typically markedly elevated, and haptoglobin is
usually undetectable.
Haemoglobinuria may be overt in a random urine sample; if it is not,
it may be helpful to obtain serial urine samples, since haemoglobinuria
can vary dramatically from day to day, and even from hour to hour (it is
more common, but not always, in the early morning: hence the adjec-
tive ‘nocturnal’). Obviously, haemoglobinuria must be distinguished
from haematuria and other causes of dark urine (Table 22.5.3.1).
Surprisingly, even today a patient may undergo extensive urological
investigations before it is realized that he/​she has PNH.
There may be free haemoglobin in the serum, and sometimes this
is so high as to interfere with clinical chemistry. These findings clearly
indicate a haemolytic anaemia with intravascular haemolysis. The
bone marrow is usually cellular, with marked to massive erythroid
Survival (%)
Median survival
allo-BMT
(Raiola et al.)
100
75
50
25
Years
1
5
10
15
20
25
Supportive care
(Hillmen et al.)
allo-BMT
(IBMTR)
allo-BMT
(IBMTR)
extrapolated
Fig. 22.5.3.1  PNH is a chronic disorder, the time course of which is often
measured in decades. From a series of 80 patients who received only
minimal supportive treatment, a median survival of about 10 years had been
estimated. Allogeneic bone marrow transplantation is still associated with
significant mortality and may have reduced the survival of some patients, but
more encouraging results have been reported on a small series from a single
centre (Raiola et al.). The time course and survival have changed markedly
with the introduction of eculizumab (Fig. 22.5.3.7), but the data shown here
are still relevant to countries where this drug is not available.
section 22  Haematological disorders
5350
hyperplasia, often with mild to moderate dyserythropoietic fea-
tures. However, at some stage of the disease the marrow may become
hypocellular or even frankly aplastic.
The definitive diagnosis of PNH must be based on demonstrating
that a substantial proportion of the patient’s red cells have an increased
susceptibility to complement due to the deficiency on their surface
of proteins (particularly CD59 and CD55) that normally protect the
red cells from activated complement. For decades this has been done
reliably by using the acidified serum (Ham–​Dacie) test. Nowadays,
the gold standard is flow cytometry that will display a bimodal (some-
times trimodal) distribution of red cells. Such analysis is also applied to
granulocytes (Fig. 22.5.3.2), revealing that the proportion of affected
granulocytes is almost always greater than that of affected red cells.
Pathophysiology
Haemolysis
Haemolysis in PNH is due to an intrinsic abnormality of the red
cell which makes it exquisitely sensitive to activated complement,
whether it is activated through the alternative pathway or through
the classic pathway. Activation through the so-​called tick-​over com-
ponent of the alternative pathway explains why there is chronic
haemolysis in PNH (Fig. 22.5.3.3, top panel). Activation through
the classic pathway, triggered by an antigen–​antibody reaction, ex-
plains why haemolysis can be dramatically exacerbated (with a con-
sequent paroxysm of haemoglobinuria) in the course of a viral or
bacterial infection. Hyper-​susceptibility to complement is due to the
deficiency of several protective membrane proteins, of which CD59
is the most important, mainly because it hinders the insertion into
the membrane of C9 polymers.
The molecular basis for the deficiency of these proteins
has been pinpointed not to a defect in any of the respective
genes, but rather to the shortage of a glycolipid molecule,
glycosylphosphatidylinositol (GPI), which through a peptide
bond anchors these proteins to the surface membrane of cells.
The shortage of GPI is due in turn to a mutation in an X-​linked
gene, called PIGA, required for an early step in GPI biosynthesis.
In virtually each patient the PIGA mutation is different. This is not
surprising since these mutations are not inherited: rather, each
one is a somatic mutation that takes place de novo in a haemopoi-
etic stem cell. As a result, the patient’s bone marrow is a mosaic
of mutant and nonmutant cells, and the peripheral blood always
contains both GPI-​negative PNH cells and GPI-​positive cells (Fig.
22.5.3.2).
Thrombosis
This is one of the most immediately life-​threatening complications
of PNH and yet one of the least understood pathogenetically. It
could be due to impaired fibrinolysis because the urokinase plas-
minogen activator receptor (uPAR) is a GPI-​linked protein, or to
activated complement causing hypercoagulability, or to activated
complement and haemoglobin in the plasma causing hyperactivity
of platelets, or any combination of these.
Table 22.5.3.1  Differential diagnosis of dark urine
Different sorts of dark urine
Causes
Additional tests
Possible diagnosis
Haematuria
Many
Clears on centrifugation
Mostly urinary tract or glomerular pathology
Myoglobinuria
Rhabdomyolysis
Ultrafiltration; spectroscopy
March myoglobinuria
Haemoglobinuria
Intravascular haemolysis
Repeat crossmatch
Incompatible blood transfusion
Donath–​Landsteiner antibody
Paroxysmal cold haemoglobinuria
G6PD activity
G6PD deficiency
Blood film for malaria parasites
Blood cultures
Red cell morphology
‘Blackwater fever’
Clostridium perfringens septicaemia
Microangiopathic haemolytic anaemia
Ham; flow cytometry for CD59
PNH
G6PD, glucose-​6-​phosphate dehydrogenase; PNH, paroxysmal nocturnal haemoglobinuria.
Erythrocytes
33%
400
Counts
300
200
100
0
67%
Granulocytes
60
40
20
15
10
5
Normal
PNH patient
Counts
Counts
Counts
15
M2
CD59–FITC
M1
M2
CD59–FITC
M1
0
CD59–FITC
400
300
200
100
0
95%
5%
CD59–FITC
M2
M1
Fig. 22.5.3.2  Flow cytometry analysis of blood cells in a patient with
PNH. On the left, red cells and granulocytes from a normal person
display a unimodal distribution of surface expression of the GPI-​linked
protein CD59, which protects red cells against complement-​mediated
lysis. On the right, a similar analysis reveals a clearly bimodal distribution
in a patient with PNH, and from this analysis the size of the PNH cell
population can be quantified. FITC, fluorescein-​isothiocyanate.
Courtesy of Dr David Araten.
22.5.3  Paroxysmal nocturnal haemoglobinuria
5351
Bone marrow failure and the relationship
between PNH and AAA
PNH has an intimate link to AAA, which manifests in several ways.
Firstly, as stated previously, PNH is often preceded by AAA, and
sometimes a patient with PNH becomes ‘less haemolytic’, ‘more
pancytopenic’, and ultimately evolves to frank AAA. Secondly, in-
tensive immunosuppression is a standard-​of-​care treatment in
AAA, and a beneficial response to the same treatment can be seen
also in patients with PNH. Thirdly, in terms of pathogenesis, AAA
is regarded as an organ-​specific autoimmune disease mediated by
‘activated’ cytotoxic (CD8+) T lymphocytes which are able to inhibit
haemopoietic stem cells. GPI-​specific CD1d-​restricted T cells have
been demonstrated in both PNH and AAA.
It therefore seems that an element of bone marrow failure is the
rule rather than the exception in PNH: an extreme view is that PNH
is a form of AAA in which bone marrow failure is masked by the
enormous expansion of the PNH clone that populates the patient’s
bone marrow. In other words, two different mechanisms cooperate
in producing PNH (Fig. 22.5.3.4):  autoimmune damage to stem
cells, and a somatic mutation in the PIGA gene. This notion is sup-
ported by two further lines of evidence: (1) by targeted inactivation
of the Piga gene in mouse embryonic stem cells one can produce
mice with a PNH cell population, but this population does not grow
further as it does in patients with PNH; and (2) by using refined flow
cytometry technology, PNH cells harbouring PIGA mutations can
be demonstrated in normal people at a frequency in the order of 10
per 1 million. Both these findings indicate that some other factor
is required, in addition to a somatic mutation in the PIGA gene, in
order to cause expansion of a PIGA mutant clone and thus PNH.
Most likely, the same cytotoxic damage to stem cells that would
otherwise cause AAA spares the PNH stem cells, thus allowing the
PNH clone to grow to the size when it gives clinical PNH. The pre-
cise mechanism whereby the PNH stem cells escape damage is not
yet known, but one possibility is that the GPI-​specific T cells men-
tioned previously damage GPI-​positive stem cells, whereas they are
harmless for the GPI-​negative stem cell from which the PNH clone
originates.
Complications
The most important complication is thrombosis, which is nearly
always venous, and can be life-​threatening especially if it affects
either the abdominal veins (Fig. 22.5.3.5) or the intracranial veins.
The Budd–​Chiari syndrome, because of its characteristic clinical
picture, is usually easy to recognize: however, in PNH it is some-
times associated with portal vein thrombosis, and this may limit
the extent of liver enlargement. Thrombosis of the splenic vein
should be suspected whenever a patient with PNH has, or develops,
splenomegaly. Thrombosis of one of the mesenteric veins is much
more difficult to diagnose clinically. Appropriate investigations
include Doppler ultrasonography, contrast-​enhanced CT, and
magnetic resonance imaging venography (for this purpose, pos-
sibly the most sensitive imaging technique). Recognizing venous
thrombosis is of great practical importance, because thrombolytic
Alternative
pathway
C5
C3
C3
Eculizumab
C5
C3
Alternative
pathway
C3
Coombs’ ++
Fig. 22.5.3.3  Red cells and complement in PNH. Top: in PNH in the
steady state, erythrocytes, by virtue of being deficient in the complement
regulators CD55 and CD59, suffer as a result of spontaneous (tick-​over)
complement activation, with consequent intravascular haemolysis
through formation of the membrane attack complex (MAC); exacerbated
haemolysis (‘paroxysm’) will result from activation of extra complement
through the classical pathway. Bottom: on eculizumab, PNH erythrocytes
are protected from haemolysis through C5 blockade, but continuing
upstream complement activation may lead to C3 opsonization (positive
Coombs’ test) and consequent extravascular haemolysis.
Modified from Luzzatto L, Risitano AM, Notaro R (2010). Paroxysmal nocturnal
hemoglobinuria and eculizumab. Haematologica, 95, 523–​6.
Target: GPI
PIGA mutation
in a HSC
Subclinical
GPI-negative
clone
Expansion of GPI-
negative clone
APLASTIC ANAEMIA
Target:
other molecules
T cell-mediated
autoimmune attack
against HSC
PNH
Fig. 22.5.3.4  The role of somatic mutation and bone marrow failure
in causing PNH. This diagram illustrates that two separate factors are
required to bring about PNH as a clinical disease. On the one hand, a
PIGA mutation on its own will produce a PNH clone, but there will be no
basis for it to expand; on the other hand, damage to haemopoietic stem
cells (HSC) can cause aplastic anaemia without PNH. When both factors
cooperate, and if the damage to HSC is GPI mediated, then there will
be selective expansion of the PNH clone. A patient with PNH often has
a history of aplastic anaemia, and whether aplastic anaemia or PNH is
diagnosed first will depend on the timing of the PIGA mutation as well as
on the kinetics of BMF and of the expansion of the PIGA mutant clone.
section 22  Haematological disorders
5352
therapy with tissue plasminogen activator has been carried out
successfully even after 6 weeks from the onset of signs and symp-
toms (Fig. 22.5.3.5).
Management
Bone marrow transplantation
Unlike other acquired haemolytic anaemias, PNH may be lifelong,
and this is important in our approach to management. The only
form of treatment that can provide a cure for PNH is allogeneic
bone marrow transplantation, which should be offered for consid-
eration to any young patient with PNH for whom a human leuco-
cyte antigen-​identical sibling is available. Results similar to those for
AAA can be expected, with long-​term disease-​free survival ranging
from 60 to 100% in the few series that have been published. By con-
trast, in a few cases in which bone marrow transplantation has been
carried out from unrelated donors the outcome in several PNH pa-
tients has not been good (Fig. 22.5.3.1).
Eculizumab
A major advance in the management of PNH has been the introduc-
tion of complement blockade by the use of a humanized monoclonal
antibody, eculizumab, which targets the C5 component of com-
plement (Fig. 22.5.3.3, bottom panel). This has proven effective in
controlling intravascular haemolysis, hence haemoglobinuria disap-
pears in most patients, and at least one-​half of the patients who were
transfusion dependent no longer require transfusions, and in others
the need for transfusions is significantly reduced (Fig. 22.5.3.6). In
addition, distressing symptoms such as abdominal pain are relieved
and hence quality of life improves. However, given its mechanism of
action, eculizumab is not a curative treatment: its benefits will last as
long as the agent is administered through an intravenous infusion at
fortnightly intervals.
Since only the distal complement pathway is blocked, in PNH
patients on eculizumab C3 fragments will bind to PNH red cells
that, being so opsonized, become susceptible to phagocytosis by
macrophages (Fig. 22.5.3.3). Thus, patients on eculizumab still have
haemolysis, but this is usually milder than what they had before
eculizumab and, being extravascular, it is far less symptomatic. It is
also important to note that (unlike intravascular haemolysis) there
is no iron loss with extravascular haemolysis, hence for the first time
PNH patients are at risk of iron overload if they still require blood
transfusion whilst on eculizumab.
As the distal complement pathway is blocked in patients on
eculizumab, they are at an increased risk for infection by menin-
gococcus, hence immunization against this organism is mandatory
before starting eculizumab. In most patients this treatment has been
remarkably free of serious side effects, but there have been a few
instances of severe infection which require immediate antibiotic
treatment.
A very significant extra advantage of eculizumab treatment is
that it also decreases the risk of thrombosis, which is especially
(a)
(b)
Fig. 22.5.3.5  Abdominal vein thrombosis in PNH can resolve with thrombolytic therapy.
(a) Extensive thrombus in the inferior vena cava in a patient with known PNH who had developed
Budd–​Chiari syndrome a few days earlier: it is not infrequent in PNH for thrombosis to involve
multiple veins in the abdomen all at once. (b) A thrombus-​free vena cava 2 days after an
intravenous infusion of tissue plasminogen activator.
Courtesy of Raymond Thertulien; see also Haematologica 97: 344, 2012.
Hgb≥11
33.3%
8≤Hgb<11
46.3%
≤50%
14.8%
50%
5,6%
n = 54
T
r
a
n
s
f
u
s
i
o
n
i
n
d
e
p
e
nd
e
n
c
e
T
r
a
n
s
f
u
s
i
o
n
d
e
p
e
n
d
e
n
c
e
Fig. 22.5.3.6  Effect of eculizumab treatment on blood transfusion
requirements in PNH. About two-​thirds of the patients are or have
become transfusion independent. The smaller sectors indicate patients
who still require blood transfusion, subdivided according to whether
the requirement is greater or less than one-​half of what it was before
eculizumab treatment.
Updated 2013 from Risitano AM et al. (2009).
22.5.3  Paroxysmal nocturnal haemoglobinuria
5353
important because patients with PNH are not fully protected
from thrombosis even by painstaking anticoagulant treatment.
Most patients on eculizumab have not only a better quality of life
but probably an increased survival (Fig. 22.5.3.7). Several suc-
cessful pregnancies have been supervised in PNH patients on
eculizumab.
Supportive care
Eculizumab is not available in many countries because it is very
expensive. It is therefore appropriate to remember that, before its
introduction, supportive management supervised by somebody
who has experience of PNH can help patients to ‘live with PNH’
for years, sometimes for decades, and sometimes with a good
quality of life.
The mainstay of support is the transfusion of filtered red cells
whenever necessary. Folic acid supplements (≥3 mg/​day) are
mandatory; the serum iron concentration should be checked
periodically and iron supplements added as indicated. There is
no evidence that prednisone (which used to be administered at a
dose of 15 to 30 mg on alternate days) decreases the rate of haem-
olysis, and long-​term administration of prednisone, even at a low
dosage, is contraindicated in view of its well-​known serious po-
tential side effects (a short course of prednisone may sometimes
appear helpful in dealing with an episode of massive haemo-
globinuria associated with intercurrent infection). Any patient
who has had a deep vein thrombosis should be given anticoagu-
lant prophylaxis.
Bone marrow failure
Eculizumab will have clearly no effect on the bone marrow failure
component of PNH. When the manifestations of bone marrow
failure predominate, the approach to treating PNH becomes similar
to that indicated for AAA: accordingly, a logical option is intensive
immunosuppressive treatment with antilymphocyte globulin (ALG)
and ciclosporin. Although no formal trial has ever been conducted,
this approach has particularly helped to relieve severe thrombo-
cytopenia and/​or neutropenia in patients in whom these were the
main problem(s): by contrast, there is often little beneficial effect
on the haemolysis itself. Thus, the therapeutic effects of ALG and
eculizumab are in a sense complementary. New approaches to in-
hibit complement in PNH are currently being investigated.
FURTHER READING
Araten D, et al. (1999). Clonal populations of hematopoietic cells with
paroxysmal nocturnal hemoglobinuria genotype and phenotype
are present in normal individuals. Proc Natl Acad Sci U S A, 96,
5209–​14.
Dacie JV (1999). The haemolytic anaemias, 3rd edition, Vol. 5.
Churchill Livingstone, London.
Dulau-Florea A, et al. (2019). Detection of paroxysmal nocturnal
hemoglobinuria (PNH) in bone marrow aspirates. Semin Hematol,
56(1), 65–8.
Gargiulo L, et  al. (2013). Glycosylphosphatidylinositol-​specific,
CD1d-​restricted T cells in paroxysmal nocturnal hemoglobinuria.
Blood, 121, 2753–​61.
Hillmen P, et al. (2006). The complement inhibitor eculizumab in par-
oxysmal nocturnal hemoglobinuria. N Engl J Med, 355, 1233–​43.
Hillmen P, et al. (2013). Long-​term safety and efficacy of sustained
eculizumab treatment in patients with paroxysmal nocturnal
haemoglobinuria. Br J Haematol, 162, 62–​73.
Kelly RJ, et al. (2015). Eculizumab in pregnant patients with parox-
ysmal nocturnal hemoglobinuria. N Engl J Med, 373, 1032–​9.
Luzzatto L, Bessler M, Rotoli B (1997). Somatic mutations in par-
oxysmal nocturnal hemoglobinuria:  a blessing in disguise? Cell,
88, 1–​4.
Luzzatto L, Gianfaldoni G, Notaro R. (2011) Management of parox-
ysmal nocturnal haemoglobinuria: a personal view. Br J Haematol,
153, 709–​20.
Parker CJ (2016). Update on the diagnosis and management of par-
oxysmal nocturnal hemoglobinuria (2016). Hematology Am Soc
Hematol Educ Program, 2016, 208–​16.
Risitano AM, Marotta S (2016). Therapeutic complement inhibition in
complement-​mediated hemolytic anemias: past, present and future.
Semin Immunol, 28, 223–​40.
Risitano AM, et al. (2009). Complement fraction 3 binding on erythro-
cytes as additional mechanism of disease in paroxysmal noc-
turnal hemoglobinuria patients treated by eculizumab. Blood, 113,
4094–​100.
Socié G, et al. (2019). Eculizumab in paroxysmal nocturnal haemo-
globinuria and atypical haemolytic uraemic syndrome: 10-year
pharmacovigilance analysis. Br J Haematol, 185(2), 297–310.
Takeda J, et al. (1993). Deficiency of the GPI anchor caused by a som-
atic mutation of the PIG-​A gene in paroxysmal nocturnal hemo-
globinuria. Cell, 73, 703–​11.
1
20
40
60
80
100
2
Time (years)
Pre-eculizumab n = 30
×  = 6.46
P = .01
On eculizumab n = 79
Cumulative % surviving
3
4
5
6
7
8
9
Fig. 22.5.3.7  Effect of eculizumab treatment on survival in PNH.
Survival curves calculated for patients before and after being on a
regular eculizumab regimen.
From Kelly RJ, et al. (2011). Long-​term treatment with eculizumab in paroxysmal
nocturnal hemoglobinuria: sustained efficacy and improved survival. Blood, 117,
6786–​92.

22.6 Erythroid disorders 5354 22.6.1 Erythropoiesi
22.6 Erythroid disorders 5354 22.6.1 Erythropoiesis 5354 Vijay G. Sankaran
CONTENTS
22.6.1	
Erythropoiesis  5354
Vijay G. Sankaran
22.6.2	
Anaemia: pathophysiology, classification, and
clinical features  5359
David J. Weatherall and Chris Hatton
22.6.3	
Anaemia as a challenge to world health  5366
David J. Roberts and David J. Weatherall
22.6.4	
Iron metabolism and its disorders  5371
Timothy M. Cox and John B. Porter
22.6.5	
Anaemia of inflammation  5402
Sant-​Rayn Pasricha and Hal Drakesmith
22.6.6	
Megaloblastic anaemia and miscellaneous
deficiency anaemias  5407
A.V. Hoffbrand
22.6.7	
Disorders of the synthesis or function of
haemoglobin  5426
Deborah Hay and David J. Weatherall
22.6.8	
Anaemias resulting from defective maturation
of red cells  5450
Stephen J. Fuller and James S. Wiley
22.6.9	
Disorders of the red cell membrane  5456
Patrick G. Gallagher
22.6.10	 Erythrocyte enzymopathies  5463
Alberto Zanella and Paola Bianchi
22.6.11	 Glucose-​6-​phosphate dehydrogenase
deficiency  5472
Lucio Luzzatto
22.6.12	 Acquired haemolytic anaemia  5479
Amy Powers and Leslie Silberstein
22.6.1  Erythropoiesis
Vijay G. Sankaran
ESSENTIALS
Biological mechanisms of erythropoiesis
Erythropoiesis is a highly regulated, multistep process in which stem
cells, after a series of amplification divisions, generate multipotential
progenitor cells, then oligo-​ and finally unilineage erythroid progen-
itors, and then morphologically recognizable erythroid precursors
and mature red cells.
Ontogeny of erythropoiesis—​this involves a series of well-​
coordinated events during embryonic and early fetal life. In the
fetus, the main site of erythropoiesis is the liver, which initially pro-
duces mainly fetal haemoglobin (HbF, α2γ2) and a small component
(10–​15%) of adult haemoglobin (HbA, α2β2), with the fraction of
HbA rising to about 50% at birth. After birth, the site of erythroid
cell production maintained throughout life is the bone marrow, with
the final adult erythroid pattern (adult Hb with <1% fetal Hb) being
reached a few months after birth.
Regulation of erythropoiesis—​the main regulator is erythropoietin, a
sialoglycoprotein that is produced by interstitial cells in the kidney in
response to tissue hypoxia and exerts its effect by binding to a spe-
cific receptor on erythroid burst-​forming units (BFU-​Es), erythroid
colony-​forming units (CFU-​Es), and proerythroblasts.
Abnormalities of erythroid production
Abnormal erythropoietin production—​anaemia can be caused by
acquired or congenital deficiency in erythropoietin production,
most commonly in chronic kidney disease. Impaired tissue oxygen
delivery is a common cause of erythropoietin-​driven secondary
erythrocytosis, often caused by chronic lung disease, sometimes by
22.6
Erythroid disorders
22.6.1  Erythropoiesis
5355
congenital heart anomalies, and rarely by haemoglobin mutations.
Some kidney cancers increase erythropoietin production and hence
cause secondary erythrocytosis.
Other causes of abnormal erythroid production—​these include
(1)  acquired and congenital defects in erythropoietin signalling;
(2) acquired and congenital defects in the transcription factors GATA1
or EKLF, which are required for activation of erythroid-​specific genes;
(3) acquired or congenital abnormalities in ribosome synthesis or
splicing factors (Diamond–​Blackfan anaemia and myelodysplastic
syndromes); and (4) factors that lead to premature red cell destruc-
tion, including inherited defects in protein structure (e.g. hereditary
spherocytosis), enzyme defects in metabolic pathways (e.g. pyruvate
kinase), and haemoglobin defects (e.g. sickle cell anaemia).
Introduction
Every second in humans, over two million red blood cells are pro-
duced in the bone marrow. This process is carefully coordinated
and is referred to as erythropoiesis. Alterations of erythropoiesis
can result in a variety of pathological conditions due to a mismatch
between the removal of red blood cells from the circulation and
the production of red cells in the bone marrow. Anaemia is very
common worldwide, with nearly 30% of the world population af-
fected with some form of this condition. Anaemia may be due to
nutritional, infectious, or systemic aetiologies, while other forms are
a consequence of intrinsic defects affecting developing precursors
or mature red blood cells directly. Erythropoiesis is frequently and
commonly perturbed in these varied aetiologies that result in an-
aemia. In this chapter, we discuss the process of erythropoiesis, how
this process varies in the course of human development, and how
erythropoiesis is regulated by both external factors and through in-
trinsic regulation. This chapter will introduce erythropoiesis as a
framework for subsequent chapters focused on the pathophysiology
of the disorders that arise as a result of its disruption.
The process of erythropoiesis
To ensure that red blood cells can be rapidly produced to balance
their continuous loss following a circulation time of approximately
120 days, an intricately regulated differentiation process has emerged
for red blood cell production from haematopoietic stem cells in the
bone marrow. Precise regulation of this process is necessary, since
it must be able to compensate rapidly for haemolysis or blood loss.
As described subsequently, this process is exquisitely sensitive to a
variety of molecules, which can regulate the rate at which erythro-
poiesis occurs. Feedback systems exist to ensure that the process of
erythropoiesis maintains a red blood cell count at homeostatic levels
in most individuals.
All blood cells arise initially from a rare population of cells that
are capable of lifelong maintenance of haematopoiesis, known as
haematopoietic stem cells (see Chapter  22.2.1). Approximately 1
in 104 to 105 cells in the bone marrow are estimated to be haem-
atopoietic stem cells. These haematopoietic stem cells can give rise
to all the differentiated cells that compose the blood, including red
blood cells, platelets, and the white blood cell lineages, including
both myeloid (granulocytes, monocytes, eosinophils, and baso-
phils) and lymphoid cells (T lymphocytes, B lymphocytes, and NK
cells). Traditional models suggest that a series of differentiation di-
visions can occur from the haematopoietic stem cells to give rise to
multilineage and then more restricted progenitors that are in turn
capable of producing the various terminally differentiated cells that
compose the blood. Work over several years has elucidated surface
markers that can allow enrichment of these progenitor populations
within the bone marrow. While substantial support exists for this
hierarchical model of haematopoiesis, more recent work suggests
that lineage commitment may occur in earlier haematopoietic stem
or progenitor cells that then give rise to each lineage through a sep-
arate differentiation process. Earlier studies may have been con-
founded by examining bulk populations of cells, rather than looking
at cells individually. It is likely that some combination of these
models exists in humans, and further work, particularly by exam-
ining patients with blood disorders, will be needed to gain further
insight into these models.
Regardless of the uncertainty of the early steps in haemato-
poietic differentiation, the subsequent stages of differentiation
that define erythropoiesis are well understood and characterized
(Fig. 22.6.1.1). The earliest committed erythroid progenitors are
the burst-​forming unit erythroid cells (BFU-​Es). These progen-
itors have traditionally been identified through the use of colony-​
forming assays in semisolid medium where a single BFU-​E is
capable of giving rise to colonies containing hundreds to thousands
ProE
BasoE
PolyE
OrthoE
Retic
RBC
Haemoglobin expression
BFU-E
CFU-E
HSC
Identified through
colony assays
Fig. 22.6.1.1  A schematic showing normal erythropoiesis. The haematopoietic stem cell (HSC) differentiates
to a series of multipotential progenitors that can then give rise to the earliest erythroid-​restricted progenitor, the
burst forming unit erythroid cell (BFU-​E). This cell can then give rise to the colony-​forming unit erythroid cell
(CFU-​E). These early erythroid progenitors can only be identified through the use of colony assays, where the form
colonies composed of numerous more differentiated erythroblasts. Subsequently, a series of morphologically
identifiable precursors are produced that include the proerythroblast (ProE), basophilic erythroblast (BasoE),
polychromatophilic erythroblast (PolyE), orthochromatophilic erythroblast (OrthoE), reticulocyte (Retic), and
mature red blood cells (RBCs). The increase in haemoglobin during this process is depicted below this scheme.
section 22  Haematological disorders
5356
of erythroblasts. The BFU-​E is thought to divide and give rise to an-
other more-​differentiated committed erythroid progenitor, known
as the colony-​forming unit erythroid cell (CFU-​E). CFU-​Es form
colonies earlier than BFU-​Es in semisolid culture medium, which
are composed of tens to hundreds of erythroblasts. Recent studies
have suggested methods that allow for the prospective enrichment
of both BFU-​Es and CFU-​Es in humans and mice using cell surface
molecules, although further validation of these methods is needed
before they can regularly supplant the traditional colony forming
assays used to assess such progenitor populations. It is important to
note that BFU-​E and CFU-​E cells in the bone marrow cannot be dis-
tinguished using morphology, since they appear identical to other
blast cells in the bone marrow.
Subsequent to the CFU-​E, a series of morphologically dis-
tinct and further differentiated erythroblast populations emerge
(Fig. 22.6.1.1). The CFU-​E initially differentiates into a proery­
throblast. The proerythroblast differentiates into the basophilic
erythroblast, which then differentiates into a polychromatophilic
erythroblast. The final stages of differentiation involve dramatically
increased haemoglobinization and nuclear condensation, which
characterizes the orthochromatic erythroblast (which is notable
for its haemoglobin-​rich cytoplasm and condensed nucleus). The
orthochromatic erythroblast then undergoes enucleation, thereby
giving rise to a reticulocyte. The reticulocyte can then enter the
circulation and terminally matures by losing the remaining ribo-
nucleic acids (RNAs) to become a mature red blood cell with the
characteristic biconcave shape. Throughout this process, with
every cell division, there is amplification in the number of cells pre-
sent, thus allowing for effective production of the numerous red
blood cells necessary to replace those continuously lost from the
circulation.
Developmental erythropoiesis
The previous framework describes the general process of erythro-
poiesis, as it occurs to maintain red blood cell numbers in the
circulation of the adult. However, erythropoiesis is critical for en-
suring sufficient oxygen transport capability throughout the body
at all stages of development. In the early stages of development,
red blood cells are produced via transient populations of progen-
itors, which are eventually supplanted by the haematopoietic stem
cells which maintain blood production throughout the life of the
organism.
In humans, the first erythroid cells are detected beginning at week
3 of gestation and continue to supply the key oxygen transport cap-
ability until week 8 of gestation. This early and transient population
of cells, which are produced in the yolk sac and enter the circulation
as nucleated cells, are known as the primitive erythroid cells. These
cells have been noted to eventually enucleate in both humans and
mice. This population has a key role in ensuring that development
can proceed effectively once passive oxygen diffusion can no longer
keep pace with the rapid growth of the embryo. The function and
development of the primitive erythroid population, which largely
occurs in the circulation of the embryo, has been studied in some
detail. While the differentiation process resembles that in the adult,
as discussed earlier, there are key differences. For example, there ap-
pears to be less expansion of this population compared to that seen
at later stages of erythropoiesis in the fetal liver or bone marrow. The
molecular regulators of this process are largely conserved, but some
differences exist here too, and in humans a unique set of embry-
onic haemoglobin genes are expressed, composing the majority of
haemoglobin at this stage.
Starting at week 7 to 8 of human gestation, the fetal liver begins
to be colonized by erythroblasts and the majority of red blood cell
production for the remainder of gestation shifts to this organ. The
enucleated red blood cells produced from these erythroid pre-
cursors enter the circulation by week 8 of gestation and fetal liver
erythropoiesis maintains the circulating population of red blood
cells for the remainder of gestation. The initial erythroid progen-
itors that seed the fetal liver are derived from a transient wave of
haematopoietic progenitors that are produced in the yolk sac. As
gestation proceeds, these progenitors are supplanted by those pro-
duced from haematopoietic stem cells, which arise in the major ar-
teries of the embryo proper (including the aorta, which is a major
site of haematopoietic stem cell production). The exact stage at
which the haematopoietic stem cell-​derived progenitors supplant
the yolk sac-​derived populations in the fetal liver is not well under-
stood, and in mouse models the yolk sac-​derived blood progenitors
can maintain sufficient blood production to allow the mouse em-
bryos to survive for the entirety of gestation. The fetal liver-​derived
erythroid progenitors display similar characteristics to their adult
counterparts and are both commonly referred to as definitive
erythroid cells. The red blood cells produced from the fetal liver ex-
press different haemoglobin subtypes to their adult counterparts, as
discussed later, and are larger in volume. In addition, the precursors
to these red cells appear to have higher cell cycle rates and may have
increased sensitivity to the key hormone erythropoietin, which is
discussed later.
While the bone marrow is seeded with haematopoietic progen-
itors in the first few weeks of gestation as rudimentary cartilage and
endothelial cells appear in this region, only a small amount of haem-
atopoiesis occurs here before birth, with the fetal liver being respon-
sible for the bulk of erythropoiesis. That which does occur in the
fetal marrow has a myeloid bias. However, after birth, haematopoi-
esis shifts primarily to the bone marrow.
Developmental haemoglobin switching
While the sites of haematopoiesis shift as noted previously, the
erythroid lineages are also noted to have unique switches in the
production of haemoglobin in these different populations. As al-
ready discussed, the yolk sac-​derived primitive erythroid cells ex-
press a variety of embryonic haemoglobins. The embryonic form
of α-​globin is known as ζ-​globin and is expressed only in the
primitive erythroid cells along with the adult α-​globin molecule.
ε-​Globin is an embryonic β-​like globin, the expression of which is
restricted to primitive erythroid cells. Subsequently, once the de-
finitive erythroid lineage is produced from the fetal liver, there is
expression of adult α-​globin along with production of the fetal β-​
like globins, known as γ-​globins, as well as a small amount of adult
β-​globin. Shortly before the time of birth, the predominance of γ-​
globin production within the definitive erythroid lineage switches
to high-​level expression of adult β-​globin—​a process termed the
fetal-​to-​adult haemoglobin switch. It should be noted that all α-​like
22.6.1  Erythropoiesis
5357
globins are encoded by genes at the same locus on chromosome 16
and similarly all β-​like globin genes are found at a single locus on
chromosome 11, where the genes are controlled by a common set
of regulatory elements. The developmental regulation of the genes
encoding these different globins has been studied extensively and
there is recent insight into specific transcriptional regulators of this
process, which is discussed later in the section on intrinsic regu-
lators of erythropoiesis. The process of haemoglobin switching can
be altered with specific mutations that occur in rare individuals and
there can be persistence of certain haemoglobins at later develop-
mental stages. For example, a variety of mutations, which primarily
occur within the β-​globin gene locus itself, can cause persistence of
fetal haemoglobin (composed of α-​ and γ-​globin) into adulthood in
a condition known as hereditary persistence of fetal haemoglobin.
The mutations that cause this condition are primarily deletions in
the β-​globin gene locus, but point mutations and alterations of regu-
latory factors have also been identified in rare cases. The regulation
of fetal haemoglobin is of clinical importance, since increased pro-
duction of γ-​globin can compensate for the defective production
of β-​globin in β-​thalassaemia and can prevent the clinical symp-
toms resulting from the mutant version of β-​globin found in sickle
cell disease (see also Chapter 22.6.7). Therefore, there has been a
long-​standing interest in attempting to stimulate fetal haemoglobin
production for therapeutic purposes in both sickle cell disease and
β-​thalassaemia.
Extrinsic regulation of erythropoiesis
As discussed earlier, the process of erythropoiesis must be carefully
coordinated to ensure that the production of red blood cells can be
appropriately tailored to the need for oxygen delivery throughout
the body. The regulation of erythropoiesis is primarily controlled by
the hormone erythropoietin (EPO), which is a single chain glyco-
protein. Camot and Deflandre initially hypothesized that a hormone
with the function of EPO existed in 1906. Work from Reissman,
Stohlman, and their colleagues in the 1950s demonstrated its ex-
istence. EPO was purified using biochemical approaches in the late
1970s, and in the early 1980s the gene encoding EPO was cloned.
This resulted in the production of recombinant forms of EPO, which
are in widespread clinical use today.
EPO has a short half-​life of less than 5 h in the circulation, as a re-
sult of degradation in the liver and urinary secretion. It is primarily
produced by a group of oxygen-​sensitive cells in the cortex and outer
medulla of the kidney and therefore its secretion can be increased
by over 100-​fold in cases of hypoxia when oxygen delivery to tissues
needs to be increased.
EPO acts by binding to the erythropoietin receptor (EPOR) that
is primarily expressed on erythroid progenitors and some early
precursors. This binding event promotes expansion and differen-
tiation of erythroid progenitors and precursors to produce mature
red blood cells. EPOR is a cytokine receptor that can be dimer-
ized by binding to a single molecule of EPO in an asymmetric
manner. This dimerization following binding triggers a conform-
ational change in the EPOR-​binding Janus kinase, JAK2, which
allows its phosphorylation and subsequent activation of a variety
of downstream signalling factors. This includes the STAT5 tran-
scription factor, which can trigger the transcriptional activation of
a variety of genes in the nucleus. In addition, JAK2 activates several
downstream cellular signalling pathways, including the mitogen
­activated protein kinase and phosphatidylinositol-​3 kinase path-
ways to mediate its activities.
In addition to the use of recombinant EPO, other therapeutic ap-
proaches have been taken to modulate this pathway. Peptides that
bind and activate the EPOR have been tested in clinical trials, but are
not currently in routine clinical use. In addition, increasing EPO ex-
pression in the kidney can be accomplished by activation of hypoxia-​
inducible factors, a group of transcription factors that stimulate EPO
gene transcription in the setting of hypoxia. A group of drugs known
as prolyl hydroxylase inhibitors are able to prevent the degradation
of the hypoxia-​inducible factors and thereby activate EPO transcrip-
tion. These drugs are in clinical trials as therapeutic agents to be used
in place of recombinant EPO.
The importance of this pathway for erythropoiesis in humans has
been elegantly demonstrated by rare human mutations that affect
this process. For example, in Chuvash polycythemia and some re-
lated conditions, the von Hippel–​Lindau (VHL) gene can be mu-
tated, resulting in excessive activation of the hypoxia-​inducible
factors that in turn cause EPO levels to increase in an unregulated
fashion. Rare mutations that cause an increased oxygen affinity in
red cells due to mutations in haemoglobin or in metabolic enzymes,
such as 2,3-​diphosphoglycerate mutase, can result in decreased
oxygen sensing by the kidney (as a result of impaired oxygen re-
lease from the red cell) and thus can cause increased EPO levels and
erythrocytosis (excessive red blood cell production). Rare mutations
in the EPOR can eliminate negative regulatory pathways and cause
excessive activation downstream from the receptor, again causing
erythrocytosis. In addition, in an acquired condition known as poly-
cythemia vera, mutations in JAK2 can result in excessive activation
downstream of EPOR and cause excessive red blood cell production
(see also Chapter 22.3.5).
Another well-​studied growth factor, stem call factor (SCF), binds
to the KIT receptor that is expressed on erythroid progenitors and
to some extent on proerythroblasts. SCF can synergize with EPO
to promote erythropoiesis. A variety of naturally occurring and in-
duced mouse mutations in either the SCF or KIT receptor cause im-
paired erythropoiesis with anaemia, macrocytosis, and a reduction
in the number of erythroid progenitors, including CFU-​Es, that are
present in the mice. However, to date, no human mutations in SCF
or KIT have been identified that affect erythropoiesis.
Recent work has also uncovered other regulatory pathways that
affect red blood cell production. During clinical testing of activin
ligand traps for use in bone disease, the effectiveness of these agents
for red blood cell production was serendipitously discovered.
Follow-​up studies showed that these ligand traps appear to block
the activity of the ligand GDF11, which may normally be a negative
regulator of erythropoiesis. Further studies are necessary to under-
stand whether this is the only target of these drugs and also how
GDF11 may normally affect erythropoiesis.
In addition, in the setting of inflammation or infections, a variety
of proinflammatory cytokines can be produced that block erythro-
poiesis. This can occur indirectly due to effects on EPO production
from the kidney or on iron availability, but this can also be due to
direct effects on the process of erythropoiesis. Proinflammatory
cytokines, including tumour necrosis factor (TNF)-​α and interferon
(IFN)-​γ, have been shown to block normal erythropoiesis.
section 22  Haematological disorders
5358
While the regulation of iron homeostasis is beyond the scope
of this chapter, it should be noted that a variety of extrinsic regu-
lators of iron homeostasis can significantly impact the iron avail-
able for erythropoiesis (and iron is absolutely essential for the
production of the haem prosthetic group found in every subunit
of haemoglobin). Most notable among these regulators is the
liver-​produced peptide hormone, hepcidin, which prevents iron
absorption by the intestine and limits iron release from macro-
phages. The regulation of iron is complex, but has a significant
impact on the availability of iron for erythropoiesis and can there-
fore impact normal red blood cell production profoundly (see also
Chapters 22.6.4 and 22.6.5).
Intrinsic regulation of erythropoiesis
While factors outside of the cell, including EPO and potentially
other factors such as GDF11, play an important role in regulating
erythropoiesis, much of the ultimate control of this process occurs at
the level of gene expression. A group of factors that regulate the pro-
cess of messenger RNA transcription plays a key role in regulating
erythropoiesis. Studies in model organisms and analysis of rare mu-
tations in humans have highlighted the importance of a few specific
factors that are necessary for erythropoiesis.
The transcription factor GATA1 has been shown to have a key
role in erythropoiesis and has therefore been termed a master regu-
lator of erythropoiesis. Most genes expressed during erythropoi-
esis require GATA1 for their expression. Mutations in GATA1 can
result in severe forms of anaemia, such as rare cases of Diamond–​
Blackfan anaemia or other forms of anaemia due to impaired
erythroid cell maturation. GATA1 plays a key role in both pro-
moting transcription of genes necessary for erythropoiesis and in
repressing other genes that would normally be expressed in other
lineages.
The transcription factor KLF1 has been shown to have key roles
in erythropoiesis, particularly in the late stages of this process.
Mutations in KLF1 in humans cause a variety of phenotypes, all
of which are characterized by late-​stage defects in erythropoiesis.
This includes a condition known as congenital dyserythropoietic
anaemia, which is characterized by impaired maturation of late-​
stage erythroblasts. In addition, some mutations in KLF1 have been
shown to result in increased fetal haemoglobin production. The ex-
tent to which this function can be separated from its role in normal
erythropoiesis needs to be studied further.
A variety of other transcription factors, including TAL1, GFI1B,
ZFPM1, LMO2, LDB1, and others, have also been implicated in
erythropoiesis, although they may also have pleiotropic roles in
other cell types. Many of these transcription factors appear to act to-
gether in complexes with GATA1 and KLF1 and this combinatorial
activity may explain how some genes are activated at different times
or in varying contexts during normal erythropoiesis.
Recent genomic studies have also highlighted the role of other
transcription factors that appear to have key roles in the produc-
tion of fetal haemoglobin. By following up on genome-​wide asso-
ciation studies, which were focused on examining common genetic
variation underlying the normal distribution of fetal haemoglobin
levels, the transcription factors BCL11A and MYB were found to
be important silencing factors for fetal haemoglobin. Humans with
reduced expression of BCL11A can have substantial persistence of
fetal haemoglobin, while erythropoiesis appears to be otherwise
essentially unperturbed. However, BCL11A and MYB have other
roles and approaches to target these and other factors are under ac-
tive investigation as potential therapies for sickle cell disease and
β-​thalassaemia. Much more work needs to be done to gain an im-
proved understanding of how factors including BCL11A, MYB, and
others act to regulate the haemoglobin genes during erythropoiesis;
this could also suggest further therapeutic avenues for stimulating
fetal haemoglobin production. Recent studies have suggested that
by simply altering the looping of regions of chromatin within a gene
locus, specific genes, such as the γ-​globin genes, can be activated in
adult erythroid cells. This suggests that factors such as BCL11A may
carry out their activity in γ-​globin silencing, at least in part, by al-
tering such looping interactions.
While a great deal is known about the transcriptional regulators
of erythropoiesis, far less is known about other processes that also
contribute to gene expression. Protein translation plays an increas-
ingly appreciated role in the control of gene expression in a variety
of settings. In erythropoiesis, the key role of translation has been
highlighted as a result of a rare disorder that results in a paucity of
erythroid progenitors and precursors in the bone marrow without
alterations in other lineages—​a condition known as Diamond–​
Blackfan or hypoplastic anaemia. The majority of mutations causing
Diamond–​Blackfan anaemia are in ribosomal protein genes that
compose the ribosome in every cell. Why mutations in these genes
could cause an anaemia is poorly understood. However, recent
work has identified rare gene mutations in GATA1 that can cause
Diamond–​Blackfan anaemia and this has led to the finding that
ribosomal protein mutations appear to impair the protein trans-
lation of GATA1. Much more work is needed to understand how
protein translation contributes to erythropoiesis both in healthy in-
dividuals and in pathological states, but this recent work suggests
that promising new insight into both normal erythropoiesis and
some forms of anaemia can be gained through such studies.
Concluding remarks
In this chapter, we have provided a brief overview of the process by
which red blood cells are produced. Erythropoiesis is critical for en-
suring that red blood cells are available at all stages of development,
and, since this process varies during gestation, we have discussed
the current understanding of its developmental regulation. In add-
ition, we have attempted to provide an overview of the key external
and intrinsic regulators of erythropoiesis. We have provided a few
examples to highlight how these processes are perturbed in disease
and subsequent chapters will build upon this framework to discuss
specific disorders that affect erythropoiesis in more detail.
FURTHER READING
An X, et  al. (2014). Global transcriptome analyses of human and
murine terminal erythroid differentiation. Blood, 123, 3466–​77.
Arlet JB, et  al. (2014). HSP70 sequestration by free alpha-​globin
promotes ineffective erythropoiesis in beta-​thalassaemia. Nature,
514, 242–​6.

22.6.10 Erythrocyte enzymopathies 5463 Alberto Zan
22.6.10 Erythrocyte enzymopathies 5463 Alberto Zanella and Paola Bianchi
22.6.10  Erythrocyte enzymopathies
5463
gene, with dominant inheritance). In both cases, abnormal cation
flux results. Diagnosis of these rare conditions rests on the iden-
tification of appropriate red cell morphology in the context of
a family history of haemolytic anaemia; confirmatory genetic
testing of candidate genes is possible, and red cell cation concen-
trations may be measured in the research setting.
Other conditions
Other conditions associated with hereditary stomatocytosis in-
clude the Rh deficiency syndromes, sitosterolaemia, and familial
deficiency of high-​density lipoproteins. The rare Rh deficiency
syndromes are associated with mild to moderate haemolytic an-
aemia and absent (Rhnull) to decreased (Rhmod) erythrocyte ex-
pression of Rh antigens associated with mutation in the RHD
and RHAG genes. Sitosterolaemia is associated with early-​onset
atherosclerosis, anaemia, and macrothrombocytopenia asso-
ciated with mutation of the ABCG5/​ABCG8 cotransporters,
leading to increased intestinal absorption and decreased biliary
elimination of sterols, particularly those derived from plants.
Familial deficiency of high-​density lipoproteins (Tangier disease,
OMIM 205400) is due to mutation of ABCA1, a protein critical
for cellular cholesterol export, leading to accumulation of tissue
cholesterol esters manifest as enlarged orange-​yellow tonsils,
hepatosplenomegaly, cloudy corneas, neuropathy, and prema-
ture atherosclerosis. Affected patients exhibit mild to moderate
haemolytic anaemia.
Acquired stomatocytosis has been observed in a large number of
conditions, particularly hepatobiliary disease and acute alcoholism.
Acquired stomatocytosis has also been seen in patients with various
malignant neoplasms, cardiovascular disease, and after the adminis-
tration of vinca alkaloids.
Acanthocytosis
Acanthocytes are dense, contracted erythrocytes with irregular
‘thorny’ projections. Acanthocytes may also been found on the
peripheral smears of patients with abetalipoproteinaemia, the
McLeod red cell phenotype, or one of the neuroacanthocytosis
syndromes. Abetalipoproteinaemia (OMIM 200100) is associated
with hypolipidaemia, fat malabsorption, progressive ataxia, retin-
itis pigmentosa, and poor growth in childhood due to the inability
to produce or secrete the B apoproteins B100 and B48, or defects
in the microsomal triglyceride transfer protein (MTTP), required
for production of apoprotein B-​containing β-​lipoproteins. The X-​
linked McLeod phenotype (OMIM 314850) is due to mutation of
XK, necessary for Kell antigen expression. Affected individuals
experience compensated anaemia, susceptibility to Kell D antigen
alloimmunization, and late-​onset myopathy and nervous system
abnormalities.
The neuroacanthocytosis syndromes are a heterogeneous group
of neurodegenerative disorders including the McLeod syndrome,
chorea-​acanthocytosis due to mutation of chorein or VPS13A,
Huntington disease-​like 2 due to mutation of junctophilin-​3, and
pantothenate
kinase-​associated
neurodegeneration
(formerly
known as Hallervorden–​Spatz syndrome and its allelic variant
HARP syndrome—​hypobetalipoproteinaemia, acanthocytosis, ret-
initis pigmentosa, and pallidal degeneration) due to mutations of
pantothenate kinase 2. The cause of the acanthocytosis in these dis-
orders is unknown.
Acquired acanthocytosis may be seen in patients with severe
hepatic disease (commonly known as spur cell anaemia), hypothy-
roidism, malnutrition, and after splenectomy.
FURTHER READING
Andolfo I, et al. (2016). New insights on hereditary erythrocyte mem-
brane defects. Haematologica, 101, 1284–​94.
Bennett V, Healy J (2008). Organizing the fluid membrane bilayer:
diseases linked to spectrin and ankyrin. Trends Mol Med, 14, 28–​36.
Beris P, Picard V (2015). Non-​immune hemolysis: diagnostic consid-
erations. Semin Hematol, 52, 287–​303.
Da Costa L, et al. (2013). Hereditary spherocytosis, elliptocytosis, and
other red cell membrane disorders. Blood Rev, 27, 167–​78.
De Franceschi L, Bosman GJ, Mohandas N (2014). Abnormal red cell
features associated with hereditary neurodegenerative disorders:
the neuroacanthocytosis syndromes. Curr Opin Hematol, 21, 201–9.
Gallagher PG (2013). Abnormalities of the erythrocyte membrane.
Pediatr Clin North Am, 60, 1349–​62.
King M-​J, et al. (2015). The International Council for Standardization in
Haematology guidelines for the laboratory diagnosis of nonimmune
hereditary red cell membrane disorders. Int J Lab Hematol, 37, 304–​25.
Landry ML (2016). Parvovirus B19. Microbiol Spectr, 4, 3.
Lusher JM, Barnhart MI (1980). The role of the spleen in the
pathoophysiology of hereditary spherocytosis and hereditary
elliptocytosis. Am J Pediatr Hematol Oncol, 2, 31.
Lux SE 4th (2016). Anatomy of the red cell membrane skeleton: un-
answered questions. Blood, 127, 187–​99.
Mohandas N (2018). Inherited hemolytic anemia: a possessive
beginner’s guide. Hematology, 30, 377–81.
Niss O, et al. (2016). Genotype-​phenotype correlations in hereditary
elliptocytosis and hereditary pyropoikilocytosis. Blood Cells Mol
Dis, 61, 4–​9.
Pasini EM, et al. (2006). In-​depth analysis of the membrane and cyto-
solic proteome of red blood cells. Blood, 108, 791–​801.
Risinger M, Glogowska E, Chonat S, et al. (2018). Hereditary xero­
cytosis: Diagnostic considerations. Am J Hematol, 93, E67–E69.
22.6.10  Erythrocyte enzymopathies
Alberto Zanella and Paola Bianchi
ESSENTIALS
Numerous enzymes, including those of the hexose monophosphate
and glycolytic pathways, are active in the red cell. They are required
for the generation of ATP (needed to supply energy for sodium ex-
trusion) and the reductants NADH and NADPH (necessary to main-
tain haemoglobin in its active ferrous atomic state, as well as for the
integrity of sulphydryl groups present on essential proteins). 2,3-​
Diphosphoglycerate (2,3-​DPG), an intermediate of glucose metab-
olism, is a key regulator of the affinity of haemoglobin for oxygen, and
accessory enzymes are also active for the synthesis of glutathione, dis-
posal of oxygen free radicals, and for nucleotide metabolism.
section 22  Haematological disorders
5464
With the exception of heavy metal poisoning and rare cases of
myelodysplasia, most red cell enzyme deficiency disorders are in-
herited. They may cause haematological abnormalities, (most
commonly nonspherocytic haemolytic anaemias, but also rarely
polycythaemia or methaemoglobinaemia, manifest with autosomal
recessive or sex-​linked inheritance), and may also be associated
with nonhaematological disease (e.g. neuromuscular symptoms)
when the defective enzyme is expressed throughout the body.
Some may mirror important metabolic disorders, without produ-
cing haematological problems, making them of diagnostic value
(e.g. galactosaemia). Others are of no known clinical consequence.
With rare exceptions, it is impossible to differentiate the enzymatic
defects from one another by clinical or routine laboratory methods.
Diagnosis depends on the combination of (1) accurate ascertain-
ment of the family history; (2) morphological observations—​these
can determine whether haemolysis is present, rule out some causes
of haemolysis (e.g. hereditary spherocytosis and other red blood cell
membrane disorders), and diagnose pyrimidine 5′-​nucleotidase de-
ficiency (prominent red cell stippling); (3) estimation of red cell en-
zyme activity; and (4) molecular analysis.
The most common red cell enzyme defects are glucose-​6-​phos-
phate dehydrogenase deficiency (see Chapter  22.6.11), pyruvate
kinase deficiency, glucose-​6-​phosphate isomerase deficiency, pyr-
imidine 5′-​nucleotidase deficiency—​which may also induced by
exposure to environmental lead—​and triosephosphate isomerase
deficiency.
Introduction
Erythrocytes contain a large number of enzymes required to carry
out a variety of metabolic processes. Some inherited deficiencies
of these enzymes are designated red cell enzymopathies. They may
cause haematological disorders, including nonspherocytic haemo-
lytic anaemias, polycythaemia, and methaemoglobinaemia. Other
deficiencies do not produce haematological disorders, but instead
mirror important metabolic disorders such as galactosaemia that are
therefore of diagnostic value.
Moreover, when the defective enzyme is expressed throughout
the body, haemolytic anaemia is associated with systemic
nonhaematological manifestations such as neurological dysfunc-
tion, intellectual disability, myopathy, and susceptibility to infec-
tions. This section deals with those red cell enzyme defects that
cause haemolytic anaemia. Many have been described; most are
rare but some are sufficiently common that several hundred or even
thousands of cases have been documented. Although the enzym-
atic bases of these defects are very different, the clinical presenta-
tion is similar and relatively nondescript. Except for pyrimidine
5′-​nucleotidase deficiency, in which red cell stippling is a prominent
feature, it is impossible reliably to differentiate the enzymatic defects
from one another by clinical or routine laboratory methods.
Red cell metabolism
The two major pathways of red cell glucose metabolism are illus-
trated in Fig. 22.6.10.1. Glucose is phosphorylated to glucose 6-​
phosphate in the hexokinase reaction. It is then either metabolized
in the anaerobic Embden–​Meyerhof pathway or is oxidized in the
glucose 6-​phosphate dehydrogenase (G6PD) reaction, entering the
hexose monophosphate pathway. Anaerobic metabolism of glucose
phosphorylates ADP to ATP, providing energy to maintain erythro-
cyte shape and to transport molecules into and out of erythrocyte.
It also reduces NAD to NADH, which serves to reduce meth-
aemoglobin to haemoglobin. The hexose monophosphate pathway
reduces NADP to the NADPH and thus serves to maintain gluta-
thione and protein sulphydryl groups in the reduced state. These
pathways are similar in red cells, other tissues, and in ‘lower’ organ-
isms. However, the 2,3-​diphosphoglycerate (2,3-​DPG) shunt is a
unique feature of the Embden–​Meyerhof pathway in erythrocytes.
This ‘energy clutch’ of erythrocyte metabolism not only allows flexi-
bility in the amount of ATP that is generated in glycolysis, but also
provides a source of 2,3-​DPG, the key modulator of haemoglobin
oxygen affinity.
There are, in addition, many other metabolic functions that the
erythrocyte must carry out. Among these are the synthesis of gluta-
thione, the synthesis and degradation of nucleotides and nucleosides,
Glucose
Glucose 6-phosphate
Fructose 6-phosphate
Fructose 1, 6-diphosphate
ATP
ADP
1, 3-Diphosphoglycerate
3-Phosphoglycerate
2-Phosphoglycerate
Phosphoenolpyruvate
2 ADP
2 ATP
Pyruvate
Lactate
Glyceraldehyde 3-
phosphate
Dihydroxyacetone-
phosphate
P
e
n
t
o
s
e
Hexokinase
Glucose phosphate
isomerase
Phosphofructokinase
Glyceraldehyde-3-phosphate dehydrogenase
Phosphoglycerate kinase
Monophosphoglycerate mutase
Enolase
Pyruvate
kinase
Lactic
dehydrogenase
DPG
mutase
2, 3 DPG
phosphatase
Aldolase
Triose phosphate
isomerase
2, 3-
Diphosphoglycerate
3PG
2PG
EM Pathway
Hexose monophosphate
shunt
Rapoport-
Luebering
shunt
ATP
ADP
2 ADP
2 ATP
1, 3-Diphosphoglycerate
NADH
NAD+
Fig. 22.6.10.1  The relationship between the main red cell Embden–​
Meyerhof glycolytic pathway (EM) and the other metabolic pathways.
The insert shows the production of 2,3-​DPG in the Rapoport–​Luebering
shunt. DPG, diphosphoglycerate; PG, phosphoglycerate.
22.6.10  Erythrocyte enzymopathies
5465
the detoxification of active oxygen radicals, and the transport of
small molecules into and out of the cell.
The lack of protein synthesis in the mature red cell means that
none of the enzymes in the metabolic pathways can be replaced
during the red cell lifespan. Over the 120 days of normal red cell sur-
vival, enzyme activities decline at variable but predictable rates. This
decline probably contributes to the ageing process of the red cell.
Many of the abnormalities that affect red cell metabolism pro-
voke haemolytic anaemia. The type (acute or chronic) and degree
of anaemia depends on the metabolic cycle involved, the relative
importance of the affected enzyme, the functional properties of
the mutant enzyme with regard to kinetic abnormalities and/​or in-
stability, and the ability to compensate for the enzyme deficiency
by overexpressing isoenzymes or using alternative pathways. The
degree of anaemia may therefore vary from mild or fully compen-
sated haemolysis to life-​threatening neonatal anaemia and jaun-
dice necessitating exchange transfusion and subsequent continuous
transfusion support. Typical clinical symptoms may also include
splenomegaly, jaundice, gallstones, and iron overload (with the
latter seen in both transfused and nontransfused patients due to the
down-​regulation of hepcidin).
Genetics
Hereditary red blood cell (RBC) enzymopathies are caused by mu-
tations in genes coding for red cell enzymes. Half of the normal
activity of red cell enzymes is generally sufficient for normal func-
tion, thus most haemolytic anaemias due to red cell enzyme defi-
ciencies are inherited as autosomal recessive conditions. The only
exception is the haemolytic anaemia associated with elevated ad-
enosine deaminase activity, which is inherited as an autosomal
dominant disorder. Only two of the deficiencies, those of G6PD
and phosphoglycerate kinase (PGK), are encoded by genes on the
X chromosome.
Extensive mutation analysis at the DNA level has been carried out
on patients with most of red cell enzyme defects, and most mutations
have been found to be missense mutations or nonsense mutations;
a few deletions, insertions, and splicing mutations have also been
described, as have occasional mutations of regulatory machinery.
Many of the mutations that affect red cell metabolism and provoke
haemolytic anaemia cause instability and premature inactivation of
the enzyme; other mutations directly affect catalytic activity. The
molecular characterization of the defective enzyme has assumed
an increasingly valuable role in the diagnosis of these disorders,
including prenatal diagnostics and genetic counselling, especially as
our understanding of the genetic basis of these disorders expands
thanks to the development of high-​throughput technologies, such as
next-​generation sequencing.
Epidemiology
Red cell enzyme defects have a worldwide distribution. For most,
there are no exact and verified figures regarding their prevalence,
in part due to the lack of certified registries, and in part due to the
diagnostic difficulty resulting from their heterogeneous and over-
lapping clinical pictures. Pyruvate kinase deficiency (PKD), which
is the most common glycolytic defect, has an estimated preva-
lence of 1:20 000 in the general Caucasian population as assessed
by gene frequency studies. Reported cases, however, are fewer than
predicted—​likely because of under-​reporting, misdiagnosis and
prenatal death in severe cases.
For some of the other rarer enzyme defects, only a few cases are
reported in the literature (Table 22.6.10.1).
Specific red cell abnormalities that may cause
haemolytic anaemia
Table 22.6.10.1 summarizes some of the clinical and genetic char-
acteristics of red cell enzyme deficiencies.
The more common red cell enzyme abnormalities
G6PD deficiency
This important enzymopathy is described in Chapter 22.6.11.
Pyruvate kinase deficiency
Pyruvate kinase (PK) catalyses the conversion of phos­
phoenolpyruvate to pyruvate, coupled with the synthesis of ATP.
PKD, which is the most common glycolytic defect, is typical of con-
genital nonspherocytic haemolytic anaemias (CNSHAs) caused by
red cell enzymopathies.
PKD has been shown to have a protective effect against rep-
lication of the malaria parasite in human red cells, although
it is not clear whether PKLR mutant alleles are more prevalent
in malaria endemic areas. Clinical manifestations of PKD com-
prise the usual hallmarks of chronic haemolysis of variable se-
verity, that is, anaemia, jaundice, and splenomegaly. The degree of
anaemia varies widely, from very mild or fully compensated to
life-​threatening neonatal anaemia and jaundice requiring ex-
change transfusion and subsequent continuous transfusion sup-
port. Hydrops fetalis has also been reported in rare cases. The
anaemia tends to improve as infants grow, whereas it is constant in
adults with exacerbations in the context of infections, stress, and
pregnancy. Since 2,3-​DPG is elevated in individuals with PKD,
the anaemia may be better tolerated than in other conditions, be-
cause the oxygen dissociation curve is shifted to favour unloading
of oxygen in the tissues.
Iron overload is common in PKD, in both chronically transfused
and untransfused individuals. In nontransfused subjects, predis-
posing factors for iron loading include splenectomy, a certain de-
gree of ineffective erythropoiesis, and coinheritance of hereditary
haemochromatosis mutations. Monitoring iron status and meas-
uring tissue iron burden by R2 (Ferriscan) or T2* magnetic reson-
ance imaging (MRI) methods is indicated.
PKD is one of the most difficult red cell enzymopathies to diag-
nose. As well as its clinical heterogeneity, the enzyme itself is com-
plex with allosteric properties. In PKD, the residual enzyme activity
is not always greatly reduced and may even be normal in some cases
(Fig. 22.6.10.2). Meaningful enzyme levels can only be achieved
after total removal of leucocytes, which have up to 300 times the
PK activity of red cells. No correlation has been observed between
the residual PK activity, the degree of haemolysis, and the overall
clinical severity. The block of glycolysis at the level of PK is associ-
ated with an elevated 2,3-​DPG concentration (two to three times
normal).
section 22  Haematological disorders
5466
Table 22.6.10.1  Main features of red cell enzyme deficiencies
OMIM
Transmission
Gene
Chromosome
No. of cases
No. of
mutations
Haematological
symptoms
Other symptoms
Response to
splenectomy
Embden–​Meyerof pathway
Hexokinase
235700
AR
HK1
10q22.1
20 cases
5
CNSHA
++
Glucose-​6-​phosphate
isomerase
172400
AR
GPI
19q13.11
50 fam
31
CNSHA
Intellectual disability?
++
Phosphofructokinase
232800
AR
PFK-​M
PFK-​L
12q13.11
21q22.3
50–​100 cases
23
Erythrocytosis
Minimal haemolysis
Muscle disease,
Tarui’s disase
(glycogenosis type VII)
0
Aldolase
103850
AR
ALDOA
16p11.2
6 cases
4
CNSHA
Intellectual disability
Dysmorphism
?
Triosephosphate isomerise
190450
AR
TPI1
12p13
50–​100 cases
18
CNSHA
Neuromuscular disease
Infections
0
Phosphoglycerate kinase
300653
X-​linked
PGK1
X13.3
40 cases
20
CNSHA
Neuromuscular disease
++
Pyruvate kinase
266200
AR
PKLR
1q22
500 fam
200
CNSHA
++
Rapoport–​Luebering shunt
Bisphosphoglycerate mutase
222800
AR
BPGM
7q33
6 fam
3
Erythrocytosis, CNSHA
Hexose-​monophosphate
shunt
Glucose-​6-​phosphate
dehydrogenase
305900
X-​linked
G6PD
Xq28
400 × 106 cases
180
CNSHA—​acute
Favism
Glucose-​6-​phosphate
dehydrogenase (class I)
305900
X-​linked
G6PD
Xq28
50 fam
60
CNSHA—​chronic
Glutathione metabolism
Glutathione synthetase
AR
GSS
20q11.22
50 fam
32
CNSHA
Metabolic acidosis (5-​oxoprolinuria)
Neurological symptoms
Drug/​infections-​induced
haemolysis
0
Glutathione reductase
138300
AR
GSR
8p21.1
2 fam
3
Induced oxidative HA,
favism, neonatal jaundice
Cataract
?
γ-​Glutamylcysteine
synthetase (glutamate
cysteine ligase)
230450
AR
GCLC
GCLM
6p12.1
1p21
12 fam
6
CNSHA,
oxidative HA
Metabolic acidosis
(5-​oxoprolinuria)
Neurological symptoms
?
Glutathione peroxidase
AR
GPX1
3p21.3
1
0
Acute intravascular
haemolysis?
?
Nucleotide metabolism
Adenosine deaminase
(hyperactivity)
102730
AD
ADA
20q13.12
3 fam
0
CNSHA
Adenylate kinase
AR
AK1
9q34.11
12 fam
7
CNSHA
Motor impairment
Intellectual disability
?
Pyrimidine-​5’-​nucleotidase
606224
AR
NT5C3A
7p14.3
60 fam
26
CNSHA
++
NADH-​cytochrome b5
reductase
230800
AR
CYB5R3
22q13.2
50 cases
45
Methaemoglobinaemia
Neuromuscular disease
Intellectual disability
AD, autosomal dominant; AR, autosomal recessive; CNSHA, congenital nonspherocytic haemolytic anaemia; fam, families; HA, haemolytic anaemia. On a scale of 0 to 4+ where 4+ is a complete response, not usually required. In many
cases data are meagre.
22.6.10  Erythrocyte enzymopathies
5467
Red cell morphology is commonly unremarkable, displaying
anisocytosis and a variable portion of spur cells or acanthocytes,
particularly after splenectomy. None of these findings is specific for
PKD. Since the younger PK-​deficient erythrocytes are selectively se-
questered by the spleen, splenectomy usually results in a marked in-
crease of reticulocytes.
More than 220 different mutations have been described in PKLR
gene, causing PKD; many individuals are compound heterozygotes.
Not surprisingly, there is enormous genetic heterogeneity between
affected subjects, reflected in the multiplicity of quantitative and kin-
etic defects. In European populations, the most common mutations
are c.1529 G>A (Arg510Gln) and c.1456 C>T (Arg486Trp). These
mutations have not been detected among Asians with PKD; among
Roma peoples, the characteristic mutation results in a deletion of
exon 11, while a mutation at c.1436G>A (Arg479His) is commonly
found in Amish populations.
No curative therapy for PKD is available to date and the treatment
is therefore based on supportive measures. Red cell transfusions
may be required in severely anaemic patients, particularly in the first
years of life; the haemoglobin then tends to stabilize in many cases at
about 80 g/​litre. Splenectomy does not arrest haemolysis, and should
be reserved for severely affected, young patients who need regular
blood transfusions, and to patients who do not tolerate anaemia. It
usually results in an increase of 10 to 30 g/​litre in haemoglobin, re-
ducing or even eliminating in most cases transfusion requirements
(Fig. 22.6.10.3). Bone marrow transplantation has been performed
in a few very severely affected children.
Other red cell enzyme abnormalities
These are summarized in Box 22.6.10.1.
Glucose-​6-​phosphate isomerase deficiency
Glucose-​6-​phosphate isomerase deficiency is a further cause of
CNSHA along with G6PD deficiency and PKD. This enzyme cata-
lyses the second step of glycolysis, the interconversion of glucose
6-​phosphate to fructose 6-​phosphate. It is also known as phos­
phohexose isomerase, phosphoglucose isomerase, autocrine motility
factor, and neuroleukin, indicating that the protein has other actions
in other cells. Most reported cases of glucose-​6-​phosphate isomerase
deficiency present with mild to moderate haemolytic anaemia, but
hydrops fetalis has also been described. In some rare deficient patients,
neurological impairment or intellectual disability has been reported.
16
16
12
12
10
8
8
6
4
4
0
1600
1200
800
400
0
0
2
4
6
8
10
12
14
0
2
4
6
PK activity (IU/gHb)
PK activity (IU/gHb)
PK activity (IU/gHb)
(a)
(b)
(c)
Hb (g/dl)
Reticulocytes (109/litre)
8
10
12
14
2
Fig. 22.6.10.2  Enzyme activity in PK deficient patients. ●, not splenectomized patients, ●, splenectomized patients.
TX/yr
2
0
18
16
14
12
10
8
6
4
2
0
Hb
(g/dl)
16
12
Children
(a)
(b)
Adults
11
10
9
8
7
6
5
4
3
Fig. 22.6.10.3  Effect of splenectomy on (a) Hb levels and on (b) transfusion needs in PKD patients. ●/​■, before splenectomy, ○/​□, after
splenectomy.
section 22  Haematological disorders
5468
About 30 different mutations have been identified, underlining
the molecular heterogeneity and consequently the clinical hetero-
geneity of this disorder. The response to splenectomy is usually sat-
isfactory although it does not fully correct the haemolysis.
Pyrimidine 5′-​nucleotidase deficiency
Pyrimidine 5′-​nucleotidase catalyses the dephosphorylation of
the pyrimidine nucleotides uridine monophosphate (UMP) and
cytidine monophosphate (CMP) in the corresponding nucleosides
uridine and cytidine. The defect of pyrimidine 5′-​nucleotidase
leads to a marked accumulation of pyrimidine nucleotides that
interferes with the adenine nucleotide pool, and results in a
haemolytic anaemia characterized by pronounced basophilic stip-
pling in the red cells (Fig. 22.6.10.4). The appearance of the blood
film is similar to that seen in lead poisoning and the haemolytic
anaemia associated with this condition is likely to be due to the
inhibition of pyrimidine 5′-​nucleotidase. In congenital forms, the
haemolysis is usually mild to moderate, although severe cases have
also been reported. The accumulation of pyrimidine nucleotides
can be easily documented by measuring the ultraviolet absorp-
tion spectrum. Splenectomy commonly results in stabilization
of the haemoglobin to higher levels. Pyrimidine 5′-​nucleotidase
activity may also be decreased in aplastic anaemia and the tran-
sient erythroblastopenia of childhood, which can potentially lead
to misdiagnosis.
Triose-​phosphate isomerase deficiency
Triose-​phosphate isomerase (TPI) catalyses the interconversion
of dihydroxyacetone-​phosphate and glyceraldehyde 3-​phosphate,
which is then metabolized along the glycolytic pathway. The en-
zyme is ubiquitously expressed. TPI deficiency, reported in more
than 50 cases, results in a multisystem disorder consisting of
chronic haemolytic anaemia, increased propensity to infections,
and severe progressive neuromuscular degeneration (the latter
likely due to the formation of toxic protein aggregates induced by
misfolded TPI). With few exceptions, death occurs usually at about
the age of 5 years, and may occur as a consequence of cardiac dys-
rhythmia. Several point mutations have been identified that lead
to this disease; the most predominant is Glu104Asp, linked by
common haplotypes suggesting descent from a common ancestor.
There is no specific treatment, but supportive care, including as-
sisted ventilation, may prolong life in some instances.
The rare red cell enzyme deficiencies
Hexokinase deficiency
Hexokinase catalyses the phosphorylation of glucose to glucose
6-​phosphate, the first step in the glycolytic pathway. The defect of
this enzyme has been recorded in about 20 cases characterized by
molecular and phenotypic heterogeneity. Most patients have mod-
erately reduced activity; complete hexokinase deficiency is prob-
ably lethal. The enzyme deficiency results in moderate to mild
anaemia, but some cases present with severe anaemia and death in
the neonatal period. Typically, reduced hexokinase activity is as-
sociated with a low concentration of 2,3-​DPG within the red cells.
Patients have lower exercise tolerance for a given level of haemo-
globin than would be expected, because of the left shift in the oxygen
dissociation curve.
Hexokinase deficiency is often difficult to diagnose because the
activity of this enzyme is much higher in young red cells than in
older erythrocytes. As a result, hexokinase activity is usually in-
creased in patients with haemolytic anaemia of any type. In patients
with hexokinase deficiency, this often gives rise to the anomalous
finding that the red cell enzyme activity in the affected patient is
normal, and indeed is usually higher than that found in the hetero-
zygous parents. A careful examination of erythrocyte enzyme activ-
ities is therefore recommended, using other age-​dependent enzymes
such as PK or G6PD as an internal control.
Phosphofructokinase deficiency
Phosphofructokinase catalyses a reaction in which fructose 6-​
phosphate is phosphorylated to fructose 1,6-​diphosphate, ATP
being the donor of the phosphate group. Under normal physio-
logical conditions, this may be the major rate-​limiting step in gly-
colysis in the red cell. Erythrocytes contain two types of genetically
distinct phosphofructokinase subunits, L (liver) and M (muscle),
organized in tetramers composed of M and L subunits; there may
be five isoenzymes composed of different numbers of L and M sub-
units. Phosphofructokinase is a homotetramer of M subunits (M4)
in muscle and of L subunits (L4) in liver. A third subunit is found
in platelets. Deficiency of the M subunit causes haemolysis, but
the haemoglobin level in the blood is often normal or even higher
than normal because of the diminished 2,3-​DPG levels that are
Box 22.6.10.1  Other red cell enzyme abnormalities
•	 Glucose-​6-​phosphate isomerase deficiency
•	 Pyrimidine 5′-​nucleotidase deficiency
•	 Triosephosphate isomerase deficiency
Very rare enzyme defects
•	 Hexokinase deficiency
•	 Phosphofructokinase deficiency
•	 Phosphoglycerate kinase
•	 Aldolase deficiency
•	 Deficiency of enzymes of glutathione cycle
•	 Adenylate kinase deficiency
•	 Red cell adenosine deaminase hyperactivity
Fig. 22.6.10.4  Peripheral blood film in pyrimidine 5′-​nucleotidase
deficiency showing basophilic stippling (arrows).
22.6.10  Erythrocyte enzymopathies
5469
characteristic of this disorder. Muscle enzyme activity is also com-
promised, and a myopathy results. It is characterized by muscle
cramps and myoglobinuria on exertion. This disorder is sometimes
designated Tarui’s disease or type VII glycogenosis. Shortened red
cell viability may be a minor component of this disease. Deficiency
of the L subunit of phosphofructokinase has also been reported, but
without any clinical consequence.
Aldolase deficiency
Aldolase catalyses the conversion of fructose 1,6-​diphosphate to
glyceraldehyde 3-​phosphate and dihydroxyacetone phosphate.
There are three aldolase isoenzymes in human tissues, A, B, and C;
A isoenzyme is expressed in red cells. The deficiency of aldolase A is
extremely rare. The defect results in haemolytic anaemia which may
be associated with intellectual disability, dysmorphic features, or
myopathy.
Phosphoglycerate kinase
PGK catalyses the reversible phosphotransfer reaction from 1,3-​
bisphosphoglycerate to 3-​phosphoglycerate with production of
ATP. PGK deficiency is an uncommon X-​linked inherited disorder.
The enzyme is monomeric and is expressed in all tissues. PGK defi-
ciency shows a wide clinical phenotype: it may present with chronic
nonspherocytic haemolysis, behavioural disturbances, neurological
impairment (intellectual disability, and ataxia), and myopathy (ex-
ercise intolerance or muscle weakness). A few affected individuals
suffer the full spectrum of symptoms, whereas some cases are de-
scribed with myopathy and no haemolysis.
Diphosphoglycerate mutase deficiency
The DPG mutase is a multifunctional enzyme catalysing the reaction
from 1,3-​DPG to 2,3-​DPG. DPG mutase deficiency more frequently
leads to erythrocytosis than haemolytic anaemia, because a lack of
this enzyme prevents the formation of 2,3-​DPG. Consequently, the
oxygen affinity of the red cells is increased, stimulating erythropoi-
esis and resulting in secondary erythrocytosis.
Enzymes of glutathione synthesis
Glutathione is the major intracellular thiol in aerobic cells, and is
equally important in the erythrocytes. It has a number of critical
functions: protecting cells against oxidative damage, participa-
tion in detoxification of foreign compounds, maintenance of pro-
tein sulphydryl groups in a reduced state, and possibly transport
of amino acids. In red cells, its main function is as an antioxidant.
Glutathione is synthesized from glutamate, cysteine, and glycine
by two ATP-​dependent reactions catalysed by γ-​glutamylcysteine
synthetase and glutathione synthetase (Fig. 22.6.10.5). Defects
of both enzymes, occurring with a recessive mode of transmis-
sion, are rare and result in chronic nonspherocytic haemolytic
anaemia with increased susceptibility to oxidative stress. Severe
deficiency leads to 5-​oxoprolinuria, metabolic acidosis, and in-
tellectual disability. Complete loss of glutathione synthesis is
probably lethal.
Oxidized glutathione is reduced to glutathione by the action of
glutathione reductase, the hydrogen donor being NADPH.
Only a single family with severe, hereditary deficiency of gluta-
thione reductase has been described. No haemolysis was present, ex-
cept after ingestion of fava beans. Low activity of red cell glutathione
reductase, a flavin enzyme, is found when the intake of riboflavin
is suboptimal, but this mild or moderate enzyme deficiency has no
clinical consequences.
Adenylate kinase deficiency
Adenylate kinase modulates the interconversion of adenine nu-
cleotides catalysing the reversible phosphoryl transfer reac-
tion from ATP to AMP and ADP. Adenylate kinase deficiency is
a rare genetic disorder, with only 12 affected families reported in
the literature. From a clinical point of view, erythrocyte adenylate
Glutamic acid    +   Cysteine
Glycine +    y-Glutamylcysteine
y-Glutamylcysteine
synthetase
Glutathione
synthetase
Glutathione
S-transferase
Glutathione
peroxidase
Glutathione reductase
NADPH MetHb
reductase
Hb Fe2+
MetHb
Reduced glutathione
Oxidized glutathione
H2O
H2O2
Catalase
NADP+
NADPH
G6PD
G6P
6PG
ATP
ATP
RSSR
RSH
ADP
ADP
Fig. 22.6.10.5  The glutathione cycle and synthetic pathway. Redox control is exercised by the glutathione cycle linked to the NADPH of the
pentose phosphate pathway by glutathione reductase.
section 22  Haematological disorders
5470
kinase deficiency is associated with moderate to severe chronic
nonspherocytic haemolytic anaemia in all cases but one. In add-
ition, psychomotor impairment has also been observed in a few
patients. The relationship between this enzyme deficiency and
haemolytic anaemia remains unclear.
Red cell adenosine deaminase hyperactivity
Adenosine deaminase (ADA) catalyses the deamination of both
adenosine to inosine and 2′-​deoxyadenosine to 2′-​deoxyinosine.
Mutations causing a decreased ADA activity do not affect RBC
metabolism, but induce a severe immunodeficiency with an auto-
somal recessive inheritance. Only the overexpression of ADA, with
an autosomal mode of inheritance, produces haemolytic anaemia.
The ADA that is formed appears to be normal and the abnormality
that causes this tissue-​specific increase in enzyme activity has not
yet been discovered.
Specific red cell abnormalities that do not cause
haemolytic anaemia
Severe deficiencies of many red cell enzymes do not produce haem-
atological abnormality or, indeed, in many cases, any clinical ab-
normality at all. Included are deficiencies of 6-​phosphogluconate
dehydrogenase, δ-​aminolaevulinic acid dehydrase, acetyl-
cholinesterase, AMP deaminase, carbonic anhydrase, catalase,
galactokinase, galactose-​1-​phosphate uridyltransferase, glutathione
peroxidase,
hypoxanthine-​guanine
phosphoribosyltransferase,
ITPase, and phosphoglucomutase. Discussion of these enzyme defi-
ciencies is beyond the scope of this chapter.
General approach in diagnosis of red cell
enzymopathies, differential diagnosis
This is summarized in Box 22.6.10.2.
The clinical and haematological features of most enzyme defects
are common to other haemolytic diseases, so that the diagnostic
process in these rare disorders is the final step of a diagnostic
workup based not only on laboratory tests but also on clinical
examination, personal and family history, and the exclusion of the
most common causes of acquired haemolytic anaemia. The diag-
nosis ultimately depends upon the demonstration of low enzyme
activity and molecular analysis. However, given the rarity and the
wide clinical heterogeneity, the diagnosis of these defects can be
difficult.
The main laboratory features include the following:
•	Increased unconjugated bilirubin and lactate dehydrogenase
levels, reduced haptoglobin, increased absolute reticulocyte
number, and negative direct antiglobulin test.
•	Normochromic anaemia, sometimes producing a slight
macrocytosis. An increased mean cell haemoglobin concentration
is occasionally seen in severe cases due to dehydration brought
about by ATP deficiency.
•	Usually unremarkable RBC morphology with anisocytosis and a
variable portion of spur cells (which are not specific), particularly
after splenectomy. The presence of prominent red cell stippling
suggests a diagnosis of pyrimidine 5′-​nucleotidase deficiency.
•	Normal osmotic fragility, normal or sometime increased fluor-
escence intensity of RBC labelled with eosin-​5-​maleimide (flow-​
cytometry EMA-​binding test).
•	Noninformative osmotic gradient ektacytometry or Laser-​assisted
Optical Rotational Cell Analyzer (LoRRca).
The coexistence of CNSHAs and neuromuscular symptoms is
suggestive of triose-​phosphate isomerase deficiency, PGK, or other
rare enzyme defects.
Specific laboratory tests include red cell enzyme activity assays
and molecular testing.
Red cell enzyme assays
The diagnosis of the causative enzyme disorder underlying CNSHAs
is best achieved by measuring red cell enzyme activity by a quan-
titative assay using spectrophotometric methods. Quantification of
rarer RBC enzyme activities is a more specialized task that can be ac-
complished by the use of standardized techniques in an experienced
laboratory. There are a number of caveats that must be taken into
account, both with respect to the performance of red cell enzyme
assay and the interpretation of the results:
•	Leucocyte contamination: in some cases (i.e. PKD), the leucocyte
isoenzyme displays much higher activity than that of red cells.
Thus, contamination of a red cell suspension with a relatively
small number of white cells may obscure the diagnosis.
•	Reticulocytosis: the interpretation of the results of a red cell en-
zyme assay may also be confounded by the fact that the blood of
patients with haemolytic anaemia is enriched with reticulocytes
and young erythrocytes. Since many of the mutations that cause
red cell enzymopathies result in the production of unstable en-
zymes, the young circulating erythrocytes may actually contain
normal or near-​normal levels of enzyme. It is therefore essential
to take into account the age of the circulating cells. Family studies
may help the diagnosis in these cases.
•	Recent transfusions: it is clearly best to wait until just before a
transfusion to draw blood for testing.
Moreover, problems in interpretation may also arise when the
activity of an enzyme as measured in vitro does not accurately re-
flect its intracellular in vivo activity. This arises because of the ne-
cessity of using exceedingly high substrate concentrations for the in
vitro assay. This can be particularly problematic in the case of PKD,
because this complex allosteric enzyme has binding sites for two
Box 22.6.10.2  General approach in diagnosis of red
cell enzymopathies
•	 The diagnostic process is the final step of a diagnostic workup based
not only on laboratory tests but also on clinical examination, and per-
sonal and family history.
•	 Due to the rarity and the wide clinical heterogeneity, the diagnosis of
these defects can be difficult.
•	 The diagnosis of red cell enzymopathies ultimately depends on the
exclusion of other causes of haemolytic anaemia and upon the dem-
onstration of defective enzyme activity.
•	 Molecular testing is strongly recommended to confirm the diagnosis,
and for genetic counselling.
22.6.10  Erythrocyte enzymopathies
5471
substrates, ADP and phosphoenolpyruvate, and also for fructose di-
phosphate, an allosteric effector.
Molecular characterization
Molecular testing now plays an increasing role in the diagnosis of
most red cell enzyme defects. Molecular characterization confirms
the diagnosis and is necessary for genetic counselling. Genotype–​
phenotype correlations have been performed in some enzyme
defects such as PK, glucose-​6-​phosphate isomerase, PGK, and pyr-
imidine 5′-​nucleotidase deficiencies. Generation and characteriza-
tion of recombinant variants helped to define the effects of amino
acid replacement on the enzyme’s functional properties, and to cor-
relate genotype to clinical phenotype. However, different putative
modifiers may alter the expression of the mutant enzyme, such as
differences in splenic function, differences in genetic background
(such as some functional polymorphisms of other glycolytic en-
zymes, e.g. HFE and UGT1A1 genotypes), compensatory expression
of other isozymes, a variable degree of ineffective erythropoiesis,
or other abnormalities in unknown regulatory regions/​unknown
genes. In case of incomplete molecular characterization (mutations
identified at heterozygous level or no mutation at all identified in
the suspected gene), other causes of haemolysis should be excluded
wherever possible.
Molecular analysis has some advantages over enzyme assay-​based
diagnosis. First, DNA is very stable, even before it is purified, so
transporting samples of blood to reference laboratories is less of a
logistical problem. Transfused red cells do not pose a problem in
performing DNA-​based diagnosis, since transfused leucocytes do
not persist in the circulation.
Complications
Iron overload may be a complication in RBC enzyme defects, not
only in chronically transfused patients but also in untransfused
subjects, particularly in PKD. Predisposing factors for iron loading
include splenectomy, a certain degree of ineffective erythropoiesis,
and coinheritance of hereditary haemochromatosis (HFE gene) mu-
tations. Ferritin levels should be monitored and R2/​T2* MRI moni-
toring may be required.
Gallstone formation occurs frequently in haemolytic anaemias
and may requires periodic abdominal ultrasound monitoring.
Extramedullary haematopoiesis is a potential complication, and
paraspinal lesions causing cord compression may arise. Leg ulcers,
similar to those reported in patients with sickle cell disease and her-
editary spherocytosis, have been reported in patients with severe
anaemia.
Management of red cell enzyme defects
The treatment of red cell enzyme defects is based on supportive
measures: folate therapy is recommended in severe and moderate
forms of haemolytic anaemia; red cell transfusions may be re-
quired in severely anaemic patients, particularly in the first years
of life, during aplastic crises, infections, and pregnancy. Some pa-
tients may require splenectomy. Considering the heterogeneous
aetiology it is not surprising that the response may be difficult to
predict. In general, splenectomy is beneficial in some patients suf-
fering from deficiencies of PK, hexokinase, glucose-​6-​phosphate
isomerase, PGK, and pyrimidine 5′-​nucleotidase. Prior to splen-
ectomy, all patients should receive immunizations according to the
Centers for Disease Control and Prevention guidelines for patients
with asplenia, including the pneumococcal, meningococcal, and
haemophilus influenzae vaccines. Postsplenectomy, these vaccines
should be boosted at regular intervals according to the same guide-
lines and any newly recommended vaccines should be administered
in children and adults. Penicillin prophylaxis generally is recom-
mend for 1 to 2 years postsplenectomy; some clinicians recommend
lifelong prophylaxis, though this is more controversial. PKD has
rarely been treated by stem cell transplantation.
Future directions for therapy
No specific drug therapies are available for the routine clinical treat-
ment of enzyme defects. A pharmacological activator of PKLR is
presently in clinical trials. A gene therapy strategy has not been so
far attempted. Studies in mice have shown prolonged expression
of PK in the blood and haematopoietic organs of mice after intro-
duction of a viral vector expressing human liver-​type PK cDNA.
Additionally, other investigators have shown increased expression of
PK in transgenic mice using a gene-​addition strategy. More recently,
healthy erythroid cells has been obtained from gene-​edited PKD-​
induced pluripotent stem cells by nonintegrative lentiviral vectors.
FURTHER READING
Ayi K, et al. (2008). Pyruvate kinase deficiency and malaria. N Engl J
Med, 358, 1805–​10.
Beutler E, et al. (1977). International committee for standardization in
haematology: recommended methods for red cell enzyme analysis.
Br J Haematol, 35, 331–​40.
Beutler E (1984). Red cell metabolism: a manual of biochemical meth-
ods, 3rd edition. Grune & Stratton, New York.
Beutler E (2007). PGK deficiency. Br J Haematol, 136, 3–​11.
Bianchi P, et al. (2003). Molecular characterization of six unrelated
Italian patients affected by pyrimidine 5’-​nucleotidase deficiency. Br
J Haematol, 122, 847–​51.
Chiarelli LR, et al. (2012). Molecular insights on pathogenic effects of
mutations causing phosphoglycerate kinase deficiency. PLoS One, 7,
e32065.
Fermo E, et al. (2012). A new variant of phosphoglycerate kinase de-
ficiency (p.I371K) with multiple tissue involvement: molecular and
functional characterization. Mol Genet Metab, 106, 455–​61.
Garate Z, et al. (2015). Generation of a high number of healthy eryth-
roid cells from gene-​edited pyruvate kinase deficiency patient-​
specific induced pluripotent stem cells. Stem Cell Reports, 5, 1053–​66.
Grace RF, et al. (2015). Erythrocyte pyruvate kinase deficiency: 2015
status report. Am J Hematol, 90, 825–​30.
Hipkins R, et  al. (2009). Images in haematology. Paravertebral
extramedullary haemopoiesis associated with pyruvate kinase defi-
ciency. Br J Haematol, 147, 275.
Koralkova P, van Solinge WW, van Wijk R (2014). Rare hereditary red
blood cell enzymopathies associated with hemolytic anemia: patho-
physiology, clinical aspects, and laboratory diagnosis. Int Jnl Lab
Hem, 36, 388–​97.

22.6.11 Glucose- 6- phosphate dehydrogenase defici
22.6.11 Glucose- 6- phosphate dehydrogenase deficiency 5472 Lucio Luzzatto
section 22  Haematological disorders
5472
Meza NW, et al. (2009). Rescue of pyruvate kinase deficiency in mice
by gene therapy using the human isoenzyme. Mol Ther, 17, 2000–​9.
Pey AL, Maggi M, Valentini G (2014). Insights into human
phosphoglycerate kinase 1 deficiency as a conformational disease
from biochemical, biophysical, and in vitro expression analyses.
J Inherit Metab Dis, 37, 909–​16.
Ryan C, et al. (2004). Myelodysplastic syndrome in a patient with her-
editary pyruvate kinase deficiency. Hematol J, 5, 91–​2.
Wax JR, et al. (2007). Pyruvate kinase deficiency complicating preg-
nancy. Obstet Gynecol, 109, 553–​5.
Zanella A, Bianchi P (2000). Red cell pyruvate kinase deficiency: from
genetics to clinical manifestations. Baillieres Best Pract Res Clin
Haematol, 13, 57–​81.
Zanella A, Bianchi P, Fermo E (2007). Pyruvate kinase deficiency.
Haematologica, 92, 721–​3.
Zanella A, Fermo E, Valentini G (2006). Hereditary pyrimidine
5′-​nucleotidase deficiency: from genetics to clinical manifestations.
Br J Haematol, 133, 113–​23.
Zanella A, et al. (2005). Red cell pyruvate kinase deficiency: molecular
and clinical aspects. Br J Haematol, 130, 11–​25.
22.6.11  Glucose-​6-​phosphate
dehydrogenase deficiency
Lucio Luzzatto
ESSENTIALS
Deficiency of the enzyme glucose-​6-​phosphate dehydrogenase
(G6PD) in red blood cells is an inherited abnormality due to mu-
tations of the G6PD gene on the X chromosome that renders the
cells vulnerable to oxidative damage. The condition is widespread
in many populations living in or originating from tropical and sub-
tropical areas of the world because it confers a selective advantage
against Plasmodium falciparum malaria.
Clinical features
G6PD deficiency is mostly an asymptomatic trait, but it predisposes
to acute haemolytic anaemia in response to exogenous triggers,
including (1) ingestion of fava beans—​favism; (2) certain bacterial
and viral infections; and (3) some drugs—​notably some antimalarials
(e.g. primaquine), some antibiotics (e.g. sulphanilamide, dapsone,
nitrofurantoin), and even aspirin in high doses. Other manifestations
include (1) severe neonatal jaundice; and (2) chronic nonspherocytic
haemolytic anaemia—​the latter is only seen with rare specific gen-
etic variants.
The acute haemolytic attack typically starts with malaise, weak-
ness, and abdominal or lumbar pain, followed by the development
of jaundice and passage of dark urine (haemoglobinuria). Most epi-
sodes resolve spontaneously.
Diagnosis, prevention, and treatment
Diagnosis relies on the direct demonstration of decreased activity
of G6PD in red cells: a variety of screening tests are available, with
(ideally) subsequent confirmation by quantitative assay. Prevention
is by avoiding exposure to triggering factors of previously screened
subjects. Prompt blood transfusion is indicated in severe acute
haemolytic anaemia and may be life-​saving.
Definition
Glucose-​6-​phosphate dehydrogenase (G6PD) is a key enzyme in
redox metabolism. G6PD deficiency (OMIM 305900) is an inherited
condition in which red cells have a markedly decreased activity of
G6PD, which predisposes to haemolytic anaemia.
Epidemiology
G6PD deficiency is distributed worldwide (Fig. 22.6.11.1). There
are areas of high prevalence (up to 10–​20% or more) in Africa,
southern Europe, the Middle East, South-​East Asia, and Oceania. In
the Americas and in parts of northern Europe, G6PD deficiency is
also quite prevalent as a result of migrations that have taken place in
relatively recent historical times.
Genetics and molecular genetics
G6PD deficiency is inherited as an X-​linked Mendelian trait. The
gene encoding G6PD maps to the subtelomeric region of the long
arm of the X-​chromosome (band Xq28). It consists of 13 exons
and spans some 18.5 kb, and is expressed in all cells. Some 190 dif-
ferent G6PD mutations have been identified in G6PD-​deficient
subjects: all are in the coding sequence, except for one that affects
splicing (Fig. 22.6.11.2). Nearly all are missense point mutations
producing single amino acid replacements in the G6PD protein: in
most cases these cause G6PD deficiency by decreasing the in vivo
stability of the protein; more rarely they affect its catalytic function
(Table 22.6.11.1). Some mutations are small in-​frame deletions of
one to eight amino acids; and in a few instances there are two point
mutations rather than one (for instance, in G6PD A–​, the variant
most commonly encountered in Africa).
In view of the multitude of mutations, it is not surprising that the
clinical manifestations associated with different G6PD deficiency
variants may differ. The most important distinction is between
those variants for which an exogenous factor is required to trigger
haemolysis, and those variants for which this is not required which
manifest with a chronic nonspherocytic haemolytic anaemia. The
latter, more severe clinical phenotype can be ascribed in most cases
to an extreme degree of instability of the enzyme; for example, a
cluster of mutations that map to the dimer interface may severely
compromise the formation of the dimer, and the result is chronic
nonspherocytic haemolytic anaemia.
Interestingly, the G6PD gene is physically close to the genes for
haemophilia A, dyskeratosis congenita, and colour blindness; and
it overlaps with the gene that is mutated in the serious skin dis-
order incontinentia pigmenti. The X-​linkage of the G6PD gene has
important implications. First, as males have only one G6PD gene
(i.e. they are hemizygous for this gene), they must be either normal
or G6PD deficient. By contrast, females, having two G6PD allelic
22.6.11  Glucose-6-phosphate dehydrogenase deficiency
5473
genes, can be either normal, or deficient (homozygous), or inter-
mediate (heterozygous). Moreover, as a result of the phenomenon
of X-​chromosome inactivation, heterozygous females are genetic
mosaics, and this in turn has clinical implications. Indeed, in most
other (autosomal) enzyme deficiencies, heterozygotes are asymp-
tomatic because cells with an enzyme level close to 50% of normal
are biochemically normal. But in the case of G6PD, as a result of X-​
inactivation, the abnormal cells of a woman heterozygous for G6PD
deficiency are just as deficient as those of a hemizygous deficient
man, and therefore just as susceptible to pathology (Fig. 22.6.11.3).
Thus, it is not correct to refer to G6PD deficiency as an X-​linked re-
cessive trait, because recessive implies, by definition, not expressed
in a heterozygote: instead, G6PD deficiency is expressed in hetero-
zygotes both biochemically and clinically (Fig. 22.6.11.3). Although
it is true that heterozygotes are generally less severely affected (than
G6PD-​deficient males), this will depend on the proportion of
G6PD-​deficient red cells, which varies in different women from 1%
(the phenotype will be normal) to 99% (the phenotype will be like
that of a G6PD-​deficient male).
Biochemistry and pathophysiology
Red cells are very vulnerable to oxidative damage for two reasons.
First, oxygen radicals are generated continuously from within the
red cells as haemoglobin cycles from its deoxygenated to its oxy-
genated form. Second, red cells are directly exposed to a variety
of exogenous oxidizing agents. Oxygen radicals produced by such
compounds are converted by superoxide dismutase to hydrogen per-
oxide, which is itself highly toxic. G6PD, the first enzyme of the pen-
tose phosphate pathway (Fig. 22.6.11.4), catalyses the conversion of
glucose 6-​phosphate (G6P) and nicotinamide adenine dinucleotide
phosphate (NADP) to 6-​phosphogluconolactone and NADPH. The
most important product of the G6PD reaction, certainly in red cells,
is NADPH. First, by producing glutathione (GSH) via GSH reduc-
tase, it is crucial for the operation of GSH peroxidase; in addition,
it stabilizes catalase. These are the two enzymes able to detoxify
hydrogen peroxide (by converting it to water). Normally, G6PD
G6PDd allele freq
Polymorphic G6PD variants
32.5%
30%
25%
20%
15%
10%
5%
0%
Malaria free
A- (202A)
A- (968C)
Aures
Canton
Kaiping
Cosenza
Mediterranean
Taipei
Union
Viangchan
Local variant
Mahidol
Santamaria
Seattle
Coimbra
Chatham
Fig. 22.6.11.1  Global distribution of G6PD deficiency. Colour shades on the map indicate the median predicted allele frequency of G6PD deficiency
in malaria endemic and malaria-​eliminating countries, according to the geostatistical model designed by Rosalind E. Howes and coworkers. Coloured
circles illustrate the geographic distribution of some polymorphic G6PD alleles present in several regions. (We have used triangles for G6PD A− (968C,
L323P), in order to distinguish it from G6PD A− (202A, V68M); note that both of these mutations are always found associated with 376G, N126D).
Dark grey circles indicate ‘local’ polymorphic variants that have been detected only in a single population.
Santamaria
Seattle
Aures
Cosenza
A - (968C)
A
s
S
Taipei
Coimbra
Chatham
Mahidol
Viangchan
Union
Canton
Mediterranean
A - (202)
Kaiping
(a)
(b)
1 kb
100 bp
1
2
2
3
4
5
6
7
8
9
10
11 12 13
3 4
5
6
7
8
9
10 11 12 13
A
a
m
t
s
Mi
U
k
C
z
a
C
i
h
m
v
t
k
A
M
C
U
∫
V
S
A h
Fig. 22.6.11.2  Heterogeneity of G6PD deficiency. The 13 exons of the
G6PD gene are drawn approximately to scale; the introns (not drawn to
scale) are shown by thin lines connecting the exons. The location of the
mutations for the variants listed in Table 22.6.11.2 are shown; plus that of
G6PD Sunderland, as example of an English sporadic variant associated
with chronic nonspherocytic haemolytic anaemia and due to a deletion
of a triplet of bases, corresponding to codon 35.
section 22  Haematological disorders
5474
activity in red cells is such that NADPH is maintained at a high level
and there is practically no NADP: the NADPH/​NADP ratio plays a
large part in the intracellular regulation of G6PD activity.
The enzymatically active form of G6PD is a dimer or a tetramer of
a single protein subunit of 514 amino acids with a molecular mass
of 59 096 Da. Some regions of the molecule critical for its functions
have been identified because they are highly conserved in evolution.
The G6P-​binding site and the active site of the enzyme are located
near lysine 205. From the three-​dimensional structure of G6PD one
sees that in the dimer structure the two subunits are symmetrically
located across a complex interface of β-​sheets. The NADP binding
site is near the N-​terminus, and bound NADP is important for the
stability of G6PD.
Acute haemolytic anaemia associated with G6PD deficiency
clearly results from the action of an exogenous factor on intrinsic-
ally abnormal red cells. Although the sequence of events ending in
haemolysis is not completely understood, we know that oxidative
agents cause GSH depletion in G6PD-​deficient red cells. This is fol-
lowed by oxidation of sulphydryl groups and consequent denatur-
ation of haemoglobin (causing Heinz bodies) and probably of
other proteins. This eventually causes irreversible damage to the
membrane of red cells and hence their destruction, partly in the
bloodstream and partly through phagocytosis by macrophages. An
important feature of haemolysis in G6PD-​deficient patients depends
on the fact that G6PD decays gradually during red cell ageing (e.g.
in normal blood, reticulocytes have about five times more activity
than the 10% of oldest red cells), and this loss of enzyme activity
is accelerated with many G6PD variants. Thus, a haemolytic attack
selectively destroys older red cells because they have a more severe
shortage of G6PD. This is why in the post-​haemolytic state there is
a significant increase in G6PD activity (hence the risk of misdiag-
nosis). By contrast, with some other variants the steady-​state level of
G6PD is so low that, even in the absence of any oxidant challenge, it
becomes limiting for red cell survival: this is the case in the patients
with chronic nonsperocytic haemolytic anaemia, who may have a
red cell lifespan of between 10 and 50 days.
Clinical manifestations
Acute haemolytic anaemia
In view of the large number of people who carry a G6PD deficiency
gene, it is fortunate that the vast majority of them remain clinically
asymptomatic throughout their lifetime. However, they are all at risk
of developing acute haemolytic anaemia in response to three types
of trigger: (1) drugs (Table 22.6.11.2), (2) infections, and (3) fava
beans. Typically, a haemolytic attack starts with malaise, sometimes
associated with more or less profound weakness, and abdominal or
lumbar pain. After an interval of several hours to 2 to 3 days, the
patient develops jaundice and may pass dark urine (haemoglobin-
uria). In the majority of cases, the haemolytic attack, even if severe,
is self-​limiting and tends to resolve spontaneously. In the absence of
additional or pre-​existing pathology, the bone marrow response is
prompt and effective. Depending on the proportion of red cells that
have been destroyed (reflected in the severity of the anaemia), the
haemoglobin level may be back to normal in 3 to 6 weeks. The most
Table 22.6.11.1  Genetic heterogeneity of G6PD deficiency
Variant
class
Clinical expression
Degree of enzyme
deficiency
Examples
Amino acid
replacements
Populations where
prevalent
Mechanism of
enzyme deficiency
I
Chronic nonspherocytic
haemolytic anaemia
Usually <10% of normal
Harilaou
216 Phe→Leu
All class I variants
are sporadic
Unstable
Barcelona
Not yet known
Abnormal kinetics
II
Acute haemolytic
anaemia triggered by
broad beans or infection
or drugs
<10% of normal
Mediterranean
188 Ser→Phe
Mediterranean,
Middle East, India
Unstable
Mahidol
163 Gly→Ser
South-​East Asia
Unstable
Canton
459 Arg→Leu
China
Unstable
Union
454 Arg→Cys
Worldwide
?
III
As for class II
10 to 50% of normal
A–​
68 Val→Met
Africa,
Unstable
126 Asn→Asp
Southern Europe
Seattle
282 Asp→His
Europe
?
IV
None
60% of normal
A
126 Asn→Asp
Africa
None
Homozygous
normal
Heterozygous
Homozygous
deficient
Autosomal
(e.g. PK deficiency)
X-linked
(e.g. G6PD deficiency)
Fig. 22.6.11.3  Somatic cell mosaicism is a characteristic feature of the
products of X-​linked genes. With an autosomal enzyme defect all red
cells in a heterozygote will have approximately 50% of normal activity,
which in most cases is amply sufficient; instead, with an X-​linked defect
(e.g. G6PD deficiency) a heterozygote will have in her blood a mixture
(mosaicism) of G6PD normal and G6PD-​deficient red cells. PK, pyruvate
kinase.
22.6.11  Glucose-6-phosphate dehydrogenase deficiency
5475
serious threat in adults is the development of acute kidney injury (this
is exceedingly rare in children). The anaemia is usually normocytic
and normochromic, and it varies from moderate to extremely severe
(haemoglobin levels of 40 g/​litre or less have been recorded); it is due
largely to intravascular haemolysis, and hence it is associated with
haemoglobinaemia, haemoglobinuria, and low or absent plasma
haptoglobin. The blood film shows anisocytosis, polychromasia,
and other features associated with acute haemolysis, including
spherocytes (Fig. 22.6.11.5a); in severe cases the poikilocytosis is
very marked, with bizarre forms, numerous red cells that appear
to have unevenly distributed haemoglobin (‘hemighosts’), and red
cells that appear to have had parts of them bitten away (‘bite cells’
or ‘blister cells’). Supravital staining with methyl violet, if done
promptly, reveals the presence of ‘Heinz bodies’, consisting of pre-
cipitates of denatured haemoglobin (Fig. 22.6.11.5b) (apart from the
rare cases when they are formed because of a genetic abnormality of
haemoglobin, Heinz bodies can be regarded as a signature of oxi-
dative damage to red cells). The white blood cell count may be ele-
vated, with predominance of granulocytes. The platelet count may
be normal, increased, or moderately decreased. The unconjugated
bilirubin is elevated but the ‘liver enzymes’ are usually normal.
Among all drugs listed in Table 22.6.11.2, primaquine was
the first to be recognized (long before G6PD deficiency was dis-
covered) as a cause of haemolytic anaemia:  so much so that the
phrase ‘primaquine sensitivity’ was coined. Today, although many
other antimalarial drugs are available, there is a resurgence in the
use of primaquine, because it is still the only drug able to eradicate
hypnozoites in Plasmodium vivax infection. The standard regimen is
30 to 45 mg/​day for 14 days, and recently—​if belatedly—​testing for
G6PD is recommended as mandatory before this is administered.
Primaquine is also the only drug that can eliminate gametocytes
of P. falciparum after a clinical attack from this parasite has been
Primaquine
O2
H2O2
H2O
GSH
GSSG
NADPH
NADP
Superoxide
Dismutase
Rasburicase
Catalase
Glutathione
Reductase
Glucose 6-phosphate
6-Phosphogluconolactone
6-Phosphogluconate
Ribulose 5-phosphate
6-Phosphogluconate
Dehydrogenase
Glucose 6-phosphate
Dehydrogenase
6-Phosphogluconolactonase
Glutathione
Peroxidase
Uric acid
Primaquine
Oxidative
damage
O2
H2O2
H2O
GSH
GSSG
NADPH
NADP
Superoxide
Dismutase
Rasburicase
Catalase
Glutathione
Reductase
Glucose 6-phosphate
6-Phosphogluconolactone
6-Phosphogluconate
Ribulose 5-phosphate
6-Phosphogluconate
Dehydrogenase
Glucose 6-phosphate
Dehydrogenase
6-Phosphogluconolactonase
Glutathione
Peroxidase
Uric acid
(a)
(b)
Fig. 22.6.11.4  (a) In G6PD-​normal red cells, G6PD and 6-​phosphogluconate dehydrogenase—​two of the first
enzymes of the pentose phosphate pathway—​provide ample supply of NADPH, which in turn regenerates GSH when
this is oxidized by reactive oxygen species (e.g. O2
− and H2O2). O2
− is one of the most reactive oxygen species that can
be generated from the metabolism of pro-​oxidant compounds such as primaquine; rasburicase, on the other hand,
produces directly hydrogen peroxide in equimolar amount to uric acid degraded. (b) In G6PD-​deficient red cells, where
the enzyme activity is reduced, NADPH production is limited and it may not be sufficient to cope with the excess of
reactive oxygen species generated in the presence of pro-​oxidant compounds.
section 22  Haematological disorders
5476
successfully treated: fortunately this requires only a single 25 ​mg
dose, which can be regarded as clinically safe for G6PD-​deficient
persons.
As stated earlier, with different G6PD variants the residual red cell
G6PD activity varies: for instance, in males with G6PD A− it is not
as low as in males with G6PD Mediterranean or G6PD Mahidol.
Presumably as a result of this, the course of primaquine-​induced
haemolytic anaemia is considerably more severe with G6PD
Mediterranean than with G6PD A−; and for this reason there has
been a tendency to regard G6PD deficiency of the A− type as ‘mild’.
However, in trials recently conducted in Africa of a combination of
chlorproguanil and dapsone for the treatment of acute P. falciparum
malaria, all G6PD-​deficient children suffered exacerbation of their
anaemia, which in numerous cases was life-​threatening (they were
probably saved by blood transfusion). All these children had G6PD
A−: therefore, the word ‘mild’ for this G6PD variant in inappropriate.
Bacterial infection has been underestimated as a trigger of haemo-
lytic anaemia in G6PD-​deficient subjects. It has been also reported
that after major trauma in persons who are G6PD deficient there is
a higher rate of infectious complications and a more marked degree
of anaemia.
Favism
This is perhaps the most dramatic form of acute haemolytic anaemia
associated with G6PD deficiency: it can occur at any age, but far
more commonly in children. The clinical picture is similar to that
described earlier, but particularly prominent is haemoglobinuria,
which often develops within 6 to 24 h from the onset of symptoms.
There may be evidence of hypovolaemic shock or, more rarely, of
high-​output heart failure: either can be life-​threatening. The cause
of favism is the presence in fava beans (or broad beans, Vicia faba)
of vicine and convicine, two β-​glycosides having as aglycones the
substituted pyrimidines divicine and isouramil, which produce free
radicals in the course of their auto-​oxidation. Thus, haemolysis is
highly specific for fava beans; other beans are safe. G6PD-​deficient
subjects (especially when they are adults) do not develop an acute
attack of favism every time they eat fava beans: the reasons for this
are not yet clear, but important factors are the quantity and quality
of fava beans consumed. On the other hand, the widespread notion
that favism occurs only with some G6PD-​deficient variants and
not with others is incorrect. For instance, favism has been docu-
mented even with G6PD Seattle, a variant associated with rather
mild enzyme deficiency (c.25% of normal). Favism is a paradigm
of gene–​environment interaction:  it is practically nonexistent in
parts of Africa where G6PD deficiency is common but fava beans
are not eaten; and if fava beans are consumed it can occur in areas
(including the United Kingdom) where G6PD deficiency is rare.
Neonatal jaundice
Not every G6PD-​deficient baby becomes jaundiced after birth; how-
ever, the risk of developing neonatal jaundice is much greater in
G6PD-​deficient than in G6PD-​normal newborns. The extent of the
association between G6PD deficiency and neonatal jaundice appears
to vary in different populations. The clinical picture of neonatal jaun-
dice related to G6PD deficiency differs from the ‘classical’ Rh-​related
neonatal jaundice in two main respects: (1) it is very rarely present
at birth, with the peak incidence of clinical onset being between day
2 and day 3; and (2) there is more jaundice than anaemia, and the
anaemia is very rarely severe. The severity of G6PD-​related neo-
natal jaundice varies enormously, from subclinical, to overlapping
with ‘physiological jaundice’, to imposing the threat of kernicterus
if not treated. The reasons for this are not clear, but prematurity, in-
fection, and environmental factors, for example, naphthalene (cam-
phor balls) used in babies’ bedding and clothing, certainly play a
part in making neonatal jaundice more severe and more dangerous.
Other things being equal, the risk of neonatal jaundice is higher in
babies who have the allele of the UGT1A1 gene (encoding a UDP-​
glucuronosyltransferase) that underlies Gilbert’s disease. From the
point of view of public health, it is important to realize that in some
parts of the world G6PD deficiency is the commonest cause of severe
Table 22.6.11.2  Drugs that can trigger haemolysis in children with G6PD
Category of drug
Definite risk
Possible risk
Antimalarials
Primaquine
Dapsone-​containing combinationsa
Chloroquine
Quinine
Analgesics
Acetanilide
Aspirin
Sulfonamides/​sulfones
Sulfamethoxazole/​co-​trimoxazole
Dapsonea
Sulfasalazine
Sulfadiazine
Quinolones
Nalidixic acid
Ciprofloxacin
Norfloxacin
Moxifloxacin
Ofloxacin
Other antimicrobials
Nitrofurantoin
Methylene blue
Chloramphenicol
Other
Niridazole
Vitamin K
Rasburicase
Ascorbic acid
Glibenclamide
Note: for all drugs, the risk of haemolysis is dose related, and so is the severity of haemolysis. For instance, aspirin
up to 20 mg/​kg is probably safe; three times that dose will almost certainly cause some haemolysis.
a Dapsone can cause haemolysis even in non-​G6PD-​deficient subjects.
Table modified from British National Formulary, 55th edition, March 2008.
22.6.11  Glucose-6-phosphate dehydrogenase deficiency
5477
neonatal jaundice which, if not correctly managed, can produce per-
manent neurological damage.
Chronic nonspherocytic haemolytic anaemia
In contrast to the large majority of G6PD-​deficient subjects who
have no appreciable haemolysis in the steady state, a very small mi-
nority have chronic anaemia of very variable severity. The patient
is virtually always a male, and in general he presents because of
unexplained jaundice. Frequently the onset is at birth, and a diag-
nosis is made of neonatal jaundice (Fig. 22.6.11.6), which may be
severe enough to require exchange transfusion. Subsequently the
anaemia recurs and the jaundice fails to clear completely; or the
patients is only reinvestigated much later in life, perhaps because
of gallstones in a child or in a young adult. The spleen is usually
moderately enlarged, but it may increase in size sufficiently to cause
mechanical discomfort, or hypersplenism, or both. The severity of
anaemia ranges in different patients from borderline to transfusion
dependent. The anaemia is usually normochromic but somewhat
macrocytic; because a large proportion of reticulocytes (up to 20%
or more) will cause an increased mean cell volume and a shifted,
wider than normal, size-​distribution curve. The red cell morphology
is not characteristic, and for this reason it is referred to in the nega-
tive as being ‘nonspherocytic’. The bone marrow is normoblastic,
unless the increased requirement of folic acid associated with the
high red cell turnover has caused it to become megaloblastic. There
is chronic hyperbilirubinaemia; the serum haptoglobin may be de-
creased, and the serum lactate dehydrogenase may be increased. In
this condition, unlike in the acute haemolytic anaemia described
previously, haemolysis is mainly extravascular. However, the red
cells of these patients are naturally also vulnerable to acute oxidative
damage, and therefore the same agents (Table 22.6.11.2) that can
cause acute haemolytic anaemia in people with the ordinary type of
G6PD deficiency will cause severe exacerbations with (sometimes
massive) haemoglobinuria in people with the severe form of G6PD
deficiency.
Laboratory diagnosis
Although the clinical picture of favism and of other forms of acute
haemolytic anaemia associated with G6PD deficiency is quite char-
acteristic, the final diagnosis must rely on the direct demonstration
of decreased activity of this enzyme in red cells. With neonatal jaun-
dice and chronic nonspherocytic haemolytic anaemia, the differen-
tial diagnosis is much wider and therefore this test is even more
(a)
(b)
Fig. 22.6.11.5  Blood film in a case of acute haemolytic anaemia in a
G6PD-​deficient patient (favism). (a) Romanowsky stain, showing marked
poikilocytosis, polychromatic macrocytes, bite cells, nucleated red cells,
and a shift to the left in the granulocytic series. (b) Supravital stain with
methyl violet, showing the characteristic Heinz bodies.
Haemoglobin (g/litre)
200
100
0
Years
10
20
30
Reticulocytes (%)
Exchange
transfusion
Splenectomy
10
8
6
4
2
0
Fig. 22.6.11.6  Clinical course of a patient with chronic nonspherocytic
haemolytic anaemia caused by severe G6PD deficiency, illustrating the
high transfusion requirement, which was alleviated after splenectomy.
section 22  Haematological disorders
5478
important. The most widely used screening test is the fluorescence
spot test which, provided it is properly standardized and subjected
to quality control, is perfectly adequate for diagnostic purposes
in patients who are in the steady state; but this semiquantitative
tests is not adequate for patients in the acute haemolytic or in the
post-​haemolytic period, or with other complications; nor can it be
expected to identify all heterozygotes. Ideally, every patient found
to be G6PD deficient by screening should then be retested for con-
firmation by a quantitative assay. In normal red cells, the range of
G6PD activity, measured at 30°C, is 7 to 10 IU/​g haemoglobin. In
G6PD-​deficient males (or homozygous females), the level of G6PD
in the steady state is, by definition, less than 50% of normal; but
with most variants it is less than 20% and with some it is almost
undetectable. In heterozygous females, the level is intermediate and
extremely variable; in some cases the diagnosis may be therefore
difficult without family studies or DNA analysis. However, for prac-
tical purposes, it is most unlikely that a woman will have clinical
manifestations if her G6PD level is more than 70% of normal.
Management
Prevention
The acute haemolytic anaemia of G6PD deficiency is largely pre-
ventable by avoiding exposure to triggering factors of previously
screened subjects. Of course, the practicability and cost-​effectiveness
of screening depends on the prevalence of G6PD deficiency in each
individual community. Favism is entirely preventable by not eating
fava beans. Prevention of drug-​induced haemolysis is possible in most
cases by choosing alternative drugs. A common practical problem is
the need to give primaquine for eradication of malaria due to P. vivax
or P. malariae; in these cases the administration of a lower dose of the
drug for a longer time is the recommended approach: this will still
cause haemolysis, but of an acceptably mild degree.
Treatment of acute haemolytic anaemia and favism
A patient with acute haemolytic anaemia may be a diagnostic
problem, that once solved, does not require any specific treatment at
all; or the patient may be a medical emergency requiring immediate
action. With severe anaemia, prompt blood transfusion is defin-
itely indicated and may be life-​saving. If there is acute kidney injury,
haemodialysis may be necessary. Recovery is the rule.
Management of neonatal jaundice
This does not differ from that of neonatal jaundice due to other
causes than G6PD deficiency. In most cases, prompt phototherapy
is highly effective and sufficient; but with bilirubin levels above
300 µmol/​L (or even less in babies who are premature, or who have
acidosis or infection), exchange blood transfusion is imperative to
prevent neurological damage.
Management of chronic nonspherocytic
haemolytic anaemia
In general terms, this does not differ from that of chronic
nonspherocytic haemolytic anaemia due to other causes (e.g. pyru-
vate kinase deficiency). If the anaemia is not severe, folic acid sup-
plements and regular haematological surveillance will suffice. It will
be important to avoid exposure to potentially haemolysis-​inducing
drugs, and blood transfusion may be indicated when exacerbations
occur, mostly in concomitance with intercurrent infection. In rare
patients, the anaemia is so severe that it must be regarded as trans-
fusion dependent. In these cases, blood transfusion will be prob-
ably needed at approximately 2-​month intervals, in order to keep
the haemoglobin in the 80 to 100 g/​litre range. A hypertransfusion
regimen aiming to maintain a normal haemoglobin level is not indi-
cated (as there is no ineffective erythropoiesis in the bone marrow).
However, in patients requiring regular transfusions, appropriate
iron chelation should be instituted by the age of 2 years, and must be
continued as long as transfusion treatment is necessary; sometimes
the transfusion requirement may decrease after puberty. Although,
unlike in hereditary spherocytosis, there is no evidence of selective
red cell destruction in the spleen, splenectomy has proven beneficial
in severe cases. When a diagnosis of chronic nonspherocytic haemo-
lytic anaemia is made, the family must be given genetic counselling,
and an effort should be made to establish whether the mother is a
heterozygote; if she is, the chance of recurrence is 1:2 for every sub-
sequent male pregnancy. Prenatal diagnosis can be made by DNA
analysis if the mutation is first identified in an affected relative.
In principle, since the clinical manifestations of severe G6PD de-
ficiency are confined to the blood, that is, chronic nonspherocytic
haemolytic anaemia, this condition could be cured by allogeneic
bone marrow transplantation, but this has never been reported. For
the same reason the condition ought to be amenable to correction by
gene transfer into haematopoietic stem cells (gene therapy): this has
been done in a preclinical mouse model.
FURTHER READING
Beutler E (1991). Glucose 6-​phosphate dehydrogenase deficiency.
N Engl J Med, 324, 169–​74.
Cappellini MD, Fiorelli G (2008). Glucose-​6-​phosphate dehydro-
genase deficiency. Lancet, 371, 64–​74.
Luzzatto L, Arese P (2018). Favism and Glucose-6-Phosphate
Dehyrogenase Deficiency. New Eng J Med, 378, 60–71.
Luzzatto L, Notaro R (2001). Malaria. Protecting against bad air.
Science, 293, 442–​3.
Luzzatto L, Poggi V (2014). Glucose 6-​phosphate dehydrogenase defi-
ciency. In: Orkin S, et al. Nathan and Oski's hematology and oncology
of infancy and childhood, 8th edition, pp. 609–​29. WB Saunders/​
Elsevier, Philadelphia.
Luzzatto L, Seneca E (2014). G6PD deficiency:  a classic example
of pharmacogenetics with on-​going clinical implications. Br J
Haematol, 164, 469–​80.
Mason PJ, Bautista JM, Gilsanz F (2007). G6PD deficiency:  the
genotype-​phenotype association. Blood Rev, 21, 267–​83.
Minucci A, et al. (2012). Glucose-​6-​phosphate dehydrogenase (G6PD)
mutations database: review of the “old” and update of the new muta-
tions. Blood Cells Mol Dis, 48, 154–​65.
Pamba A, et al. (2012). Clinical spectrum and severity of hemolytic
anemia in glucose 6-​phosphate dehydrogenase-​deficient children
receiving dapsone. Blood, 120, 4123–​33.
Spolarics Z, et al. (2001). Increased incidence of sepsis and altered
monocyte functions in severely injured type A–​ glucose-​6-​phosphate
dehydrogenase-​deficient African American trauma patients. Crit
Care Med, 29, 728–​36.

22.6.12 Acquired haemolytic anaemia 5479 Amy Power
22.6.12 Acquired haemolytic anaemia 5479 Amy Powers and Leslie Silberstein
22.6.12  Acquired haemolytic anaemia
5479
Uyoga S, et al. (2015). Glucose-​6-​phosphate dehydrogenase deficiency
and the risk of malaria and other diseases in children on the coast
of Kenya: a case-​control and a cohort study. Lancet Haematol, 2,
e437–​44.
22.6.12  Acquired haemolytic anaemia
Amy Powers and Leslie Silberstein
ESSENTIALS
Premature destruction of red cells occurs through two primary
mechanisms: (1) decreased erythrocyte deformability that leads to
red cell sequestration and extravascular haemolysis in the spleen
and other components of the reticuloendothelial system—​may be
caused by membrane defects, metabolic abnormalities, exogenous
oxidizing agents, or pathological antibodies; and (2) red cell mem-
brane damage and intravascular haemolysis—​may be caused by
exposure to pathological antibodies, activated complement, mech-
anical forces, chemicals, and infectious agents.
Clinical features—​general aspects
These include (1) increased red cell production—​manifestations in-
clude reticulocytosis, polychromasia, macrocytosis, erythroid hyper-
plasia, and bone changes; (2) increased red cell destruction—​features
include decreased haemoglobin levels, fragmented red cells, decreased
haptoglobin levels, increased unconjugated bilirubin levels, increased
plasma lactate dehydrogenase levels, haemoglobinaemia, haemo-
globinuria, haemosiderinuria, and splenomegaly.
Congenital haemolytic anaemias
Congenital disorders resulting in a haemolytic anaemia include (1) dis-
orders of the red cell membrane such as hereditary spherocytosis
and hereditary elliptocytosis; (2) disorders of red cell enzymes such
as glucose-​6-​phosphate dehydrogenase deficiency and pyruvate
kinase deficiency; and (3)  disorders of globin structure. See also
Chapters 22.6.7, 22.6.10, and 22.6.11 for further discussion.
Acquired immune haemolytic anaemias
Immune haemolysis may occur when IgG, IgM, or IgA antibodies
and/​or complement bind to the erythrocyte surface. The direct anti-
globulin test (DAT) or direct Coombs’ test detects the presence of
IgG antibody or complement on the red cell surface. IgM and IgA
antibodies are not directly detectable with standard testing reagents.
Autoimmune haemolytic anaemias—​these are best classified ac-
cording to the temperature at which the antibody optimally binds
to the erythrocyte:  (1) warm autoimmune haemolytic anaemia—​
typically IgG; symptomatic patients present with anaemia, jaundice,
and splenomegaly; often associated with lymphoid malignancies;
first-​line treatment is with corticosteroids. (2)  Cold agglutinin-​
mediated autoimmune haemolytic anaemia—​autoantibodies are
typically IgM and are most active at low temperatures; seen in
younger patients following infection with Mycoplasma pneumoniae
or infectious mononucleosis and in older patients idiopathically or
in association with lymphoproliferative diseases. (3) Paroxysmal cold
haemoglobinuria. (4) Mixed type autoimmune haemolytic anaemia—​
both IgG and complement are present on the red cells; may be idio-
pathic or secondary (often to systemic lupus erythematosus).
Drug-​induced haemolytic anaemia—​haemolysis can be caused by
drugs that induce a positive DAT. Drug-​induced antibodies may be
drug dependent or drug independent depending on whether the
presence of the drug is required for their detection.
Alloimmune haemolytic anaemias—​these include (1) acute haemo-
lytic transfusion reactions—​may begin after the infusion of as little as
10 ml of incompatible blood, with symptoms and signs including
chest or flank pain, nausea, vomiting, fever, chills, hypotension, re-
spiratory distress, and haemoglobinuria. Despite immediate stopping
of the transfusion and optimal supportive care, patients can develop
renal failure, disseminated intravascular coagulation, and even die.
(2) Other conditions—​these include delayed haemolytic transfusion
reactions, passenger lymphocyte haemolysis, and haemolytic dis-
ease of the newborn (caused by RhD incompatibility, ABO incom-
patibility, or other blood group incompatibility).
Acquired nonimmune haemolytic anaemias
Common or important causes include (1) infections (e.g. malaria, ba-
besiosis); (2) drugs and chemicals (e.g. nitrofurantoin); (3) mechan-
ical (e.g. incompetent prosthetic heart valves); (4) thermal (e.g. faulty
blood warmer); and (5) venom.
Microangiopathic haemolytic anaemia (MAHA)
MAHA describes the anaemia observed in a spectrum of disorders
including haemolytic uraemic syndrome, thrombotic thrombo­
cytopenic purpura, and complement-​mediated thrombotic micro­
angiopathies. These anaemias may be a component of a congenital
or acquired disorder and may result from both immune and
nonimmune mechanisms.
Introduction
Mechanisms of haemolysis
After release into the circulation, normal red blood cells (RBCs)
survive for approximately 120 days. As the circulating red cell mass
decreases (anaemia), less oxygen is transported from the lungs to
other tissues. In response, the kidneys increase their synthesis and
secretion of erythropoietin, which stimulates erythropoiesis, in
order to restore normal red cell mass and oxygen delivery (see also
Chapter 22.6.1). A deficient red cell mass results from inadequate
production (hypoplasia), loss (haemorrhage), or premature destruc-
tion (haemolysis) of the red cells. In cases where red cell survival is
reduced by haemolysis to such an extent that normal bone marrow
cannot compensate, a haemolytic anaemia results. The haemolytic
anaemias are either genetically determined or acquired.
Consequences of haemolysis
The clinical and laboratory changes associated with haemolysis
reflect the physiological mechanisms responsible for restoring
red cell mass and removing free haemoglobin from the plasma.
These changes are outlined in Box 22.6.12.1. Within several days
section 22  Haematological disorders
5480
of the onset of haemolysis and the development of anaemia, in-
creased erythropoiesis results in erythroid hyperplasia (decreased
myeloid/​erythroid ratio) in the bone marrow and reticulocytosis
(polychromasia and macrocytosis) in the peripheral blood. The per-
ipheral blood film may also exhibit microspherocytes, fragmented
RBCs, and nucleated RBCs. If the haemolysis and anaemia begin
early in life and persist, extramedullary erythropoiesis can develop
in the spleen, liver, and lymph nodes. Chronic anaemia and the re-
sulting marrow hyperplasia can also result in long-​bone deform-
ities. Free haemoglobin in the circulation binds to the serum protein
haptoglobin. Haptoglobin–​haemoglobin complexes are removed
from the intravascular space by the reticuloendothelial system. If
the rate of haemolysis is greater than the liver’s ability to synthesize
haptoglobin, serum haptoglobin levels fall. In patients with severe
haemolysis, haemoglobinaemia and haemoglobinuria may develop.
At low plasma haemoglobin levels, much of the free haemoglobin
is reabsorbed in the proximal renal tubules. The renal tubular cells
catabolize the haemoglobin, converting iron into haemosiderin,
which is eventually shed along with renal tubular cells into the urine
resulting in haemosiderinuria. Haemosiderinuria is a reliable in-
dicator of chronic intravascular haemolysis. At higher levels, free
haemoglobin is found in the urine. Within the reticuloendothelial
system, haemoglobin is metabolized and released into the serum as
unconjugated bilirubin. The bilirubin is conjugated in the liver, ex-
creted in the gut, converted to faecal urobilinogen, partially reab-
sorbed, and excreted by the kidneys as urinary urobilinogen. The
intracellular enzyme lactate dehydrogenase is released from lysed
red cells into the plasma.
Congenital haemolytic anaemias
Congenital haemolytic anaemias result from inherited defects in
the red cell membrane, red cell enzymes, or haemoglobin. Inherited
defects in the red cell membrane include hereditary spherocytosis,
elliptocytosis,
pyropoikilocytosis,
spherocytic
elliptocytosis,
stomatocytosis, and xerocytosis. Inherited disorders of red cell
enzymes include glucose-​6-​phosphate dehydrogenase (G6PD)
deficiency and pyruvate kinase deficiency. Inherited defects in
haemoglobin synthesis include the haemoglobinopathies and
thalassaemias. These inherited anaemias are reviewed in depth in
Chapters 22.6.7, 22.6.10, and 22.6.11.
Acquired haemolytic anaemias
Immune haemolytic anaemias
Immune haemolysis may occur when IgG, IgM, or IgA antibodies
and/​or complement bind to the erythrocyte surface. The red cell-​
bound antibodies may induce extravascular haemolysis, intra-
vascular haemolysis, or both. Red cells coated with IgG typically
undergo extravascular haemolysis during their transport through
the reticuloendothelial system. Interactions between the Fc portion
of IgG and surface Fc receptors allow the macrophages to phago-
cytose the coated erythrocytes. IgM, IgA, and, occasionally, IgG ac-
tivate and fix complement to the erythrocyte surface. Macrophages
also have receptors for the activated complement component C3b
and likely phagocytose red cells through this pathway. The fixed
complement can also induce intravascular haemolysis through ac-
tivated membrane complex-​mediated lysis.
The direct antiglobulin test (DAT) or direct Coombs’ test detects
the presence of IgG antibody or complement on the red cell surface.
IgM and IgA antibodies are not directly detectable with standard
testing reagents. Rather, their presence may be indirectly demon-
strated by the detection of complement on the erythrocyte. In rare
cases, the haemolytic anaemia is due to noncomplement-​fixing IgM
or IgA antibodies. In this situation, the DAT will be falsely negative.
Eluates can be obtained from the antibody-​coated red cells to de-
termine the specificity of the antibody. Alternatively, the antibody
may be free in the serum and its specificity determined by the in-
direct antiglobulin test or indirect Coombs’ test. The presence of
antibody or complement on the red cell, however, need not reflect
ongoing haemolysis. Rather, the diagnosis of haemolytic anaemia
rests on clinical findings and other laboratory data, such as red cell
morphology, haemoglobin, bilirubin, haptoglobin, lactate dehydro-
genase levels, reticulocyte count, and the presence or absence of
haemoglobinaemia, haemoglobinuria, or haemosiderinuria. The
serological findings provide information as to whether an immune
basis exists and what type of immune haemolytic anaemia may be
present. Autoantibodies, alloantibodies, and drugs may induce im-
mune haemolytic anaemias.
Autoimmune haemolytic anaemia
Haemolytic antibodies directed against the individual’s own red
cells may arise as a primary/​idiopathic event or may be secondary to
lymphoid malignancies, connective tissue disorders, and infection.
Autoimmune haemolytic anaemia is best classified according to the
temperature at which the antibody optimally binds to the erythro-
cyte. The four major types of autoimmune haemolytic anaemia are
warm autoimmune haemolytic anaemia, cold agglutinin-​mediated
autoimmune haemolytic anaemia, paroxysmal cold haemoglobin-
uria, and mixed-​type autoimmune haemolytic anaemia. The classic
Box 22.6.12.1  The main features of haemolytic anaemia
Increased red cell production
•	 Reticulocytosis
•	 Polychromasia
•	 Macrocytosis
•	 Erythroid hyperplasia
•	 Bone changes
Increased red cell destruction
•	 Decreased haemoglobin levels
•	 Increased unconjugated bilirubin levels
•	 Decreased haptoglobin levels
•	 Increased faecal and urinary urobilinogen
•	 Haemoglobinaemia
•	 Haemoglobinuria
•	 Haemosiderinuria
•	 Splenomegaly
•	 Increased plasma lactate dehydrogenase levels
•	 Microspherocytes
•	 Fragmented RBCs
•	 Nucleated RBCs
22.6.12  Acquired haemolytic anaemia
5481
clinical and serological findings of these anaemias are shown in
Table 22.6.12.1.
Warm autoimmune haemolytic anaemia
Aetiology  The offending antibody in warm autoimmune haemo-
lytic anaemia is typically a polyclonal IgG (but can be IgM or IgA)
and can be found on the red cell, in the serum, or both. The exact
specificity of the antibody is often difficult to determine. With
rare exceptions, warm-​reactive autoantibodies bind to all red cells
tested, while others appear to have broad specificity within the
Rhesus (Rh) system. Occasionally, warm reactive autoantibodies
will exhibit specificity against an individual antigen such as Rh(D),
Rh(C), or Kell.
Clinical features  Warm autoimmune haemolytic anaemia can be
idiopathic or secondary to an underlying infection, malignancy, or
autoimmune disease. This disease can arise at any age but is more
common in older individuals, probably because of its association
with lymphoid malignancies. Women are affected slightly more
often than men. Clinical signs and symptoms can range from mild
to life-​threatening and are related to the severity of the anaemia and
ongoing haemolysis. The DAT is positive for IgG and/​or comple-
ment. In its mildest form the DAT is positive but red cell survival
is not significantly affected. Symptomatic patients present with an-
aemia, jaundice, and splenomegaly. Most patients with warm auto-
immune haemolytic anaemia have a chronic, stable anaemia. In its
severest form, patients present with fulminant intravascular haem-
olysis, progressive anaemia, congestive heart failure, respiratory
distress, and neurological abnormalities. As with other haemolytic
anaemias, the peripheral smear often demonstrates anisocytosis and
reticulocytosis with spherocytes and macrocytes (Fig. 22.6.12.1.)
The platelet count is usually normal except in patients with Evans’
syndrome where the autoantibody destroys both red cells and plate-
lets. Rarely, patients with clinical and laboratory findings consistent
with a warm autoimmune haemolytic anaemia may have a negative
DAT. These cases have been attributed to IgA, IgM, or low affinity
IgG antibodies. Alternatively, bound IgG has been reported below
the level of routine DAT detection. In the latter situation, an eluate
may demonstrate a pan-​agglutinating antibody.
Treatment  Corticosteroids, which presumably block macrophage
Fc receptor activity and inhibit antibody production, are the primary
therapy for idiopathic warm autoimmune haemolytic anaemia.
Prednisone at a dose of approximately 1 mg/​kg/​day is effective in
most patients. Higher doses rarely provide additional benefit, but
do increase the number and severity of side effects. Treatment con-
tinues until a haemoglobin level greater than 100 g/​litre is reached.
The initial dose of prednisone can then be tapered to 20 to 30 mg/​
day within a few weeks. Thereafter, the dose is reduced by 2.5 to
5.0 mg/​day per month with careful monitoring of the patient’s la-
boratory parameters. Once the patient has been in remission for 3
to 4 months at a dose of 5 mg/​day, withdrawal of steroids may be
considered.
Splenectomy and rituximab are considered second-​line ther-
apies. Rituximab (chimeric anti-​CD20 monoclonal antibody)
has been demonstrated to have a favourable response rate and
safety profile in patients with autoimmune haemolytic anaemia.
Splenectomy should be performed only in steroid-​refractory pa-
tients or patients requiring unacceptably high doses of prednisone
to maintain remission. Alternative therapies include cyclophos-
phamide, danazol, alemtuzumab, ciclosporin, and mycophenolate
mofetil. These therapeutic options should be reserved for patients
unfit for splenectomy or who have failed to respond to steroids,
rituximab, and surgery.
The decision to transfuse patients with warm autoimmune haemo-
lytic anaemia requires careful consideration. Due to the panreactive
nature of most warm autoantibodies, all cross-​matched RBCs for
transfusion will appear incompatible. Transfusion of ABO-​ and
Table 22.6.12.1  Classic clinical and serological findings observed in the immune haemolytic anaemias
WAIHA
CAIHA
Mixed type AIHA
PCH
Clinical
presentation
Extravascular Haemolysis
Haemagglutination and vascular
obstruction; intra-​ and extravascular
haemolysis
Combined warm and cold
autoimmune haemolysis
Acute, severe, intravascular
haemolysis
Autoantibody
type
IgG (rare IgM or IgA)
IgM
IgG & IgM
Biphasic IgG
(Donath–​Landsteiner antibody)
Autoantibody
specificity
Broadly reactive; relative specificity
for Rh antigens may be observed
I or I; rare Pr
Usually unclear
P antigen
DAT
IgG and/​or C3
C3
IgG and C3
C3
AIHA autoimmune haemolytic anaemia; CAIHA cold agglutinin autoimmune haemolytic anaemia or cold agglutinin disease; PCH paroxysmal cold haemoglobinuria; WAIHA warm
autoimmune haemolytic anaemia.
Fig. 22.6.12.1  The peripheral blood changes in warm autoimmune
haemolytic anaemia. There is marked anisocytosis and anisochromia
with many macrocytes and microspherocytes. The macrocytes reflect the
reticulocytosis. Magnification ×1000, Leishman stain.
section 22  Haematological disorders
5482
Rh-​compatible blood should not be withheld because of this sero-
logical incompatibility if clinically indicated for a patient with symp-
tomatic anaemia. Active serum autoantibodies can, however, mask
the presence of clinically significant alloantibodies. Therefore, the
most important consideration before transfusion is to confirm the
presence or absence of alloantibodies in the patient’s serum. Various
autologous and allogeneic red cell absorption techniques exist
to remove the autoantibody from a sample of the patient’s serum
and allow identification of any existing alloantibodies. If clinically
significant alloantibodies are present, red cells lacking the corres-
ponding antigen(s) should be selected for transfusion. If possible, it
is recommended that patients be antigen typed and provided with
RBCs which are matched for all clinically significant antigens in
order to prevent subsequent alloimmunization and delayed haemo-
lytic transfusion reactions. Transfusions in life-​threatening situ-
ations should not be delayed, however, if the above-​mentioned tests
are not readily available or completed.
Cold agglutinin-​mediated autoimmune haemolytic anaemia
Aetiology  Cold agglutinin-​mediated autoimmune haemolytic an-
aemia accounts for approximately one-​quarter of all cases of auto-
immune haemolytic anaemia. The autoantibodies are typically
IgM and are most active at low temperatures (4°C); however, rare
examples of IgG and IgA cold-​reactive autoantibodies have been
reported. In the lower temperatures of the peripheral circulation,
the IgM autoantibodies bind to red cells and activate complement.
In warmer areas of the circulation, the IgM dissociates from the
erythrocyte leaving activated complement fixed to the red cell sur-
face. The major mechanism of haemolysis in stable disease is thought
to be extravascular clearance of C3b-​coated erythrocytes in the liver.
However, intravascular haemolysis also occurs. The autoantibody
specificity is usually anti-​I. Anti-​i specificity is associated with in-
fectious mononucleosis. Other specificities, including against the Pr
protein, have been reported but are rare.
Cold agglutinin-​mediated autoimmune haemolytic anaemia can
be classified as primary and secondary. The primary or idiopathic
form, referred to as cold haemagglutinin disease (CHAD), is a chronic
condition typically characterized by the presence of a monoclonal
IgM-​κ, usually with specificity for the I-​antigen. Recent studies have
found evidence of bone marrow clonal lymphoproliferation in most
of these patients. The secondary form, referred to as cold agglutinin
syndromes, may be acute or chronic. Chronic cold agglutinin syn-
drome is most often seen in the setting of malignancy. Acute cold
agglutinin syndrome is most classically associated with mycoplasma
pneumonia and Epstein–​Barr virus infections.
Clinical features  In cold agglutinin disorders, the signs and symp-
toms of disease result from either the agglutination of red cells or
from haemolysis. Acute cold agglutinin syndrome is commonly
seen in adolescents and young adults following infection with
Mycoplasma pneumoniae or infectious mononucleosis. Haemolysis
occurs approximately 1 to 2 weeks after infection and is most com-
monly associated with a rise in polyclonal anti-​I IgM antibody with
M.  pneumonia or polyclonal anti-​i IgM antibody with infectious
mononucleosis.
Chronic cold autoimmune haemolytic anaemia occurs most com-
monly in older people, either idiopathically (primary) or associated
with an underlying condition such as chronic lymphocytic leu-
kaemia or Waldenström macroglobulinaemia (secondary). Patients
may experience chronic intravascular haemolysis and anaemia that
are exacerbated by cold temperature. Patients are often also plagued
by episodes of cold-​induced acrocyanosis and Raynaud’s phenom-
enon. Exacerbation of the haemolytic anaemia may be triggered
by a febrile infectious illness or major trauma. This exacerbation
has been attributed to the repletion of complement levels during
acute phase reactions in patients who are normally complement de-
pleted during their steady state of cold agglutinin disease. Repletion
of complement levels is thought to increase complement-​mediated
haemolysis during these episodes. Monoclonal IgM antibodies
with κ light chains and anti-​I specificity usually cause the red cell
agglutination and haemolysis in primary cold agglutinin disease.
In typical cases the cold agglutinin titre is very high (>1:105).
The thermal amplitude of the cold agglutinin, not the titre of the
antibody, most accurately predicts the severity of the disease.
Examination of the peripheral smear in patients with cold agglu-
tinins shows red cell agglutination (Fig. 22.6.12.2). The DAT is
positive for complement.
(a)
(b)
Fig. 22.6.12.2  (a, b) The peripheral blood smear changes noted in a patient with cold agglutinins. Magnification (a) ×40; (b) ×400 Wright
Giemsa stain.
22.6.12  Acquired haemolytic anaemia
5483
Treatment  Primary chronic CHAD can be managed through
both pharmacological and nonpharmacological methods. Patients
should avoid cold exposure whenever possible. In situations of
symptomatic anaemia, blood transfusions can be given provided
some precautions are taken. Antibody screening and compatibility
testing should be performed at 37°C. Blood should be given slowly
through a blood warmer. Hypothermia must be avoided during sur-
gery (especially surgical procedures involving extracorporeal cir-
cuits). Plasma exchange may be helpful as a temporizing measure in
acute situations due to the primarily intravascular location of IgM.
Corticosteroids and splenectomy are rarely effective. Rituximab, an
anti-​CD​20 monoclonal antibody, has demonstrated some clinical
effectiveness in published cases of cold agglutinin disease, however
complete remissions are uncommon. Higher response rates and
frequency of complete remission have been reported with a com-
bination of fludarabine and rituximab, suggesting that this com-
bination therapy may be preferred. Limited data suggests a role for
eculizumab, a monoclonal anti-​C5 antibody, in the treatment of
CHAD. However, further studies are needed to better elucidate the
efficacy of this agent.
There is no evidence-​based therapy for the treatment of secondary
cold agglutinin syndrome, and treatment of the underlying disease
is important when possible. Acute CHAD following a viral infection
is always self-​limited. Supportive measures, including transfusions
and avoidance of cold, may suffice. Treatment of the mycoplasma
infection shortens the duration and severity of the haemolysis.
Corticosteroids are usually unhelpful, and splenectomy is almost
never indicated.
Paroxysmal cold haemoglobinuria
Aetiology  Paroxysmal cold haemoglobinuria (PCH) is the rarest
form of autoimmune haemolytic anaemia. The disorder is caused
by the polyclonal complement-​fixing Donath–​Landsteiner IgG
antibody. In the cold, this antibody binds to, and irreversibly fixes,
complement to the red cell membrane. Upon return to warmer tem-
peratures, the antibody dissociates from the red cell and intravas-
cular haemolysis occurs by activation of the complement cascade.
The Donath–​Landsteiner antibody appears to have anti-​P specifi-
city. The P antigen is a high-​incidence antigen allowing it to bind to
practically all red cells.
In the past, PCH was commonly associated with congenital or
tertiary syphilis. Presently, PCH is most commonly seen in chil-
dren during or after a viral or bacterial infection, including measles,
mumps, Epstein–​Barr virus, varicella, cytomegalovirus, influenza,
mycoplasma, and Haemophilus influenzae. Rarely, adults may
present with PCH after a viral infection, or in association with a
lymphoma or myeloproliferative disorder.
Clinical features  Patients present with symptoms of acute intravas-
cular haemolysis, and the physical exam shows signs of jaundice and
anaemia. Laboratory abnormalities will reflect acute intravascular
haemolysis, and all complement levels may be low after haemolytic
episodes due to consumption. Aside from findings commonly seen
with haemolytic anaemias, review of the peripheral blood smear
may show neutrophil erythrophagocytosis. During or shortly after
a haemolytic episode, the DAT may be positive for complement but
will be negative for IgG. The DAT, however, may be negative in some
cases, and patients with a Coombs’ negative haemolytic anaemia
should be worked up further with the Donath–​Landsteiner test. The
Donath–​Landsteiner test is specific for PCH.
Treatment  No specific therapy for paroxysmal cold haemo-
globinuria exists. Most postinfectious cases of paroxysmal cold
haemoglobinuria are self-​limited and require only supportive care,
including avoiding exposure to the cold. Transfusion is indicated
only for severe haemolysis and life-​threatening anaemia. Crossmatch
compatible P-​antigen positive blood may be used for transfusion.
Transfusions with extremely rare P-​antigen-​negative blood should
be reserved only for those patients who do not respond to random
donor blood. The use of a blood warmer should be considered.
As most cases of acute PCH are self-​limited, immunosuppres-
sive or other pharmacological therapies are typically not indicated.
Steroids have not been shown to be useful. Data on the use of thera-
peutic plasma exchange, rituximab, and eculizumab are limited.
Mixed-​type autoimmune haemolytic anaemia
Aetiology  Rarely, autoimmune haemolytic anaemias may be classi-
fied as mixed type. This diagnosis should be reserved for those auto-
immune haemolytic anaemias involving both warm autoantibodies
and cold autoantibodies that have pathological serological features.
This is in contrast to the estimated 30% of warm autoimmune
haemolytic anaemias that are associated with the presence of benign
cold agglutinins (titre ≤64 with a thermal reactivity <30°C). The
warm-​reactive IgG autoantibodies are indistinguishable from anti-
bodies encountered in warm autoimmune haemolytic anaemia. The
IgM autoantibodies may be of high titre and high thermal amplitude
like those seen in cold-​agglutinin syndrome or may have low titres
at 4°C and have high thermal amplitudes, reacting at 30°C or above.
These IgM autoantibodies usually have no distinguishable specifi-
city, but on occasion have I or i specificities.
Clinical features  Mixed-​type autoimmune haemolytic anaemia
may be idiopathic or secondary. Frequent association with systemic
lupus erythematosus has been reported. The haemolytic anaemia is
often severe and chronic with intermittent exacerbations. Exposure
to cold has been reported to increase haemolysis in some cases.
The DAT is typically positive for both IgG and complement, and
the eluate will contain a warm reactive IgG autoantibody. Complex
serum reactivity will be noted in all phases of testing.
Treatment  Steroids, splenectomy, rituximab, or cytotoxic agents
may provide therapeutic benefit in mixed-​type autoimmune haemo-
lytic anaemia. If blood transfusions are necessary, selection of blood
should adhere to transfusion guidelines outlined earlier for warm
autoimmune haemolytic anaemia. Administration of blood through
a warmer should be considered.
Drug-​induced immune haemolytic anaemia
Drugs may induce antibodies to bind to the erythrocyte surface
resulting in a positive DAT or haemolysis. Drug-​induced anti-
bodies may be classified as either drug-​dependent antibodies or
drug-​independent antibodies. Drug-​dependent antibodies may
only be detected if the drug is present in the test system, while
drug-​independent antibodies do not require the in vitro addition of
the drug for detection. Both types of antibodies may be associated
with haemolysis. Patients with an unexplained haemolytic anaemia
should have a complete medication history taken with each drug
section 22  Haematological disorders
5484
considered as a potential aetiology. Clinical evidence supporting the
role of a particular drug in a haemolytic anaemia includes the asso-
ciation of haemolysis with drug initiation and resolution of haem-
olysis with drug cessation.
Drug-dependent antibodies
Certain drugs bind to the red cell membrane with a high affinity.
Association of the drug with the membrane constituents allows the
drug to act as a hapten. The antibodies produced are commonly IgG
and are directed predominantly against the drug. The drug-​coated
red cells undergo extravascular destruction via Fc receptor-​mediated
recognition by splenic macrophages. The extravascular haemolysis
can develop gradually, but may be life-​threatening if left untreated.
After the offending drug is identified and withdrawn, haemolysis
will often resolve quickly. However, the positive DAT and the haem-
olysis may persist for weeks if the drug is bound firmly enough to
the RBC membrane to prevent its clearance from the circulation.
Laboratory workup typically reveals a DAT that is positive for IgG.
The eluate and serum will not react with RBCs unless they are coated
with drug. Penicillin and some of the cephalosporins are the most
notorious examples of this phenomenon. Approximately 3% of pa-
tients receiving large doses of penicillin (millions of unit per day)
intravenously will develop a positive DAT. Only rarely do patients
develop haemolytic anaemia.
Other drugs are not thought to strongly adhere to the RBC
membrane. The mechanism by which these drugs induced a
haemolytic anaemia is controversial and has previously been re-
ferred to as the immune complex mechanism of drug-​induced
haemolytic anaemia. In this scenario, the drug may loosely bind
to the RBC membrane and induce the binding of IgM or IgG anti-
bodies that activate complement and cause intravascular haem-
olysis. Alternatively, it is hypothesized that IgM antibodies bind to
circulating drug to form immune complexes which loosely adhere
to RBCs and induced complement mediated intravascular haem-
olysis. The DAT is often positive for complement. The patient’s
serum and eluate will only show reactivity if drug is present in
the reaction mixture. The onset of the haemolysis is often abrupt
and resultant anaemia can be severe. Haemoglobinaemia, haemo-
globinuria, and renal failure are common. Once the offending
drug is withdrawn, the haemolysis stops. Antibodies induced by
second-​ and third-​generation cephalosporins are thought to act
via the mechanism.
Drug-​independent antibodies
Some drugs stimulate the synthesis of red cell autoantibodies. The
mechanisms of antibody stimulation are not well understood, but
may include immune dysregulation, molecular mimicry, and/​or
drug adsorption causing alteration of RBC membrane antigens.
Patient serum and red cell eluates react with normal red cells in the
absence of the drug. The autoantibodies are indistinguishable from
those found in warm autoimmune haemolytic anaemia. The DAT
usually becomes positive after 3 to 6 months of drug administration.
The haemolysis typically ceases within 2 weeks after the withdrawal
of the drug, but the DAT can remain positive for up to 2 years. α-​
Methyldopa, L-​dopa, procainamide, mefenamic acid, fludarabine,
and sulindac are examples of drugs that can stimulate the produc-
tion of red cell autoantibodies.
Nonspecific protein adsorption
A drug-​induced positive DAT may also reflect nonimmunological
adsorption of protein, including immunoglobulins. Haemolysis due
to nonimmunological protein adsorption has been associated with
β-​lactamase inhibitors, platinum-​based chemotherapeutic agents,
and cephalosporins. These drugs may alter RBC membranes so that
immunoglobulins nonspecifically adhere to their surfaces, causing a
positive DAT. Although previously not thought to cause haemolysis,
these bound immunoglobulins may mediate RBC destruction in
some cases.
Alloimmune haemolytic anaemias
Acute haemolytic transfusion reactions
Aetiology  Catastrophic cases of alloimmune haemolysis may occur
following the transfusion of ABO-​incompatible red cells. Naturally
occurring IgM anti-​A and anti-​B antibodies bind to the incompat-
ible red cells and activate complement resulting in intravascular
haemolysis. Human error leading to the mis-identification of pa-
tients, their blood samples, or the units of red cells to be transfused,
is responsible for virtually all cases of ABO incompatibility. Other
non-​ABO IgG alloantibodies can cause acute, severe haemolysis,
and acute haemolysis may also be seen following the administration
of ABO-​incompatible plasma containing components. Although
apheresis platelets may frequently be transfused across ABO groups
without adverse consequences, high levels of anti-​A, anti-​B, or anti-​
A,B in these units have been reported to cause acute haemolytic re-
actions in some recipients.
Clinical features  Acute haemolytic transfusion reactions pre-
sent within 24 h of transfusion. Symptoms of an acute haemo-
lytic transfusion reaction may begin after the infusion of as little
as 10 ml of incompatible blood. The signs and symptoms include
fever, chills, nausea, vomiting, hypotension, respiratory distress,
haemoglobinuria, and chest, flank, back, or infusion site pain.
Despite treatment, acute haemolytic transfusions reactions can
result in renal failure, disseminated intravascular coagulation,
and even death.
When a possible acute haemolytic transfusion reaction is first
recognized, the transfusion must be immediately stopped and a
full investigation should be undertaken. All labels, paperwork,
and the patient’s identification band should be rechecked for ac-
curacy. The blood bank paperwork and workup should also be re-
viewed. All units previously cross-​matched or dispensed but not
yet transfused must be retrieved to prevent any additional reac-
tions. A post-​transfusion blood sample should be obtained to de-
termine if a haemolytic reaction has occurred. A positive DAT or
the visual evidence of haemolysis in the serum is supportive of the
diagnosis of acute haemolysis, particularly if neither of these is ob-
served on a pre-transfusion blood sample. Further evaluation may
involve repeat ABO and Rh typing, antibody identification, and
crossmatches using both pre-​ and post-​transfusion specimens to
determine the identity of the causative antibody. In some cases,
nonimmune-​related haemolysis may instead be the cause of the
acute reaction. Overheating of blood in a blood warmer, attempts
to transfuse blood rapidly through a small-​bore needle, and con-
comitant administration of hypotonic solutions and drugs have
been associated with haemolysis.
22.6.12  Acquired haemolytic anaemia
5485
Treatment  Once an acute haemolytic transfusion reaction is sus-
pected, the blood transfusion should be stopped immediately, as the
mortality rate is correlated with the amount of incompatible blood
that is transfused. Treatment should be guided by the clinical con-
dition of the patient. Intravenous access should be maintained for
aggressive treatment of hypotension with intravenous fluids. Pressor
agents (low-​dose dopamine) have been used for mitigation of renal
complications, although the effectiveness of this intervention is con-
troversial. Other critical measures include monitoring the urine
output and promoting renal blood flow with diuretics (furosemide
or mannitol). Transfusion of platelets, plasma, or cryoprecipitate
may be necessary for the treatment of life-​threatening bleeding sec-
ondary to disseminated intravascular coagulation. Heparin has also
been used in the treatment of disseminated intravascular coagula-
tion; however, caution is urged due to the potential for haemorrhage.
Limited data suggest that eculizumab may be effective in the treat-
ment of an acute ABO-​mediated haemolytic transfusion reaction if
given shortly after the causative transfusion, presumably due to its
ability to inhibit complement-​mediated intravascular lysis.
Delayed haemolytic transfusion reactions
Aetiology  Delayed haemolytic transfusion reactions typically
occur in patients who have been alloimmunized to RBC antigens
by previous transfusions or pregnancies. Approximately 2 to 3% of
transfusion recipients become alloimmunized to non-​ABO red cell
antigens, although this figure may be much higher in certain patient
populations, such as those with sickle cell disease. Haemolysis is not
generally seen during the primary immune response since the trans-
fused red cells often disappear from the circulation before antibody
titres reach clinically significant levels. In the absence of further anti-
genic stimuli, antibody titres may diminish to undetectable levels.
Subsequent transfusion of red cells possessing the offending antigen,
however, will induce an anamnestic response with reappearance of
the IgG antibodies within hours to days. Binding of the IgG anti-
body to the transfused antigen-​positive red cells results in a posi-
tive DAT and possibly mild to moderate extravascular haemolysis.
Although numerous specificities are described, antibodies against
the Rh, Kidd, Duffy, Kell, and MNS system antigens are commonly
implicated in delayed haemolytic transfusion reactions.
Clinical features  Most patients experiencing a delayed haemolytic
transfusion reaction present with fever, jaundice, and decreasing
haemoglobin levels 1 to 2 weeks after the transfusion of incompat-
ible red cells. Delayed haemolytic transfusion reactions are often
discovered during evaluation for fever of unknown origin or when
the haemoglobin level fails to increase following transfusion.
Treatment  Treatment is rarely necessary; acute kidney injury or
disseminated intravascular coagulation is uncommon. If a delayed
haemolytic transfusion reaction is suspected, both the patient’s
serum and an eluate from the circulating red cells should be tested
for alloantibodies. If alloantibodies are present, their specificities
should be determined. Donor red cell units lacking the offending
antigen should be selected for all subsequent transfusions, even if
the antibody is no longer detectable on routine antibody screens.
Passenger lymphocyte haemolysis
Aetiology  Recipients of a haematopoietic or a solid-​organ trans-
plant may experience delayed extravascular haemolysis. In this
circumstance, lymphocytes of donor origin produce haemolytic
antibodies against ABO or other red cell antigens possessed by the
recipient.
Clinical features  Haemolysis due to passenger lymphocytes is
most commonly seen in out-​of-​group yet ABO-​compatible liver and
bone marrow transplants (group A or group B recipients of group
O tissue) but can also occur in recipients of lung, heart, and kidney
transplants. A positive DAT and haemolysis can begin within several
days after the transplant and continue for several months.
Treatment  If significant ABO haemolysis occurs, patients should
be transfused with group O red cells. If non-​ABO haemolysis is pre-
sent, elution of the patient’s red cells may help to identify the anti-
body specificity and allow transfusion of antigen-​negative red cells.
Haemolytic disease of the newborn
Haemolytic disease of the newborn occurs when maternal IgG anti-
bodies cross the placenta and bind to fetal red cells resulting in extra-
vascular haemolysis. Usually these antibodies possess specificities
within the Rh or ABO blood group systems. Occasionally the anti-
bodies are directed against other red cell antigens such as the Kell,
Kidd, and Duffy. In the mildest cases, anaemia develops after birth
and is of little clinical consequence. In more severe cases the neo-
nate develops progressive anaemia and jaundice within the first
week of life. If left untreated, bilirubin may reach levels associated
with kernicterus causing brain damage and death. In the most severe
cases, the fetus develops profound anaemia during gestation and
may be stillborn or delivered grossly oedematous (hydrops fetalis).
An infant with hydrops fetalis also has ascites, hepatosplenomegaly,
and erythroblastosis and usually dies shortly after birth.
Rh incompatibility  Although incompatibility for the A  and B
blood group antigens is now the most common cause of haemo-
lytic disease of the newborn, the most severe cases have historic-
ally been attributed to the anti-​Rh(D) antibody. In the majority of
these cases, haemolytic disease of the newborn occurs in Rh(D)-​
negative women carrying a Rh(D)-​positive fetus. The mother de-
velops anti-​D IgG antibodies following exposure to the D antigen
during a previous pregnancy, or as a result of the transfusion of
D-​antigen-​positive red cells. Rh(D) alloimmunization may be due
to transplacental haemorrhage from the fetus at the time of de-
livery. Spontaneous transplacental haemorrhage can also occur
during gestation, particularly during the third trimester, as well
as with ectopic pregnancy, spontaneous or therapeutic abortion,
chorionic villus sampling, amniocentesis, caesarean section, and
trauma. Approximately 16% of untreated Rh(D)-​negative women
who deliver a Rh(D)-​positive child will become alloimmunized to
the D antigen if not treated with prophylactic Rh immune globulin.
Exposure of a Rh(D)-​negative mother to as little as 0.1 ml of fetal D-​
positive blood can result in sensitization.
It is essential to identify pregnant women at risk for Rh(D)
haemolytic disease of the newborn to prevent sensitization. All
pregnant women should have their ABO and Rh types identified as
early as possible. Their serum should be screened for alloantibodies
against the D antigen and other red cell antigens. Pregnant women
who are D-​antigen negative and have an initial negative antibody
screen should have their serum retested for alloantibodies at 28
weeks’ gestation. If the initial antibody screen is found positive,
section 22  Haematological disorders
5486
antibody titres should be followed at 2-​ to 4-​week intervals to de-
termine whether further sensitization is occurring. A rising titre of
anti-​D antibody or other clinically significant red cell alloantibodies
indicates ongoing sensitization and possible haemolytic disease of
the fetus and newborn. The presence of an antibody, however, does
not indicate ongoing haemolysis in all cases. Naturally occurring
IgM antibodies are common during pregnancy but do not cross the
placenta. Furthermore, fetal red cells may lack the antigen corres-
ponding to the mother’s antibody. Molecular typing of fetal DNA is
available for many red cell antigens including D, E/​e, C/​c, Jka/​Jkb,
and K1/​K2.
Middle cerebral artery peak systolic velocity measured by
Doppler ultrasonography is a noninvasive and accurate tool that
can be used to monitor and assess the severity of haemolysis. If the
fetus is experiencing significant haemolysis and anaemia, clinical
intervention must be prompt. Before 34 weeks of gestation, intra-
uterine transfusion with leucoreduced and irradiated blood lacking
the offending antigen should be performed. After 36 weeks’ ges-
tation, induced labour should be considered. Upon birth of an
‘at-​risk’ fetus, a sample of cord blood should undergo a DAT and
have measurements of haemoglobin and bilirubin performed. If the
DAT on the cord blood sample is positive and the mother’s antibody
screen remains negative, haemolytic disease secondary to ABO in-
compatibility or antibodies against low-​incidence red cell antigens
should be considered.
In affected infants, phototherapy can be used to decrease bili-
rubin levels. Infants with severe anaemia or severe jaundice should
undergo exchange transfusion. A  nonsensitized Rh(D)-​antigen-​
negative mother’s blood should also be tested to determine the
amount of fetomaternal haemorrhage at delivery. Administration
of 300 µg of IgG anti-​D (RhIg) within 72 h of delivery will protect
up to 99% of D-​antigen-​negative mothers from developing anti-​D
antibodies. Prophylactic administration of RhIg at 28 weeks’ ges-
tation and following invasive procedures or traumatic events will
virtually eliminate the chance of alloimmunization. Patients with
large transplacental haemorrhages quantitated by the Kleihauer–​
Betke acid-​elution technique should receive additional RhIg at a
dose equivalent to 300 µg for every 15 ml of fetal RBCs or 30 ml of
fetal blood.
ABO incompatibility  Although 20% of pregnancies are ABO in-
compatible, severe haemolytic disease of the newborn due to ABO
incompatibility is rare. Group A  and group B infants of group
O mothers are at greatest risk, due to the IgG antibody anti-​A,B
made by group O individuals. Unlike with the Rh(D) antigen,
ABO-​haemolytic disease of the newborn occurs during the first
pregnancy as often as subsequent pregnancies. Most cases are
asymptomatic to mild, and exchange transfusion with group O red
cells is rarely required. The decrease in severity observed in cases
due to ABO incompatibility may be due to decreased surface ex-
pression of the A and B antigens on fetal cells, and the presence of
A and B antigens on many tissues leading to dilution of the anti-
body effect.
Nonimmune acquired haemolytic anaemias
Red cell survival may also be reduced by a number of noninherited,
nonimmune mechanisms. As red cells circulate, they are vulner-
able to a variety of insults that may cause structural or metabolic
alterations. These changes generally result in reduced red cell
deformability leading ultimately to extravascular haemolysis.
These insults include infection, mechanical trauma, and exposure
to chemicals, heat, or venom. They often also cause intravas-
cular haemolysis by directly lysing the red cell membrane. Other
causes of acquired nonimmune haemolytic anaemias are listed in
Box 22.6.12.2.
Infection
Infectious causes of haemolysis are primarily parasites and bacteria.
Direct parasitization of red cells by Plasmodium falciparum, P. vivax,
and P. malariae causes both intravascular haemolysis due to direct
membrane destruction and extravascular haemolysis due to mem-
brane alteration and activation of the reticuloendothelial system.
Infrequently, in utero infection of the fetus with Toxoplasma gon-
dii resembles severe haemolytic disease of the newborn. Infants are
born hydropic and severely anaemic. Premature delivery and still-
birth are common. Babesia microti, endemic in areas of the Northeast
and Midwest of the United States of America, is transmitted by
ticks and causes severe haemolysis during the erythrocytic phase
of its life cycle. Bacterial infections, particularly Gram-​negative
organisms which produce endotoxin or proteolytic enzymes, may
produce mechanical haemolysis by inducing disseminated intra-
vascular haemolysis or red cell membrane damage via degradation
of membrane phospholipids and proteins. Bartonella bacilliformis
endemic to western South America causes Oroya fever character-
ized by fever, chills, musculoskeletal pain, and acute intravascular
haemolysis. Clostridium perfringens releases enzymes that degrade
the RBC membrane, and clostridial sepsis has been associated with
rapid, massive haemolysis.
Chemical
Drugs and chemicals known to cause haemolysis through direct
oxidative damage are summarized in Tables 22.6.12.2 and
22.6.12.3. In most cases, the strong oxidant activity of these chem-
icals overwhelm normally functioning reduction mechanisms re-
sponsible for protecting haemoglobin and the red cell membrane.
Variability in the absorption of the chemical or its metabolism
Box 22.6.12.2  Other causes of acquired haemolytic anaemia
•	 Paroxysmal nocturnal haemoglobinuria
•	 Lipid disorders
•	 Liver disease:
—	 Hepatitis
—	 Cirrhosis
—	 Gilbert’s disease
•	 Chronic alcoholism (Zieve syndrome)
•	 Wilson’s disease
•	 Vitamin E deficiency
•	 Hypersplenism
•	 Hyperbaric oxygen therapy
•	 Total body irradiation
•	 Chronic large granular lymphocytic leukaemia
•	 Renal disease
•	 Cardiopulmonary bypass
•	 Freshwater/​saltwater drowning
22.6.12  Acquired haemolytic anaemia
5487
determines whether a particular individual will develop chemical-​
induced haemolytic anaemia. Often it is the chemical’s metabolite
that is responsible for inducing haemolysis. The red cells of new-
borns do not have functional reduction mechanisms and thus are
more sensitive to oxidant activity.
Mechanical
Mechanical fragmentation of erythrocytes can occur when for-
eign material is placed within the vasculature, when fibrin strands
or platelet thrombi obstruct small blood vessels, or when direct
physical forces compress superficial blood vessels. Thrombotic
microangiopathies, which result in a microangiopathic haemolytic
anaemia, are discussed in the following microangiopathic haemo-
lytic anaemias section.
Foreign material
Mechanical haemolysis occurs most commonly with artificial
valvular prostheses, particularly when accompanied by turbu-
lent blood flow. Bacterial endocarditis and associated valvular
vegetations can also cause fragmentation of red cells. Haemolysis
also occurs in up to 10% of patients with transjugular intrahepatic
portosystemic shunts. Increased cardiac output as a result of an-
aemia, exercise, or medications can increase the rate of red cell
fragmentation. The peripheral smear usually demonstrates
schistocytes and microspherocytes. Severe haemolysis usually re-
quires surgical repair.
March haemoglobinuria
Haemoglobinuria can occur in soldiers or joggers following ex-
tended periods of marching or running on a hard surface, or in
karate or conga drummer enthusiasts following practice. This mech-
anical haemolysis appears to be the result of red cell compression
in superficial blood vessels during the period of contact between
the extremity and the hard surface. The peripheral smear is normal.
Treatment is unnecessary as the syndrome is otherwise symptomless
and lacks significant clinical sequelae.
Thermal haemolysis
Normal red cells undergo fragmentation and lysis when heated to
temperatures of 49°C or higher. The two most common clinical situ-
ations associated with heat-​induced red cell lysis are the use of faulty
blood warmers during transfusion or patients who have sustained
extensive burns.
Venom
Haemolysis has been observed following bee and wasp stings, spider
bites, and snake bites. The haemolysis occurs secondary to dissem-
inated intravascular coagulation or as a result of proteolytic enzymes
contained within the venom.
Microangiopathic haemolytic anaemia
MAHA is a descriptive term for a haemolytic anaemia due to
intravascular RBC fragmentation which produces schistocytes.
MAHA is a component of numerous congenital and acquired
immune and nonimmune disorders. Due to its diverse aetiolo-
gies, MAHA is described separately in this text. Thrombotic
microangiopathies (TMAs) are a frequent cause of MAHA;
however, MAHA may also be seen in numerous other settings,
including the presence of intravascular devices as previously de-
scribed. Thrombotic microangiopathies are characterized by the
presence of MAHA and thrombocytopenia and have a variety of
Table 22.6.12.2  Some drugs that may induce haemolysis in
G6PD-​deficient individuals
Antimalarials
Sulfones
Primaquine
Thiazolesulfone
Pamaquine
Dapsone
Pentaquine
Nitrofurans
Chloroquine
Nitrofurantoin
Quinidine
Nitrofurazone
Quinine
Furazolidone
Quinacrine
Antipyretics/​analgesics
Sulfonamides
Acetanilide
Sulfanilamide
Aspirin
Sulfacetamide
Paracetamol (acetaminophen)
Sulfapyridine
Phenacetin
Sulfamethoxazole
Aminopyrine
Sulfafurazole
Other drugs
Sulfamethoxypyridazine
Methylene blue
Sulfoxone
Nalidixic acid
Sulfadiazine
Chloramphenicol
Sulfamerizine
Doxorubicin
Sulfisoxazole
Dimercaprol
Sulfadimidine
Probenecid
Other chemicals
Vitamin K analogues
Napthalene
Phenazopyridine
Trinitrotoluene
p-​Aminosalicylic acid
Toluidine blue
Ciprofloxacin
Norfloxacin
Note: see also Chapter 22.6.11.
Table 22.6.12.3  Chemicals that cause haemolysis
Oxidative haemolysis
Nitrofurantoin
Arsine gas
Sulfonamides
Chlorate
Sulfones (dapsone)
p-​Aminosalicylic acid
Phenazopyridine
p-​Nitroaniline
Phenacetin
Nitrobenzene derivatives
Phenylhydrazine
Vitamin K analogues
Phenothiazine
Paraquat
Isobutyl nitrate
Naphthalene (mothballs)
Amyl nitrite
Hydrogen peroxide
Nonoxidative haemolysis
Copper
Lead
section 22  Haematological disorders
5488
aetiologies. In thrombotic microangiopathies, the red cells are
fragmented during their forced passage through vessels con-
taining microthrombi. The degree of anaemia is variable. The per-
ipheral smear reveals findings typical of mechanical haemolysis
including schistocytes, microspherocytes, and a reticulocytosis
(Fig. 22.6.12.3). Varying degrees of thrombocytopenia are also
observed.
The primary TMAs include thrombotic thrombocytopenic pur-
pura (TTP), shiga toxin-​mediated haemolytic uraemic syndrome,
complement-​mediated TMA, drug-​induced TMAs, and other rare
hereditary disorders of vitamin B12 metabolism and coagulation.
Systemic disorders may also lead to the development of MAHA,
including preeclampsia, HELLP, malignant hypertension, severe
infections, disseminated malignancies, DIC, autoimmune dis-
orders and haematopoietic progenitor cell or solid-​organ transplant.
Systemic disorders and mechanical causes of MAHA must be dis-
tinguished from the primary TMA syndromes, as therapy for the
former is typically directed at the underlying disorder. Several of the
primary TMA syndromes will be discussed further in the following
sections.
Shiga toxin-​mediated haemolytic uraemic syndrome
Shiga toxin-​mediated haemolytic uraemic syndrome is primarily,
but not exclusively, a disease of childhood. The disorder consists
of widespread damage to the vascular endothelium and fibrin de-
position. These pathological changes are frequently most severe in
the renal arterioles and glomerular capillaries. The disorder usu-
ally develops following a febrile illness. Numerous reports have
documented the development of haemolytic uraemic syndrome
following infections with toxin-​secreting strains of Escherichia
coli (strain O157:H7) or shigella. Initial nausea, vomiting, and
diarrhoea can develop into severe abdominal pain and bloody
diarrhoea. Acutely, the child may develop hypertension, oliguria,
purpura, bleeding, and anaemia. If left untreated, convulsions,
coma, and death may occur. Mortality rates in young children
are reported to be low (3%), however the disease is more severe
with a higher mortality in adults. The peripheral smear exhibits
schistocytosis and thrombocytopenia. Therapy consists mainly
of supportive care, transfusion, control of blood pressure, and
dialysis.
Thrombotic thrombocytopenic purpura
TTP is caused by either a congenital deficiency of or an acquired
inhibitor to a serum metalloprotease (ADAMTS13) which is re-
sponsible for cleaving unusually large multimers of von Willebrand
factor. Left uncleaved, the large von Willebrand factor multimers in-
duce TTP by causing the agglutination of circulating platelets. TTP
occurs mainly in adults and more commonly involves the central
nervous system, although renal abnormalities can occur. Patients
may present with fever, purpura, petechiae, anaemia, thrombocyto-
penia, and neurological abnormalities. The neurological sequelae
include convulsions, coma, paralysis, delirium, and stroke. The per-
ipheral smear demonstrates schistocytes, thrombocytopenia, and
a reticulocytosis. Congenital TTP may be treated with plasma in-
fusions. In acquired TTP, front-​line therapy includes steroids and
daily plasma exchange with plasma or virally-​inactivated solvent-​
detergent plasma. Plasma exchange accomplishes one or more of the
following: (1) removal of the antibody to the protease; (2) removal of
large multimers of von Willebrand factor; and/​or (3) replenishment
of normal protease.
In patients who do not initially respond to plasma exchange, a
thorough re-​evaluation for alternative diseases should be under-
taken. The addition of rituximab should be considered, although
the optimal dose and schedule have not been established. Other op-
tions include the use of cryo-​poor supernatant as the replacement
fluid. Cryo-​poor supernatant contains markedly reduced levels of
normal von Willebrand factor which is believed to enhance the
formation of microthrombi in some patients. Additional reported
therapies in refractory or relapsing patients have included Cytoxan,
ciclosporin, vincristine, bortezomib, mycophenolate mofetil, N-​
acetylcysteine, and splenectomy. Clinically significant bleeding
is uncommon in TTP, and prophylactic platelet transfusions are
typically unnecessary unless the patient requires an invasive pro-
cedure where the risk of bleeding is significant. Although there is
some concern that platelet transfusions may ‘fuel the fire’ in pa-
tients with TTP, evidence supporting this assumption is lacking
and platelets should not be withheld in the setting of clinically sig-
nificant bleeding.
Complement-​mediated TMA
Complement-​mediated TMA is the result of the uncontrolled ac-
tivation of the alternative complement pathway due to mutations
or antibodies to complement regulatory proteins. Mutations may
include loss of function mutations in regulatory genes (CFH, CFI,
or CD46) or gain-​of-​function mutations in effector genes (CFB or
C3). Functional deficiencies in factor H have also been reported
due to CFH antibodies. Patients with complement-​mediated TMA
present with acute kidney injury and hypertension, MAHA, and
thrombocytopenia. The ADAMTS13 level is typically not severely
deficient and the stool is negative for the Shiga toxin. As normal
plasma levels of C3, C4, CFB, CFH, and CFI do not exclude the
diagnosis of complement-​mediated haemolytic uraemic syndrome,
screening for mutations and antibodies to the complement proteins
is required. Initial treatment of complement-​mediated haemolytic
Fig. 22.6.12.3  The peripheral blood changes in microangiopathic
haemolytic anaemia. This patient had recurrent thrombocytopenic
purpura and the marked fragmentation of the red cells together with
microspherocytosis is evident on the blood film. Magnification × 1000,
Leishman stain.
22.6.12  Acquired haemolytic anaemia
5489
uraemic syndrome is supportive. Eculizumab can be considered
a first-​line therapy if available. Plasma exchange may be useful in
some patients, and immunosuppressive therapy should also be
considered in patients with antibody-​mediated disease. Combined
liver–​renal transplant is curative for complement-​mediated muta-
tions of CFH, CFI, CFB, and C3.
FURTHER READING
Berentsen S, Randen U, Tjonnfjord (2015). Cold agglutinin-​mediated
autoimmune hemolytic anemia. Hematol Oncol Clin N Am, 29,
455–​71.
Crowther M, et al. (2011). Evidence-​based focused review of the treat-
ment of idiopathic warm immune haemolytic anemia in adults.
Blood, 118, 4036–​40.
Davenport RD (2012). Hemolytic transfusion reactions. In Popovsky
M (ed) Transfusion reactions, pp 1–​51. AABB Press, Bethesda, MD.
Davidson RJL (1969). March or exertional hemoglobinuria. Semin
Hematol, 6, 150.
Fung MK, et al. (ed.) (2014). Technical manual, 18th edition. American
Association of Blood Banks, Bethesda, MD.
Furlan M, et al. (1998). Von Willebrand factor-​cleaving protease in
thrombotic thrombocytopenic purpura and the hemolytic-​uremic
syndrome. N Engl J Med, 339, 1578–​84.
George JN, Nester CM (2014). Syndrome of thrombotic
microangiopathy. N Engl J Med, 371, 654–​66.
Hows J (1986). Donor-​derived red blood cell antibodies and immune
hemolysis after allogeneic bone marrow transplantation. Blood, 67,
177–​81.
Judd WJ (2001). Practice guidelines for prenatal and perinatal
immunohematology, revisited. Transfusion, 41, 1445–​52.
Marsh GW, Lewis SM (1969). Cardiac hemolytic anemia. Semin
Hematol, 6, 133–​45.
Naik R (2015). Warm autoimmune haemolytic anemia. Hematol Oncol
Clin N Am, 29, 445–​53.
Petz LD, Garratty G (2004). Immune hemolytic anemias. Churchill
Livingstone, Philadelphia.
Ramsey G (1991). Red cell antibodies arising from solid organ trans-
plants. Transfusion, 31, 76–​86.
Price EA, Schrier SL (2013). Extrinsic nonimmune haemolytic
anemias. In: Hoffman R, et al. (eds) Hematology: basic principles and
practice, pp. 628–​37. Elsevier Saunders, Philadelphia.
Shanbhag S, Spivak J (2015). Paroxysmal cold hemoglobinura. Hematol
Oncol Clin N Am, 29, 473–​8
Shirey RS, et al. (2002). Prophylactic antigen-​matched donor blood
for patients with warm autoantibodies: an algorithm for transfusion
management. Transfusion, 41, 1435–​41.
Tsai H-​M, Lian EC-​Y (1998). Antibodies to von Willebrand factor-​
cleaving protease in acute thrombotic thrombocytopenic purpura.
N Engl J Med, 399, 1585–​94.
Weinstock C, et al. (2015). Successful use of eculizumab for treatment
of an acute haemolytic reaction after ABO-​incompatible red blood
cell transfusion. Transfusion, 55, 605–​10.

22.6.2 Anaemia pathophysiology, classification, an
22.6.2 Anaemia: pathophysiology, classification, and clinical features 5359 David J. Weatherall† and Chris Hatton
22.6.2  Anaemia: pathophysiology, classification, and clinical features
5359
Basak A, et al. (2015). BCL11A deletions result in fetal hemoglobin
persistence and neurodevelopmental alterations. J Clin Invest, 125,
2363–​8.
Chen MJ, et al. (2011). Erythroid/​myeloid progenitors and hemato-
poietic stem cells originate from distinct populations of endothelial
cells. Cell Stem Cell, 9, 541–​52.
Dussiot M, et al. (2014). An activin receptor IIA ligand trap cor-
rects ineffective erythropoiesis in beta-​thalassemia. Nat Med, 20,
398–​407.
Hu J, et  al. (2013). Isolation and functional characterization of
human erythroblasts at distinct stages:  implications for under-
standing of normal and disordered erythropoiesis in vivo. Blood,
121, 3246–​53.
Iolascon A, et al. (2013). Congenital dyserythropoietic anemias: mo-
lecular insights and diagnostic approach. Blood, 122, 2162–​6.
Kuhrt D, Wojchowski DM (2015). Emerging EPO and EPO receptor
regulators and signal transducers. Blood, 125, 3536–​41.
Li J, et al. (2014). Isolation and transcriptome analyses of human eryth-
roid progenitors: BFU-​E and CFU-​E. Blood, 124, 3636–​45.
Ludwig LS, et al. (2014). Altered translation of GATA1 in Diamond-​
Blackfan anemia. Nat Med, 20, 748–​53.
McGrath KE, et al. (2015). Distinct sources of hematopoietic progen-
itors emerge before HSCs and provide functional blood cells in the
mammalian embryo. Cell Rep, 11, 1892–​904.
Mettananda S, Gibbons RJ, Higgs DR (2015). Alpha-​globin as a mo-
lecular target in the treatment of beta-​thalassemia. Blood, 125,
3694–​701.
Musallam KM, et al. (2013). Clinical experience with fetal hemoglobin
induction therapy in patients with beta-​thalassemia. Blood, 121,
2199–​212.
Nienhuis AW, Persons DA (2012). Development of gene therapy for
thalassemia. Cold Spring Harb Perspect Med, 2, a011833.
Notta F, et  al. (2016). Distinct routes of lineage development re-
shape the human blood hierarchy across ontogeny. Science, 351,
aab2116.
Orkin SH, Zon LI (2008). Hematopoiesis: an evolving paradigm for
stem cell biology. Cell, 132, 631–​44.
Palis J (2014). Primitive and definitive erythropoiesis in mammals.
Front Physiol, 5, 3.
Sankaran VG, Orkin SH (2013). The switch from fetal to adult hemo-
globin. Cold Spring Harb Perspect Med, 3, a011643.
Sankaran VG, Weiss MJ (2015). Anemia: progress in molecular mech-
anisms and therapies. Nat Med, 21, 221–​30.
Sankaran VG, et al. (2008). Human fetal hemoglobin expression is
regulated by the developmental stage-​specific repressor BCL11A.
Science, 322, 1839–​42.
Sankaran VG, et al. (2011). MicroRNA-​15a and -​16-​1 act via MYB to
elevate fetal hemoglobin expression in human trisomy 13. Proc Natl
Acad Sci U S A, 108, 1519–​24.
Sankaran VG, et al. (2012). Exome sequencing identifies GATA1 mu-
tations resulting in Diamond-​Blackfan anemia. J Clin Invest, 122,
2439–​43.
Suragani RN, et al. (2014). Transforming growth factor-​beta super-
family ligand trap ACE-​536 corrects anemia by promoting late-​stage
erythropoiesis. Nat Med, 20, 408–​14.
Ulirsch JC, et al. (2014). Altered chromatin occupancy of master
regulators underlies evolutionary divergence in the transcrip-
tional landscape of erythroid differentiation. Plos Genet, 10,
e1004890.
Yoder MC (2014). Inducing definitive hematopoiesis in a dish. Nat
Biotechnol, 32, 539–​41.
22.6.2  Anaemia: pathophysiology,
classification, and clinical features
David J. Weatherall† and Chris Hatton
ESSENTIALS
Anaemia is defined by the World Health Organization as a reduction
of the haemoglobin concentration to less than 130 g/​litre (males) or
less than 120 g/​litre (females). It is a common problem, with a preva-
lence around 3% for middle-​aged men and 14% for middle-​aged
women in the United Kingdom, and is seen more frequently with
age. There is a much greater prevalence in the developing world.
Adaptation to anaemia
Reduction in delivery of oxygen to the tissues triggers a var-
iety of compensatory mechanisms, including (1)  modulation of
oxygen affinity—​largely mediated by an increase in red blood cell
2,3-​biphosphoglycerate; (2) increased production of erythropoietin—​
the main growth factor for red blood cell production; (3)  redis-
tribution of flow to benefit the myocardium, brain, and muscle;
(4) increase in cardiac output; and (5) reduction of mixed venous
oxygen tension to increase the arteriovenous oxygen difference.
Clinical manifestations
The clinical picture depends on whether anaemia is of rapid or in-
sidious onset. Acute blood loss presents with features of intravas-
cular volume depletion. Anaemia of gradual onset may (if mild) be
asymptomatic or simply manifest as slight fatigue and pallor, or (if
more severe) present with features including exertional dyspnoea,
tachycardia, palpitations, angina, light-​headedness, faintness, and
signs of cardiac failure.
Causes and classification
Anaemia can be caused by the defective production of red cells or
an increased rate of loss of cells, either by bleeding or premature
destruction (haemolysis).
The causes of defective production of red cells include (1) de-
ficiency of key haematinics, including iron, vitamin B12, or folate;
(2) anaemia of chronic disease (see also Chapter 22.6.5); (3) reduced
erythropoietin production typically seen in chronic kidney disease;
and (4) primary diseases of the bone marrow.
Haemolytic anaemias may be classified as either (1)  genetic—​
including membrane defects, haemoglobin disorders, and en-
zyme deficiencies; or (2)  acquired—​including autoimmune and
nonimmune disorders.
Clinical approach
The key issues are to determine (1) the degree of disability caused by
the anaemia and hence how quickly treatment must be started, a key
question being ‘Is blood transfusion required?’, and (2) the cause of
the anaemia.
† It is with great regret that we report that David J. Weatherall died on 8
December, 2018.
section 22  Haematological disorders
5360
The main causes of anaemia can usefully be classified ac-
cording to the associated red cell morphological changes:
(1) hypochromic, microcytic—​including iron deficiency (the com-
monest cause of anaemia) and thalassaemia (common in some
populations); (2) normochromic, macrocytic—​vitamin B12 or folate
deficiency, alcohol, and myelodysplasia; (3)  polychromatophilic,
macrocytic—​haemolysis; (4)  normochromic, normocytic—​chronic
disorders, renal failure, and diseases of the bone marrow; and
(5)  leucoerythroblastic—​myelofibrosis, leukaemia, and metastatic
carcinoma. The reticulocyte count also provides a good way of
thinking about the underlying cause of anaemia—​whether due to
defective production or excess consumption of red cells.
Definition of anaemia
The main function of the red blood cells is oxygen transport. Hence,
a functional definition of anaemia is ‘a state in which the circulating
red cell mass is insufficient to meet the oxygen requirements of the
tissues’. However, many compensatory mechanisms can be brought
into play to restore the oxygen supply to the vital centres, and there-
fore in clinical practice this definition is of limited value. For this
reason, anaemia is usually defined as ‘a reduction of the haemo-
globin concentration, red cell count, or packed cell volume to below
normal levels’.
It has been extremely difficult to establish a normal range of
haematological values, and hence the definition of anaemia usu-
ally involves the adoption of rather arbitrary criteria. For example,
the World Health Organization recommends that anaemia should
be considered to exist in adults whose haemoglobin levels are
lower than 130 g/​litre (men) or 120 g/​litre (women). Children aged
6 months to 6 years are considered anaemic at haemoglobin levels
below 110 g/​litre, and those aged 6 to 14 years below 120 g/​litre. The
disadvantage of such arbitrary criteria for defining anaemia is that
there may be many apparently normal individuals whose haemo-
globin concentration is below their optimal level. Furthermore, the
published ‘normal values’ for adults indicate that there is such a large
standard deviation that many women must be considered ‘normal’
even though they have haemoglobin levels below 120 g/​litre.
Prevalence of anaemia
The prevalence of anaemia has been studied in many populations,
but it is difficult to compare data from different sources because of
variations in methodology and criteria. Certain patterns emerge,
however. An early survey carried out in the United Kingdom estab-
lished that haemoglobin levels were low in a significant proportion
of the population, particularly susceptible groups being children
under the age of 5 years and pregnant women. A later random popu-
lation study, also in the United Kingdom, reported a prevalence of
anaemia of 14% for women aged 55 to 64 years and 3% for men aged
35 to 64 years. These and similar studies have shown that anaemia
is most common in women of child-​bearing age and that it then be-
comes relatively less frequent, although the prevalence increases
again in the 65-​and-​over age group. Interestingly, it is only in the last
group that the prevalence in men and women is almost the same.
Where the cause of the anaemia has been analysed in these surveys,
most cases have been due to iron deficiency. No doubt these preva-
lence data vary considerably between the developed countries, but it
is clear that nutritional anaemia is relatively common in most popu-
lations at certain periods during development and late in life.
Adaptation to anaemia
The function of the red cell is to carry oxygen between the lungs and
the tissues. However, tissue oxygenation is the result of a complex
series of interactions of different organ systems, of which the red
cell is only one (Table 22.6.2.1). Obviously the cardiac output, ven-
tilatory function, and state of the capillaries are of great importance
as well. Each of these oxygen supply systems is regulated differently.
Ventilation responds to changes in pH, CO2, and hypoxia. Cardiac
output responds to the amount of blood entering the heart, and this
is regulated mainly by the effects of tissue metabolism as it modifies
the resistance to blood flow in the microvasculature. The erythron
itself responds to changes in haemoglobin concentration, arterial
oxygen saturation, and the oxygen affinity of the circulating haemo-
globin. Thus a decreased capacity of any of these components may
be compensated for by increased activity of the others in an attempt
to maintain tissue oxygenation.
Oxygen diffuses across the alveolar membrane and into the
blood, which equilibrates with the alveolar gas; the approximate
oxygen tension is 75 to 100 mmHg (11–​13 kPa). As blood is pumped
through the tissue capillaries, oxygen diffuses out. Although the
venous oxygen tension varies between organs, the oxygen tension of
the pooled venous blood in the pulmonary artery, the ‘mixed venous
oxygen tension,’ is remarkably constant at 40 mmHg (c.5.5 kPa). By
reducing the oxygen-​carrying capacity of blood, anaemia tends to
reduce the arteriovenous oxygen difference, and this may be com-
pensated for by the following mechanisms:  (1) modulation of
oxygen affinity; (2) increased production of erythropoietin (EPO);
(3) redistribution of flow between different organs; (4) increase in
cardiac output; and (5) reduction of mixed venous oxygen tension
to increase the arteriovenous oxygen difference.
Intrinsic red cell adaptation
The consequences of anaemia on the normal oxygen-​binding curve
of blood are shown in Fig. 22.6.2.1. Anaemia, by lowering the
Table 22.6.2.1  The steps involved in the transport of oxygen to the
tissues
Steps
Factors involved
Ambient O2 tension
Altitude
↓
Ventilation
Alveolar ventilation
↓
Gas-​to-​blood diffusion
Ventilation/​perfusion ratio
Anatomical shunt
Circulation
Cardiac output
↓
Blood: haemoglobin concentration, oxygen
dissociation curve
Tissue diffusion
Intercapillary distance
22.6.2  Anaemia: pathophysiology, classification, and clinical features
5361
haemoglobin concentration, proportionately reduces the oxygen-​
carrying capacity of the blood. As a response to this, there is an in-
crease in the 2,3-​biphosphoglycerate (2,3-​BPG) concentration in the
red cell, shifting the dissociation curve to the right and significantly
enhancing tissue oxygen delivery (Fig. 22.6.2.1).
With increasing severity of anaemia there is a progressive increase
in 2,3-​BPG, which may increase oxygen delivery by as much as 40%
for the same haemoglobin concentration. It should be noted, how-
ever, that a consequence of this adaptation is a lower venous oxygen
content and hence a lower reserve of oxygen available for a further
increase in oxygen demand, as might occur during exercise, for ex-
ample. Hence the increase in 2,3-​BPG in anaemia tends to ameli-
orate the effects of the diminished oxygen-​carrying capacity of the
blood, so reducing the adaptation required by other steps involved
in tissue oxygen delivery (Fig. 22.6.2.2). 2,3-​BPG levels vary in a
variety of other clinical conditions, some of which are summarized
in Box 22.6.2.1.
Erythropoietin
EPO is the major hormone involved in the regulation of erythro-
poiesis. Interaction of EPO with its receptor on red cell precursors
results in the stimulation of erythroid-​cell division, differentiation,
and the prevention of the apoptosis of erythroid progenitors. The
hormone is produced primarily in the kidney in adult life and in the
liver during fetal development. EPO production is increased by a
hypoxic stimulus secondary to anaemia.
A nucleotide sequence close to the EPO gene, the hypoxia regula-
tory element, is responsible for hypoxic regulation of the EPO gene
transcription. This, in turn, is controlled by the transcription factor
hypoxia inducible factor-​1 (HIF-​1). HIF-​1 is part of a widespread
oxygen-​sensing mechanism and is found in many cell types that do
not express the EPO gene. It is made up of two subunits, HIF-​1α
and HIF-​1β; only the former is regulated by hypoxia. HIF-​1 protein
levels are increased by hypoxia and return to normal with adequate
oxygenation. In the presence of oxygen HIF-​1α is hydroxylated by
an oxygen-​sensitive proline hydroxylase. Hydroxylated HIF-​1α be-
comes a target for interaction with the von Hippel–​Lindau protein
that initiates the rapid destruction of HIF-​1α. In essence, this com-
plex constitutes the oxygen sensor. Thus, variation in the produc-
tion of EPO in various conditions, particularly renal disease, may
have profound effects on adaptation to anaemia. There is evidence
that, at least in some forms of anaemia, there may be a decline in
EPO production at a given haemoglobin level with age (see also
Chapter 22.6.5). Further details of the mechanisms of action of EPO
are given in Chapter 22.3.5.
Local changes in tissue perfusion
The total blood volume does not change greatly in anaemia and
therefore increased tissue perfusion has to be achieved by shunting
blood from less to more vital organs. There is vasoconstriction of
the vessels of the skin and kidney; this mechanism has little effect on
Oxygen content (vol. %)
20
15
10
5
0
0
20
40
60
80
100
Venous
Arterial
3.3 vol.%
4.5 vol. %
Fig. 22.6.2.1  Enhancement of oxygen loading by decreased red cell
oxygen affinity in a patient with anaemia. An anaemic patient with a 50%
reduction in haemoglobin concentration has only a 27% reduction in
oxygen unloading.
Based on Klocke RA (1972). Oxygen transport and 2,3-​biphosphoglycerate (BPG).
Chest, 62 5 Suppl, 795–​855.
ANAEMIC
NORMAL
40–45%
6.0
5
10
15
Haemoglobin g/100 ml
Heart rate
Stroke volume
BPG
25%
2.6
A–V
Cardiac
index
pV
Fig. 22.6.2.2  The changes in factors involved in oxygen delivery
with progressive anaemia. As anaemia becomes more severe, cardiac
compensation becomes more significant. P(V)O2, mixed venous
oxygen tension.
From Bellingham AJ (1974). The red cell in adaptation to anaemic anoxia.
Clin Haematol, 3, 577–​94.
Box 22.6.2.1  Some conditions in which there is a change in red
cell 2,3-​BPG levels leading to modification of oxygen transport
Increased 2,3-​BPG; increased P50, reduced whole-​blood
oxygen affinity
•	 Anaemia
•	 Alkalosis
•	 Hyperphosphataemia
•	 Renal failure
•	 Hypoxia
•	 Pregnancy
•	 Cyanotic congenital heart disease
•	 Thyrotoxicosis
•	 Some red cell enzyme deficiencies
Decreased 2,3-​BPG; decreased P50, increased whole-​blood
oxygen affinity
•	 Acidosis
•	 Cardiogenic or septic shock
•	 Hypophosphataemia
•	 Hypothyroidism
•	 Hypopituitarism
•	 Following replacement with stored blood
section 22  Haematological disorders
5362
renal function. The organs that gain from the redistribution seem to
be mainly the myocardium, brain, and muscle.
Cardiovascular changes
It seems likely that mild anaemia is compensated for by shifts in
the oxygen dissociation curve. Overall, oxygen consumption is un-
changed in anaemia. However, when the haemoglobin level falls
below 70–​80 g/​litre, there is an increase in cardiac output, both at
rest and after exercise (Fig. 22.6.2.2). The stroke rate increases and a
hyperkinetic circulation develops, characterized by tachycardia, ar-
terial and capillary pulsation, a wide pulse pressure, and flow mur-
murs. The circulation time is shortened, left ventricular stroke work
is increased, and coronary flow is increased in proportion to the
increased cardiac output. It has been found that there is an acute
reversal of the high-​output state of chronic anaemia in response to
orthostatic stress or pressor amines. This suggests that redistribution
of blood volume and vasodilatation with reduced afterload play a
dominant role in the hyperkinetic circulatory responses to chronic
anaemia. The mechanism of the vasodilatation is not known; it may
be a direct result of tissue hypoxia. An additional factor that may be
of some importance in increasing cardiac output is the reduction in
blood viscosity produced by a relatively low red cell mass.
Although the normal myocardium may tolerate sustained hyper-
activity of this type indefinitely, patients with coronary artery disease
or those with extreme anaemia may have impaired oxygenation of
the myocardium. In such cases, cardiomegaly, pulmonary oedema,
ascites, and peripheral oedema may occur, and a state of high-​output
cardiac failure is established. At this stage, the plasma volume is al-
most always increased.
Pulmonary function
As blood, regardless of its oxygen-​carrying capacity, is almost com-
pletely oxygenated in the lungs, the oxygen pressure of arterial
blood in an anaemic patient should be the same as that in a normal
individual, and hence an increase in respiratory rate should not
improve the oxygenation of the tissues. Curiously, however, severe
anaemia is associated with dyspnoea. Although in some patients
this may be related to incipient cardiac failure, in most cases it ap-
pears to be an inappropriate response to hypoxia which is centrally
mediated.
Clinical manifestations and classification
of anaemia
Clinical effects of anaemia
As anaemia reduces tissue oxygenation it is not surprising that it
is associated with widespread organ dysfunction and hence an ex-
tremely varied clinical picture. The picture depends on whether the
anaemia is of rapid or more insidious onset.
After acute blood loss the red cell mass and plasma volume are
reduced proportionately and the symptoms are mainly of volume
depletion. Depending on the amount of fluid replacement there
may be a small fall in the packed cell volume during the first 10 h;
volume replacement by the influx of albumin from the extravascular
compartment takes between 60 and 90 h. Hence the picture of rapid
blood loss is characterized by the typical syndrome of shock, with
collapse, dyspnoea, tachycardia, a poor volume pulse, reduced blood
pressure, and marked peripheral vasoconstriction.
With anaemia of a more insidious onset, the compensatory mech-
anisms outlined earlier have time to come into play. In mild an-
aemia there may be no symptoms or simply increased fatigue and a
slight pallor. As the anaemia becomes more marked, the symptoms
and signs gradually appear. Pallor is best discerned in the mucous
membranes but tends to be an unreliable clinical sign in all but the
most severe anaemia. Pallor of the nail beds and palmar creases may
also be assessed. Cardiorespiratory symptoms and signs include
exertional dyspnoea, tachycardia, palpitations, angina or claudica-
tion, night cramps, increased arterial pulsation, capillary pulsation,
a variety of cardiac bruits, reversible cardiac enlargement, and, if
cardiac failure occurs, basal crepitations, peripheral oedema, and as-
cites. Neuromuscular involvement is reflected by headache, vertigo,
light-​headedness, faintness, tinnitus, roaring in the ears, cramps, in-
creased cold sensitivity, and haemorrhages in the retina. There may
be a low-​grade fever.
In older people, in whom associated degenerative arterial disease
is common, anaemia may present with the onset of cardiac failure.
Alternatively, previously undiagnosed coronary narrowing may be
unmasked by the onset of angina. Other symptoms of arterial degen-
erative disease may also be exacerbated or unmasked, for example,
intermittent claudication and a variety of neurological pictures as-
sociated with cerebral arteriosclerosis. It is important that anaemia
is recognized as a contributing factor to the symptoms of these de-
generative diseases as its correction may bring about considerable
symptomatic improvement.
Causes and classification of anaemia
A reduction in the red cell mass can result from either the defective
production of red cells or an increased rate of loss of cells, by either
premature destruction or bleeding. Decreased production of red
cells may result from a reduced rate of proliferation of precursors
in the bone marrow or from failure of maturation leading to their
intramedullary destruction: that is, ineffective erythropoiesis. Based
on this approach, we can derive a very simple pathophysiological
classification of anaemia, as shown in Box 22.6.2.2, in which the
causes are divided into failure of red cell proliferation, defective mat-
uration, haemolysis, and blood loss.
Anaemia due to defective proliferation of red cell precursors
The major causes of this group of anaemias are an inadequate supply
of iron, primary diseases of the bone marrow that involve stem cells
or later erythroid precursors, and a reduction in the amount of EPO
reaching the red cell precursors (Box 22.6.2.3).
Box 22.6.2.2  The main groups of anaemias classified
according to the underlying cause
•	 Reduced red cell production:
—​	 Defective precursor proliferation
—​	 Defective precursor maturation
—​	 Defective proliferation and maturation
•	 Increased rate of red cell destruction:
—​	 Haemolysis
•	 Loss of red cells from the circulation:
—​	 Bleeding
22.6.2  Anaemia: pathophysiology, classification, and clinical features
5363
Red cell precursors require adequate iron supplies for normal
proliferation, and the anaemia of iron deficiency tends to be
hypoproliferative as well as dyserythropoietic. Chronic inflammatory
disorders and related conditions also interfere with the iron supply to
erythroid precursors, probably mainly due to hepcidin blocking the
release of catabolized red cell iron from reticuloendothelial macro-
phages. Hepcidin, a peptide hormone produced by the liver, blocks
ferroportin, which normally mobilizes iron from macrophages and
gut endothelial cells onto transferrin for transportation to the bone
marrow (see Chapters 12.7.1 and 22.6.4). Hepcidin levels in the blood
are increased in chronic inflammatory disorders. Therefore, the fun-
damental defect in iron deficiency anaemia and the anaemia of in-
flammation is similar, in that the supply of iron is inadequate to meet
the requirements for erythropoiesis.
Defective proliferation of red cell precursors can result from
any of the causes of bone marrow failure, including infiltration
with leukaemic or other neoplastic cells, damage due to ionizing
radiation, drugs, or infection, and various intrinsic lesions of
the stem cells or red cell precursors. The intrinsic disorders in-
clude the congenital hypoplastic anaemias, involving either all
the myeloid elements or the red cell precursors alone (see also
Chapter 22.5.1).
Finally, decreased proliferation of the red cell precursors may re-
sult from EPO deficiency. The most common cause is chronic renal
failure. A similar mechanism may be involved in conditions in which
the tissue requirement for oxygen is reduced. These include various
endocrine disorders such as hypothyroidism and hypopituitarism.
It may also explain the mild anaemia associated with haemoglobin
variants with decreased oxygen affinity.
As a group, the hypoproliferative anaemias are associated with
a low reticulocyte count and defective proliferation of the bone
marrow precursors. The red cells are usually normochromic
and normocytic, although there may be a mild macrocytosis. If
the anaemia is due to iron deficiency, the cells are hypochromic.
If granulopoiesis is normal, the defect in red cell proliferation
is reflected by an increase in the marrow’s myeloid:erythroid
(M:E) ratio.
Defective red cell maturation
Defects of red cell maturation may involve primarily nuclear or
cytoplasmic maturation (Box 22.6.2.3). Those involving nuclear
maturation include vitamin B12 and folic acid deficiency and
other causes of megaloblastic anaemia (see also Chapter 22.6.6),
and some of the primary marrow disorders including erythro-
leukaemia. The important causes of defective cytoplasmic mat-
uration include the inherited disorders of globin synthesis, the
thalassaemia syndromes (see also Chapter 22.6.7), and the gen-
etic and acquired defects of iron metabolism that characterize the
sideroblastic anaemias. There are other genetic defects of red cell
maturation, the congenital dyserythropoietic anaemias, in which
the aetiology is becoming clearer. Furthermore, agents such as
drugs, chemicals, and infections may interfere with erythroid
maturation.
The main pathological mechanism common to all the anaemias
that result from maturation abnormalities is ineffective erythro-
poiesis. In other words, there is marked erythroid proliferation
but many of the precursors are destroyed in the bone marrow be-
fore they enter the circulation. Hence, the characteristic finding is
marked erythroid hyperplasia with a reduction in the M:E ratio,
associated with a low reticulocyte count. Because of the significant
intramedullary destruction of precursors there is usually an elevated
level of bilirubin and lactate dehydrogenase. Furthermore, there are
nearly always morphological abnormalities of the red cell precursors.
The anaemias that are associated with abnormal nuclear matur-
ation, such as those due to vitamin B12 and folic acid deficiency, are
characterized by megaloblastic erythropoiesis and macrocytic red
cells, while those caused by abnormal cytoplasmic maturation are
characterized by normoblastic hyperplasia and hypochromic and
microcytic red cells. However, even in these last conditions, there is
marked anisocytosis and there may be a proportion of macrocytes in
the peripheral circulation.
Blood loss
Anaemias due to chronic blood loss may develop very insidi-
ously and cause considerable diagnostic problems (see also
Chapter 22.6.4). Chronic blood loss from the gastrointestinal
tract or uterus of more than 15 to 20 ml/​day produces a state of
negative iron balance. Assuming that the patient starts with a
normal body store of iron, which is usually in the region of 1 g,
the bone marrow will be able to maintain a normal haemoglobin
Box 22.6.2.3  Main causes of anaemia due to reduced or
defective production of red cells
Reduced proliferation of precursors
•	 Iron deficiency anaemia
•	 Anaemia of chronic disorders:
—​	 Infections, malignancy, collagen disease, etc.
•	 Reduced erythropoietin production:
—​	 Renal disease
•	 Reduced oxygen requirements:
—​	 Hypothyroidism
—​	 Hypopituitarism
•	 Reduced oxygen affinity of haemoglobin
•	 Primary disease of the bone marrow:
—​	 Aplastic anaemia:
•	 Primary
•	 Secondary to drugs, irradiation, chemicals, toxins, etc.
•	 Pure red cell hypoplasia
•	 Infiltrative disorders:
—​	 Haematological malignancy
—​	 Secondary carcinoma
—​	 Marrow fibrosis
Defective maturation of precursors
•	 Nuclear maturation:
—​	 Vitamin B12 deficiency
—​	 Folate deficiency
—​	 Erythroleukaemia
•	 Cytoplasmic maturation:
—​	 Iron deficiency
—​	 Disorders of globin synthesis
—​	 Disorders of haem and/​or iron metabolism
—​	 Disorders of porphyrin metabolism
•	 Other mechanisms:
—​	 Congenital dyserythropoietic anaemias
—​	 Infection
—​	 Toxins and chemicals
section 22  Haematological disorders
5364
level until the iron stores are totally depleted. At this stage there
is no demonstrable iron in the bone marrow and the plasma iron
level starts to fall, but the patient is not anaemic. With a fur-
ther fall in the plasma iron level, the haemoglobin level starts
to fall, although at this stage the erythrocyte morphology may
be relatively normal, as are the red cell indices. It is only when
iron deficiency anaemia is well established that the typical mor-
phological appearances of the red cells develop, and only after
extreme periods of iron depletion that the tissue changes of iron
deficiency become manifest. Consequently, the often-​cited clin-
ical signs of iron deficiency (such as koilonychia) are exceed-
ingly uncommon.
From these considerations it is apparent that there may be pro-
longed blood loss before a patient presents with the symptoms and
signs of anaemia. During the earlier stages, the peripheral blood film
may not be helpful in diagnosis even though the serum iron level may
be extremely low. Indeed, sometimes a dimorphic blood picture with
normochromic and hypochromic cell populations may be seen. With
chronic blood loss there is quite often a persistent thrombocytosis,
and a hypochromic blood picture with thrombocytosis should al-
ways raise the possibility of chronic bleeding. In practice, the most
common sites of such bleeding are a hiatus hernia, peptic ulcer, the
large bowel, or the uterus; malignancy of the gastrointestinal or gy-
naecological tract must be considered.
Haemolytic anaemia
When the lifespan of red cells is shortened there is a reduction in the
circulating red cell mass, which leads to relative tissue hypoxia. This
in turn causes an increased output of EPO with stimulation of the
bone marrow and an increased rate of red cell production. This is re-
flected by a raised reticulocyte count and a mild macrocytosis due to
the presence of young red cells in the peripheral circulation. Due to
the increased rate of red cell destruction, there is an increased pro-
duction of bilirubin (from haem destruction, via biliverdin) which
leads to mild icterus and the presence of increased amounts of
urobilinogen in the urine and stool. Thus the haemolytic anaemias
(Box 22.6.2.4) are characterized by a variable degree of anaemia, a
reticulocytosis, and hyperbilirubinaemia. Their pathophysiology is
considered in detail in subsequent chapters.
Red cells are prematurely destroyed either because of an intrinsic
lesion or as a result of the action of an extrinsic agent. The intrinsic
abnormalities of the red cells that lead to their premature removal
are nearly all genetic defects of the cytoskeleton/membrane, haemo-
globin, or metabolic pathways. The extrinsic agents that may cause
premature destruction of the cells include a variety of antibodies,
chemicals, drugs, and toxins, or bacteria and parasites. In addition,
red cells may be damaged by direct trauma in the microcirculation or
on body surfaces.
Premature destruction of red cells may take place either intra-
vascularly or extravascularly, or, as occurs more commonly, in
both sites. The site of destruction depends on the type and degree
of damage to the red cell. For example, complement-​damaged cells
develop large holes in the membrane due to the membrane attack
complex, and are destroyed in the circulation, whereas IgG-​coated
cells are removed mainly by the Fc receptor-​bearing cells of the re-
ticuloendothelial system.
Clearly, there are numerous causes of premature destruction
of red cells. These will be considered in detail in later chapters in
this section. Usually it is easy to recognize that a particular an-
aemia has a haemolytic basis, by virtue of the reticulocytosis and
mild macrocytosis associated with erythroid hyperplasia of
the bone marrow, hyperbilirubinaemia, and increased urinary
urobilinogen. However, it should be remembered that many
anaemias associated with the abnormal proliferation or matur-
ation of red cells have a haemolytic component. For example, there
may be a slightly shortened red cell survival in patients with per-
nicious anaemia or thalassaemia and yet there may be a very poor
reticulocyte response. Similarly, there is a haemolytic component
in the anaemia due to inflammation or malignancy but again the
marrow response is poor.
General approach to the anaemic patient
Clinical assessment
The clinical assessment of patients with anaemia has two main ob-
jectives. First, it is essential to determine the degree of disability
caused by the anaemia and hence how quickly treatment must be
started. Second, as much information as possible about the likely
cause of the anaemia must be obtained from a detailed clinical his-
tory and physical examination. There is no place for the attempted
‘blind’ treatment of anaemia without first establishing the cause,
­except in the most urgent clinical settings.
In assessing the severity of the anaemia and how urgently treat-
ment should be instituted, a detailed history of the patient’s exercise
tolerance must be obtained. This should include a specific enquiry of
symptoms suggestive of cardiac complications including angina, dys-
rhythmias, positional dyspnoea, cough, or ankle swelling. The clin-
ical examination should include a careful assessment of the degree of
pallor, the position of the neck veins, whether there are warm extrem-
ities and a bounding pulse with a large pulse pressure, the presence of
ankle or sacral oedema, and whether there are basal crepitations on
respiratory examination. Severely ill patients with profound anaemia
require immediate treatment in an environment where they can be
under constant observation, have regular measurements of their cen-
tral venous pressure, and be managed by experienced clinical and
nursing staff. The risk of precipitating cardiac overload in these cases
is such that transfusion must be undertaken slowly and carefully.
Box 22.6.2.4  General classification of haemolytic anaemia
Genetically determined
•	 Defects involving the structure and/​or metabolism of the membrane
•	 Haemoglobin disorders
•	 Enzyme deficiencies involving the main metabolic pathways
Acquired
•	 Immune (iso-​ or auto-​)
•	 Nonimmune:
—​	 Trauma
—​	 Membrane defects
—​	 Drugs, chemicals, toxins
—​	 Bacteria, parasites
—​	 Hypersplenism
A more detailed description is given in Chapters  22.6.8, 22.6.9, 22.6.10,
22.6.11, and 22.6.12.
22.6.2  Anaemia: pathophysiology, classification, and clinical features
5365
It cannot be emphasized too strongly that in many cases the an-
aemia is a symptom of a nonhaematological disorder. A detailed his-
tory and clinical examination will often provide a clue as to the likely
cause of the anaemia, and which laboratory investigations are likely
to be most productive for confirming the diagnosis.
Haematological investigation
A preliminary blood count and blood film examination should
classify anaemia into hypochromic-​microcytic, and macrocytic
or normochromic, normocytic varieties (Box 22.6.2.5). In young
women with a history of heavy menstrual loss, it is reasonable to
assume that a hypochromic anaemia is due to iron deficiency, and
to focus investigation and subsequent management on the gy-
naecological tract. However, hypochromic anaemia in men, or in
postmenopausal women, suggests blood loss requiring urgent in-
vestigation until proven otherwise. Serum ferritin, serum iron
level, and total iron-​binding capacity should be assayed to con-
firm a diagnosis of iron deficiency. Hypochromic anaemia with
a normal serum ferritin may suggest the anaemia of chronic dis-
ease, or defects in haemoglobin synthesis such as thalassaemia and
sideroblastic anaemia; however, it important to be certain that iron
deficiency is not being masked by the acute phase response, which
will raise the ferritin level in many cases. Distinguishing between
the anaemia of chronic disease and iron deficiency, especially in the
elderly comorbid population, can be challenging, and a therapeutic
trial of iron may occasionally be required.
The diagnosis of a macrocytic anaemia always requires fur-
ther investigation. Simple blood tests to exclude readily remedi-
able causes such as vitamin B12 and folate deficiency should be
performed in the first instance; a reticulocyte count will also be
an indication of whether haemolytic anaemia needs to be in-
cluded in the differential diagnosis. Where these do not suggest
an explanation, a bone marrow examination should be considered.
A macrocytosis with a normoblastic bone marrow may result from
alcohol abuse, haemolysis, or, occasionally, one of the refractory
anaemias with hyperplastic bone marrow. Macrocytic anaemias
with megaloblastic bone marrows are usually due to vitamin B12
or folate deficiency and should be investigated accordingly (see
Chapter 22.6.6).
The normochromic, normocytic anaemias often cause more diag-
nostic difficulty. Some help can be gained from a determination of
whether the white cell and platelet counts are normal. If there is asso-
ciated neutropenia and thrombocytopenia, a primary disease of the
bone marrow is likely; hence, bone marrow examination should be
made to determine whether there is hypoplasia of the various pre-
cursor forms, hypoplastic or aplastic anaemia, or whether the pan-
cytopenia results from malignant infiltration of the bone marrow.
If there are nucleated red cells or immature myeloid cells on the
peripheral film (i.e. a leucoerythroblastic picture), a bone marrow
examination is essential, as this type of reaction usually indicates in-
filtration of the bone marrow with abnormal cells, either as part of a
primary marrow disease such as leukaemia, or metastatic carcinoma.
In the normochromic, normocytic anaemias in which the white cell
count and platelet count are normal, assessment of the patient’s renal
function is important, along with consideration of the possibility of
the anaemia of chronic disease. After these conditions have been ex-
cluded, there remain the chronic anaemias associated with endocrine
deficiencies or the primary red cell hypoplasia.
The management of anaemia
The management of specific forms of anaemia is described in detail
in subsequent chapters. However, a few principles can be outlined
here. In general, a cause should always be sought before treatment
is instituted. As mentioned previously, most cases of iron deficiency
anaemia require further investigation for a source of blood loss. If
there is a clear-​cut history of poor diet, multiple pregnancies, or ob-
vious heavy menstrual bleeding, it is reasonable to start iron therapy
and observe the haemoglobin level both during the period of treat-
ment and for some months after iron therapy has been stopped.
A rise in the haemoglobin level of approximately 10 g/​litre per week
indicates a full haematological response. For the megaloblastic
anaemias, blood samples should be obtained for serum folate and
vitamin B12 levels. Should haematinic supplementation be needed,
a brisk reticulocyte response 5 to 7 days after initiating therapy sug-
gests that there will be a full restoration of the haemoglobin level to
normal.
Blood transfusion should always be avoided unless the haemo-
globin level is dangerously low, in which case it is reasonable to
transfuse the patient up to a safe level and then allow the haemo-
globin to return to normal following appropriate treatment of the
underlying cause. The decision whether to transfuse an anaemic
Box 22.6.2.5  The main causes of anaemia classified
according to the associated red cell changes
Hypochromic–​microcytic (reduced mean red cell volume (MCV),
mean corpuscular haemoglobin (MCH), and mean corpuscular
haemoglobin concentration (MCHC))
•	 Genetic:
—​	 Thalassaemia
—​	 Sideroblastic anaemia
•	 Acquired:
—​	 Iron deficiency
—​	 Sideroblastic anaemia
—​	 Chronic disorders (mildly hypochromic, occasionally)
​Macrocytic (increased MCV)
•	 With megaloblastic marrow:
—​	 Vitamin B12 or folate deficiency
•	 With normoblastic marrow:
—​	 Alcohol, myelodysplasia
Polychromatophilic–​macrocytic (increased MCV)
•	 Haemolysis
Normochromic–​normocytic (normal indices)
•	 Chronic disorders:
—​	 Infection, malignancy, collagen disease, rheumatoid arthritis
•	 Renal failure
•	 Hypothyroidism, hypopituitarism
•	 Aplastic anaemia or primary red cell hypoplasia
•	 Primary disease of bone marrow, leukaemia, myelofibrosis, infiltration
with other tumours
Leucoerythroblastic (indices usually normal)
•	 Marrow fibrosis
•	 Haematological malignancies
•	 Metastatic carcinoma

22.6.3 Anaemia as a challenge to world health 5366
22.6.3 Anaemia as a challenge to world health 5366 David J. Roberts and David J. Weatherall†
section 22  Haematological disorders
5366
patient depends mainly on the severity of the anaemia and its cause.
For example, a young patient with a haemoglobin level of 50 g/​litre
who is shown to have an active duodenal ulcer should probably be
transfused because they would be at severe risk from a further brisk
bleed from the ulcer. On the other hand, a patient of similar age with
a similar haemoglobin level due to chronic nutritional iron defi-
ciency might well be allowed to restore their haemoglobin level by
oral iron therapy.
Occasionally, patients present in gross congestive cardiac failure
with profound anaemia. This picture is usually seen in elderly pa-
tients with long-​standing pernicious anaemia or iron deficiency.
This type of condition still carries a high mortality and requires
urgent treatment. Such profoundly anaemic patients require trans-
fusing up to a safe level, that is, a haemoglobin value of 60 to 80 g/​
litre. This can usually be achieved by the slow transfusion of one or
maybe two units of packed red cells. A very careful check on the
neck veins and lung bases should be made throughout the period
of transfusion. Ideally, a central venous pressure line should be in-
serted before the transfusion is started.
FURTHER READING
Koury MJ (2005). Erythropoietin: the story of hypoxia and a finely
regulated hematopoietic hormone. Exp Hematol, 33, 1263–​70.
O’Donnell A, et al. (2007). Age-​related changes in adaptation to severe
anemia in childhood in developing countries. Proc Natl Acad Sci U
S A, 104, 9440–​4.
Prchal JT (2010). Clinical manifestations and classification of erythro-
cyte disorders. In: Kaushanasky K, et al. (eds) Williams’ hematology,
8th edition, pp. 455–​62. McGraw-​Hill Medical, New York.
22.6.3  Anaemia as a challenge
to world health
David J. Roberts and David J. Weatherall†
ESSENTIALS
Anaemia is a very common problem in low-​ and middle-​income
countries (LMICs):  27% of the world’s population or 1.93 billion
people are affected by anaemia (2013) and more than 90% of people
with anaemia live in the developing world. Preschool children and
women of reproductive age are particularly affected by anaemia and
more 60% of anaemia is caused by iron deficiency.
Causes of anaemia in LMICs—​this is often multifactorial, with
causes including (1) nutritional deficiencies—​iron, folate, vitamin B12;
(2) chronic infection—​including malaria, tuberculosis, AIDS; (3) blood
loss—​hookworm, schistosomiasis; (4)  protein–​energy malnutrition;
(5) malabsorption—​for example, tropical sprue; (6) hereditary—​for
example, thalassaemias, haemoglobin variants, glucose-​6-​phosphate
dehydrogenase deficiency.
A series of vicious cycles exist in LMICs—​maternal anaemia due to
iron or folate deficiency and chronic malaria is associated with the
birth of underweight infants who frequently have low iron stores, may
also be folate deplete, and are usually anaemic from about 6 months
of age. Such infants are prone to infection, particularly gastrointes-
tinal, and may be further depleted of iron or folate by inappropriately
prolonged breastfeeding or weaning onto an inadequate diet. They
are exposed to hookworm infection as soon as they start to crawl,
malaria becomes an important problem after 6 months, and in many
populations the increasingly common haemoglobinopathies are a
further cause of anaemia after the first few months of life.
Introduction
Despite improvements in nutrition and hygiene, which have re-
duced childhood mortality in many low-​ and middle-​income coun-
tries (LMICs), anaemia continues to be an important problem in the
health of the world’s population. It is not, of course, a disease in its
own right but simply a by-​product of a wide variety of different dis-
orders, most of which are described in detail elsewhere. However,
because of its importance as a source of chronic ill health in many
populations, the global aspects of the aetiology and manifestations
of anaemia are summarized briefly in this chapter. Readers who wish
to learn more of the complex literature on this important topic are
referred to the extensive reviews cited at the end of the chapter.
Definition and prevalence
It has been very difficult to produce an adequate definition of ­anaemia.
‘Normal’ haematological values vary with age, between sexes, at dif-
ferent altitudes, and, possibly, between races. On the other hand, it is
helpful to have a standard set of haemoglobin levels at different ages
below which ‘anaemia’ is defined. The World Health Organization
(WHO) has attempted to set out criteria of these types, summarized
in Table 22.6.3.1. Despite their many shortcomings, including meth-
odological vagaries, they at least provide a way of obtaining an ap-
proximate comparison of the distribution and frequency of anaemia
among the different countries of the world and still remain valid.
The global prevalence of anaemia was first estimated in the 1980s.
A  review of the epidemiological data available at this time sug-
gested that about 1.3 billion people were affected by anaemia, par-
ticularly in LMICs. Infants, young children, menstruating women,
Table 22.6.3.1  Definition of haemoglobin levels below which
anaemia is said to exist in populations at sea level (WHO 1968)
Haemoglobin (g/​litre) below
Children, 6 months–​6 years
110
Children, 6–​14 years
120
Adult males
130
Adult females (nonpregnant)
120
Adult females (pregnant)
110
† It is with great regret that we report that David J. Weatherall died on 8
December, 2018.
22.6.3  Anaemia as a challenge to world health
5367
and, especially, pregnant women were the most severely affected
groups (Table 22.6.3.2). More recent estimations suggest that al-
though the prevalence of anaemia has declined, nearly 2 billion
people worldwide are affected by anaemia. Using publicly available
data, Kassenbaum and colleagues estimated mild, moderate, and
severe anaemia from 1990 to 2010 for over 180 countries, by sex and
well-​defined age groups and attributed the cause of anaemia using
data from the Global Burden of Diseases, Injuries and Risk Factors
(GBD) 2010 Study.
Global anaemia prevalence in 2010 was 32.9%, causing 70 mil-
lion years lived with disability (YLDs), which amounted to just
under 10% of health loss for all conditions. The prevalence of an-
aemia declined for both sexes from 1990 to 2010, particularly for
males. Anaemia is most common in South Asia and sub-​Saharan
Africa, while the greatest reduction in the burden of anaemia over
this period was in Asia. Iron deficiency anaemia was the top cause
globally but over 15 diseases were considered to contribute to an-
aemia and such a survey could not describe the multifactorial nature
of anaemia in individuals. About one-​fifth of perinatal mortality and
one-​tenth of maternal mortality in LMICs are attributable to iron
deficiency. In total, 0.8 million deaths worldwide are now attribut-
able to iron deficiency, that is, about 1.3% of all male deaths and 1.8%
of all female deaths.
These global surveys also highlight the more complex aetiology of
anaemia in older age groups. Gynaecological causes, led by uterine
fibroids, are important causes of anaemia in females from menarche
to menopause in all settings. Chronic kidney disease and gastro-
intestinal disorders increase sharply in older individuals to become
the most important causes of anaemia in many areas.
The complex and multiple aetiology of anaemia
in low-​ and middle-​income countries
The main causes of anaemia in LMICs are summarized in Box
22.6.3.1. It is very difficult to determine their relative import-
ance, particularly in LMICs. Most surveys have focused on one
particular mechanism (e.g. iron or folate deficiency). To obtain a
true picture of the cause of anaemia in a particular population it
is essential to obtain consecutive data over a long period. For ex-
ample, work in the Gambia has shown that the haemoglobin levels
in children vary significantly at different times of the year; anaemia
is much more common in the wet season when malaria transmis-
sion is at its highest. To complicate matters, this is also the time
when diarrhoea and malnutrition are most common. Heavy rains
after many dry months have profound effects on the community;
sanitation measures are disrupted and food stores are at the lowest
level in the annual cycle (Fig. 22.6.3.1).
These observations underline the multifactorial aetiology of an-
aemia across the world. Nonetheless, it is clear that iron deficiency,
which probably affects at least 15% of the world’s population, is the
most important factor; the many other diseases that can exacer-
bate anaemia are often operating on a  background of low body iron
stores.
Table 22.6.3.2  Estimated prevalence of anaemia by region and sex
Region
Percentage anaemic
Children
Women 15–​49 years
Men
0–​4 years
5–​12 years
Pregnant
All
15–​59 years
LMICs
51
46
59
47
26
HICs
12
  7
14
11
  3
World
43
37
51
35
18
LMICs, low- and middle-income countries; HICs, high-income countries.
Data from DeMaeyer EM, Adiels-​Tegman M (1985). The prevalence of anemia in the world. World Health Statist Quart, 38, 302–​16.
Box 22.6.3.1  Important causes of anaemia in LMICs
Acquired
•	 Nutritional:
—	 Iron, folate, vitamin B12
•	 Chronic infection:
—	 Malaria, leishmaniasis, schistosomiasis, tuberculosis, HIV
•	 Blood loss:
—	 Hookworm
—	 Schistosomiasis
•	 Protein–​energy malnutrition
•	 Malabsorption:
—	 Tropical sprue and related disorders
Hereditary
•	 Thalassaemias
•	 Haemoglobin variants
•	 Glucose-​6-​phosphate dehydrogenase deficiency
•	 Ovalocytosis
Number of cases
120
100
80
60
40
20
0
Rains
Severe anaemia
Malnutrition
Gastroenteritis
Jan
Apr
Jul
Oct
Rains
Rains
Jan
Apr
Jul
Oct
Jan
Apr
Jul
Oct
1990
1989
1988
Fig. 22.6.3.1  Admissions to the children’s ward in a hospital in the
Gambia over dry and rainy seasons.
Data from Brewster DR, Greenwood BM (1993). Season variation of paediatric
disease in The Gambia, West Africa. Ann Trop Paediatr, 13, 133.
section 22  Haematological disorders
5368
Iron deficiency
The causes of iron deficiency anaemia are extremely complex and
vary widely among different populations (see Chapter 22.6.4). The
absorption of nonhaem iron, except from breast milk, is compara-
tively restricted, and the content of iron in breast milk is very low.
Iron deficiency is particularly common in communities in which
food is predominantly of vegetable origin. The three great staples in
these populations are rice, wheat, and maize. Sorghum and millet
are also important in parts of Africa and Asia. Soy and similar leg-
umes are an important source of protein in many countries. The iron
content of these diets is generally low, and, furthermore, absorption
is inhibited by fibre, phytates, phosphates, and polyphenols, all of
which occur in high levels in vegetarian diets. Populations who have
remained as hunter-​gatherers, and pastoralists who eat meat, appear
to have a lower frequency of iron deficiency anaemia.
The body’s response to infection may also reduce iron stores and
iron utilization. Hepcidin is regulated by proinflammatory medi-
ators, such as tumour necrosis factor and interleukin-​6, which are
elevated in a wide variety of infections. High hepcidin levels stimu-
lated by malaria infection or bacterial infections reduce absorption
of iron from the gut and also reduce incorporation of iron into red
cells. Iron may act as a growth factor for malaria parasites, and so
raised hepcidin and reduction in available iron may form part of a
protective innate response to malaria infection. However, this pro-
tective response may contribute to functional iron deficiency and
anaemia in endemic areas.
Against this background of deficient or borderline dietary iron
intake, there are several other factors which may exacerbate iron
deficiency. Iron requirements are greatly increased during preg-
nancy because of the expansion of the maternal red cell mass
(c.500 mg), iron transport to the fetus (c.300 mg), and the consti-
tution of the placenta (c.25 mg), together with any blood loss at
birth. Although there is some compensation by the cessation of
iron loss due to menstruation (c.200 mg), the total requirements
for a single pregnancy are more than 1 g. Iron is also excreted in
breast milk and although the concentration is low, this loss, par-
ticularly with prolonged breastfeeding, places a further burden on
maternal iron stores.
In many tropical countries, there are important sources of patho-
logical iron loss due to parasitic infection. Hookworm infestation
affects millions of people worldwide. These parasites attach them-
selves to the mucosa of the intestinal tract. With a worm load of
1000 eggs/​g faeces, the intestinal blood loss averages about 2.5 ml/​
day, representing 1 mg of iron. Although some of this is reabsorbed,
perhaps up to 40%, hookworm infestation is an important source of
iron imbalance. Infection with Schistosoma mansoni results in intes-
tinal blood loss, while S. haematobium results in chronic haematuria.
In Kenyan children, for example, mean iron losses in those infected
with S. haematobium varied from 149 to 652 μg/​day, according to the
magnitude of the egg counts.
Finally, it should be remembered that chronic ill health due to
protein–​calorie malnutrition or chronic infection may, by its effect
on a patient’s appetite, result in further depletion of iron intake.
It must be emphasized that many surveys for assessing body iron
stores have used methods which are confounded by associated in-
flammatory disease or other disorders. These problems are particu-
larly germane to surveys which have been based on serum iron or
ferritin levels. More recently, screening methods based on estima-
tion of transferrin receptor levels and/or hepcidin have been devel-
oped but their application to large populations is, as yet, limited.
Folate deficiency
Folate deficiency is thought to be the second most frequent cause of
nutritional anaemia in the world’s population. The mechanisms are
complex and differ widely between different populations depending
in the way in which food is prepared, in particular the temperature
at which it is cooked. It is also clear that dietary folate deficiency is
not the whole story. Research in Africa suggests that the continuous
anorexia which accompanies recurrent infections such as malaria
or tuberculosis is a major cause of folate deficiency in children.
Postinfective malabsorption and the tropical sprue syndrome are
also important causes of folate deficiency, particularly in the Indian
subcontinent. Folate requirements may be increased in patients with
erythroid hyperplasia secondary to chronic haemolytic anaemia
(e.g. sickle cell anaemia, or chronic malarial infection). They also
increase markedly during pregnancy. In women with low baseline
folate stores, megaloblastic anaemia in pregnancy or the puerperium
is particularly common.
Vitamin B12 deficiency
Nutritional vitamin B12 deficiency is uncommon, although it is
observed in true vegans, particularly in the Indian subcontinent.
Infants born of mothers with sprue or postinfective malabsorption
who are fed on breast milk or goat’s milk containing insufficient
vitamin B12 may develop megaloblastic anaemia with locomotor
complications during the early months of life.
Infection
Almost any chronic infection may produce anaemia. Globally,
the most important are the parasitic disorders, malaria, visceral
leishmaniasis (kala-​azar), schistosomiasis, and some forms of
trypanosomiasis.
Malaria is still the most important parasitic illness of humans.
Currently it is estimated that it has a global incidence of about
200 million cases per year, with over 450 000 deaths. Its trans-
mission and clinical manifestations are considered in Chapter
8.8.2. Profound anaemia is a major cause of mortality and mor-
bidity during acute attacks of Plasmodium falciparum malaria
in nonimmune individuals, but, from the perspective of health
in LMICs, chronic infection with this organism in childhood
is an extremely common cause of anaemia. This is most com-
monly seen in areas of high malarial transmission and is also a
growing problem in regions of lower transmission because the
rise in antimalarial drug resistance prolongs the average duration
of infection. The anaemia of chronic malaria has a complex basis
involving haemolysis, hypersplenism, and a suboptimal bone
marrow response, often set against a background of iron or folate
deficiency. In some populations, notably those of Africa, India,
and parts of South-​East Asia, chronic malarial infection may be
complicated by the hyper-​reactive malarial splenomegaly syn-
drome, in which hypersplenism plays a major role in the gener-
ation of chronic anaemia.
The haematological manifestations of the other common parasitic
illnesses in the tropics are considered elsewhere in this text.
22.6.3  Anaemia as a challenge to world health
5369
Malabsorption
Many people in tropical climates, both indigenous populations and
expatriates who have worked in rural areas, have abnormalities of
the intestinal mucosa, often associated with impairment of absorp-
tion. These structural and functional alterations of the gut have been
called ‘tropical enteropathies’ (see Chapter 15.10.8). It is likely that
they result from adaptation to life in the contaminated environment
of the tropics, with frequent gastrointestinal infections and differ-
ences of diet.
More severe malabsorption syndromes, called sprue and
postinfective malabsorption, are associated with chronic diarrhoea,
wasting, and a variable degree of anaemia. The pathophysiology and
world distribution of these syndromes are considered in Section
15. They are nearly all associated with anaemia, which has a com-
plex aetiology including folate deficiency and, in some cases, iron
deficiency.
It should also be remembered that in a tropical setting malabsorp-
tion can also result from colonization of the small bowel by specific
parasites, including Giardia lamblia, Strongyloides stercoralis, crypto-
sporidium, and others. Abdominal tuberculosis with malabsorption
is also common. In Africa, HIV infection is now an important cause
of malabsorption and bone marrow suppression.
Inherited anaemias
The inherited haemoglobin disorders are becoming an increasingly
common cause of anaemia, particularly in LMICs countries. They
are described in detail in Chapter 22.6.7.
Due to heterozygote advantage against P.  falciparum malaria,
the important inherited haemoglobin disorders, notably sickle cell
anaemia and the thalassaemias, have a high frequency throughout
tropical populations of Africa and Asia (Table 22.6.3.3). Sickle
cell anaemia and its variants are particularly common in Africa,
some Mediterranean populations, and throughout the Middle
East and parts of India. They also occur at a high frequency in the
Caribbean and in other regions with large African populations.
The thalassaemias occur at a high frequency in parts of Africa,
the Mediterranean, the Middle East, the Indian subcontinent, and
throughout South-​East Asia. There is now clear evidence that these
conditions will produce a major public health problem in these
countries in the future. Children with sickle cell anaemia are highly
susceptible to systemic bacterial infection and without neonatal
screening for this disease and delivery of vaccination, prophylaxis,
and good medical care, up to half of children born with HbSS may
die within the first 3 years of life. As poorer countries go through
the demographic transition, resulting from better hygiene and con-
trol of infectious illness, infants with these genetic anaemias are now
surviving long enough to present for diagnosis and treatment. The
relatively high frequency of consanguineous marriages in many
LMICs also plays an important role in maintaining the frequency of
these recessively inherited diseases.
The effect that a high frequency of a disease such as thalassaemia
can have on the health economy of an emerging country was shown
graphically in the case of Cyprus after it passed through the demo-
graphic transition in the 1950s. It was estimated that if every patient
with this disease was treated with regular blood transfusion and ap-
propriate medication, within 15 years the management of this one
condition would consume up to 40% of the island’s health budget.
Recent studies in Indonesia indicate that, at a minimum estimate,
approximately 1.25 million units of blood will be required each year
to treat a proportion of the thalassaemic population in future years.
In many populations, there are hundreds of thousands of car-
riers for β-​thalassaemia or the more common severe forms of
α-​thalassaemia. Although they are asymptomatic they have
haemoglobin values which are, on average, 10 to 15 g/​litre below
normal. During pregnancy, they retain this difference so that in the
midtrimester they have haemoglobin values of approximately 80 g/​
litre or less. They have increased folate requirements and, in some
populations, there appears to be an increased frequency of folate de-
ficiency in pregnancy.
It should be remembered that the inherited anaemias may be ex-
acerbated by other illnesses which are widespread in tropical coun-
tries. Folate requirements are increased in all these conditions and
secondary folate deficiency is extremely common. They may also
be exacerbated by malaria; children may develop malarial infec-
tion from infected blood donors. There is also a high frequency of
other blood-​borne infections, particularly hepatitis C and, in some
populations, HIV. Furthermore, there is clear evidence that sickle
cell anaemia and thalassaemia can render children more prone to
infection. In short, like all forms of anaemia in the tropical world,
the inherited disorders of haemoglobin may present with a complex
series of complications due to a background of nutritional deficiency
and a wide variety of infections.
These complex interactions have a dominant effect on the prog-
nosis for the important inherited haemoglobin disorders. Early
studies in Africa reported a marked paucity of patients with sickle
cell anaemia despite a very high carrier frequency, indicating that
very few patients with this disorder were surviving beyond early
childhood. This may still be the case in parts of rural Africa. On the
other hand, in more developed countries, and with a high quality of
medical care, patients with this disease are regularly surviving into
adult life; the mean survival time in the United States of America is
now approximately 42 years, with many patients surviving to old
age. A similar situation exists for β-​thalassaemia. In poorer coun-
tries, supplies of blood may be limited, there may be difficulties in
screening blood for agents such as hepatitis C and HIV, and the pro-
hibitive cost of iron-​chelating agents means that even children who
Table 22.6.3.3  Annual births of severe disorders of haemoglobin
Sickle cell anaemia
Sub-​Saharan Africa
240 900
Elsewhere
93 000
HbSC disease
54 700
Thalassaemia
β-​Thalassaemia major
23 300
HbE β-​thalassaemia
20 600
HbH disease
14 500
HbS β-​thalassaemia
12 300
Hb Bart’s hydrops
5200
These data have to be viewed with caution because in many cases they are based
on small surveys from a small number of centres in individual countries; there is
evidence that the distribution of these conditions in different countries is extremely
heterogeneous. The data are based on Modell and Darlinson (2008), Pielet et al. (2010),
and Weatherall (2011), together with a variety of personal communications to the
author.
section 22  Haematological disorders
5370
receive transfusion die from iron loading before they reach the age
of 20 years.
There are other inherited anaemias which are particularly common
in tropical countries due to heterozygote advantage against malaria.
Glucose-​6-​phosphate dehydrogenase deficiency is estimated to
occur in some 100 million individuals worldwide. Its clinical and
haematological manifestations are discussed in Chapter  22.6.11.
They include haemolytic reactions to a wide variety of drugs, and,
of particular public health significance, to certain foods (favism).
There is a form of ovalocytosis, particularly common in Melanesia,
which is associated with a mild and well-​compensated haemolytic
anaemia. Recent studies have shown that carriers of Melanesian
ovalocytosis are protected against cerebral malaria.
Consequences of anaemia
The results of many studies directed at determining the functional
consequences of anaemia are still controversial. It is often difficult
to distinguish between the effects of anaemia per se and the conse-
quences of iron or folate deficiency on other physiological functions.
Whatever the mechanism, chronic anaemia is associated with di-
minished function.
Many studies have suggested that even mild anaemia may re-
duce near-​maximal work capacity. The WHO has recently stressed
the increasing evidence that iron deficiency in children may reduce
learning ability, intelligence, and, in extreme cases, may lead to intel-
lectual disability. There is no doubt that anaemia increases maternal
mortality and morbidity. There is a very large literature on the effect
of iron deficiency on resistance to infection, as mediated through
either immune function or the bacteriostatic and bactericidal roles
of iron-​containing proteins such as transferrin and lactoferrin. The
complex relationship between iron status and susceptibility of infec-
tion requires further work. It is clear that folate deficiency is asso-
ciated with an increased prevalence of obstetric complications and
fetal malformation, although its effect on intellectual and immune
function is less clear.
In short, because of the remarkable ability of otherwise healthy
individuals to adapt to moderate anaemia it seems likely that many
of the associated manifestations which have been observed result
from the effects of different deficiency states on other physiological
functions rather than the anaemia per se. On the other hand, chronic
severe anaemia, particularly in childhood, results in a wide variety
of complications including failure of growth and development and
possibly susceptibility to infection.
Prevention
Anaemia was responsible for roughly 8% of all nonfatal health loss
for all diseases in 2013, counting some 70 million YLDs. To put this
in perspective, it is roughly the same as the health loss caused by
depressive disorders (61.6 million YLDs) but is greater than dis-
ability due to asthma, diabetes, and cardiovascular disease com-
bined (61.3 million YLDs). However, a considerable proportion of
the burden of disease from anaemia has the potential for remedy.
It is beyond the scope of this brief review to discuss the pro-
tean aspects of the prevention of anaemia, particularly in poorer
countries. As outlined, the prevalence of anaemia is a reflection of
gross poverty, particularly as manifested by nutritional deficiency,
infection, and malabsorption. Its control requires action on many
different fronts, including improvements in diet, fortification of
commonly eaten foods with iron, the use of modified milk formulae
for infants, malaria and hookworm control, iron and folate supple-
mentation in pregnancy, and all-​round improvements in hygiene.
Good antenatal care helps to prevent anaemia in childhood by re-
ducing prematurity, increasing average birth weight, and improving
the nutritional status of the newborn.
Widespread and indiscriminate iron supplementation where mal-
aria and bacterial infections are common is potentially harmful and
certainly controversial. Current research is addressing when and
how to give iron safely. Progress in reducing the prevalence of an-
aemia has been substantial where comprehensive, intersectoral ap-
proaches to prevent anaemia have been implemented. For example,
in Bangladesh, anaemia prevalence fell by more than 40% between
1990 and 2013.
FURTHER READING
Beales PF (1997). Anaemia in malaria control: a practical approach.
Ann Trop Med Parasitol, 91, 713–​8.
DeMaeyer EM, Adiels-​Tegman M (1985). The prevalence of anemia in
the world. World Health Stat Q, 38, 302–​16.
DeMaeyer EM, et al. (1989). Preventing and controlling iron-​deficiency
anaemia through primary health care. World Health Organization,
Geneva.
Drakesmith H, Prentice AM (2012). Hepcidin and the iron-​infection
axis. Science, 338, 768–​72.
Gallacher PG, Ehrenkranz RA (1995). Nutritional anaemias in in-
fancy. Clin Perinatol, 22, 671–​92.
Global Burden of Disease Study (2015). Global, regional, and national
age–​sex specific all-​cause and cause-​specific mortality for 240 causes
of death, 1990–​2013: a systematic analysis for the Global Burden of
Disease Study 2013. Lancet, 385, 117–​71.
Hercberg S, Galan P (1992). Nutritional anaemias. Clin Haematol, 5,
143–​68.
Kassebaum NJ, et al. (2014). A systematic analysis of global anemia
burden from 1990 to 2010. Blood, 123, 615–​24.
Khusun H, et al. (1999). World Health Organization hemoglobin cut-​
off points for the detection of anemia are valid for an Indonesian
population. J Nutr, 129, 1669–​74.
Murray CF, et al. (2013). Global, regional, and national incidence and
mortality for HIV, tuberculosis, and malaria during 1990-​2013: a
systematic analysis for the Global Burden of Disease Study. Lancet,
384, 1005–​70.
Modell B, Darlison M (2008). Global epidemiology of haemoglobin
disorders and derived service indicators. Bull World Health Organ,
86, 480–​7.
Piel FB, et al. (2010). Global distribution of the sickle cell gene and
geographical confirmation of the malaria hypothesis. Nat Commun,
1, 104.
Weatherall DJ, Kwiatkowski D, Roberts D (2009). Hematologic
manifestations of systemic diseases in children of the developing
world. In: Orkin SH, et al. (eds) Nathan and Oski’s hematology of
infancy and childhood, 7th edition, pp. 1741–​68. WB Saunders,
Philadelphia.
Weatherall DJ (2011). The challenge of haemoglobinopathies in
resource-​poor countries. Br J Haematol, 154, 736–​44.

22.6.4 Iron metabolism and its disorders 5371 Timo
22.6.4 Iron metabolism and its disorders 5371 Timothy M. Cox and John B. Porter
22.6.4  Iron metabolism and its disorders
5371
World Health Organization (1968). Nutritional anaemias. Technical
Report Series No. 405. World Health Organization, Geneva.
World Health Organization (2002). The world health report 2002—​
reducing risks, promoting healthy life. World Health Organization,
Geneva.
World Health Organization (2015). The world malaria report. World
Health Organization, Geneva. http://​apps.who.int/​iris/​bitstream/​
10665/​200018/​1/​9789241565158_​eng.pdf
22.6.4  Iron metabolism and
its disorders
Timothy M. Cox and John B. Porter
ESSENTIALS
Iron deficiency and iron storage disease—​the latter principally due
to inherited and acquired anaemias such as thalassemia—​are dis-
orders of massive clinical significance across the globe. Iron de-
ficiency is the most common cause of anaemia, affecting about 1
billion people; about 0.75 million people have thalassaemia. Largely
neglected by health services in rich and resource-​poor countries
alike, disorders of iron metabolism, whether inherited, nutritional,
or otherwise, represent a long-​standing public health challenge.
Improved screening methods for detection, diagnosis, and appro-
priate supplementation—​as well as genetic counselling—​can offer a
great deal to relieve the burden in stricken communities. Advances
in chelation therapy have improved the survival of patients with
iron-​loading anaemias and transfusion-​related haemochromatosis,
and better understanding of the molecular pathophysiology of iron
homeostasis now offers the prospect of definitive therapies to con-
trol pathological erythropoiesis and the inappropriate drive to ac-
quire lethal quantities of toxic iron.
Iron homeostasis
Iron is an essential nutrient, the recommended daily allowance being
10 to 20 mg, depending on the bioavailability of food iron con-
stituents. Iron requirements are greater during periods of growth in
childhood and adolescence and in pregnancy, also in menstruating
women and during lactation. Patients with chronic haemorrhage and
intravascular haemolysis need sufficient iron to compensate for in-
creased losses.
Iron is a critical component of haem proteins and nonhaem en-
zyme systems required for oxygen transport, mitochondrial respir-
ation, and essential enzymatic reactions. High-​affinity iron-​binding
proteins have evolved to facilitate iron transport and delivery to
sites of storage and utilization, and especially for haem biosynthesis.
Most iron in the body is coordinated in protoporphyrin IX as haem
(ferroprotohaem). Small amounts of iron circulate in the plasma,
bound in the ferric form to the glycoprotein, transferrin. Iron is stored
in the mononuclear phagocyte system principally as intracellular fer-
ritin and its proteolytic degradation product, haemosiderin.
Iron absorption occurs in the duodenum and upper jejunum.
The processes are complex, but the following proteins are in-
volved: the divalent metal transporter protein (DMT1), ferrireductase,
hephaestin, ferroportin 1 and a haem transporter.
Capacity for safe storage of iron in intracellular deposits (mainly
in tissue macrophages) is limited, and the ability to dispose of ex-
cess iron in the body by excretion is negligible. Iron balance is
maintained by rigorous control of absorption from the diet, which
is principally orchestrated by the actions of hepcidin. The mech-
anisms by which hepcidin release is controlled by iron status and
in response to activity of the erythroid marrow are only now being
understood.
Iron deficiency
Causes—​iron-​poor diets alone are a rare cause of iron deficiency an-
aemia, except in growing children. It is critical to consider (1) en-
hanced loss of iron—​due to excess menstruation, pregnancy, and
losses from the gastrointestinal tract (hookworms, ulcerating lesions,
angiodysplasia); and (2) malabsorption of iron (e.g. after gastrointes-
tinal surgery and or coeliac disease). At least one rare inherited dis-
ease causes iron-​refractory iron deficiency anaemia (IRIDA).
Clinical features—​apart from symptoms of anaemia and pallor, the
most common are restless legs, angular cheilosis, atrophic glossitis,
hair loss, and dystrophy of the nails with longitudinal ridging and
koilonychia.
Investigation and diagnosis—​iron deficiency causes microcytic
anaemia, usually in association with reduced serum transferrin sat-
uration with iron (<16%), a raised total plasma transferrin, and a re-
duced serum ferritin concentration (<12 µg/​litre). In many cases,
the cause will be obvious (e.g. menorrhagia in a young woman),
but in others, diligent investigation will be required (e.g. to diag-
nose or exclude gastrointestinal cancer and other sources of occult
blood loss).
Management—​apart from dealing with the underlying cause, this
involves iron-​replacement therapy, which should normally be ad-
ministered orally, although parenteral preparations may (rarely) be
necessary.
Iron storage disease (haemochromatosis, iron overload)
Ferrous and ferric ions are chemically reactive so that excess of free
iron is toxic; tissues with elevated concentrations of the metal show
functional impairment and structural injury leading to ‘iron storage
disease’ (haemochromatosis).
Causes—​(1) genetic, hereditary, or primary haemochromatosis; or
(2) secondary haemochromatosis, including (a) diseases character-
ized by ‘ineffective’ or disordered erythropoiesis (e.g. β-​thalassaemia)
which induce inappropriate absorption of iron by the intestine;
(b) repeated blood or red cell transfusion; and (c) administration of
iron in quantities or formulations that overwhelm natural homeo-
static mechanisms.
Clinical features—​these depend on the rate and distribution of ex-
cess iron and include (1) heart disease—​cardiomyopathy is a leading
cause of death in refractory anaemias such as β-​thalassaemia;
(2)  fibrotic liver disease and cirrhosis, sometimes complicated by
hepatocellular cancer; (3) endocrine failure (e.g. hypogonadotropic
hypogonadism, diabetes mellitus, and mineralocorticoid failure);
(4) skin and joint manifestations; and (5) microbial infection.
section 22  Haematological disorders
5372
Investigation and diagnosis—​iron storage disease is usually sus-
pected on the basis of (1)  raised serum ferritin measurements,
and (2)  when the saturation of serum transferrin exceeds 60%.
Definitive diagnosis requires confirmation of the genetic mutation
(most commonly C282Y for the HFE gene in persons of European
descent) and of iron overload, either by noninvasive iron deter-
mination by liver or cardiac magnetic resonance imaging or by
liver biopsy with specific elemental analysis and/​or histochemical
staining for iron.
Management—​in the early stages, potentially fatal sequelae of
iron toxicity can be prevented by prompt institution of measures
to deplete iron:  (1) where erythropoiesis is normal—​by repeated
venesection; and (2) for patients with disordered (ineffective) haem-
atopoiesis or inadequate venous access—​iron chelators. Hitherto,
parenteral administration of desferrioxamine has provided the
best standard of care, but orally active chelators (deferiprone and
deferasirox) have better acceptability and efficacy.
Future developments—​striking advances in understanding the
control of iron balance by the hormone hepcidin, the molecular
pathogenesis of iron loading in dyserythropoietic anaemias, and
of the role of the transforming growth factor-​β superfamily of pro-
teins and hypoxia-inducible factors in the regulation of erythropoi-
esis, are driving innovative clinical research. These advances offer a
good chance of introducing transformative new treatments for life-​
shortening iron-​loading anaemias due to ineffective erythropoiesis.
Introduction
Disorders of iron metabolism frequently contribute to disease and
are of immense significance for global health. Iron deficiency is the
dominant cause of anaemia and affects more than 1 billion people;
about 0.75 million people have thalassaemia and about 0.33 mil-
lion have sickling disorders and other haemoglobinopathies (see
also Chapter 22.6.3). These conditions erode working capacity
as well as well-​being. Moreover, the iron-​loading anaemias and
sickling disorders are expensive to treat, in some regions affecting
the economies of whole societies.
Iron deficiency is rife: it affects women in rich and deprived
countries alike, but is most common among the poor, in in-
fants and the elderly, those who eat little or no meat, and those
with hookworm infestation. At the same time, the prevalence of
haemoglobinopathies and other anaemias such as myelodysplasia
that cause inappropriate absorption of iron and ineffective
erythropoiesis, with or without the need for red cell transfusions,
mean that iron storage disease is also a massive challenge in global
health. In addition, as discussed in Chapter  12.7.1, Hereditary
haemochromatosis occurs at a high gene frequency in certain
populations, including those of European descent (adult haemo-
chromatosis, due to mutations in HFE) and of sub-​Saharan
African origin (African iron overload), in whom the nature of
the predisposing gene is unknown. Other rare genetic forms of
haemochromatosis (‘juvenile’) in which endocrine failure and
cardiomyopathy due to iron toxicity occur in childhood or early
adult life are due to defects in the complex signalling pathways
that control iron homeostasis.
General aspects of iron metabolism and disease
Iron is the fourth most common element in the Earth’s crust; while
essential for aerobic life, its electrochemical properties also pose
challenges for living organisms. The metal assumes two readily inter-
convertible redox states (divalent ferrous iron, Fe2+, and trivalent
ferric iron, Fe3+) which are highly reactive. In the environment, iron
exists principally in the oxidized ferric state, which, under neutral
conditions, is then rapidly hydrolysed to insoluble polyhydroxide
complexes, which cannot be assimilated. High-​affinity iron-​binding
proteins, which form stable ferric complexes, have evolved to facili-
tate iron transport and delivery to sites of storage and utilization.
Iron is essential for human health
As a component of metalloenzymes and complexed to form haem,
iron participates in the transport of oxygen by haemoglobin and
myoglobin and harvesting metabolic energy obtained via the elec-
tron transport chain by the agency of cytochromes. In addition,
iron-​dependent enzymes are required for critical reactions (e.g.
xenobiotic metabolism) as well as the biosynthesis of dopamine,
catecholamines, and melanin. Iron deficiency, which affects infants,
children, young adults, and older people in all populations, is prob-
ably the most frequent organic illness worldwide; its high prevalence
is evidence of the critical availability of iron as a key nutrient.
Iron deficiency is associated with anaemia and nonhaematopoietic
manifestations that impair work efficiency and contribute to chronic
ill health, as well as loss of mucosal integrity; iron deficiency an-
aemia is associated with pica, in effect an obsessive craving to eat
or chew unusual materials which are often of no nutritional value.
Pica, which reflects mental changes induced by iron deficiency, has
important socioeconomic, environmental, and behavioural associ-
ations with poverty and hookworm infection. The causal role of iron
deficiency is confirmed by a rapid response to iron supplementation.
Iron and infection
The study of iron metabolism in animals and plants is a vast and
continually expanding field in biology. Since the need for and avail-
ability of iron for growth of pathogens is critical for host defence
as well as microbial virulence, deeper understanding of iron physi-
ology has become an imperative in pathogen research. Strong se-
lective pressures throughout evolution have resulted in adaptations
to cellular iron deficiency as well as mechanisms to sequestrate bio-
available iron—​so limiting microbial invasion and growth.
In contrast, not only are patients with iron overload susceptible to
certain bacterial and fungal infections, but indiscriminate iron sup-
plementation may have catastrophic effects and exacerbate infec-
tions. The protozoal infection, malaria, provides an example of this
effect, despite its frequent occurrence in populations where anaemia
is rife. Recent studies in which the ferroportin gene was deleted se-
lectively in murine erythroid cells showed accumulation of excess
intracellular iron, cellular damage, and haemolysis—​as well as a
poor outcome from experimental malaria. In humans, a prevalent
mutation in ferroportin (glutamate replacing histidine at position
248) impairs the normal turnover of ferroportin when iron is defi-
cient and increases abundance of the iron exporter. Not only is there
is evidence that inheritance of the variant protein protects against
severe malaria, in African populations this mutant allele of human
22.6.4  Iron metabolism and its disorders
5373
ferroportin is over-​represented—​presumably as a consequence of
evolutionary selection. Interrelations between iron and infection
are increasingly important in clinical medicine but space does not
permit further detailed discussion here.
Iron toxicity
‘Free’ iron that is not coordinated by an iron-​binding molecule
interchanges between ferric and ferrous oxidation states, thereby
promoting the formation of damaging free radicals; these mediate
the injury to cells and tissues that characterizes iron storage disease.
Hydroxyl radicals, the most damaging reactive oxygen species, are
generated by interactions between superoxide and ferric ions; they
catalyse the modification of cell membrane lipids—​a common fea-
ture of iron-​induced tissue injury. There are clear clinical parallels
between secondary haemochromatosis related to the iron-​loading
anaemias and the severe genetic forms of this disease, usually
termed juvenile haemochromatosis. After many years of detailed
study, these subjects are converging also with the expectation of
credible therapeutic development for the anaemias, which remain a
worldwide challenge in medical practice.
When iron accumulates excessively, the iron-​binding capacity of
plasma transferrin is often exceeded, so that a fraction of the iron
present in the blood occurs as low molecular weight species that are
only bound loosely to plasma proteins. This nontransferrin-​bound
iron in human plasma induces a distinct pattern of tissue iron up-
take compared with physiological transferrin-​mediated delivery: it
catalyses the peroxidation of unsaturated lipids and generates re-
active complexes which damage DNA.
Iron-​mediated injury to DNA
Iron-​mediated DNA modification is potentially mutagenic and, as
in the inherited copper toxicity disease, Wilson disease, is likely to
contribute to the known association between iron excess and cancer.
These disorders, characterized by excess hepatic deposition of multi-
valent transition metals, induce oxidative stress and increase the risk
of liver cancer. The frequency of p53 mutated alleles in noncancerous
tissue may be a biomarker of genomic injury and identify individ-
uals at increased cancer risk: a higher frequency of transversions at
codon 249 have been reported in liver samples from patients with
haemochromatosis compared with tissue from control subjects who
do not have excess iron. These findings suggest that the generation
of reactive free radical species from iron overload lead to mutations
in the p53 tumour suppressor gene that may contribute to the greatly
increased risk of hepatocellular cancer.
Body iron composition and evaluation
of iron status
Body iron composition
The total amount of iron in the adult body is between 3 and 4 g, most
of which is coordinated to protoporphyrin IX as haem (Fig. 22.6.4.1).
Haem is found principally as haemoglobin and myoglobin; appre-
ciable quantities are found also in the viscera, especially the liver,
kidney, and intestine, where it is present in cytochromes and other
iron proteins. Cytochromes of the electron transport chain and of
the P450 system for the metabolism of xenobiotics are abundant in
these organs and, notably in specific regions of the brain. In an adult,
about 2.5 g of iron is complexed in haemoglobin, with an additional
0.5 g of iron in myoglobin, principally in muscles.
Plasma iron
Very small amounts of iron circulate in the plasma: here it is bound
in the ferric form to the glycoprotein, transferrin, which is normally
only about one-​third saturated with iron, so that with a mean con-
centration of 3 g/​litre for a protein of molecular weight 80 000, the en-
tire plasma compartment contains less than 2 mg of elemental iron.
In health, the concentration of another iron protein, ferritin, in
the plasma does not exceed about 250 µg/​litre. Plasma or serum fer-
ritin does not itself contain appreciable iron; rather, serum ferritin
concentrations usually reflect the stores of iron in the body. Serum
ferritin concentrations below the healthy reference range nearly al-
ways indicate iron deficiency. In contrast, a high concentration of
serum ferritin may reflect high iron stores but also occurs in inflam-
matory states, liver disease (e.g. from hepatic steatosis), or certain
cancers (e.g. Hodgkin lymphoma).
Storage compartment
Iron is stored in the mononuclear phagocyte system (previously
known as the reticuloendothelial system) principally as intra-
cellular ferritin and its cognate proteolytic degradation product,
haemosiderin. In healthy men, body iron stores do not exceed 1.5 g
and are usually 0.5 g or less in healthy women. Deposits of nonhaem
iron that serve as stores in the iron-​rich tissues can be revealed by
staining with Perls’ reagent (acid potassium ferrocyanide), with
which they give a strong Prussian blue reaction. Faint staining
with Perls’ reagent is seen in healthy parenchymal liver cells; but
Plasma
4 mg
Erythrocytes
2500 mg
5 mg daily
20 mg daily
20 mg daily
RBC
destruction
Absorption
1–2 mg
daily
Monocyte-macrophage
system
Loss
1–2 mg
daily
Bone marrow
RBC
production
Body stores
1000 mg
Myoglobin
and respiratory
enzymes
300 mg
Fig. 22.6.4.1  Daily flux of iron through storage and transport
compartments.
section 22  Haematological disorders
5374
the principal deposits of storage iron are observed in bone marrow,
macrophages of the spleen, and in Kupffer cells, specialized macro-
phages of the liver.
Evaluation of iron status
The manifestations of iron deficiency and overload are discussed
in the following sections. Laboratory investigation of iron status
includes measurement of transferrin saturation and serum ferritin
(Box 22.6.4.1). In health, transferrin is about one-​third saturated
with iron. In acute (or chromic) inflammation, increased hepcidin
biosynthesis brought about by inflammatory mediators such as
interleukin (IL)-​6, decreases transferrin saturation. However, the total
plasma transferrin concentration is decreased or within the healthy
reference range. In contrast, in iron deficiency, low serum iron is
typically accompanied by increased total transferrin, and hence a
low transferrin saturation.
The presence of hypochromic microcytic red cells supports the
diagnosis of iron deficiency, although it is important to remember
that small red cells occur also in thalassaemia carriers. Low serum
ferritin concentrations are a useful confirmatory biomarker of iron
deficiency. An elevated concentration of serum ferritin implies in-
creased body iron but serum ferritin is an acute phase reactant and
is increased in acute or chronic inflammation, so that an elevated
serum ferritin is not specific. Serum ferritin is persistently raised
in patients with hyperferritinaemia-​cataract syndrome (see later);
although they do not develop systemic iron excess. If mistakenly
treated by phlebotomy, the serum ferritin (principally containing
light-​chain subunits) does not fall, and may even rise paradoxically.
Serum ferritin concentrations are also raised independently of
iron overload in malignant disease (including common cancers
as well as lymphomas including Hodgkin lymphoma), or released
from the liver in hepatitis—​including steatohepatitis. Again, it is im-
portant to emphasize that while low serum ferritin almost invariably
indicates iron deficiency, a raised serum ferritin does not always in-
dicate iron overload.
Additional diagnostic blood tests
It may sometimes be necessary to undertake additional tests to
confirm the presence of iron deficiency or iron overload. While
in most cases blood film examination is avoided, it is surprising
as to how informative this simple preparation and microscopy
can be—​particularly for detecting the host response to blood loss
(polychromasia with modest macrocytosis with a reticulocytosis)
and the presence of red cell fragments that indicate intravascular
haemolysis.
There are advocates for the measurement of free circulating trans-
ferrin receptors, which may be determined by immunoassay: ex-
pression of soluble transferrin receptor protein is enhanced under
conditions of iron deficiency and plasma concentrations are ele-
vated in the presence of functionally iron-​deficient erythropoiesis.
However, serum transferrin receptor concentrations are elevated
under conditions of erythroid hyperplasia in the bone marrow and
especially when ineffective erythropoiesis occurs (megaloblastic an-
aemia, haemoglobinopathies, sideroblastic anaemia).
Direct estimation of storage iron
Staining of iron stores in the bone marrow with Perls’ reagent is a
time-​honoured, robust, but mildly invasive method for resolving
difficulties that arise in the investigation of patients with suspected
iron deficiency anaemia. Although an examination of the amount
of iron (usually graded semi-​quantitatively on a scale from 0 to 4,
reflecting the strength of Prussian blue staining) does not provide
any information as to the availability of the iron for haemoglobin
formation, it does provide useful information as to the appropri-
ateness of iron therapy for hypochromic anaemia. Bone marrow
examination, moreover, may be diagnostic in patients suffering
from hypochromic anaemias due to primary or sideroblastic change
in the marrow, since the characteristic ring sideroblasts, with or
without other myeloblastic changes, will be apparent.
Liver iron concentration obtained from liver biopsies has been
used to quantify iron overload and evaluate responses to chela-
tion treatment but is invasive. More recently, determination of
liver iron concentration by magnetic resonance imaging (MRI)
is being used to confirm iron overload in the case of unexplained
hyperferritinaemia (raised serum ferritin concentration). This non-
invasive approach is slowly replacing liver biopsy for confirming
iron overload; it is also used to assess cardiac iron in transfusional
iron overload (see ‘Transfusional iron overload’). This analysis re-
quires a calibrated and externally validated method using T2* or R2
measurements but can be performed on most 1.5-​Tesla MRI ma-
chines. Low concentrations of hepatic iron ascertained by MRI in
the presence of hyperferritinaemia indicate liver disease or an in-
flammatory process—​thus meriting investigations for these condi-
tions rather than iron overload.
Human iron physiology
Erythropoiesis and recycling of iron
Iron is essential for the biosynthesis of haem and haemoglobin
formation during the maturation of red cell precursors. The circu-
lating iron-​binding plasma glycoprotein, transferrin is the principal
physiological mediator of iron delivery and transport about the body.
For assimilation into haem, iron is transported and taken up from
plasma transferrin by the recycling endocytosis of the iron-​bound
ligand via transferrin receptor type 1; the iron moiety is incorporated
Box 22.6.4.1  Laboratory investigations for iron
deficiency anaemia
•	 Serum ferritin concentration (low)
•	 Serum iron concentration (low)
•	 Transferrin saturation (low)
•	 Soluble transferrin receptora (elevated)
•	 Serum hepcidin concentrationb
Note: in addition to haemoglobin concentration, check red cell indices (mean
cell volume and mean cell haemoglobin content). A reticulocyte count and
film examination are desirable to judge the extent of compensatory erythro-
poiesis for haemorrhage (or haemolysis) and may support the diagnosis of
haemoglobinopathy, sideroblastic anaemia, or coeliac disease (dimorphic
picture with hyposplenic features).
a More reliable for diagnosis of iron deficiency than transferrin parameters but
not universally available.
b Highly specialized investigation: if inappropriately elevated, suggests an-
aemia of chronic disorders.
22.6.4  Iron metabolism and its disorders
5375
into protoporphyrin IX by mitochondrial ferrochelatase which util-
izes ferrous ions that are generated by chemical reduction within
the cell.
The lifespan of the red cell is up to 120 days and about 1% of
the steady-​state haemoglobin pool turns over daily. This requires
de novo synthesis of approximately 6 g of haemoglobin into which
20 mg of iron is incorporated in the haem moiety.
Most of the iron required for haemoglobin synthesis in the basal
state is recycled from senescent red cells after their destruction by
macrophages (Fig. 22.6.4.1). Iron is delivered to the erythron in the
plasma by transferrin that binds to transferrin receptors on eryth-
roid precursors The iron-​transferrin–​receptor complex is internal-
ized and, after acidification in endosomes, its iron is released. After
recycling to the cell surface, the apotransferrin, which has a low
affinity for the receptor at neutral pH, is released and can thus be
reutilized.
Increased delivery of iron occurs in association with erythroid ex-
pansion and disturbed maturation of red cell precursors that express
cell surface transferrin receptors under the influence of the renal
hormone, erythropoietin (see ‘Erythropoietin’). Erythropoiesis is
stimulated in the presence of hypoxia, after bleeding and haemolysis,
as well as in dyserythropoietic conditions (such as thalassaemia and
megaloblastic anaemia), and net assimilation of iron from the diet
is increased.
Storage of iron
Ferritin, the principal protein for the safe storage of iron, is widely
distributed in nature. As a large 24-​subunit multimer, ferritin com-
prises heavy and light polypeptide chains. Differences in the sub-
unit composition influence the rates of iron uptake and release in
different tissues. The main function of ferritin is the sequestration
of ferric iron in a soluble, accessible, but nontoxic form. Partial pro-
teolytic breakdown of iron-​loaded ferritin molecules leads to the
formation of haemosiderin, which exists as relatively insoluble ag-
gregates in storage cells of the macrophage series; both proteins can
release stored iron but the turnover of iron in haemosiderin is re-
tarded compared with that intracellular ferritin.
The heavy-​chain component of ferritin has antioxidative prop-
erties. Transcription (in contrast to iron-​regulated translational
expression) of the ferritin heavy-​chain gene, FTH1, located on
chromosome 11q12.3, is increased by the regulatory factor, NF-​κB.
This appears to inhibit tumour necrosis factor (TNF)-​α-​induced
apoptosis by suppressing iron-​mediated generation of reactive
oxygen species.
Iron homeostasis
Iron, an essential nutrient, is fastidiously conserved by the body and
only a fraction of that which is utilized in the bone marrow and other
tissues is lost by obligatory processes involving exfoliation of epi-
thelia and intercurrent blood loss, such as that incurred by trauma or
menstruation. Net iron balance is ultimately controlled at the level of
absorption of organic and inorganic iron from the diet by the small
intestine.
Assimilation of iron
Iron is absorbed from dietary sources principally in the duodenum
and proximal jejunum. The amount of iron available varies greatly,
and even under optimal circumstances only a small fraction is nor-
mally absorbed: in adult men, the daily requirement is on average
0.8 mg, whereas adult women of reproductive age need about 2 mg
daily. Depending on the bioavailability of the constituent iron, the
recommended daily allowance of iron in the diet is 10 to 20 mg.
Digestion of food iron
Inorganic ferric ions are released by digestion and reduced by lu-
minal factors such as ascorbate, as well as the action of ferrireductase
(DCYTB); Fe2+ ions are taken up via the divalent metal transporter
(DMT1) in the brush-​border membrane. Iron complexed to haem
enters enterocytes through a distinct uptake pathway. Within en-
terocytes, the membrane-​bound copper protein, hephaestin, is im-
plicated in reoxidation of ferrous to ferric ions; Fe3+efflux across
the basolateral membrane and delivered to unbound plasma
apotransferrin is mediated by the transporter, ferroportin (FPN1).
Haem iron may be more readily absorbed than inorganic iron in
the human intestine and, depending principally on the content in
meat, may constitute an important source of iron in nonvegetarians.
Dietary phytates and medication including antacids and tetracyc-
line antibiotics, as well as proton pump inhibitors, H2 antagon-
ists, and prior upper gastrointestinal surgery, strongly influence
the intraluminal bioavailability and hence absorption of inorganic
food iron.
Meeting increased needs for iron
The requirement for iron is greater in patients with recurrent
bleeding or in those who are blood donors; iron requirements are
also greater during periods of growth in childhood, adolescence,
and during pregnancy, when the daily requirement may be as much
as 5 mg.
Given its iron content, loss of 1 ml of blood represents approxi-
mately 0.5 mg of iron; this relationship simplifies estimates of iron
required to meet that incurred by blood losses, for example, by
menorrhagia (>80 ml/​month) or from other sources. Depending on
peripartum blood loss, each pregnancy represents a maternal invest-
ment of up to 1.5 g iron, which greatly exceeds savings due to the
cessation of menstruation.
Iron absorption and release from tissue stores
In health, iron absorption in the duodenum and upper jejunum is a
finely regulated process that matches the acquisition of iron from the
diet to body requirements for erythropoiesis; this compensates for
daily obligatory losses of iron (e.g. from the exfoliation of epithelia
or menstruation).
Iron uptake
Genetic studies of mutant strains of mice with abnormalities of iron
metabolism have shed light on the iron-​absorption mechanism. The
divalent metal transporter protein, DMT1, which is expressed in the
upper small intestine and cells of the erythron, is essential for uptake
of ferrous ions. The human DMT1 gene maps to the long arm of
chromosome 12 and encodes a 12-​membrane-​spanning protein that
is expressed in the apical membrane of the upper intestine. DMT1
is also produced in developing erythroid cells, in which it is respon-
sible for the intracellular delivery of iron derived from transferrin
after endocytosis for haemoglobin synthesis.
section 22  Haematological disorders
5376
Reduction of inorganic iron
A ferrireductase (cytochrome b reductase) is encoded by DCYTB;
the cognate protein reduces ferric to ferrous ions and is localized
to the intestinal brush-​border membrane and present in red cell
precursors.
Mucosal ferrireductase occurs at the apical microvillous mem-
brane of mammalian intestinal mucosa and this activity is increased
by nutritional iron deficiency. Dietary iron is found in haem proteins
as a component of haemoglobin, myoglobin, and tissue cytochromes.
Haem iron
Haem iron occurs principally but not exclusively in meat, which is
an important facultative source of iron in the diet of many humans.
Haem uptake by enterocytes requires a membrane protein, but the
identity of this putative carrier remains controversial. After uptake
as the intact molecule, the haem moiety is opened up by the action
of haem oxygenases in the intestinal mucosa to release iron (and
carbon monoxide).
Efflux of iron to the portal plasma
Ferric ions derived from the dietary sources are exported from en-
terocytes by the carrier molecule ferroportin, which is localized to
the basolateral membrane of intestinal epithelial cells. Ferroportin is
also responsible for the efflux of iron liberated from the iron stores in
the macrophage compartment; it is the principal site for the regula-
tion of iron flux and distribution and its activity is controlled by the
master regulatory hormone, hepcidin.
Heterozygous deficiency of ferroportin gives rise to a dominantly
inherited pathological storage of iron in macrophages and macro-
phage-​rich organs such as the liver and spleen (see Chapter 12.7.1).
Homozygosity for such mutations would probably be incompatible
with life.
Intracellular oxidation of ferrous iron in the epithelium
A putative copper-​binding protein, hephaestin, which maps to the X
chromosome and has sequence similarity to caeruloplasmin, appar-
ently mediates oxidation of ferrous iron and cooperates functionally
with ferroportin in promoting the transepithelial transport of iron in
enterocytes. Inherited defects of hephaestin in experimental mice lead
to iron deficiency anaemia; none have yet been identified in humans.
Iron release from tissue stores
Most storage iron is present in macrophages in the bone marrow,
spleen, and Kupffer cells of the liver. Ferric ions released by break-
down of ferritin and haemosiderin are exported by the action of
ferroportin and bind the plasma protein transferrin.
Adaptive responses to iron status
In iron deficiency, or even when iron stores are modestly depleted,
more of the bioavailable food iron is absorbed. Hypoxia also en-
hances the absorptive capacity of the small intestine and induces the
ferrous iron carrier, divalent metal transporter, DMT1, as well as the
accompanying ferrireductase, encoded by DCYTB.
Dysregulation of iron absorption in iron-​loading anaemias
For reasons that are only now being understood, certain anaemias,
particularly those associated with dyserythropoiesis and hence in-
effective erythropoiesis, also increase absorption of iron in the
intestine. Where the anaemia is long-​standing (e.g. congenital or ac-
quired sideroblastic anaemia), or in haemoglobinopathies such as
β-​thalassaemia, inappropriate intestinal absorption of iron accom-
panies massive expansion of the erythropoietic marrow; hepcidin
biosynthesis in the liver is suppressed and toxic iron overload re-
sults. This secondary haemochromatosis can occur in the absence of
iron loading that results from multiple red cell transfusions.
Oral iron toxicity
Maintenance of iron balance by the intestine normally protects the
body from the potentially toxic effects of iron-​rich diets. Only under
exceptional circumstances, such as the ingestion of alcoholic beverages
containing abundant iron due to peculiarities of manufacture (e.g. the
kaffir beers that are fermented in iron pots by the South African Bantu),
does excess dietary iron lead to iron storage disease. It seems probable
that those individuals who develop iron storage disease in the context
of long-​standing excessive ingestion of highly available iron, harbour
genetic cofactors such as mutant alleles of the adult haemochromatosis
gene, HFE, or because of an underlying haematological disorder, such
as α-​ or β-​thalassaemia trait or a cell-​intrinsic haemolytic anaemia.
There are numerous reports of haemochromatosis associated with red
cell disorders, including the congenital dyserythropoietic disorders
(see Chapters 22.6.7 and 22.6.8) and red cell enzyme defects, such
pyruvate kinase deficiency (see Chapter 22.6.10).
Physiological regulation of iron status
While iron is critical for health, its essential electrochemical prop-
erties also confer great danger in its use, handling, and storage.
Elaborate mechanisms exist to maintain the balance of supply (ac-
quired by the small intestine from dietary sources) and requirements
without incurring tissue injury. Body iron balance is regulated at the
level of acquisition from the intestine: in operational terms, net ab-
sorption of dietary iron is modulated by the iron status of the tis-
sues (‘storage regulator’) and controlled by activity of erythropoiesis
in the bone marrow (‘erythroid-​regulator’). These aspects are dis-
cussed in more detail in following sections.
Hepcidin—​an iron-​regulatory hormone
Hepcidin plays a critical role in the control of iron homeostasis.
This polypeptide (molecular mass c.2800) is a member of a family
of cysteine-​rich peptides with antimicrobial activities. Hepcidin
serves as an ‘iron regulatory hormone’, which inhibits efflux of iron
from macrophages and enterocytes. Hepcidin is present in serum
and urine but is generated in, and secreted by the liver. Hepcidin
biosynthesis and release by the liver is induced by iron loading and
suppressed by anaemia, hypoxia, and by inflammatory cytokines—​
including IL-​6, which also activates the hepcidin promoter through
signal transduction and transcriptional regulation.
Mode of action
A critical action of hepcidin is attributed to its capacity to bind to
ferroportin—​the transport protein which mediates export of ferric
ions from mammalian cells and which is expressed particularly on the
basolateral membrane of enterocytes and the plasma membrane of
macrophages. Binding of hepcidin to ferroportin on the plasma mem-
brane induces ferroportin internalization and degradation in the lyso-
some. Thus, hepcidin inhibits net absorption of dietary iron as well as
its release from the macrophage storage compartment (Fig. 22.6.4.2).
22.6.4  Iron metabolism and its disorders
5377
Dominant mutations in the SLC40A1 gene that encodes
ferroportin lead to an adult form of haemochromatosis (HFE4)
in males and females associated with prominent deposition of
iron in macrophages present in the marrow, liver, and spleen (see
Chapter  12.7.1). Recently, ferroportin was found to be highly
abundant in mature red cells and the capacity to export iron is
suppressed by iron supplementation. This may in part explain the
adverse effects of indiscriminate mass administration of iron in
populations with a high level of malaria exposure.
Congenital deficiency of hepcidin due to recessively transmitted
mutations in the HAMP gene that maps to chromosome 19q23 is
also associated with greatly enhanced absorption of iron and ju-
venile haemochromatosis (HFE2B—​see Chapter  12.7.1). In con-
trast, excess synthesis and secretion of hepcidin by the liver inhibits
the release of iron from the storage compartment and also decreases
net absorption of iron from the diet by the small intestine. The main
action of hepcidin has been verified in mice, and humans, in whom
there is a hereditary deficiency of the peptide: intestinal absorption
of iron proceeds in an unrestrained manner and thus iron circu-
lates that is unbound to transferrin. This reactive pool of plasma
iron is responsible for unregulated distribution of the iron and is
associated with rapid parenchymal injury in many tissues (human
hepcidin deficiency is a subtype of juvenile haemochromatosis (see
Chapter 12.7.1)).
Biosynthesis of hepcidin is transcriptionally regulated in the liver in
response to the concentrations of extracellular and intracellular iron.
This is achieved by formation of a macromolecular complex of bone
morphogenetic protein receptors and their iron-​specific ligands as well
as modulators and molecules that serve as iron sensors (Fig. 22.6.4.3).
Overall, molecular regulation of hepcidin appears to be orches-
trated through the bone morphogenetic protein/​smooth mothers
against decapentaplegic homologue (BMP/​SMAD) signalling
pathway (see later). BMP6 expression is upregulated in response to
iron and induces hepcidin through phosphorylation of the SMAD1/​
5/​8 complex (see Chapter 12.7.1).
Haemojuvelin (HJV), the hereditary haemochromatosis type 2A
associated protein, and the transferrin receptor-​2 isoform, which is
mutated in another rare adult form of haemochromatosis (HFE3),
also interact with this multimolecular complex on the plasma mem-
brane. While the details of the molecular interactions of HJV with
its partners are yet to be fully worked out, as with hepcidin encoded
by HAMP1, mutations in the HJV gene on chromosome 1q21 cause
a severe, juvenile form of iron overload—​indicating its critical role
in regulating net iron balance. HJV interacts with hepcidin and
Dietary
iron
Hepcidin
Hepcidin
Plasma transferrin
Fe Tfn
Ferroportin
Fe3+
Haem
Fe3+
Fe2+
Spleen
Macrophage
storage
compartment
Upper
small
intestine
Marrow
macrophages
Reduction
(DCyt B)
Uptake
(DMT1)
Fe Tfn
Ferroportin
Ferroportin
Fe Tfn
Inflammation
IL6 etc
Erythropoiesis
Transferrin
saturation
Liver
Kupffer
cells
+
+
–
Hepcidin
HAMP1 gene
transcription
Fig. 22.6.4.2  Hepcidin regulates iron balance by controlling ferroportin expression. Hepcidin, HAMP, is a regulatory polypeptide hormone
synthesized and released by hepatocytes according to systemic iron status and erythropoietic activity; hepcidin release is also increased in the presence
of inflammation. Hepcidin reduces the availability and assimilation of iron for erythropoiesis and storage by regulation of iron acquisition from the
diet and release of iron from the storage and recycling compartment. Hepcidin release is controlled negatively by the so-​called storage and erythroid
regulators. Inflammatory signals, mediated by interleukin-​6 and Janus kinase/​signal transducers and activators of transcription (JAK-​STAT) pathway,
enhance release. Hepcidin binds to ferroportin on plasma membranes and promotes internalization and degradation of this transporter: this reduces
iron efflux from storage macrophages and across the basolateral membrane of enterocytes in the small intestinal mucosa.
section 22  Haematological disorders
5378
Liver
–
Receptor
Spleen and
other storage
macrophages
Erythroferrone and
other red cell factors
EPO
Kidney
Lungs
Erythron
Red cell mass
Upper small
intestine
mucosa
Ferroportin
Ferroportin
Fe3+
Tfn
Fe3+
Tfn
O2
Hepcidin
Hepcidin
Pa O2
JAK2 - STAT
HIFS
Heart
Dietary
iron
+
+
(a)
Fig. 22.6.4.3  (a) Homeostatic control and the interplay of erythropoiesis and iron supply. Erythroid drive is the dominant influence on iron
homeostasis. This comprises several physiological pathways that ensure a balanced supply of iron for haemoglobin formation sufficient to meet
tissue requirements for oxygenation. Ineffective erythropoiesis profoundly disrupts this tightly regulated control system and leads to inappropriate
iron loading with tissue injury (secondary haemochromatosis). Tissue hypoxia inhibits the action of Fe2+-​dependent prolyl hydroxylases which
normally initiate proteasomal destruction of hypoxia-​inducible factor (HIF)-​1α and HIF2α by rendering them susceptible to ubiquitination by the von
Hippel–​Lindau complex (see text). When tissue Po2 is lowered, as in anaemia or high altitude, HIF-​2α induces erythropoietin (EPO) transcription. On
binding to its receptor (EPOR), in colony-​forming units EPO activates the JAK2/​STAT5 pathway for erythroid gene expression in the marrow. As the red
cell population expands, erythroferrone and other soluble red cell factors are released: erythroferrone attenuates hepcidin synthesis and coordinates
the iron supplied by plasma transferrin with the demands for haemoglobin formation.
Ferroportin
Hormone
release
Ferroportin
Iron storage
macrophage
Fe
Fe
DMT1
Ferrireductase
Fe3+
Iron
transferrin
TfR1
BMP receptor
SMADS
Iron
transferrin
TfR2
BMP6
HFE
Fe3+
Hepcidin
Hepcidin
Proximal intestine
Intestinal mucosal epithelium
Fe3+ transferrin
Fe3+ transferrin
Hepcidin
mRNA
Hepatocyte
Tf
Tf
Haemojuvelin
+
+
+
HAMP1 gene
+
(c)
Fig. 22.6.4.3  (b) Hypoxia and iron deficiency increase transcription of TMPRSS6, encoding the transmembrane protease 6 (matriptase-2), which
cleaves membrane-bound haemojuvelin (HJV), a bone morphogenetic protein (BMP)-6 coreceptor essential for suppressing hepcidin when iron is in
excess; matriptase-2 attenuates the regulation of hepcidin synthesis by erythroferrone.
Fig. 22.6.4.3  (c) Molecular control of hepcidin by systemic iron-sensing negative transcriptional control of HAMP1 encoding hepcidin by iron
status is mediated by the canonical bone morphogenetic protein (BMP) receptor kinase pathway. This involves principally BMP6, BMP2, and smooth
mothers against decapentaplegic (SMAD) protein signalling (SMAD4). Cell-surface BMP ligand interactions are modulated by iron transferrin and
multiprotein receptor interactions in hepatocytes which include transferrin receptors 1 and 2 (TFR1 and TFR2), the HFE1 protein, and haemojuvelin,
HJV. Mutations in these proteins and ferroportin (FPN1) lead to distinct forms of hereditary haemochromatosis; TMPRSS6 mutations impair BMP/
SMAD signalling and cause iron-refractory iron deficiency anaemia (IRIDA) due to excessive hepcidin action.
Fe3+
Fe3+
Fe3+
Fe3+
TFR2
Iron transferrin
TFR1
BMP6
BMPR
HJV
TMPRSS1
HFE
Soluble haemojuvelin
Ferroportin
Ferroportin
Hormone
release
Iron storage
macrophage
Proximal intestine
Intestinal mucosal epithelium
ApoTf
Fe3+ transferrin
Fe3+ transferrin
Fe3+
DMTI
Ferrireductase
Food iron
Fe3+
Fe3+
ApoTf
Hepcidin
Hepcidin
DCytB
Fe2+
Fe3+
Hepcidin
mRNA
HAMP1 gene
SMAD4
SMAD4
SMADS Pi
P
Tf
Tf
Liver parenchymal
cell
SMADS
(b)
section 22  Haematological disorders
5380
thus serves as a coreceptor for BMP signalling in the regulation of
hepcidin expression.
Activin receptors
Hepcidin expression is influenced by BMPs and requires activa-
tion of serine/​threonine kinases present on the surface of liver
parenchymal cells. This control pathway responds to BMP signal-
ling that involves activin through the participation of the two type
I activin receptor serine kinases and two type II receptor kinases—​
respectively named ALK2 and ALK3 and the ActRIIA and BMPR2
receptor pairs.
Downstream signalling events of the BMP pathway that regu-
lates hepcidin expression ultimately induces phosphorylation of
intracellular transducing molecules, termed receptor-​interacting or
R-​SMADs—​a complicated hybrid term derived from the nomen-
clature of conserved homeobox genes that regulate development
in Drosophila and vertebrates. Cell-​surface binding of the activin-​
related receptor serine kinases leads to phosphorylation of SMAD4
which is a nuclear transcription factor. SMAD4 forms heteromeric
molecular complexes which influence proliferative and/​or differ-
entiation functions and have been implicated in the regulation of
hepcidin.
Erythropoietin
Erythropoietin is a major determinant of the rate and amount of iron
used in the body (see Chapter 22.3.5). It is a secreted glycoprotein
that coordinates red cell formation with the availability of oxygen to
the tissues. A negatively regulated feedback system ensures there is
sufficient oxygen transported by haemoglobin to meet the demands
of aerobic metabolism. Erythropoietin is, in effect, a hormone that
controls the production of red cells in the bone marrow in response
to changes in tissue oxygenation reflected by the Pao2: it thus affects
the demand for iron by the erythron as well as the flux of iron bound
to transferrin in and out of the plasma compartment.
Understanding the regulation and mode of action of erythropoi-
etin is essential to understanding the pathophysiology of secondary
haemochromatosis, which is the principal cause of death in patients
with the iron-​loading anaemias.
The stimulus for erythropoietin secretion is reduced capillary par-
tial pressure of oxygen in the juxtamedullary renal cortex, which op-
erationally reflects the oxygenation status of the entire body. When
tissue oxygen tension (Po2) decreases, release of erythropoietin
from peritubular interstitial fibroblasts in the adult kidney (or, in the
fetus, by hepatic perisinusoidal cells) is increased.
Erythropoietin stimulates the differentiation and maturation
of red cell precursors in response to hypoxia (and blood loss); the
hormone also stimulates expansion of the erythroid progenitor
population in the marrow. The effects are coordinated at the level
of transcription by the hypoxia-​inducible factor-​2 (HIF-​2α), which
induces expression of the erythropoietin gene in the adult kidney
(and fetal liver) and is itself regulated by oxygen-​sensitive iron
enzymes.
Erythropoietin represses the default programme of cell death
(apoptosis) in erythrocyte progenitors and serves as a driving
erythropoietic stimulus which, in cooperation with other growth
factors such as IL-​3, IL-​6, and stem cell factor, ensures erythroid-​
cell lineage development from stem cells. Under the influence of
erythropoietin, erythroid burst-​forming units (BFU-​Es) and eryth-
roid colony-​forming units (CFU-​Es), proliferate. These rapidly dif-
ferentiate to form proerythroblasts, basophilic erythroblasts, and
normoblasts; these cells all express erythropoietin receptors until
the reticulocyte stage is reached, since reticulocytes lack nuclei and
hence all transcriptional activity.
Control of erythropoietin expression
Erythropoietin expression is driven by binding of the oxygen-​
labile HIF-​α subunits, HIF-​1α, HIF-​2α, and HIF-​3α, to hypoxia-​
responsive elements in the erythropoietin gene (see Chapter 22.3.5).
There is evidence that HIF-​2α is the principal factor controlling
physiological erythropoietin expression as well as the accompanying
adaptations in expression of DMT1 and ferrireductase (DCYTB) in
the small intestine—​which support enhanced iron absorption in
response to hypoxia and iron deficiency.
The preferential use and specificity of HIF-​2α for regulation of
the principal haematological responses to hypoxia are impaired
when the von Hippel–​Lindau (VHL) system and other regulatory
pathways are activated by disease-​related mutations. In the pres-
ence of iron, oxygen-​dependent degradation domains (containing
specific proline residues modified by at least three Fe2+-​dependent
prolyl-​4-​hydroxylases that require oxoglutarate) decrease the
abundance of the HIF-​α subunit. Prolyl hydroxylation of the HIF
proteins promotes their binding by the VHL protein complex,
which serves as an ubiquitin ligase and promotes their degradation
by proteasomes.
An additional level of control by molecular oxygen is exerted
on HIF-​α: its transcriptional activity is modulated by hydroxyl-
ation of a specific asparagine residue in the C-​terminal transactiva-
tion domain. This factor inhibiting HIF (FIH) is also an iron-​ and
2-​oxoglutarate-​dependent dioxygenase; its action interferes with
recruitment of promiscuous coactivators (CREB-​binding protein
and p300) to the HIF transcriptional complex. Under conditions
of hypoxia, reciprocal inactivation of FIH enhances recruitment
of CREB-​binding protein and so induces expression of HIF-​2α
target genes.
Expression of the prolyl and asparaginyl hydroxylases is itself con-
trolled by the HIFs, so that high abundance of HIF-​α under stress
is self-​limiting and partially declines during sustained periods of
hypoxia.
These oxygen-​sensitive enzymes are Fe2+-​dependent and 2-​
oxoglutarate-​dependent oxygenases and they incorporate atomic
oxygen into the target protein. Under hypoxic conditions, the hy-
droxylation is suppressed, and HIF-​α subunits serve only as a basally
active transcriptional complex. There is evidence in human cells
that HIF-​2α is principally responsible for the physiological con-
trol of erythropoietin expression—​the affinity of the prolyl system
for oxygen has been estimated in the healthy arterial Po2 range
(c.150 mmHg) but the reciprocal asparaginyl system has a greater
oxygen affinity (Po2 c.60 mm).
Mode of action of erythropoietin
Erythropoietin binds to its homodimeric receptor which is abun-
dant on the surface of erythroid progenitors, thereby inducing
conformational changes that activate the associated cytoplasmic
Janus-​activated kinase 2 (JAK2) by autophosphorylation of target
22.6.4  Iron metabolism and its disorders
5381
tyrosine residues. Phosphorylation leads to the engagement of adap-
tors and effectors, including the signal transducers and activators of
transcription factor-​5 (STAT-​5). STAT-​5 activates diverse molecular
targets including the extracellular signal regulated kinases: Jun N-​
terminal kinases, p38 mitogen-​activated protein kinase (MAPK),
and phosphoinositide-​3 kinase p38. Activation of phosphoinositide-​
3 kinase/​protein kinase B pathway via JAK2 also leads to phosphor-
ylation of the key transcription factor, GATA-​1, that is critical for
programming erythroid differentiation.
Effectors of erythropoietin action
STAT5 signalling in the nucleus drives transcriptional expression of
several genes, including those encoding:
•	the antiapoptotic mitochondrial protein, B-​cell lymphoma-​extra
large, which promotes erythroid cell survival by binding free
cytochrome C
•	transferrin receptor 1, which mediates uptake of protein-​bound
iron in plasma
•	the iron-​responsive element 2 binding protein, that binds to a
stem-​loop sequence in the 5ʹ-​untranslated regions of L-​chain fer-
ritin (inhibiting translation) and to the 3ʹ-​untranslated region of
transferrin receptor mRNA (to prevent its degradation).
Binding of erythropoietin to its receptor also suppresses the action
of death receptors such as the Fas ligand receptor, TNFα, and apop-
tosis that is mediated by the TNF-​related apoptosis-​inducing ligand
(TRAIL).
Homeostatic control systems in iron metabolism
The fastidious relationship between iron supplied from the diet,
and that required for erythropoiesis, haem protein biosynthesis,
and compensation for insensible losses, depends on the recycling
of haemoglobin iron and iron balance physiologically controlled
by the regulatory hormone, hepcidin. In health, the magnitude of
the tissue iron stores and rate of red cell formation affect net iron
absorption and hepcidin activity is influenced by two principal dy-
namic factors:
1.	 An entity that regulates the pool of stored and circulating iron
(the so-​called ‘storage’ regulator named by Clement Finch,
and representing a hormonal feedback loop). Fundamentally
the storage regulator allows the requirements for growth and
physiological iron losses to be met. This control system has a
limited capacity to influence net acquisition of iron (maximum
about 2 mg daily).
2.	 A further conceptual entity, the ‘erythroid regulator’—​a high-​
capacity control system which serves to support the drive for red
cell production in the face of limited supplies of iron irrespective
of body iron balance. The erythron accounts for about 80% of
plasma iron flux: in iron deficiency related to haemorrhage, daily
absorption of dietary iron may rise from less than 1 to 30 to 40
mg but even in iron-​overloaded patients with ß-​thalassaemia,
daily acquisition of 3 to 4 mg iron is often long sustained.
Under basal and physiological conditions, both the iron storage
and the erythroid regulators influence the expression, secretion,
and action of the principal iron hormone, hepcidin. This ensures an
adequate supply of iron, while avoiding its gratuitous accumulation
of this reactive and potentially toxic metal.
Storage regulator
The ‘iron-​storage regulator’ reflects the overall stimulation hepcidin
secretion by hepatocytes in response to adequate concentrations
of intracellular iron. A  plausible candidate has been postulated
to be BMP6, but other members of this protein family may be in-
volved. In effect, hepatic expression of hepcidin correlates with
nonparenchymal iron stores and iron saturation of transferrin.
A feedback loop that incorporates the control of ferroportin func-
tion mediated by hepcidin would ensure the retention of intracel-
lular iron in storage macrophages.
Erythroid regulator
The concept of the more elusive ‘erythroid regulator’ (or regulators)
reflects the indirect signal that downregulates hepcidin secretion by
the liver: this would serve to meet the requirements for iron during
erythroid expansion in the bone marrow driven by erythropoietin.
When additional iron bound to transferrin for haemoglobin syn-
thesis in the marrow is required, immature erythroid cells are postu-
lated to use a signal to downregulate hepcidin secretion by the liver,
so enhancing entry of iron to the plasma from macrophage storage
compartment with supplementation of new incoming iron assimi-
lated from the diet in the small intestine.
Several molecular candidates for the erythroid regulator have
been proposed, including (1) cytokine growth differentiation factor
15 (GDF15), a member and modulator of the TGFβ superfamily;
(2) twisted gastrulation (TWSG1), which suppresses hepcidin ex-
pression and in human cells has been reported to inhibit secretion
of hepcidin indirectly; but perhaps the most plausible candidate is
(3) erythroferrone (ERFE, also termed FAM132B), a member of the
TNFα family which is released from the spleen and bone marrow
of mice within a few hours after experimental removal of blood
and immediately precedes the physiological decrease of circulating
hepcidin. Administration of erythropoietin induces release of this
putative hormonal factor only from bone marrow and spleen.
The action of erythroferrone in murine erythroblasts is influ-
enced by the systemic erythropoietin and its cognate receptor-​
mediated signalling through the Jak2/​Stat5 pathway. Mice
engineered genetically to lack the Fam132b protein, while ap-
parently healthy in the steady state, have defects in their red cell
maturation. After phlebotomy or administration of erythropoi-
etin, Fam132b−/​− mice have exaggerated hepcidin expression, so
that compensatory acquisition of iron in these conditions of ‘stress
erythropoiesis’ is severely impaired. These characteristics strongly
support the putative physiological role for this protein as the
‘erythroid regulator’ in mammalian iron homeostasis. If this is, as
expected, the case, the protein represents a factor of major biomed-
ical significance, especially for therapeutic exploitation.
Other factors
Erythroblasts also sense the availability of iron by circulating iron
bound to plasma transferrin via the second transferrin receptor
(TFR2) that is mutated in rare forms of haemochromatosis (HFE3)
(Chapter 12.7.1). This iron signal appears to attenuate the sensi-
tivity of the erythroid precursors to erythropoietin. By inhibiting
section 22  Haematological disorders
5382
transcription of hepcidin in the liver, erythroferrone enhances the
activity of ferroportin and so promotes release of iron from the
macrophage storage compartment and net absorption of iron in the
small intestine. Erythroferrone mediates some actions of erythro-
poietin, which allows the body to compensate for blood loss as well
as failure of adequate oxygen delivery to the peripheral tissues in-
duced by hypoxia.
This adaptive response is a physiological adaptation to iron defi-
ciency, after blood loss and under hypoxic conditions; red cell for-
mation in the bone marrow influences iron homeostasis. The release
of erythroferrone from erythroblasts would reduce hepcidin release
by the liver and so increase acquisition of iron for erythropoiesis.
The nature of the erythroid regulator of iron homeostasis has been
unclear but the properties of erythroferrone render it an attractive
candidate to explain the pathological and inappropriate increase
in the acquisition of iron in conditions such as ß-​thalassaemia and
other iron-​loading anaemias. The common factor is ineffective
erythropoiesis and erythroid hyperplasia driven by the action of
erythropoietin.
In operational terms, ineffective erythropoiesis disrupts the
regulatory mechanisms that maintain systemic iron balance:  the
pathological marrow due to erythroid expansion driven in part
by erythropoietin and tissue hypoxia overcomes the ‘storage regu-
lator’ and causes unregulated net absorption and delivery of iron in
the face of adequate stores. Hepcidin is inappropriately and mark-
edly suppressed if there is no prior burden of iron from red cell
transfusions in dyserythropoietic anaemias such as ß-​thalassaemia
intermedia, absorption of iron can enhanced up to 10-​fold, which
is grossly in excess of requirements for erythropoiesis; iron toxicity
and secondary haemochromatosis are the result. Finally, in mouse
models at least, increasing concentrations of hepcidin or transferrin
may alleviate anaemia and dyserythropoiesis and restrict iron up-
take by erythroblasts.
Iron deficiency
General aspects
Iron deficiency is a major challenge for global health and the most
common cause of anaemia worldwide (see also Chapter 22.6.3).
The deficiency reflects requirements for iron that exceed that
obtained from the diet (often in a diet that is otherwise adequate),
the effects of chronic blood loss, and malabsorption. Contributory
causes include the loss of iron by excess exfoliation of cells, as
in inflammatory diseases of the gastrointestinal tract, and (more
rarely) urinary iron loss due to intravascular haemolysis associ-
ated with haemoglobinuria. The availability of iron modulates
erythropoiesis: iron restriction reduces the synthesis of haem and
α-​globin chains in ß-​thalassaemia.
About 30% of the world’s population (estimated by the World
Health Organization to be 7.6 billion at the end of 2017)—​more
than 2.25 billion people—​are anaemic and about 50% will have
iron deficiency anaemia. Even in rich countries, such as the
United States of America and those in Europe, up to 20% of
menstruating women have signs of iron deficiency. In children
and young adults, there is a frequency of between 5 and 10% of
iron deficiency anaemia, particularly in deprived socioeconomic
groups.
Iron deficiency anaemia impairs well-​being, cognitive and phys-
ical performance, as well as working capacity, and to a considerable
extent these factors contribute to poverty. Administration of iron
during pregnancy improves maternal, neonatal, infant, and even
long-​term outcomes in children. Moreover, there is evidence sug-
gesting that in children, judicious iron supplementation may be able
to improve cognitive, psychomotor, and physical development.
Many epidemiological studies have been based on the erroneous
attribution of microcytic anaemia to iron deficiency. Diverse condi-
tions, including the anaemia of chronic disorders and genetic dis-
eases with haemoglobinopathies such as β thalassaemia trait, are
characterized by hypochromic or microcytic red cell indices, and
iron deficiency is but one cause. Population surveys based on the
detection of iron-​deficient erythropoiesis, especially those using de-
termination of free red cell zinc protoporphyrin concentrations by
fluorimetry, enhance the detection of true iron deficiency anaemia;
determinations of serum ferritin concentrations can also help to
discriminate between the anaemia of chronic disease and true iron
deficiency.
Causes of iron deficiency
Nutritional factors
Iron deficiency anaemia is often attributed to an iron-​poor diet,
but in the absence of significant blood loss or intestinal parasites
including hookworm, even the most iron-​poor diets rarely cause
iron deficiency anaemia, except in growing children. The amount
of iron required to repair obligatory losses is very small, so that at
least 90% of the iron required for de novo haemoglobin formation
in erythropoiesis is retrieved from senescent erythrocytes broken
down by the mononuclear phagocyte system.
Furthermore, once iron deficiency develops, striking adaptive
changes occur in the absorptive mechanism in the upper small in-
testine so that assimilation of bioavailable iron in the diet is greatly
enhanced. Iron deficiency, and the response to the removal of a
unit of blood, may increase the overall absorptive efficiency of the
intestine for iron up to 10-​fold—​thus greatly enhancing the in-
corporation of dietary iron. Iron deficiency is associated with the
recruitment of a greater length of mucosal surface for participation
in the absorption of luminal iron in the upper small intestine. In
experimental animals with iron deficiency, mucosal expression of
DMT1 on the brush-​border membrane of the intestinal epithelium
is induced. Iron deficiency is also associated with enhanced intes-
tinal expression of mucosal ferrireductase activity, thus increasing
absorptive capacity for inorganic ferric iron released by digestion of
food. The extent to which the absorptive pathway for organic iron
complexed as haem is modified by iron deficiency and/​or anaemia
is less well studied.
Composition of foods
Alcoholic beverages may be a significant source of iron, and the
absorption of haem iron present in red meat, poultry, and fish is
usually between 15 and 35%. Between 2 and 20% of nonhaem iron
present in fruit and vegetable sources is absorbed. Natural enhan-
cers of iron absorption such as ascorbic acid, which maintains
22.6.4  Iron metabolism and its disorders
5383
ferrous iron in its reduced form in the intestinal lumen, promote
direct uptake by DMT1. Fructose and other organic compounds
of low molecular weight also form soluble ferrous complexes after
release from nonhaem sources in food. Healthy individuals in rich
countries, ingest between about 10 and 15 mg of iron in the so-​
called Western diet, daily. Adult men with normal iron stores ab-
sorb approximately 2% of the nonhaem iron ingested, whereas men
with iron deficiency can absorb more than 20% of iron from this
source in the diet; the comparable figures for haem iron are 26 and
47%, respectively.
Components that interfere with the availability of food iron
Many compounds present in the diet  also inhibit or impede the
absorption of iron released by digestion in the lumen. These com-
pounds include tannin, especially present in tea; phytates, present in
bran and nuts; dietary fibre; and other inhibitory factors including
drugs such as tetracycline antimicrobials, proton pump inhibitors,
and alkalis. Some vegetarians of Asian origin ingest large amounts
of phosphate and phytates, which inhibit the absorption of iron pre-
sent in diets that may contain up to 30 mg of assayable total iron
each day. Another example is spinach—​although very rich in iron,
much of this is not readily available. When ingested, spinach induces
formation of odiferous black stools: the passage of unabsorbed iron
through the small intestine and its delivery to the colon leads to the
formation of insoluble ferrous sulphide complexes due to the pres-
ence of sulphur-​reducing bacteria.
Loss of iron
Menstrual losses
Women in the reproductive age group lose iron regularly as a re-
sult of menstrual bleeding. The recommended daily allowance for
women is higher than in all other groups: this must supply sufficient
bioavailable iron to meet the increased needs. The average require-
ment for healthy menstruating women is approximately 1.4 mg of
iron daily to replace losses, compared with men, who lose about 0.8
to 0.9 mg of iron per day.
Menorrhagia is a frequent, and a frequently invoked cause of iron
deficiency due to excess losses of iron compared with iron intake in
women. In the United Kingdom, 5% of women aged 30 to 49 years
consult their general practitioners each year with menorrhagia and
similar figures apply to other developed countries. Established men-
orrhagia (defined as the loss of more than 80 ml of blood each normal
cycle) causes anaemia in two-​thirds of women. Benign leiomyomas
(‘fibroids’) occur in about 10% of women with menorrhagia overall,
but in 40% of those with severe menorrhagia (>200 ml per cycle).
Iron deficiency anaemia occurs in two-​thirds of women with proven
menorrhagia.
Quantitative determination of menstrual loss is often imprac-
tical in everyday clinical work; to resolve this matter in practice,
Duckitt has suggested that menorrhagia may be defined oper-
ationally as a complaint of regular excessive menstrual blood
loss that interferes with the physical, emotional, social, and ma-
terial quality of life. Menorrhagia may occur alone or with other
symptoms. It is important to be cautious if other potential causes
of iron deficiency anaemia are to be sought since about half of
women having a hysterectomy for menorrhagia are found to have
an apparently normal uterus; however, this does not necessarily
imply that the procedure was not warranted. A prevailing view
is that ‘idiopathic ovulatory menorrhagia’ reflects disordered
endometrial synthesis of prostaglandins—​this may account for
the beneficial effects of nonsteroidal anti-​inflammatory drugs in
the management of menorrhagia (with or without symptoms of
dysmenorrhoea).
Although the rates of hysterectomy are decreasing, about 20% of
British women, and 35% of those in North America have under-
gone the procedure before the age of 60 years; in at least half of these
women, menorrhagia is the principal complaint.
Pregnancy
Pregnancy is associated with iron deficiency, particularly during the
mid and last trimester, when growth of the fetus is rapid. Twin preg-
nancies and frequent childbirth, especially in women of low socio-
economic groups, often cause iron deficiency anaemia. Although
anaemia is an important influence on maternal health and resili-
ence, a large study has reported no reliable association between
maternal anaemia and the complications of pregnancy, including
preterm labour.
Pregnancy itself is associated with the development of adap-
tive responses in the intestine and iron transport proteins that
enhance the avidity of the gastrointestinal tract for bioavailable
food iron. Clearly, socioeconomic and sociopolitical consider-
ations are likely to influence the population occurrence of iron
deficiency in women of the reproductive age group, particularly
since the investment of about 1 to 1.5 g of iron occurs with each
pregnancy carried to term. This estimate includes blood loss asso-
ciated with the birth and the investment of iron placed in human
milk, which contains up to 0.5 mg/​litre of iron bound to a whey
protein, lactoferrin.
Intestinal parasites—​hookworm
More than 400 million people are estimated to suffer from hook-
worm infestation. The helminth causes ill-​health and misery associ-
ated with fatigue and inefficiency across the globe. The two common
hookworms of humans are Ancylostoma duodenale and Necator
americanus. These helminths attach themselves to the lining of the
small intestine by their buccal capsules and cause chronic blood loss
by sucking blood from the intestinal villi.
Hookworm is widely distributed in Southern Europe, Africa,
the Middle East, the Indian subcontinent, East Asia, and the New
World, especially Brazil and the Southern United States of America.
The infestation may be light, so that iron loss is not sufficient to
cause frank iron deficiency. In hookworm disease, involving Old
World and New World hookworms, heavy infestation occurs as a
result of repeated exposure of the skin to soil contaminated by inva-
sive hookworm larvae. Mucosal immunity may also be reduced in
the susceptible host. Although it is not known exactly what hook-
worms remove from human blood, intact red cells transit through
the nematode gut: each Anclyostoma worm induces the loss of up to
300 µl of blood daily, whereas each Nectator worm causes the loss of
up to 50 µl.
Hookworm anaemia is often regarded as a disease of poor
farmers who have only poor sanitation; but hookworm infestation
section 22  Haematological disorders
5384
has a very wide distribution with long-​standing industrial conno-
tations beyond subsistence agriculture. Hookworm is well known
as an occupational disease in miners, particularly deep miners of
metal ore rather than colliers. The disease is widely reported in Swiss
tunnel workers and North as well as South Europe, California, and
Queensland. Notorious as ‘miners’ anaemia’ and with typically florid
‘bunches’ (cutaneous larva migrans) in those who worked in the
deep copper/​tin mines in Cornwall, the link between occupational
skin exposure to hot and humid, mineral-​rich soils and wholly inad-
equate access to sanitation has been established since the early years
of the 20th century.
Clearly, induction of frank anaemia in hookworm infestation
will be dependent on the iron content of the diet, the extent of
tissue iron stores, and the duration and intensity of the mucosal
helminth infestation itself. The heaviest infections usually affect
rural workers in agricultural communities or miners where re-
peated exposure occurs in isolated locations and where crops or
minerals are harvested under conditions of poor sanitation. The
iron deficiency anaemia of hookworm disease may be difficult to
diagnose when the mucosal inflammation that accompanies heavy
infestation is associated with reduction in serum proteins such as
albumin and transferrin; this, combined with an acute phase re-
sponse, may at first lead to a mistaken diagnosis of the anaemia of
chronic disorders.
Hookworm infestation is an under-​recognized cause of
maternal anaemia, and this has precluded the use of anthel-
mintic treatment in health provision for pregnant women. In
sub-​Saharan Africa, it has been estimated that nearly 40  mil-
lion women of reproductive age are infected with hookworm;
of these, about 7 million were pregnant in 2005. As expected,
increasing intensity of infestation was associated with worsening
anaemia in pregnant women living in poor countries. Given the
number of pregnant women at risk of preventable hookworm-​
related anaemia, there is an urgent unmet need to determine the
benefits of anthelmintic treatment and to develop safe preventive
methods against this infestation. These measures will then need
effective introduction—​without the fear of damaging the fetus or
mature offspring—​to the benefit of the rural poor who are most
at risk.
Since up to two-​thirds of the haemoglobin iron released by the
worms can be reabsorbed in the intestine, significant anaemia oc-
curs only when there is a heavy parasite load. Patients with hook-
worm disease experience fatigue, dyspnoea, palpitations, and
mental changes—​including pica related to severe iron deficiency.
Nonspecific abdominal pain occurs and radiographic examin-
ation of the intestine or endoscopy may reveal duodenitis with a
punctate inflammation associated with partial villus atrophy of the
duodenojejunal mucosa. Oedema may result from cardiac failure
in severe cases, but is more frequently due to hypoalbuminaemia
caused by parasite-​related protein-​losing enteropathy. Hookworm
disease may be associated with other opportunistic helminth infec-
tions such as ascariasis or strongyloidiasis (the latter with a risk of
the fatal hyperinfection syndrome). From many aspects, hookworm
infestation contributes to a vicious cycle of poverty due to incapacity
for work as result of illness and the preferential use of poor labour in
rural environments or in deep mining, where the risk of hookworm
invasion is greatest.
Intrinsic gastrointestinal disease (see Section 15)
The gastrointestinal tract is a key and often cryptic source of blood
loss which should always be considered in patients with iron defi-
ciency anaemia. Ulcerating lesions of the small and large intestine—​
including cancers—​are often responsible for iron deficiency
anaemia. Chronic intermittent bleeding can also arise from unusual
sources such as Meckel’s diverticula, strictures, angiodysplastic le-
sions, hamartomas, and other benign but ulcerating tumours, such
as leiomyomas. Gastric ulcers cause chronic intermittent bleeding,
but duodenal ulcers rarely cause chronic gastrointestinal blood loss;
these typically cause episodes of acute bleeding.
Oesophageal ulceration and inflammatory lesions cause iron
deficiency anaemia, but caution is needed in attributing blood
loss sufficient to cause iron deficiency to such a source, especially
hiatal hernia, unless other potential sites of bleeding have been
excluded.
Unusual sources of gastrointestinal bleeding include multiple
telangiectatic lesions of Osler–​Rendu–​Weber disease (hereditary
haemorrhagic telangiectasia)—​in which bleeding may occur any-
where from the nasal or oropharynx down to the stomach and
upper intestine. The connective tissue disease, pseudoxanthoma
elasticum, is also associated with recurrent, often severe, upper
gastrointestinal haemorrhage. The blue bleb naevus syndrome,
Peutz–​Jeghers syndrome, and other hereditary gut polyposes are
rare causes of chronic gastrointestinal bleeding. Inflammatory dis-
ease of the lower small intestine and colon such as Crohn disease
and ulcerative colitis, usually associated with chronic intestinal
blood loss, may present with an abdominal history in which iron
deficiency anaemia is prominent.
Miscellaneous causes of blood loss
Very occasionally, iron deficiency anaemia due to self-​bleeding may
have to be considered: blood may be removed from almost any site
but bizarre tactics may be adopted to conceal the process, thus re-
quiring considerable ingenuity, and often intensive detective work,
to identify the source.
Bronchial or pulmonary blood loss  The striking appearance of
expectorated blood means that the iron deficiency anaemia associ-
ated with frank haemoptysis will demand little diagnostic skill, but
disease-​related haemorrhages sufficient to induce chronic anaemia
are very rare. In contrast, recurrent intra-​alveolar pulmonary haem-
orrhage may be massive but is often cryptic, even though it may
cause unexplained illness and severe anaemia—​as in Goodpasture
syndrome.
Urinary tract blood loss  With respect to haematuria, it should
be remembered that not all red discoloration of urine is due to red
cells:  where there is doubt, haemoglobinuria and myoglobinuria
should be considered. In rare circumstances, the differential diag-
nosis should include other pigments such as those derived from
beetroot, or the oxidized pyrrole, porphobilin, derived from the
porphyrin precursor, porphobilinogen. Occasionally, alkapton, the
oxidized product of homogentisic acid in alkaptonuria may take on
a red hue rather than black and give rise to confusion—​as may the
presence of anthocyanins and phenolphthalein (the latter usually in
nonfresh, alkaline urine), in individuals who use these substances
as purgatives.
22.6.4  Iron metabolism and its disorders
5385
Haemolysis  Iron can be lost in the urine through the kidney in
conditions where chronic intravascular haemolysis occurs, and this
may be sufficient to induce iron deficiency in the absence of overt
changes in urine colour. Often the loss of iron is chronic and oc-
curs through the exfoliation of haemosiderin-​rich tubular epithe-
lial cells into the urine; under these circumstances the urine is not
discoloured.
Patients with haemolysis due to prosthetic cardiac valve malfunc-
tion, paraprosthetic leaks, or valvular defects causing shear stress
or other mechanical effects may have frank haemoglobinuria and
methaemoglobinuria. Testing the urine for free haem and protein by
stick urinalysis and examination of the blood film for characteristic
red cell fragments, strongly suggests the diagnosis.
Haemoglobinuria  In paroxysmal nocturnal haemoglobinuria,
chronic intravascular haemolysis causes sustained urinary iron
loss with or without visible haemoglobinuria. In these circum-
stances, free haemoglobin is released, which quickly saturates the
capacity of the plasma proteins haptoglobin and haem-​binding
protein, haemopexin; free haemoglobin thus appears in the glom-
erular filtrate from which it is endocytosed by proximal tubular
cells and degraded. After degradation to haemosiderin, iron is lost
in the urine as the iron-​loaded epithelium is exfoliated; consequen-
tial haemosiderinuria is readily detected by diagnostic microscopy
of the centrifuged urine deposit after reaction with Perls’ reagent
(Prussian blue granular staining). While haemosiderinuria is a diag-
nostic sign, it does not immediately follow a bout of haemolysis, ap-
pearing only after 2 to 3 days, although it may persist for several
weeks after a haemolytic attack.
In march haemoglobinuria, mechanical injury to erythrocytes in
the circulation of the feet may induce similar features with conse-
quential iron deficiency, for example, in service recruits and high-​
performance athletes. In recent years, an often misdiagnosed but
closely related syndrome, aptly termed ‘foot-​strike haemolysis’ by
American physicians, has been identified in marathon runners and
regular ‘joggers’. This condition, which occurs in both sexes but par-
ticularly in young athletic women, is characterized by signs of iron
deficiency with hypochromia, polychromasia, and macrocytosis.
These changes are due to accelerated erythropoiesis as well as red cell
fragmentation (visible on blood film examination). The true source
of iron loss—​through the renal glomeruli—​often escapes detection,
unless haemosiderinuria is specifically sought at times close to the
period of exercise.
Malabsorption of iron
The inability to release and absorb adequate amounts of iron from
the diet is an important but frequently overlooked cause or con-
tributor to iron deficiency. Diseases of the stomach, duodenum, and
upper jejunum may be responsible for the malabsorption of food
iron, but simple radioactive tracer measurements may fail to iden-
tify the absorptive defect. On the other hand, properly conducted
radioactive food labelling studies show that there is malabsorption
of nonhaem and haem food iron after gastric bypass surgery or in-
testinal resection.
Acquired defects of the intestinal mucosa other than inflam-
matory disorders may contribute to malabsorption of therapeutic
iron. Young children with iron deficiency anaemia refractory to oral
therapy that was corrected by parenteral supplementation have been
reported. Careful investigation in some has revealed an absorptive
defect for iron which was corrected itself by systemic iron supple-
mentation, raising the possibility that severe iron deficiency itself
prejudices the ability of the mucosal epithelium in the upper small
intestine to carry out its normal absorptive function. However, no
further investigations to identify the nature of this acquired meta-
bolic defect are available.
Rarely, iron deficiency may result from inflammatory disease of
the upper intestine that causes malabsorption. Coeliac disease in in-
fants and adults may be responsible, and the iron deficiency is often
combined with deficiency of folic acid. Sometimes large pharma-
cological doses of iron with or without folic acid may overcome the
anaemia caused by coeliac disease, but unless a strict gluten-​free
diet is adhered to after the iron supplements cease, the anaemia re-
curs. Although malabsorption of food iron contributes to the iron
deficiency associated with coeliac disease, loss of iron exacerbates
the effects of malabsorption, coexisting iron loss being related to
increased epithelial exfoliation with crypt hyperplasia and at times
bleeding due to local ulceration.
Malabsorption of food iron due to abnormal motility and
maldigestion associated with upper gastrointestinal surgery is com-
pounded by anacidity caused by gastritis or acid-​suppressing agents,
which—​if administered for long periods—​lead to gastric atrophy.
Long-​term administration of alkalis and certain iron-​chelating
drugs such as the tetracycline antimicrobials can also impair iron
absorption. Dietary factors, such as ingestion of food containing
excess phytate compared with bioavailable inorganic iron, can crit-
ically reduce gastrointestinal absorption of iron. Bariatric surgery
which leads to diminished gastric acid secretion and bypasses the
duodenum is complicated by iron deficiency anaemia; prophylactic
supplementation is thus recommended after this procedure, par-
ticularly in menstruating women.
Geophagia
Geophagia—​the deliberate consumption of non-​nutritive earth,
soil, chalk, or clay—​has an under-​recognized association with iron
deficiency. The behaviour occurs principally in certain rural or poor
urban populations and the relationship with malabsorption of iron
and iron deficiency is complex. The condition is an extreme form of
pica and has a cultural history extending from ancient times in the
records of several Roman physicians but also in the palaeontology
of Africa. Geophagia may be restricted to individuals or be prac-
tised by family or community groups. It is a classical feature of iron
deficiency with pregnancy in which perverted taste perceptions are
often present (pica). In adults, the condition has been associated
strongly with poverty, and while it occurs in individuals it has also
attracted attention as a cult behaviour most often affecting women,
who may as a result regularly ingest large quantities of calcium, so-
dium, or potassium salts, often contaminated with lead, together
with silicates present in earth and clay.
Colonial African doctors in the 19th century noted wide-
spread geophagia in slaves. Here, geophagia was linked to poor
health and declining work output; the African explorer, David
Livingstone, described earth-​eating among slaves in Zanzibar
but apparently rejected poverty as the explanation. Geophagia
section 22  Haematological disorders
5386
was reported in African slaves transported to Brazil as well as
southern parts of North America, where it was known by the
term ‘cachexia Africana’. The disorder occurs contemporaneously
in Georgia and Louisiana, sub-​Saharan as well as urban South
Africa, but has been reported in individuals and groups at various
times from nearly every part of the world, including India and the
Far East. In some individuals, famine or malnutrition is present,
but geophagia may accompany group religious practices or reflect
a form of ritualistic purification.
Geophagia occurs most typically in black women belonging to
African American communities in the south-​eastern United States
of America and in rural as well as urban areas in South Africa
and Nigeria. The behaviour is prevalent among pregnant women
in Africa, more than half of whom may report the practice: re-
cently, a high frequency of the disorder has been noted in pregnant
Hispanic women in Mexico and the south-​western United States
of America.
Cultural perspective  As a cult behaviour, geophagia is often
considered to represent a psychiatric state, but this view is less
easy to sustain in whole communities where it resembles a cul-
tural practice that occurs within the context of poverty all over
the world. The phenomenon noted historically in slaves subjected
to cruel and inhuman mistreatment often seems to have persisted
in their disadvantaged latter-​day descendants. Reports after the
2008 earthquake on Haiti and the consequential socioeconomic
disaster, indicated an outbreak of geophagia with widespread
consumption of ‘bon bons de terres’—​biscuits made from soil,
vegetable oil, and crude salt. In summary, whatever the psycho-
pathological causes, the association of geophagia with starvation
and deprivation—​as well as postcolonial failures of international
politics—​is striking.
The complex behaviours and at times frank psychopathology of
geophagia is repeatedly emphasized in reports of compulsive pica
and earth consumption by disturbed patients residing in long-​
term psychiatric institutions, often with bizarre consequences.
Despite apparent associations with migration, slavery, famine,
poverty, other pica-​like behaviours, and iron deficiency in many
societies, the cultural, psychiatric, and nutritional factors that
contribute to geophagia are neither constant by association nor
inevitably driven by single psychosocial or pathophysiological
mechanisms.
Effects of geophagia  The main clinical consequences of geophagia
relate to an entrained pica behaviour aggravated by the chronic
neuropsychiatric effects of iron deficiency; there is symptomatic an-
aemia and often weight loss. In South Africa the condition occurs in
women who purchase earth which may be contaminated with appre-
ciable quantities of lead. Other toxic minerals such as arsenic can be
accidentally coingested, especially in China and parts of the Indian
subcontinent. Ingestion of large quantities of indigestible material
may cause intestinal bloating or, occasionally, life-​threatening ob-
struction and perforation.
A more frequent association is a parasitic helminth infestation
such as ascariasis, and the dog tapeworm, toxocara. Demographically
and geographically in the New World, the condition may be asso-
ciated with endemic hookworm anaemia, typically Necator ameri-
canus but also Strongyloides spp., parasites that are not acquired by
oral ingestion but by environmental exposure and skin invasion.
In the Old World, Africa, and Europe, geophagia is highly associ-
ated with iron deficiency and geohelminth infections, most notably
with Ascaris lumbricoides and Trichuris trichiura. Children who
pursue geophagia acquire these infections and also the dog parasite,
Toxocara canis. A notable feature, revealed by a longitudinal study in
more than 800 pregnant women from western Kenya, was the high
frequency of geohelminth reinfection in the months after effective
anthelmintic treatment at term. Reinfection was strongly associated
with the practice of geophagia.
Genetic causes of iron deficiency anaemia
While illustrative of the importance of molecular components in-
volved in iron metabolism, so far as can be determined, mono-
genic causes of iron deficiency are exceptionally rare. Mutations
affecting serum transferrin, the proteins involved in transferrin
receptor cycle, enzymes that bring about the formation of the first
committed precursor of haem biosynthesis (5-​aminolaevulinic
acid), and proteins of the iron sulphur cluster may induce
hypochromic anaemia, usually accompanied by iron overload.
Hepcidin plays an indirect role in erythropoiesis by controlling
the availability of iron in the plasma. Inappropriate elevation of
hepcidin concentrations occur in the rare genetic iron-​refractory
iron deficiency anaemia (IRIDA) and in the anaemia of chronic
disease.
Iron-​refractory iron deficiency anaemia
Original studies in members of a Sardinian family with microcytic
anaemia due to defective iron absorption and utilization identified
the molecular basis of an apparently ultra-​rare condition that may
in fact account for many instances of hitherto undiagnosed iron
deficiency anaemia unresponsive to oral iron and with limited re-
sponsiveness to parenteral iron. At first thought to be inherited
as a recessive trait, it is now clear that heterozygotes can also be
affected.
After excluding the involvement of known genes implicated
in iron metabolism, a genome-​wide search identified a locus
encompassing the matriptase-​2 gene, TMPRSS6 (also known as
transmembrane protease, serine 6), which is located on human
chromosome 22q. In the original pedigree, affected patients were
found to harbour a homozygous splicing mutation which is pre-
dicted partially to inactivate protease function. Plasma and urinary
hepcidin concentrations were later shown to be inappropriately ele-
vated. The corresponding murine gene (Tmprss6) has been shown
to be an essential component of a pathway that is sensitive to iron
lack and suppresses the release of hepcidin.
The type II transmembrane serine protease, TMPRSS6 or
matriptase-​2, cleaves HJV present on the plasma membrane, thus
attenuating its action on the hepcidin promoter mediated by BMP6
signalling through the transducer, SMAD4 in the liver.
BMP6 and related ligands bind to the BMP receptor and
coreceptor, membrane HJV, to induce phosphorylation of the
SMAD proteins and formation of heteromeric complexes with the
common mediator SMAD4. The complexes stimulate transcription
of the HAMP gene to induce synthesis of hepcidin. Matriptase-​
2 activity further downregulates hepcidin transcription because
the release of soluble HJV serves as an antagonist of the bone
22.6.4  Iron metabolism and its disorders
5387
morphogenetic pathway by competing with the membrane form
for BMP ligands.
In summary, impaired TMPRSS6 (matriptase-​2) activity en-
hances hepcidin biosynthesis with elevated plasma concentra-
tions of this master hormonal regulator: the outcome is refractory
microcytic anaemia and persistent functional iron deficiency
due to suppressed iron absorption adversely complicated by im-
paired release of iron from the exiguous stores in macrophage
compartment. This genetic disease is richly informative for under-
standing iron physiology and in particular opens up a potential
avenue for therapeutic development in secondary iron storage
disease (Box 22.6.4.2).
Clinical manifestations of iron deficiency
Symptoms of iron deficiency include fatigue, pallor, sore tongue,
palpitations, irritability, and little-​recognized mental changes, such
as pica. Iron deficiency is a notable cause of the restless legs syn-
drome (Willis–​Ekbom syndrome), sometimes known as delusional
parasitosis.
Iron deficiency induces behavioural changes in experimental ani-
mals, but lethargy in humans is often the only symptom apart from
pica, which includes craving for soils and the ingestion of silica-​
rich earths, sometimes viewed as a cult practice, geophagia. Other
variant cravings are sufficiently characteristic to be dignified by spe-
cial terms, such as ryzophagia (rice), amylophagia (other starches,
including potato), lithophagia (stones or gravel), cautopyreiophagia
(burned matches), and trichophagia (hair); but milk, salty and sour
foods, sweets, and dates are all recorded from different regions and
probably reflect specific cultural familiarities. Pagophagia (craving
for ice or cold drinks), noted in Hippocratic times, is strongly linked
to iron deficiency, especially in younger subjects. The pain associ-
ated with biting into and chewing solid ice, sometimes leading to
fractured teeth, only emphasizes the compelling nature of this
capricious behavioural fixation. Pagophagia has been reported as a
manifestation of iron deficiency 2 or 3 years after Roux-​en-​Y gastro-
intestinal bypass surgery for pathological obesity management; it
responds rapidly to parenteral iron replenishment. Iron deficiency
combined with the abnormal taste preferences may account for the
bizarre food craving that constitutes part of the folklore of preg-
nancy in many cultures, and in women from economically deprived,
as well as rich nations.
There may be a complaint of dysphagia associated with the de-
velopment of an oesophageal web (Patterson–​Brown–​Kelly or
Plummer–​Vinson syndrome), which usually occurs in elderly or
middle-​aged women with chronic iron deficiency.
Clinical signs of severe iron deficiency include pallor and
nonerythropoietic manifestations:  angular cheilosis, atrophic
glossitis, and dystrophy of the nails with longitudinal ridging or
koilonychia (which has a predilection for the nails of elderly women
with long-​standing iron deficiency). Moderate hair loss may also be
a feature of integumental iron deficiency. Severe iron deficiency may
occasionally coexist with splenomegaly, which resolves after iron
treatment.
Signs of underlying disease include peripheral oedema
(hypoalbuminaemia associated with massive hookworm infec-
tion) and oronasal and palatal telangiectasia associated with
Osler–​Rendu–​Weber disease (hereditary haemorrhagic telangi-
ectasia). Papilloedema has also been reported and may reverse after
treatment of the anaemia. Finally, care should be taken to search
for peripheral signs of systemic diseases such pseudoxanthoma
elasticum, hereditary haemorrhagic telangiectasia, and Peutz–​
Jeghers syndrome: these hereditary causes of bleeding are readily
missed but their identification has material significance in clinical
management.
Diagnosis of iron deficiency
Routine haematological parameters will reveal microcytic an-
aemia, usually in association with an unequivocal reduction in
serum transferrin saturation (<16%) and a reduced serum ferritin
concentration (<12 µg/​litre) (Box 22.6.4.1). Blood film will con-
firm microcytosis but may also show increased variation in red
cell size and shape often with atypical forms such as ‘target’ forms
and ‘pencil’-​shaped cells. The absence of these features and of an
acute phase reactive response may suggest dyserythropoietic or
sideroblastic anaemia or thalassaemia trait. ß-​thalassaemia trait is
also associated with a raised HbA2 while α-​thalassaemia trait is not.
Patients with iron deficiency consequent upon intravas-
cular haemolysis (e.g. related to mechanical heart valves or
march haemoglobinuria) also show red cell fragmentation and
polychromasia with macrocytosis due to compensatory erythro-
poiesis. Paroxysmal nocturnal haemoglobinuria similarly induces
iron deficiency with net renal excretion of iron also due to urinary
shedding of haemosiderin-​loaded tubular epithelial cells. Lead
poisoning may be associated with iron-​deficient indices, with or
without full-​blown sideroblastic changes. A bone marrow aspirate
stained with Perls’ reagent for iron in marrow macrophages will
rapidly confirm reduced or absent stainable iron in the storage
compartment: the presence of ring sideroblasts, dyserythropoietic
features, and/​or megaloblastic change would also guide differen-
tial diagnosis.
Box 22.6.4.2  Modulators of erythropoiesis to treat
iron-​loading anaemias
•	 Inhibitors of Janus-​activated kinase-​2 (JAK2) signalling:
—​	 Targets erythropoietin, the master regulator of erythropoiesis.
—​	 Binding of erythropoietin to its receptor activates the cytoplasmic
kinase JAK2 and signal transducers and activators of transcription
to effect hypoxic responses.
—​	 Clinical JAK2 inhibitors include ruxolitinib, which is used clinically
for myeloproliferative diseases.
•	 Attenuating TGFβ signalling with ligand traps:
—​	 Activin signalling involves the transforming growth factor-​β (TGFβ)
family and other components of erythropoietin-​induced erythroid
proliferation, such as GDF11, which prevents differentiation. Excess
activin signalling is a feature of dyserythropoiesis.
—​	 Sotatercept and ACE-​536 are undergoing trials to reduce activin
I and II signalling thereby inducing apoptosis of progenitors and
stimulate erythroid differentiation.
•	 Hepcidin augmentation:
—​	 Use of hepcidin mimetics or ‘minihepcidins’ but small peptides
have a short half-​life.
—​	 Transgenic overexpression of hepcidin—​challenging to obtain
appropriate regulation.
—​	 Inhibition or repression of TMPRSS6, the upstream regulator of
hepcidin expression.
section 22  Haematological disorders
5388
The presence of immunoreactive serum transferrin receptors may
provide additional evidence in favour of iron deficiency anaemia,
but because an increased concentration of these receptors occurs in
several marrow disorders and the enzyme-​linked immunosorbent
assay tests are relatively expensive, the role of this analyte in the
routine diagnosis of iron deficiency is not yet established. Red cell
zinc protoporphyrin concentrations greater than 35 µg/​dl of whole
blood are usually observed in patients with iron deficiency; values
greater than 100 µg/​dl are generally associated with lead toxicity.
Extremely high values may indicate the presence of erythropoi-
etic protoporphyria or lead poisoning in the latter, free, rather than
zinc protoporphyrin IX accumulates. Modest elevations in erythro-
cyte protoporphyrin can be observed in patients with haemolytic
anaemias, sideroblastic anaemia, and occasionally, the anaemia of
chronic disorders.
Investigation of iron deficiency
The identification of iron deficiency anaemia must be regarded as
an illness description rather than a satisfactory diagnosis for any pa-
tient in its own right: management should always include a serious
attempt to determine its root cause. Common errors occur when
iron deficiency is ascribed to the presence of other facile causes such
as ‘poor diet’ or menorrhagia—​ostensibly factors that are challen-
ging to quantify. All too often, mild oesophagitis or gastritis re-
ported at endoscopy placates the incurious investigator when the
underlying cause is bleeding from a coincidental source such as a
cryptic gastrointestinal cancer for which a diligent search is often
required. Malabsorption of iron as a result of, for example, coeliac
disease, or chronic urinary loss of iron consequent upon intravas-
cular haemolysis, are also important causes that are frequently not
even considered until late into the illness.
History
A full evaluation of the patient with iron deficiency should include
a detailed and systematic dietary history, including the consump-
tion of drugs, such as aspirin and nonsteroidal anti-​inflammatory
agents which may be responsible for gastrointestinal bleeding.
Additional gastrointestinal symptoms should be explored (e.g.
change in bowel habit), together with other evidence of blood loss
(e.g. presence of melaena or rectal bleeding). Enquiry should be
made to quantify the extent of menstrual loss, if the bleeding has
been ascribed, as is often the case, to menorrhagia in women of
reproductive age.
Attention should be paid to the family history and a travel history
to consider causes such as hereditary haemorrhagic telangiectasia,
Peutz–​Jeghers syndrome, familial polyposis coli, or hookworm dis-
ease. Patients with malabsorption of iron often have accompanying
nutritional deficiencies; coeliac disease has several well-​known asso-
ciations with autoimmune disorders such as type 1 diabetes mellitus
and is most common but not at all exclusive to patients with Irish
ancestry.
Examination
Clinical examination should extend from detailed enquiry about
previous gastrointestinal disease or surgery to an examination for
visceral enlargement, abdominal lymphadenopathy, splenomegaly,
masses, and other features suggestive of intra-​abdominal pathology
such as portal hypertension and abdominal cancer. Hereditary
haemorrhagic telangiectasia, pseudoxanthoma elasticum, or Peutz–​
Jeghers syndrome may be suspected in the presence of subtle or lo-
calized cutaneous, oronasal, or palatal lesions.
Investigations for source(s) of bleeding
The presence of iron deficiency anaemia demands a convincing
explanation and a robust causal diagnosis: while malabsorption
may be neglected, occult haemorrhage is usually the most im-
portant to identify. The gastrointestinal system is the most fre-
quent source of bleeding but can present a laborious challenge
for diagnostic pursuit. Patients over the age of 60 years should be
evaluated promptly, with upper and lower gastrointestinal endos-
copy carried out as soon as reasonably possible. Iron deficiency
anaemia in premenopausal women is often due to menorrhagia,
dietary deficiency, hiatal hernia, or a combination of these factors,
but endoscopy should not be neglected in men and women under
60 years in the presence of features such as weight loss or change
in bowel habit.
With patients in whom the cause of the iron deficiency is not
apparent, intensive studies may be needed to confirm the pres-
ence and identify the source of gastrointestinal bleeding, including
detection of occult blood in several consecutive samples of stool.
Sophisticated endoscopic and radiographic studies of the gastro-
intestinal tract and serological studies for the presence of coeliac
disease may be required, and occasionally there is a need to quan-
tify the amount of blood loss daily in the faeces or during menstrual
flow by using radiolabelled chromium red cell studies.
In difficult cases, percutaneous visceral angiography of the coeliac
and mesenteric arteries has proved useful for detecting sites of ac-
tive gastrointestinal bleeding, due, for example, to angiodysplasia
that are beyond the reach of conventional endoscopic procedures. In
those patients who are actively bleeding, such a procedure can iden-
tify local sites of blood loss greater than 0.5 to 1.0 ml/​min.
The recent introductions of fibreoptic double-​balloon enteroscopy
and wireless capsule endoscopy offer powerful, largely noninvasive
means to examine the entire small-​intestinal mucosa extensively for
the presence of bleeding lesions. Additional procedures dependent
on well-​resourced radiological facilities include CT enterography
and small bowel MRI.
Meckel’s diverticulum is a potential cause of obscure gastro-
intestinal bleeding in young adults and children. Some Meckel’s
diverticula can be diagnosed by scintigraphic studies using
technetium-​99m labelled pertechnetate, which may be concentrated
in the ectopic gastric mucosa. Meckel’s diverticulum and intestinal
strictures, particularly in the ileum, may occasionally be revealed by
retrograde colonic contrast radiographic studies.
Other diagnostic tests include searching for endomysial
(transglutaminase) antibodies, with confirmatory duodenojejunal
biopsy to detect coeliac disease. Examination of the urine and
sometimes sputum may be required to detect occult iron loss in ex-
foliated macrophages or proximal tubular cells, respectively, where
intrapulmonary haemorrhage or renal iron loss from glomeruli is
suspected.
Sometimes extensive diagnostic procedures fail to identify the
cause of iron deficiency when occult gastrointestinal bleeding is re-
sponsible. Under these circumstances it is sometimes appropriate
22.6.4  Iron metabolism and its disorders
5389
for an experienced surgeon to conduct a diagnostic laparotomy to
try to identify the bleeding lesion, although this is rarely required
when there is access to the full range of modern imaging techniques.
Occasionally, in younger adults and children, diagnostic laparotomy
is indicated to identify Meckel’s diverticula, intestinal stricture, and
congenital abnormalities such as duplications that serve as occult
sources of blood loss.
The patient with recurrent chronic iron deficiency anaemia
often presents a formidable challenge. First and foremost, an un-
equivocal diagnosis of iron deficiency is needed, and—​so far as
possible—​blood loss, even from cryptic sources, should be ex-
cluded. While losses of iron through bleeding are frequently re-
sponsible, iron deficiency due to malabsorption of iron and even
urinary losses as a result of chronic compensated intravascular
haemolysis may need to be considered. Expert examination of the
blood film with a reticulocyte count and review of present and
past measures of iron status, as well as dietary review and other
measures of nutritional status, are fundamental. There is a need
to capture the past travel and surgical history, as well as a compre-
hensive list of prescribed and other drugs. Experienced physicians
may need to seek interdisciplinary expertise in radiology, nuclear
medicine, and surgery before prematurely abandoning the search
for the causal lesion.
Management of iron deficiency
General aspects
As a general rule, iron should only be recommended as a treatment
for iron deficiency where that diagnosis is established beyond rea-
sonable doubt: the presence of other causes of anaemia (e.g. defi-
ciency of folic acid or haemoglobinopathy) can be easily missed
to the detriment of the patient, and positive harm can be done by
gratuitous iron supplementation. Many microbes require iron and
there is more than a hypothetical risk of infection if iron is given
unnecessarily.
Treatment of causes of anaemia, including bleeding, is clearly
a critical aspect of the management of iron deficiency anaemia.
Bleeding lesions in the gastrointestinal tract may require specific
treatment; coeliac disease should be treated with a gluten-​free
diet. Occasionally, patients with a chronic bleeding disorder for
which surgery is not effective, such as hereditary haemorrhagic
telangiectasia, may require long-​term iron supplementation at
doses less than that required to treat the acute iron deficiency
state. In such circumstances, periodic monitoring is required to
ensure that the level of iron replacement is adequate to meet the
demands of the bone marrow for de novo haem synthesis and that
iron overload is not occurring. It should be recognized that relief
of iron deficiency will improve many symptoms suffered by a pa-
tient even though they may suffer from an incurable underlying
disease.
Treatment with iron should continue until iron stores are replen-
ished: there is no excuse for inadequate therapy, especially in those
patients who are likely to suffer recurrent bleeding. Particular atten-
tion is needed for iron-​deficient patients who have had episodes of
acute bleeding treated by blood transfusion and who at the time of
therapy are not anaemic. These patients require appropriate iron re-
placement to replenish iron stores for their long-​term restitution of
health. Because iron therapy leads to a reduction in the avidity of the
transport system of the intestine for iron, it should be continued for
several months after the anaemia has been corrected to re-​establish
appropriate iron stores, ideally as reflected by a serum ferritin deter-
mination within the normal range.
Iron should be replaced not only to restore the normal haemo-
globin concentration but to replenish body iron stores. It is neces-
sary to replace iron depleted in somatic tissues such as the muscles,
where it is an essential component of mitochondrial cytochromes
and other proteins critical for optimal aerobic metabolism.
Occasionally, a therapeutic trial of oral iron for a defined period
is justified to verify a suspected diagnosis of iron deficiency
anaemia.
The effects of therapeutic iron supplementation should be moni-
tored: a reticulocyte response is normally observed in peripheral
blood, peaking 7 to 10 days after initiating treatment, and with a
significant increase in blood haemoglobin concentration apparent
within 2 to 4 weeks. If there is no evidence of continued blood loss,
the haemoglobin concentration should come within the healthy
­reference range within 2 months. Failure to meet these expectations
suggests either that the anaemia is not caused by iron deficiency,
or that there is continued depression of bone marrow function, or
that there is persistent blood loss—for which further investigation
is needed. Malabsorption of dietary iron is rarely severe enough
to compromise the haematological response to pharmacological
supplementation and an adequate response to oral iron does not
preclude the existence of impaired assimilation of physiologically
available iron in the small intestine.
Oral delivery of iron
Iron salts are best administered by mouth unless there are over­
whelming reasons for using the parenteral route—​parenteral prepar-
ations of iron are associated with a greatly increased risk of toxicity.
The outdated iron—​dextran complex as well as newer iron–​sucrose
preparations are associated with hypersensitivity, including se-
vere anaphylactoid reactions. Oral ferrous salts are better absorbed
than ferric salts, but in practice show little difference among pre-
parations in terms of rate of repair of anaemia at a given dosage of
elemental iron.
It is usual to treat iron deficiency anaemia with preparations of
oral iron that contain 100 to 200 mg of elemental iron daily. For full-​
blown iron deficiency anaemia, ferrous sulphate 200 mg is typically
administered three times daily (equivalent to 3 × 65 mg of elemental
iron). Some patients are unable to tolerate such a dose of iron be-
cause of constipation, diarrhoea, or abdominal pain and flatulence;
the presence of tarry, black stools with a sulphurous odour further
impair acceptability and the required persistence with therapy.
Under these circumstances, the dose of iron may be reduced and
this, rather than a change of iron salt preparation, usually improves
tolerability. The frequency of unwanted effects with ferrous sulphate
is generally similar to that of other iron salts when comparable quan-
tities of elemental iron are ingested.
Once established, the optimal therapeutic response to oral iron
increases the blood haemoglobin concentration by 1–​2 g/​litre per
day. Replenishment of iron has a slow effect on the epithelial changes
of iron deficiency: the atrophic glossitis may take several months to
improve as iron stores are replenished. The nonhaematological ef-
fects of iron deficiency in skeletal and cardiac muscle are also slow
section 22  Haematological disorders
5390
to respond. In contrast, the behavioural manifestations, for example,
pica syndromes, often respond to iron therapy within a few days.
Slow-​release oral preparations of iron are available, which
the manufacturers often claim release sufficient iron over a 24-​h
period for optimal haematological responses after once daily dos-
ages. However, these preparations are likely to distribute the iron
beyond the upper jejunum and thereby bypass those regions of
the intestine in which iron absorption is most avid. Compound
preparations of iron including B vitamins and folic acid are avail-
able, but there is little justification for prescribing these except
for prophylactic use in pregnancy (see following sections). In
infants and children, sugar-​free preparations of iron complexes
are available in the form of polysaccharide iron or iron–​sodium
EDTA (sodium iron edetate) complexes, which can be used safely
as recommended by the manufacturer. In premature infants, up
to 2.5 ml of syrup containing approximately 5 mg/​ml may be used
twice daily; up to 5 ml three times daily may be given to children
aged 6 to 12 years.
Pregnancy
Prophylactic iron is recommended in pregnant women who have
risk factors for iron deficiency such as a diet that lacks bioavailable
iron (often with low or no meat consumption) or prior menorrhagia.
Prophylactic iron is also used in the management of infants of low
birth weight, including premature babies, twins, and infants de-
livered by Caesarean section.
Compound preparations of iron with folic acid are used for the
treatment and prevention of iron and folic acid deficiencies in preg-
nancy. To prevent neural tube defects in women planning a preg-
nancy, the United Kingdom Department of Health advises that a
medicinal or food supplement of 400 µg/​day of folic acid is taken
before conception and during the first 12 weeks of pregnancy.
Parenteral delivery of iron
Provision of iron by the parenteral route does not normally lead to
more rapid repair of anaemia than when adequate oral iron prepar-
ations are used. Given its potential toxicity, the only justification for
the use of parenteral iron is in patients who are unable to cooperate
with or tolerate oral iron therapy; those with severe gastrointestinal
disease that causes malabsorption or continuing severe blood loss;
or in the management of anaemia in advanced chronic kidney dis-
ease where the bone marrow works best (with or without admin-
istration of exogenous erythropoiesis-​stimulating agents) when
serum ferritin is elevated to a supranormal level, which cannot be
attained with oral iron supplementation.
Iron dextran was withdrawn in 1992. Current preparations (ferric
carboxymaltose (Ferrinject), where 1 ml of solution contains 50 mg
of iron for injection/​infusion, and iron sucrose (Venofer), where 1
ml of solution contains 20 mg of iron (as iron(III)-​hydroxide sucrose
complex)), are less likely to cause severe ‘anaphylactoid’ reactions
than the now obsolete iron dextran (Imferon) preparations. Care
should be taken to exclude patients with a history of hypersensitivity
reactions and intravenous preparations should only be administered
where true iron deficiency has been confirmed.
The main difficulty that appears to arise with these and related
preparations is that the first pharmaceutical iron products for par-
enteral use (e.g. Imferon) were based on high molecular weight iron
dextran; this agent was withdrawn from the market due to manufac-
turing difficulties. Recent preparations have been lower molecular
weight iron dextrans (Infed/​Cosmofer), but while safer than the
high-​molecular forms these still carry an appreciable risk of severe
sensitivity reactions, which occur despite prior test dosing that is re-
commended before the full-​dose administration.
Other authorized products for intravenous use such as iron
gluconate (Ferrlecit) and iron sucrose (Venofer) apparently con-
tain loosely bound iron and hence are administered only in rela-
tively low doses of, say, 100 mg total infusion. Thus there remains
a need for effective preparations of iron for intravenous infusion
which are safe and allow larger corrective and preventive dosing
for patients with marked iron deficiency and with no other op-
tion for treating it. Latterly, several innovative new formulations
of iron preparations have been introduced including ferumoxytol
(Feraheme) and ferric carboxymaltose (Ferinject); the most recent
is iron isomaltoside 1000 (Monofer). This preparation is composed
of iron and chemically modified isomalto-​oligosaccharides with a
mean molecular weight of 1000 Da and principally 3–​5 bound glu-
cose units; the drug is finding acceptance and is now authorized
in European countries. The drug allows for a high-​dose iron infu-
sion that will replete iron stores in many patients at a single visit;
the formulation contains 100 mg iron/​ml but is not entirely free of
sensitivity reactions.
Unwanted and toxic effects of parenteral iron preparations
A history of allergic disorders including asthma, eczema, and prior
anaphylaxis are regarded as contraindications to the use of paren-
teral iron, as is liver disease and concurrent infection. Moreover,
these drugs cannot be recommended for children. Side effects in-
clude nausea, vomiting, taste disturbances, hypotension, paraesthe-
siae, abdominal disorders, fever, flushing, anaphylactoid reactions,
and the reactivation of inflammatory arthropathy. Injection site re-
actions, including phlebitis occur. Parenteral iron should probably
be avoided in patients with pre-​existing cardiac disease including ar-
rhythmias or angina. Parenteral iron is contraindicated in children
below the age of 14 years.
The Committee for Orphan Medicinal Products of the European
Medicines Agency continues to emphasize that intravenous iron
products should be administered when staff trained to evaluate
and manage anaphylactic/​anaphylactoid reactions as well as re-
suscitation facilities are immediately available. Patients should
be monitored for signs of hypersensitivity during and for at least
30 min after each administration of an intravenous iron product.
It is important to note that the committee considers that the risk
of hypersensitivity is increased in patients with known allergies
(including drug allergies) and in patients with immune or inflam-
matory conditions (e.g. systemic lupus erythematosus, rheumatoid
arthritis), as well as in patients with a history of severe asthma,
eczema, or other atopic allergy. In these patients, intravenous iron
products should only be used if the benefit is clearly judged to out-
weigh the potential risk and in full knowledge and consultation
with the recipient.
Having considered the overall regulatory data, the Committee for
Medicinal Products for Human Use under the European Medicines
Agency has concluded that the benefits of intravenous iron-​
containing medicinal products continue to outweigh the risks in the
22.6.4  Iron metabolism and its disorders
5391
treatment of iron deficiency situations when the oral route is insuffi-
cient or poorly tolerated.
A recent review by experienced North American haematologists
is relatively sanguine about the rarity of serious adverse events with
contemporary parenteral iron products. Of importance, however,
the United States Food and Drug Administration note that the
agency received 49 reports of death temporally associated with ad-
ministration of intravenous iron during the 5 years 2011 to 2016, 30
of which were adjudicated and determined to be anaphylaxis.
The development of porphyria cutanea tarda in patients receiving
renal replacement therapy is almost invariably an indication and
consequence of iatrogenic iron overload.
Administration of parenteral iron
Iron–​sucrose complex is given by slow intravenous infusion. Iron
carboxymaltose can be administered undiluted as a slow intra-
venous injection (infusion time dependent on dose) or diluted as
a slow infusion. The maximum single dose is 20 mg iron/​kg body
weight, not exceeding 1000 mg of iron. Doses are calculated, ac-
cording to the manufacturer’s instructions, from the haemoglobin
concentration, which should be reassessed no earlier than 4 weeks
after the final administration to allow time for a haematological
response.
At least 15 min of close observation should elapse after the test
dose before the therapeutic dose is administered. Iron isomaltose
offers convenient dosing options up to 20 mg iron/​kg body weight
with, at the time of writing, no test dose recommended by the manu-
facturer. Where less than 1 g is required, infusion should be under-
taken slowly over at least 15 min; infusion of higher total doses
should extend for at least 30 min. However, these recommendations
are not always successful and careful studies on the nature of iron-​
induced hypersensitivity reactions suggest that most are comple-
ment mediated and not true anaphylactic reactions. A convincing
case for managing these reactions in at-​risk patients by reducing the
rate of administration has been made by Szebeni and colleagues, to
whom the reader is referred.
Unusual syndromes with iron-​deficient erythropoiesis
Congenital deficiency of transferrin
There are a few reports of deficiency or virtual absence of serum
transferrin in infants with disturbed growth, marked hypochromic
anaemia, and disordered iron metabolism associated with systemic
iron storage leading to tissue injury. This disease is extremely rare
but holds great fascination for those investigators with an interest
in the pathophysiology of iron metabolism. Profound deficiency of
serum transferrin disturbs the normal ligand–​receptor signalling
mechanisms indicated in the overall control of body iron balance
and absorption in the intestine. Hypo-​ or atransferrinaemia in hu-
mans appears to be inherited as an autosomal recessive trait; the
gene encoding human serum transferrin maps to chromosome
3. Infusions of serum transferrin or plasma restore normal growth
and improve the abnormalities of iron homeostasis; iron-​deficient
erythropoiesis is also corrected, with resolution of the anaemia. The
half-​life of transferrin in the plasma is 5 to 10 days and so infusions
of plasma or purified preparations enriched with transferrin can be
administered at intervals. Since most individuals with transferrin
deficiency produce limited amounts of the protein antigen, immune
reactions to exogenous human transferrin appear to be either mild
or rare.
Absolute deficiency of transferrin receptors, for example, as oc-
curs in mouse embryos generated as a result of gene disruption
technology in embryonic stem cells, is incompatible with normal
development beyond the late embryo stage.
Acquired defects in the transferrin receptor
There is at least one well-​documented instance of an acquired de-
fect of iron delivery associated with signs of iron-​deficient erythro-
poiesis caused by loss of human transferrin receptor function. This
condition was associated with the development of antinuclear factor
and other autoantibodies as part of an autoimmune illness in an
adult woman with hypochromic anaemia. Autoantibodies directed
against the transferrin receptor were identified in the serum of the
patient, but the anaemia, with its attendant sideropenia, ultimately
responded to a combination of steroids and azathioprine therapy,
and the titre of transferrin receptor autoantibodies of peripheral
blood cells diminished. The extent to which this phenomenon oc-
curs generally during the course of autoimmune disorders associ-
ated with anaemia is unknown.
Other causes of refractory iron-​deficient erythropoiesis
Iron-​refractory iron deficiency anaemia
The IRIDA syndrome was originally reported as a rare autosomal
recessive iron metabolism disorder characterized by iron deficiency
anaemia (hypochromic, microcytic) that is often unresponsive to
oral iron intake and partially responsive to parenteral iron treat-
ment. The disorder is due fundamentally to excess action of the
iron-​regulatory hormone, hepcidin, and is caused by mutations that
impair the action of matriptase-​2, a transmembrane serine protease
encoded by the TMPRSS6 gene. As explained earlier, this protease
controls release of hepcidin by the liver and contributes to the main-
tenance of iron homeostasis.
IRIDA shows a mixed pattern of autosomal inheritance with di-
verse expressivity in affected pedigrees; there is usually milder ex-
pression in heterozygous individuals who inherit only one copy of
a mutant TMPRSS6 allele from a parent. Although it has been only
recently recognized, over 50 cases in families of diverse ethnic origin
are reported.
Although present lifelong, in most patients the manifestations
of anaemia are not severe and development is not impaired during
childhood. Women harbouring TMPRSS6 mutations are generally
more severely affected, as would be expected in conditions associ-
ated with iron deficiency.
Investigations reveal a hypochromic, microcytic anaemia with
very low serum iron and transferrin saturation; if measured, serum
hepcidin concentration may be elevated, but given that reliable clin-
ical assays are not always available, low hepcidin determinations are
not always found—​the reference range is broad and serum hepcidin
concentrations fluctuate. The ratio between the iron saturation of
serum transferrin and immunoreactive hepcidin has been proposed
to give better discrimination but has yet to be widely accepted. A fur-
ther confounding feature is that serum ferritin concentrations may
be within the normal range and are often modestly elevated after
treatment with intravenous iron. While the diagnosis is suggested
by the history of long-​term anaemia and parental consanguinity, it
section 22  Haematological disorders
5392
is likely that many patients with this condition escape diagnosis and
that artefactual anaemia or frank malingering may be considered.
The only definitive method for making the diagnosis of IRIDA syn-
drome is molecular analysis of the TMPRSS6 gene.
Unexplained iron deficiency
Despite increasing awareness of the need to determine its cause, un-
explained iron deficiency in children and adults is a frequent occur-
rence worldwide. In some patients in whom intensive investigation
fails, the expected parameters of iron deficiency associated with
iron-​deficient erythropoiesis are present after a failure to respond
to generous oral supplementation with iron salts; administration of
parenteral iron, however, leads to reticulocytosis with resolution of
iron-​deficient red cell indices.
At the time of writing, in such patients no convincing molecular
lesions have been identified in DMT1, ferroportin, hephaestin, or
uncharacterized moieties involved in the transport of iron across
the intestine. However, despite the absence of wide-​scale systematic
studies, it seems likely that IRIDA, in rare homozygotes and more
frequent heterozygotes, may be responsible for numerous undiag-
nosed patients with iron-​deficient erythropoiesis that responds only
to parenteral iron supplementation. With the advent of better diag-
nostic facilities in centres where advanced techniques are used in
the management of malignant diseases of the blood, molecular diag-
nosis of IRIDA may become incorporated into clinical practice and
there is a case for it to be provided by diagnostic genetic services.
Determining the frequency of pathological mutations in TMPRSS6
in different populations would go far in justifying the provision
of such diagnostic tests for patients with unexplained sideropenic
anaemia.
Secondary iron storage disease (secondary
haemochromatosis)
Primary iron storage disease (hereditary or genetic haemochroma-
tosis) occurs when excess iron accumulates as a result of heredi-
tary defects that lead to enhanced net absorption of iron. Usually
tissue injury and iron storage occur slowly, so that in most cases
damage from excess iron takes over two decades to manifest itself.
However, the increasing recognition of patients with genetic forms
of juvenile haemochromatosis accompanied by avid net accumu-
lation of iron, even during childhood, has strong clinical parallels
with aggressive secondary haemochromatosis due to the iron-​
loading anaemias. In both cases, early parenchymal iron loading,
from multiple sources, is associated with injury to the endocrine
system and heart. Juvenile and adult haemochromatosis is de-
scribed in detail in Chapter 12.7.1.
General aspects
Iron storage disease results from repeated blood transfusions or
sustained increases in iron absorption accompanying a primary
disorder of haematopoiesis with anaemia; these two principal
causal influences may coexist in the same patient. Each transfused
unit of blood contains about to 225 mg of iron as haemoglobin,
so patients receiving repeated blood transfusions to support an-
aemia typically accumulate iron at about 10 times the rate that
occurs from conditions associated with chronically increased iron
absorption.
While about 0.75 million people have thalassaemia, there are 332 000
conceptions annually estimated by the World Health Organization to
be affected by diseases affecting globin chains. About 275 000 have a
sickle-​cell disorder and need early intervention; 56 000 infants have
a major thalassaemia, including at least 30  000 who need regular
transfusions and 5500 who die with the complications of thalassaemia
major in the first months of life.
Iron storage disease occurs in patients who have received oral iron
therapy over many years as medicinal tonics or as treatment for re-
fractory anaemia. However, it is unknown if this occurs in the con-
text of other coexisting disorders or the presence of mutant alleles of
the HFE gene, underlying bone marrow disease, or the coinheritance
of a thalassaemia trait or red cell disorder such as pyruvate kinase de-
ficiency. Conversely, iron excess may develop spontaneously in pa-
tients with haemolytic (and especially dyserythropoietic) anaemias
alone through the recently identified mechanisms of hepcidin sup-
pression by a bone marrow-​derived factor (e.g. erythroferrone).
However, iron overload most commonly results from repeated
blood transfusion with or without underlying expansion of eryth-
roid marrow (Box 22.6.4.3).
Based on understanding of the molecular cell biology of iron
homeostasis and the control of erythropoiesis, several treatments
are being evaluated. These potential therapies interrogate respect-
ively, the pathological action of hepcidin and the action of members
of the transforming growth factor-​ß ligand superfamily on activin-​
like receptors in erythropoiesis: both avenues have reached a prom-
ising stage of clinical investigation.
Whatever the physiochemical basis, the mechanisms of iron-​
mediated toxicity are probably shared between the iron storage
syndromes:  primary genetic haemochromatosis and the sec-
ondary haemochromatosis associated with blood transfusion and
the iron-​loading anaemias certainly have many clinical features in
common.
Causes of iron overload
Transfusional iron overload
Each millilitre of whole donor blood contains the equivalent of 0.47
mg/​ml of elemental iron complexed with protoporphyrin, hence as
stated previously, a ‘unit’ of transfused red cells derived from an ori-
ginal donation of 475 ml thus contains about 225 mg of elemental
Box 22.6.4.3  Anaemias associated with iron storage disease
from transfusion and/​or increased iron absorption
•	 Congenital dyserythropoietic anaemia types I and II
•	 β-​Thalassaemia including transfusion-​dependent thalassaemia major
and the intermediate phenotype (nontransfusion dependent)
•	 Sideroblastic anaemia (congenital or acquired)
•	 Hereditary spherocytosis (when in association with one or more mu-
tant alleles of the HFE gene)
•	 Megaloblastic anaemia
•	 α-​Thalassaemia (haemoglobin H disease) (rarely)
•	 Pyruvate kinase deficiency (from transfusion and increased iron
absorption)
•	 Diamond–​Blackfan and aplastic anaemia (from transfusion)
22.6.4  Iron metabolism and its disorders
5393
iron, which is eventually retrieved after red cell breakdown in the
macrophage system as a result of the actions of haem oxygenase,
which releases bilirubin, carbon monoxide, and one atom of iron per
haem molecule linked to each globin subunit, thus each molecule of
haemoglobin A yields four iron atoms.
There is no physiological mechanism by which excess iron ac-
quired from transfusions can be excreted: after phagocytosis of the
senescent red cells by macrophages, the iron is initially retained by
the mononuclear phagocyte system. Continued delivery of iron de-
rived from transfused red cells leads to the excess of iron-​loaded
ferritin and its breakdown product, haemosiderin, in parenchymal
cells throughout the body, with ensuing tissue injury and functional
impairment. Initially this occurs preferentially in hepatocytes but
subsequently extends to the endocrine system and myocardium.
After the transfusion of 15 to 20 units of blood (representing
about 5 g of elemental iron), iron toxicity becomes evident. After 100
to 200 transfused units, myocardial iron accumulation and severe
toxicity is inevitable.
Iron-​loading anaemias
In dyserythropoietic anaemias not requiring regular blood trans-
fusion, such as milder thalassaemia phenotypes (intermedia) and
some sideroblastic anaemias, symptoms and signs of iron storage
disease are caused by excessive and sustained absorption of iron
from the diet by the intestine. Thus, although some patients with
β-​thalassaemia intermedia are treated by occasional transfusion,
much of the excess iron stored in the body originates from ingested
rather than transfused iron and is initially deposited in periportal
hepatocytes. Net absorption of dietary iron in β-​thalassaemia inter-
media may be increased more than fivefold above that in healthy
age-​matched subjects.
In regularly transfused patients with β-​thalassaemia major,
including the prevalent iron-​loading condition, haemoglobin E/​
ß-​thalassaemia, massive expansion of the erythropoietic marrow
is suppressed to render absorption of iron normal or near normal.
However, in patients with thalassaemia who are intermittently
transfused, erythroid hyperplasia persists and excessive absorption
of iron from the diet contributes significantly to the iron storage de-
rived from transfused cells; several grams of additional iron may be
acquired each year by this route.
Patients with hypochromic anaemias due to sideroblastic change
in the marrow are particularly at risk because they are often wrongly
diagnosed as chronic or recurrent iron deficiency anaemia; they
thus receive long-​term iron supplementation that serves merely to
exacerbate the iron-​loading state. It is noteworthy, however, that pa-
tients with haemolytic anaemia due to sickle cell-​haemoglobin C
disease do not commonly develop marked iron overload as a result
of enhanced iron absorption: full-​blown iron storage disease ap-
pears mainly in transfused patients with chronic anaemias.
Particular difficulties arise in refractory anaemias and haemo­
globinopathies in which there is a hyperplastic bone marrow with
ineffective erythropoiesis that drives the inappropriate absorption
of iron by the intestine. Patients with a rare inherited anaemia,
Diamond–​Blackfan anaemia, have few or no red cell precursors so
that iron bound to transferrin is not cleared by the bone marrow; the
plasma concentration of free, nontransferrin iron rises, and in trans-
fused patients poses a high risk of systemic iron overload.
In the South African Bantu and related sub-​Saharan African
populations, excess iron is ingested in an unusually bioavailable
form in beers and other alcoholic drinks prepared by fermentation
in iron pots (e.g. kaffir beers). Soluble complexes of readily bioavail-
able iron in these drinks contribute to secondary haemochromatosis
with frank scurvy in young-​ and middle-​aged men. Although much
of the iron is at first detected in the mononuclear phagocyte system
(and is seen particularly in Kupffer cells on liver biopsy), associated
hypogonadism and vitamin C deficiency later induce scurvy and
osteoporosis. Dietary adjustment and iron chelation therapy may
relieve the disorder, which is becoming less common after its rec-
ognition in the early 1950s. Family studies point to a genetic com-
ponent which predisposes individuals to this secondary iron storage
disease within given pedigrees, but the causal gene or genes have not
been identified.
Pathophysiology of iron loading in ß-​thalassaemia
In ß-​thalassaemia, imbalanced globin-​chain biosynthesis leads to
anaemia that is characterized by the deposition of aggregated free
α-​globin chains in the cytoplasm. Excess hemichrome pigment with
increased nonhaem iron is also present, and this is implicated in re-
active oxygen-​mediated oxidative stress and Heinz body formation.
Examination of the marrow shows marked expansion but eryth-
roid differentiation does not proceed to full maturation and there
are signs of extensive programmed death of red cell precursors.
Dyserythropoiesis in ß-​thalassaemia is thus characterized by several
abnormalities: expansion (often massive) of erythroid progenitors,
accelerated differentiation of the erythroid precursor cell population
(to the point of the development of polychromatophilia), and ar-
rested red cell maturation.
In operational terms, ineffective erythropoiesis disrupts the
homeostatic mechanisms of systemic iron balance:  the patho-
logical marrow signal overcomes the control exerted by the physio-
logical ‘storage regulator’ so that net absorption and delivery of iron
proceeds in an unregulated manner, even in the face of adequate
or increased iron stores. Hepcidin is inappropriately and pro-
foundly suppressed. If there is no prior burden of iron from red cell
transfusions in dyserythropoietic anaemias such as ß-​thalassaemia
intermedia, absorption of iron can be enhanced as much as ten-​
fold and is greatly in excess of requirements for erythropoiesis. The
pathological excess of labile iron readily induces further cytotoxicity
and the consequential effects of secondary haemochromatosis.
Inherited disorders of ferritin
Deficiency of the ferritin heavy chain
Studies in members of a large Japanese pedigree with autosomal
dominant iron-​storage disease resembling adult-​onset genetic
haemochromatosis, HFE1, found affected patients to harbour a mu-
tation (A49U) in the iron-​responsive element of H ferritin mRNA.
This results in a deficiency of the H-​ferritin protein component of
the ferritin multimer with impaired ferroxidase activity and capacity
for iron storage.
Inherited defects in the ferritin light chain
Two classes of informative mutations have been reported in the
ferritin light-​chain gene. Rare frameshift mutations which dis-
tort the C-​terminus of the L-​ferritin subunit cause a dominant
section 22  Haematological disorders
5394
neurodegenerative disease with adult-​onset dementia due to oxi-
dative injury and iron deposition in neurons. The casual mutations
impede assembly of the 24 mixed subunits of the multimeric ferritin
molecule and markedly impair its iron-​storage efficiency.
In contrast, point mutations that affect only the untranslated 5ʹ
iron-​responsive element in the cognate FTL mRNA cause a domin-
antly transmitted form of cataract with presentation in childhood.
This condition is known as the hyperferritinaemia-​cataract syn-
drome, in which excess free light chains precipitate in the lens of
the eye without evidence of systemic disease or iron excess, but with
markedly elevated serum ferritin concentrations which are often er-
roneously ascribed to iron overload or even haemochromatosis.
Clinical features of iron overload
Many of the clinical features of established secondary iron storage
disease are similar to those observed in the hereditary forms of ju-
venile haemochromatosis (see Chapter 12.7.1). The consequences
of transfusional iron overload and the benefits of chelation treat-
ment are probably best documented in patients with ß-​thalassaemia
major. Initially, iron derived from transfused red cells accumulates
as storage iron in the macrophage system; subsequently excess iron
appears in hepatocytes and ultimately in the endocrine system
and heart.
This distribution pattern reflects iron transferred from molecular
species other than transferrin: there is a striking susceptibility of par-
ticular cell populations within the tissues that show the most prom-
inent signs of iron-​related injury to this form of iron. For example, in
the entire adenohypophysis it is the limited population of but a few
hundred gonadotroph cells that accumulate iron, ultimately leading
to hypogonadotropic hypogonadism as the first manifestation of an-
terior pituitary failure. This endocrinopathy causes arrested sexual
development and infantilism. Iron also accumulates preferentially
in the β-​cells of pancreatic islets, leading to diabetes mellitus; in the
zona glomerulosa of the adrenal glands, with adrenal failure due to
mineralocorticoid deficiency; and in the chief cells of the parathy-
roid glands, ultimately causing hypoparathyroidism.
Transfusional iron overload, as in established hereditary forms
of haemochromatosis (especially juvenile haemochromatosis), has
a striking predilection for the myocardium, which is also attrib-
uted to unregulated uptake of nontransferrin-​bound iron, with risk
increasing with the total number of units transfused. Myocardial
disease (cardiomyopathy) can cause sudden death from about
15 years of age in untreated patients caused by tachyarrhythmias
and/​or heart block due to injury to the cardiac conducting system.
Refractory heart failure due to extensive cardiac myocyte injury and
fibrosis is also frequent. Modern iron-​chelation regimens are able
to prevent and to some extent reverse iron-​related cardiomyopathy,
but the effects of iron toxicity in the endocrine system are largely
irreversible.
Susceptibility to infection
Iron overload is associated with a greatly increased the risk of micro-
bial infection as free iron (i.e. that which unbound to transferrin),
is readily utilized, and is a competitive trophic and survival factor.
Indeed, infection is the second most common cause of death in pa-
tients with ß-​thalassaemia who receive regular red cell transfusions.
The following microbial pathogens have been associated with in-
fections in patients with iron overload: Yesinia enterocolitica, Listeria
monocytogenes, Plasmodium falciparum, noncholera vibrios (e.g.
Vibrio vulnificus), Mycobacterium tuberculosis, and Mycobacterium
avium complex. Fungal pathogens include Candida albicans,
Aspergillus spp., and the agents of mucormycosis.
It is noteworthy that patients receiving desferrioxamine, a nat-
ural high-​affinity iron-​binding molecule obtained from the Gram-​
positive bacterium, Streptomyces pilosus, remain at risk from severe
infections (e.g. Yersina enterocolitica and Vibrio vulnificus). Many
iron-​chelators, including desferrioxamine, are related to natural
siderophores which are produced by microbes and harnessed to
scavenge environmental iron in stable complexes that are subse-
quently taken up after binding to dedicated microbial receptor
complexes; these systems serve as important microbial virulence
factors. Pharmacological use of iron-​chelating drugs may bypass the
requirement for endogenous siderophores to compete for scarce en-
vironmental iron and thus render the treated host more susceptible
to invasive infections.
Diagnosis and monitoring
In transfusional iron overload (without chelation therapy), the
amount of excess body iron can be estimated quite reliably from
the transfusion history. The extent of life-​threatening systemic
(extrahepatic) iron deposition increases when the saturation of
serum transferrin by iron is greater than 60%. Above 70% sat-
uration, significant and potentially damaging concentrations of
nontransferrin iron (‘unbound’) are likely to be present, and iron
uptake is altered categorically from the physiological distribution
mediated by the transferrin receptor. When nontransferrin bound
iron is present, ‘free’ iron in the plasma is independently transferred
to tissues, which occurs in an unregulated and potentially toxic
manner.
Serum ferritin
The serum glycoprotein form of ferritin is a moderately reliable bio-
marker of iron overload. While most ferritin in the serum is unsat-
urated and has an infinitesimal capacity to transport iron, secretion
of the protein occurs in response to hepatic iron store. Serum ferritin
concentration—​readily determined by immunoassay—​is in wide
use for monitoring excess iron storage and its treatment.
Serum ferritin may either under-​ or overestimate the burden
of iron in the body and is a poor surrogate biomarker for risk
management in relation to haemochromatosis. For example, in
nontransfusion-​dependent ß-​thalassaemia, ferritin provides under-
estimates of liver iron concentrations, whereas in the presence of
pre-​existing liver disease such as hepatitis C, or in the presence of
fatty liver, serum ferritin alone may cause the degree of iron overload
to be overestimated. Hyperferritinaemia occurs in unrelated dis-
eases including the hyperferritinaemia-​cataract syndrome, Hodgkin
lymphoma, and other malignancies. Vitamin C deficiency tends to
depress serum ferritin concentrations.
A downward trend in serum ferritin concentration generally
indicates negative iron balance (more iron removed than trans-
fused) and an upward trend probably indicates inadequate dosing
or inadequate adherence to chelator therapy. However, coexisting
liver damage or inflammatory conditions can be responsible for
increasing serum ferritin concentrations. For these reasons, it is ad-
visable to obtain independent evidence of iron overload, its distribu-
tion, and its response to treatment.
22.6.4  Iron metabolism and its disorders
5395
Liver iron concentration
Liver iron concentration is the most reliable predictor of total body
iron, typically by way of the formula of Angelucci (derived from
liver biopsy data) in which body storage iron in mg iron/​kg body
weight = 10.6 × the liver iron concentration expressed as mg/​g dry
weight of tissue. However, with the advent of validated MRI tech-
niques to measure tissue iron, biopsy for tissue iron quantification
and histological examination is rarely carried out today, especially in
rich ‘developed’ countries.
The upper reference limit of liver iron concentration is about 1.8
mg/​g dry tissue weight; values above 7 mg/​g indicate inadequate
chelation and increased risk of tissue injury; and values greater
than 15 mg/​g (1.5% of dry liver weight) are associated with severe
extrahepatic iron deposition and myocardial injury.
Although noninvasive radiological techniques usually obviate the
need for liver biopsy and direct chemical quantification of tissue iron,
in some cases examination of liver biopsy samples may allow staging
of the disease, particularly in relation to coincidental viral hepatitis
in which fibrosis and cirrhosis combined with iron deposits in the
parenchymal cell can confound the clinical pathological findings. In
such circumstances, biopsy may contribute information of authentic
and valuable diagnostic utility with regard to iron overload.
Cardiac iron overload
While serial determinations of liver iron concentration enable ab-
solute iron overload and the direction of change to be monitored,
liver iron concentration is a relatively unreliable predictor of car-
diac iron deposition and associated injury with impaired function.
Myocardial biopsy is both invasive and unreliable as a means to in-
vestigate suspected myocardial iron overload.
The use of cardiac MRI, most notably with the T2* technique, is a
desirable tool for identifying the presence of increased cardiac iron
as well as the direction of change with chelation therapies. Patients
with a T2* less than 20 ms have increased cardiac iron while those
with T2* values less than 10 ms are at high risk of developing heart
failure within the next year if chelation is not intensified. Where
possible, it is recommended that centres monitoring patients with
transfusional iron overload should systematically undertake such
assessments in patients who are at risk. The frequency of monitoring
will depend on the degree of systemic and heart iron overload and
the underling haematological condition responsible for the iron
storage. Generally it is recommended that patients with transfusion-​
dependent ß-​thalassaemia have cardiac T2* carried out annually
unless the signal has been repeatedly found to be within the healthy
reference range and iron loading is otherwise well controlled.
Management—​iron-​chelating drugs
Until now, the use of iron-​selective chelating agents has been the
mainstay of management in patients with iron overload. This
stratagem has been the subject of intensive clinical research since
the introduction of desferrioxamine (dissociation constant for ferric
iron, 10−31 M) more than 50 years ago. The life-​saving effects of intra-
muscular desferrioxamine in ß-​thalassemia were shown in 1981,
and over the last 30 years the field has expanded with the develop-
ment of safe, orally active iron-​chelating agents.
Chelatable pools of iron in the tissues are derived from red cell
breakdown in the macrophage compartment as well as proteolysis of
ferritin in the liver. These operationally described pools of iron are
dynamic but also finite: they are generated continuously, so that the
most effective removal of toxic iron deposits depends on continuous
exposure to chelating agents. At any moment, only a fraction of the
body iron is accessible to chelating agents and thus the process of
depleting the excessive iron is slow. However, given the extraor-
dinary affinity of the iron chelators in clinical use, even in the pres-
ence of excess systemic iron accumulation, care is needed to reduce
the dose of the agent to minimize toxicity as the iron is removed.
The primary goal of chelation therapy is to remove body iron to
match the iron accumulation rate, or if substantial concentrations
of tissue iron have already accumulated, to induce negative iron
balance with the prevention or amelioration of iron-​induced tissue
injury. Chelation therapy is effective in transfusional iron overload
and prolongs life expectancy; it prevents heart disease, endocrine
failure, and hepatic fibrosis.
While the best evidence of the benefit of long-​term chela-
tion accrued from studies of patients with transfusion-​dependent
ß-​thalassaemia, other anaemias requiring transfusion, such as
myelodysplasia, Diamond–​Blackfan anaemia, and pyruvate kinase
deficiency, benefit from this treatment. Patients with sickle cell dis-
ease who have received multiple transfusions may be at less risk of
myocardial iron overload compared with those with thalassaemia
at apparently similar burdens of excess iron, but there is a risk of
liver injury if no chelation is given. Use of chelation for patients with
transfusion-​dependent myelodysplasia needs to take into account
the age and prognosis of the patient.
Compelling evidence gathered over 30 years shows that survival
in patients with transfusion-​dependent β-​thalassaemia is improved
by treatment with subcutaneous desferrioxamine, which prevents,
and can reverse, the cardiac manifestations of iron-​storage disease.
It must be noted, however, that full compliance with this demanding
treatment is required for benefit to accrue, which requires equal
commitment from the patient and the medical and nursing per-
sonnel who provide care.
Three iron chelators are licensed for the treatment of iron over-
load:  desferrioxamine, which must be administered parenterally
due to poor gastrointestinal absorption, and the two orally active
chelators, deferiprone and deferasirox. The chemical characteris-
tics, pharmacokinetics, routes of iron excretion, and recommended
doses of these agents are summarized in Table 22.6.4.1.
Guidelines for starting chelation therapy are founded on long
clinical experience with desferrioxamine. With this agent, overnight
subcutaneous infusions at least 5 nights a week were typically started
when the serum ferritin concentration exceeded 1000 µg/​litre or
after 20 transfusion episodes. While this paradigm has also been
used for other iron chelators, it is probably overcautious because the
ototoxicity, retinal toxicity, and bone growth abnormalities observed
at low serum ferritin concentrations and at desferrioxamine doses
higher than 40 mg/​kg are unlikely to occur with other chelators. For
example, serum ferritin concentrations as low as 500 µg/​litre are
often achieved with the orally absorbed iron chelator, deferasirox.
Desferrioxamine
Desferrioxamine is a hexadentate iron chelator which binds ferric
iron (iron(III)) in a 1:1 molar/​atomic ratio. Its relatively high mo-
lecular weight limits gastrointestinal iron absorption and this, to-
gether with a short systemic half-​life, means that desferrioxamine
section 22  Haematological disorders
5396
has to be given by continuous infusion. Desferrioxamine promotes
urinary excretion of iron that is derived from red cell catabolism; an
approximately equal amount of iron is depleted from hepatocellular
stores and excreted into the faeces via the biliary system.
Administration and dosing
The usual route for desferrioxamine administration is by slow sub-
cutaneous infusion over 8–​12 h, five to seven times per week; this is
can be done on an ambulatory basis in adults, but nocturnal admin-
istration is more often used, particularly in children. Nocturnal ad-
ministration relies on the use of slow clockwork, battery-​operated,
or balloon infusion devices. Although electrical syringe pumps are
in common use, smaller (and conveniently quieter) infusion devices
are now available. Light, prefilled balloon pumps, though expensive,
are also in use.
Patients with a high requirement for transfusion (>0.5 mg/​kg day
of iron accumulation) generally require higher doses than those
with low transfusion requirements. In patients without cardiac dis-
ease, the daily administration of oral ascorbic acid at 2 to 3 mg/​kg
increases iron excretion as ferrioxamine. In patients with heart failure
or high myocardial iron, continuous intravenous desferrioxamine
(maximum 60 mg/​kg), delivered through an indwelling line can
often relieve acute heart failure and also slowly decrease the loading
of iron in the myocardium. Desferrioxamine has sometimes been
given intravenously together with blood transfusion, but the im-
pact on iron balance is small; the drug should not be added directly
to the blood, but to avoid potentially toxic bolus administration of
desferrioxamine should be coadministered through a separate intra-
venous route through the same cannula.
Side effects and monitoring
The most important unwanted effects of desferrioxamine are ototox-
icity and retinal toxicity. These are more likely at higher doses and
where iron overload is less marked, and children are more suscep-
tible to these toxicities than adults. It is prudent to monitor visual
acuity and auditory function at intervals during treatment. The
daily dose in children is 20 to 40 mg/​kg of body weight and doses
should not exceed 40 mg/​kg because of risks to growth and bone
Table 22.6.4.1  Characteristics of iron-​chelating drugs
Compound
Desferrioxamine
Deferasirox
Deferiprone
Molecular weight (Da)
560
373
139
Route of absorption
Subcutaneous, intravenous,
intramuscular
Oral
Oral
Half-​life of iron-​free chelator
20–​30 min
12–​16 h
3–​4 h
Maximum plasma concentration
of iron-​free drug (µM)
7–​10
80
90–​450
Minimum plasma concentration
with daily dosing (µM)
0
20
0
Elimination of iron complex
Urine = faeces
Iron complex removed more
slowly than free drug
Mainly faeces
Mainly urine
Metabolism
Mainly in liver to iron-​binding
metabolites
Mainly in liver to iron binding
glucuronides
90% eliminated in faeces, 60%
unmetabolized.
Glucuronide formed in liver does not bind
iron
Recommended dose mg/​kg per
day titrated for rate and level of
iron loading
30–​60
5–​7 ×/​week
20–​40 once daily
75–​100 in 3 divided doses
Main adverse effects
Ocular, auditory, bone growth
retardation, local reactions,
allergy
Gastrointestinal, increased creatinine,
proteinuria
Hepatitis
Gastrointestinal, arthralgia,
agranulocytosis/​neutropenia
Potential drug interactions
Vitamin C (in doses >200 mg)
Prochlorperazine
—​Inducers of uridine diphosphate
glucuronyl transferase
—​Bile acid sequestrants
—​Substrates of cytochrome P450 (CYP)-​
3A4/​5, CYP2C8, or CYP1A2
Drugs inducing neutropenia
Aluminium-​based antacids
Vitamin C?
Licensed indications <6 yearsa
First line for thalassemia major
age 2–​6
First line—​age 2–​6 years USA
Second line—​age 2–​6 years Europe
Insufficient information for licensing
Licensed indication <6 years
First line in TDT
First-​line TDT. First line NTDT
Other chelation not tolerated
Chelator toxicity monitoring
Pure tone audiometry yearly
Serum creatinine—​before starting,
then weekly for first month—​thereafter
monthly
Neutrophil count 1–​2 weekly
Retinal assessment yearly
Urine protein/​creatinine ratio monthly
Aspartate alanine transaminase monthly
Aspartate alanine transaminase monthly
NTDT, nontransfusion-​dependent thalassaemia; TDT, transfusion-​dependent thalassaemia.
a Licensing in children differs between USA and Europe.
22.6.4  Iron metabolism and its disorders
5397
development of higher doses. In adults with established iron over-
load, the effective dose is 40 to 50 mg/​kg five to seven times per week
(licensed up to 60 mg/​kg in the United States of America). The dose
must also be reduced when serum ferritin concentration decreases,
because the risk of retinal or auditory toxicity increases when the
dose, relative to serum ferritin, is high.
Rarely, minor gastroenterological disturbances, myalgia, and very
rarely anaphylaxis may occur. Desferrioxamine interacts unfavour-
ably with phenothiazines and coma may result, especially in patients
with modest iron overload. Apart from minor localized skin reac-
tions, desferrioxamine is usually otherwise well tolerated. These re-
actions can usually be controlled by reducing the concentration of
the drug in the infusion (and always <10% weight/​volume) and by
alternating the infusion sites. Hydrocortisone in doses of up to 100
mg has been reported to reduce severe cutaneous reactions.
Some patients receiving desferrioxamine develop infections with
microorganisms such as yersinia and fungi, including Candida
and Mucor spp., that have fastidious requirements for iron. Iron-​
overloaded patients may also develop other systemic microbial in-
fections and are particularly susceptible to fulminating sepsis caused
by the marine vibrio, V. vulnificus. It seems likely that under these
circumstances the ferrioxamine complex may serve as nature in-
tended, that is, as an available iron ligand for uptake by microbial
siderophore systems.
Despite the inconvenience of its use, long-​term studies of patients
receiving desferrioxamine for iron storage disease in homozygous
ß-​thalassaemia syndromes show that it is largely safe; moreover,
desferrioxamine improves cardiac function and life expectancy and
arrests hepatic fibrosis in secondary haemochromatosis. The intro-
duction of desferrioxamine has also been associated with decreasing
rates of endocrine failure such as diabetes, hypothyroidism, and
hypoparathyroidism. Should these complications develop, prompt
replacement of deficient hormones (or vitamin D analogues in
hypoparathyroidism) should be introduced. Sex-​steroid hormone
replacement, for patients developing hypogonadotropic hypo-
gonadism, should improve growth, sexual development, bone
density, and self-​esteem.
Deferasirox
For many patients, the orally active tridentate ferric iron che-
lator, deferasirox, offers an attractive option for once-​daily therapy
without the discomfort and limitations of continuous subcutaneous
infusions. Once-​daily administration is effective as a consequence of
the long plasma half-​life of the chelator. Deferasirox is able to attain
sufficient trough concentrations to complex labile iron species that
are present in the plasma.
Deferasirox chelates the same pools of iron as desferrioxamine,
but—​unlike the latter—​the iron complex is excreted almost entirely
in the faeces. In Europe, the drug is indicated for the treatment of
chronic iron overload in adults and children over the age of 6 years
with thalassaemia major who receive frequent blood transfusions
(equivalent to >7 ml packed red cells/​kg per month). Deferasirox is
also licensed for other forms of transfusional iron overload in which
desferrioxamine is either contraindicated or inadequate.
Administration and dosing
As with desferrioxamine, the effective dose depends on the rate of
iron accumulation derived from the breakdown of transfused red
cells. For a patient receiving between 0.3 and 0.5 mg/​kg per day
of iron loading, a dose of 20 to 30 mg/​kg once daily is sufficient to
balance input and excretion. For patients in whom a negative iron
balance is needed as a result of pre-​existing marked iron overload,
or who have a transfusional iron loading rate estimated to be greater
than 0.5 mg/​kg per day, a dose of up to 40 mg/​kg per day may be
given. Dose increments up to this value can also be given in patients
whose serum ferritin concentrations fail to decrease, according to
the extent of iron overload (as judged by transfusion history and
serum ferritin concentration). Dose adjustments should be made at
intervals of 3–​6 months.
At 20 to 30 mg/​kg deferasirox per day, liver iron concentrations
and serum ferritin concentrations decrease without impaired safety
over a follow-​up period as long as 5 years in ß-​thalassaemia major
and sickle cell disease. Progressive removal of cardiac iron over
3 years of follow-​up has been shown: normalization of cardiac iron
over this period occurs in patients with moderate myocardial iron
loading (myocardial T2* of 10–​20 ms), and significant improvement
has been shown in patients with more severe myocardial T2* of be-
tween 6 and 10 ms. Improvement or stabilization in liver fibrosis has
been demonstrated in a 3-​year prospective study.
Deferasirox has recently been licensed for the treatment of
nontransfusional iron overload in ß-​thalassaemia intermedia pa-
tients where low doses of 5 to 10 mg, carefully titrated against serum
ferritin and liver iron, are effective and well tolerated.
A newer formulation of deferasirox, in a film-​coated tablet, is now
available; the correct dose of this formulation is 0.7 times the effective
dose of the deferasirox dispersible tablet because of its increased ab-
sorption. This formulation appears to be well-​tolerated and gener-
ally more acceptable to patients than the dispersible preparation.
Side effects and monitoring
Monitoring of baseline and monthly hepatic and renal function
(including tests for proteinuria) is required weekly for the first
month of treatment with deferasirox, or after dose increments, and
monthly thereafter. Modest increases in serum creatinine concen-
tration (about 30%) occur in about one-​third of patients, but these
rarely progress. Since acute kidney injury has been occasionally re-
ported, in patients whose serum creatinine concentrations rise, or
where the serum creatinine exceeds the upper healthy reference
limit, dose reduction, or temporary interruption is recommended.
Annual ear and eye examinations, as well growth parameters and
sexual development in children, should be carefully monitored, but
inner ear and retinal toxicities are very rare. Gastrointestinal effects
are common but usually mild to moderate and readily controlled.
About 10% of patients experience a skin eruption soon after starting
treatment: mild to moderate skin eruptions can initially controlled
by dose reduction, but with a severe skin reaction the treatment
must be stopped and—​after healing—​carefully reintroduced at a low
dose followed by cautious increments.
Deferiprone
Deferiprone, a bidentate oral iron chelator, has been licensed in
Europe and, more recently, the United States of America for treat-
ment of iron overload in patients with thalassaemia over 6 years of
age unable to tolerate desferrioxamine or where desferrioxamine
is contraindicated. This drug, of the hydroxypyridone class, has
a short plasma half-​life due to its rapid glucuronidation in the
section 22  Haematological disorders
5398
liver; it is used at a dose of 75 to 100 mg/​kg per day of body weight
daily in three divided doses. Iron is excreted almost entirely in
the urine.
Direct, truly randomized ‘head-​to-​head’ comparisons between
deferiprone and deferasirox monotherapy have yet to be undertaken.
Administration and dosing
Deferiprone monotherapy induces negative body iron balance in
about one-​third of patients with severe homozygous β-​thalassaemia
at 75 mg/​kg per day, with attendant reductions in serum ferritin
concentrations. In the remaining patients, where negative iron
balance is not achieved or maintained, desferrioxamine is often
added at conventional doses between twice and five times a week to
achieve this goal—​a stratagem referred to as combination therapy
(see ‘Combinations of iron-​chelating agents’). High doses (100 mg/​
kg per day) of deferiprone may be more effective at improving ab-
normal myocardial T2* than desferrioxamine alone when given at
standard doses five times a week.
Side effects and monitoring
Deferiprone may cause serious toxicity, including neutropenia and
the occasional incidence of agranulocytosis; weekly monitoring
of the neutrophil count is recommended. Before the drug was ap-
proved, there was controversy about the progression or lack of pro-
gression of liver fibrosis on this treatment.
Combinations of iron-​chelating agents
Combinations of iron-​chelating drugs, although not specifically li-
censed, have been widely used when the desired therapeutic effect
with monotherapy has not been not achieved, or where the patient
finds it difficult to adhere to chelation monotherapy with sufficient
frequency. In principle, combinations may work either by increasing
total exposure to chelating molecules or by true synergism, where
one chelator with rapid kinetic access to iron (e.g. deferiprone) shut-
tles iron onto a ‘sink’ chelator with higher iron affinity but slower
kinetic access (e.g. desferrioxamine).
The most commonly used combination has been deferiprone
with subcutaneous desferrioxamine, but using diverse regimens
and dosing, especially where iron balance was not achieved with
deferiprone monotherapy or when use of desferrioxamine was insuf-
ficient and thus ineffective. Subcutaneous desferrioxamine, given at
night, combined with daily deferiprone renders synergy improbable,
as with such a regimen the chelators are unlikely to coexist usefully
in the tissues. However, near 24-​h exposure to chelating molecules
can be achieved with this combination if desferrioxamine is given
every night. In a randomized prospective trial, this stratagem has
also been shown to be more effective at decreasing myocardial
iron deposits (estimated by MRI (T2*)), than standard doses of
desferrioxamine as a monotherapy.
Deferasirox has been combined with desferrioxamine in patients
with severe myocardial siderosis (T2* 5–​10 ms): it brought about
rapid removal of iron from the liver as well as the heart, and deteri-
oration of left ventricular function was arrested. In one randomized
clinical study, deferiprone combined with deferasirox was reported
to be effective at removing heart and liver iron.
No new toxicities have so far been reported with the combined
therapies.
Management—​other aspects of care
The single most important aspect of care is adherence to iron che-
lation therapy and monitoring, especially for infants and other
young patients with iron-​loading anaemias such as ß-​thalassaemia.
Regular attendance of special clinics is advisable so that wide-​
ranging professional support from familiar personnel can reinforce
medical care delivered with attention to continuity and nurturing
independence.
In ß-​thalassaemia major, the pre transfusion blood haemoglobin
concentration should be maintained between 95 and 105 g/​litre with
a mean interval concentration of 120 g/​litre. If transfusion require-
ments increase in the context of an enlarging spleen, intensification
of the transfusion regimen may decrease spleen size, and this ultim-
ately will reduce the transfusion requirement. In patients suffering
from ß-​thalassaemia, except as a last resort, the high risk of throm-
bosis and infection normally should preclude splenectomy.
Patients with secondary iron overload should be monitored not
only for the progression of their iron storage as determined by
parameters of iron metabolism, but also clinically for the presence
of iron-​mediated tissue injury. Monitoring the effectiveness of che-
lation treatment is most frequently undertaken by serial determin-
ation of serum ferritin concentrations. Trends in ferritin identify
under-​ or overtreatment and also identify poor adherence to therapy
at an early point. The relative value of liver iron determinations and
cardiac T2* MRI were discussed earlier in this chapter. Regular clin-
ical monitoring and assessment of cardiac, hepatic, and endocrine
function assist in the assessment of iron storage disease and thera-
peutic efficacy of iron chelation therapy. Regular echocardiography
and electrocardiography are also essential aspects of management
and therapeutic monitoring.
Endocrine disease
Infantilism with hypogonadotropic hypogonadism are frequent
manifestations in patients with secondary iron storage disease and
contribute greatly to psychosocial difficulties and social exclusion,
as well as depression, in adolescents and children. Prompt diagnosis
and institution of appropriate sex hormone replacement is a crit-
ical component of holistic care. Consideration of puberty induction
using recombinant gonadotropins and advice about later fertility
management are key to the well-​being of these patients, particularly
since long-​term survival is increasingly a reality of contemporary
haematological care.
Hormone determinations and careful clinical monitoring are
essential to search for the presence of other aspects of endocrine
failure, including hypoparathyroidism and adrenocortical failure,
which may be very difficult to detect but critically important to treat.
Vigilance should be maintained for the development of diabetes
mellitus.
Cardiac disease
Decreasing ejection fraction of the left ventricle usually precedes the
functional complications of myocardial iron overload; it thus carries
a poor prognosis and mandates urgent intensification of the iron-​
chelating regimen, often to include use of continuous infusion of
desferrioxamine.
Desferrioxamine given continuously through a permanent
indwelling portable catheter safely sited in the superior vena cava,
22.6.4  Iron metabolism and its disorders
5399
with careful attention to preventing sepsis, is a satisfactory method
for securing reversal of cardiac disease in high-​risk patients with
serum ferritin concentrations that persist at greater than 2500 µg/​
litre or in whom hepatic iron concentrations exceed 1.5% of dry liver
weight. Histological studies in postmortem cardiac tissue obtained
from patients with severe cardiac iron deposition and heart failure
leading to death or transplantation showed limited interstitial fi-
brosis and no evidence of the extensive replacement cardiac fibrosis
that characterized the disease before chelation therapy; this accords
with the view that heart failure associated with cardiac haemo-
chromatosis is potentially reversible.
Continuous intravenous infusions of desferrioxamine not ex-
ceeding 50 to 60 mg/​kg daily are now recommended for patients in
whom the left ventricular function has deteriorated below reference
values or where there is evidence of very severe myocardial iron
loading (T2* ≤6 ms). This regimen is often administered with sup-
plemental deferiprone at standard doses to accelerate the rate of iron
removal. High-​dose intravenous infusions may cause unacceptable
toxic injury, especially in the retina and inner ear. Improved out-
comes occur with the use of anticoagulation induced by warfarin,
and scrupulous attention to cutaneous needle resiting and skin care
is necessary if the risk of thrombosis and complicating infections is
to be kept to a minimum.
Skeletal disease
Patients with stunted growth and infantilism frequently develop
skeletal disease beyond that related to their expanded bone marrow,
and investigations should be carried out to search for osteopenia and
osteoporosis for which additional therapy will be needed. Bone dis-
ease and growth arrest may be caused by the overenthusiastic use of
desferrioxamine in young infants: the daily dose of desferrioxamine
should be reduced to below 40 mg/​kg, which usually restores normal
growth velocity.
Pregnancy
Desferrioxamine is not recommended for use in during pregnancy,
but despite this many successful pregnancies have been reported
without fetal injury. Where possible, the drug should be avoided
during the middle trimester and should almost certainly be avoided,
because of unknown teratogenicity, in early pregnancy or at the time
of any planned conception. Nonetheless, it may be reasonable to re-
start desferrioxamine therapy in the final trimester of pregnancy if
the risks to the mother from iron storage disease are high.
No information is available on the use of deferasirox or deferiprone
and these drugs cannot be recommended in pregnancy until more
experience is forthcoming.
Psychological matters
Psychological difficulties, sometimes accompanied by disturbed
behaviour, are prevalent in children and adolescents receiving
iron-​chelation therapy. Transfusion for chronic anaemias and ap-
propriate counselling is needed over long periods to build trust with
the patients and their families, thereby to support engagement with,
and adherence to, this essential treatment.
Prognosis
The principal causes of death in secondary iron storage disease are car-
diac failure and arrhythmia, endocrine failure, and the consequences
of diabetes mellitus, infection, and hepatocellular carcinoma. When
associated with transfusion therapy and intestinal hyperabsorption
of iron in the chronic anaemias with dyserythropoiesis, secondary
haemochromatosis is rapidly fatal unless it is treated. However, the
outcome of iron storage disease in patients with chronic anaemia
is now greatly improving, with enhanced life quality and duration.
Of patients with β-​thalassaemia major who are unable to
comply with iron chelation therapy, less than one-​third survive to
25 years of age. One study has indicated that 95% of patients with
β-​thalassaemia who administer desferrioxamine subcutaneously
more than 250 times each year will survive to 30 years, whereas only
12% of those who do not achieve this dosing will survive to that age.
In the United Kingdom, the overall survival is 50% at 35 years; but
the actuarial survival at 40 years was 80% in more than 100 patients
treated at one specialist centre.
Continuous intravenous desferrioxamine can reverse life-​
threatening arrhythmias in cardiac iron overload and also improve
or reverse left ventricular or biventricular heart failure in most pa-
tients. The actuarial survival of patients with β-​thalassaemia and
life-​threatening iron-​storage disease has been reported to be greater
than 60% at 13 years when so treated, emphasizing the benefits of
care administered at a dedicated treatment centre. From 1980 to
1999, there were 12.7 deaths from all causes per 1000 patient-​years.
In 2000 to 2003, the death rate from all causes fell significantly to 4.3
per 1000 patient-​years and mortality attributable to iron overload
decreased from 7.9 to 2.3 deaths per 1000 patient-​years. Moreover,
effective iron chelation reduces the frequency of hypogonadism,
diabetes, and growth retardation.
As cardiac disease is becoming less frequent in patients with thal-
assaemia major, many are now reaching their sixth decade of life.
The risks of liver disease, including hepatocellular carcinoma, re-
main to be determined. Delayed consequences on health of other
potential sequelae, such as endocrine failure, should also benefit
from better treatment, but the consequences of delayed puberty or
infantilism are hard to shake off in the surviving, late-​treated adult.
Chelation therapy improves the quality of life as well as survival in
β-​thalassaemia and has salutary effects that are comparable in other
forms of transfusional iron overload. An important aspect of the
new chelation regimens, either as monotherapy or in combination,
will be to determine whether it will be possible safely to achieve
lower serum ferritin concentrations than those typically obtained
with desferrioxamine monotherapy. Since this may further reduce
the adverse consequences of iron overload, the best means to opti-
mize contemporary therapy still requires intensive clinical explor-
ation and comprehensive evaluation of the effects on quality of life
measures.
Future perspectives
Chelation is an effective and proven therapy but the optimal solution
for iron overload would be its prevention. While regular transfusion
and iron chelation are the mainstay of treatment for severe forms
of thalassemia, this is in effect a therapeutic supportive stratagem
which does not definitively address the underlying pathophysiology
of the diseased marrow and persistent drive to toxic iron loading.
Patients with this disease still have large unmet needs, which im-
pair their capacity for physical fulfilment and an enriching quality
of life that is free of the need for intensive medication and clinical
monitoring.
section 22  Haematological disorders
5400
Cell and genetic therapies
Haematopoietic stem cell transplantation and lentiviral-​mediated
haematopoietic stem cell gene therapy are attractive developments
for definitive correction of some forms of marrow disease (e.g. the
thalassaemia syndromes and sickling disorders). However, these
emerging treatments are not available worldwide and are often com-
plicated by the pre-​existing effects of ineffective erythropoiesis, es-
tablished iron overload, and transfusion-​related immunization. In
particular, these prior conditions increase the risk of infection and in
particular susceptibility to cardiotoxic marrow conditioning drugs.
Several trials of gene transfer using third-​generation lentiviral vec-
tors to correct haemoglobinopathies, including sickle cell anaemia and
ß-​thalassaemia syndromes, are underway. These developments reflect
ambitious desires to introduce a single curative step which would be
superior to haematopoietic stem cell (bone marrow) transplantation.
In June 2019, the EMA conditional approved Zynteglo, a therapy based
on re-infusing endogenous CD34+ stem-cells transduced ex-vivo with
the wild type beta-globin gene in transfusion-dependent patients who
are 12 years or older. The main limit to gene therapy seems to be the con-
ditioning regimen, and at present, ex vivo gene transfer to endogenous
stem cells does not avoid this requirement. Experimental gene editing
in animal models is encouraging, but routine use in human recipients
has yet to be proven safe, specific, and effective.
Other potential treatments
Identification of pathological disturbances in dyserythropoiesis that
drive iron overload immediately point to alternative targets for the
treatment of this disease. The experimental agents under investiga-
tion are based on contemporary molecular understanding of iron
homeostasis and its relationship to erythropoiesis. Specific treat-
ments in late-​phase clinical exploration include (1) the clinical use of
selective Janus kinase 1/​2 inhibitors, (2) parenteral administration of
long-​acting minihepcidins, (3) activin receptor type IIA and/​or IIB
ligand trap biological agents that are predicted to attenuate patho-
logical SMAD2/​3 activation in erythroid precursors, and (4) puta-
tive inhibitors of erythroferrone or its release. In a cognate field, the
anaemia of renal failure due to erythropoietin deficiency, a novel
HIF prolyl hydroxylase inhibitor (roxadustat) effectively treats the
anaemia, with improved iron utilization, hepcidin concentrations
and erythropoeitin—without the increased risk of stroke or death
associated with exogenous erythropoetin therapy.
JAK2 inhibitors in β-​thalassemia
Preclinical studies using convincing disease models in living rodents
have shown that a JAK2 inhibitor markedly decreased spleen size and
attenuated ineffective erythropoiesis, thus clinical use of JAK2 inhibi-
tors may correct splenomegaly and obviate the need for splenectomy
and frequent blood transfusion, and with these effects management of
the iron-​loading anaemia is likely to be greatly improved. Ruxolitinib, a
JAK2 inhibitor, has shown modest effects on spleen size in small phase
2 studies in patients with transfusion-dependent thalassaemia.
Activins, TGFβ signalling, and hepcidin regulation
Increased haemoglobin concentrations were a chance finding in ex-
ploratory clinical studies of ACE-​536, an activin receptor II ligand trap,
conducted in healthy volunteers and women with postmenopausal
osteoporosis. The observation led to a careful review of the role of
activins in the development of the bone marrow and stimulated an
examination of the haematological effects of these agents in disordered
erythropoiesis. Two activin receptor II ligand traps (sotatercept and
luspatercept) that block GDF-​11, and ACE-​536 (a recombinant pro-
tein containing a modified activin receptor type IIB), are being inves-
tigated for the treatment of ineffective erythropoiesis in thalassaemia
and myelodysplasia. ACE-​536 promotes late-​stage erythroid differen-
tiation by binding to TGFβ superfamily ligands, thereby inhibiting sig-
nalling through the SMAD2/​3 transcription factors.
The challenge
Credible scientific advancement in this field is long overdue, but with
continuing refinement of practice, the availability and experience of
oral iron chelators, the outlook for patients with otherwise fatal iron-​
loading anaemias is improving. Most patients are located in resource-​
poor regions, but even where access to emerging molecular therapies
is restricted, patients with limited capacity to pay for novel agents
can look forward to increasingly competitive pricing of the second-​
generation oral iron-​chelating drugs. The ultimate challenge will be
to match longer survival with improvement in quality of life.
FURTHER READING
Anderson GJ, et al. (2005). Mechanisms of haem and non-​haem iron
absorption: lessons from inherited disorders of iron metabolism.
Biometals, 18, 339–​48.
Anderson LJ, et al. (2004). Myocardial iron clearance during reversal of
siderotic cardiomyopathy with intravenous desferrioxamine: a pro-
spective study using T2* cardiovascular magnetic resonance. Br J
Haematol, 127, 348–​55.
Andrews NC, Schmidt PJ (2007). Iron homeostasis. Annu Rev Physiol,
69, 69–​85.
Auerback M, Adamson JW (2016). How we diagnose and treat iron defi-
ciency anemia. Am J Hematol, 91, 31–​8.
Babitt JL, et  al. (2006). Bone morphogenetic protein signaling by
hemojuvelin regulates hepcidin expression. Nat Genet, 38, 531–​9.
Bain BJ, Bates IM, Laffan MA (2016). Dacie and Lewis, Practical
Haematology, 12th edition. Elsevier Health Sciences, London.
Baksi AJ, Pennell DJ (2014). Randomized controlled trials of iron che-
lators for the treatment of cardiac siderosis in thalassaemia major.
Front Pharmacol, 5, 217.
Benoist B, et al. (2008). Worldwide prevalence of anaemia 1993–​2005.
World Health Organization, Geneva.
Borgna-​Pignatti C, et al. (2004). Survival and complications in patients
with thalassemia major treated with transfusion and deferoxamine.
Haematologica, 89, 1187–​93.
Bothwell T, et al. (1989). Nutritional iron requirements and good iron
absorption. J Intern Med, 226, 357–​65.
Bowes O, et al. (2014). Hereditary hyperferritinaemia cataract syn-
drome. Lancet, 383, 1520.
Breda L, Rivella S (2014). Modulators of erythropoiesis:  emerging
therapies for hemoglobinopathies and disorders of red cell produc-
tion. Hemato Oncol Clin North Am, 28, 375–​86.
Bryant BJ, et al. (2013). Ascertainment of iron deficiency and depletion
in blood donors through screening questions for pica and restless
legs syndrome. Transfusion, 53, 1637–​44.
Camaschella C, Nai A (2016). Ineffective erythropoiesis and regula-
tion of iron status in iron loading anaemias. Br J Haematol, 172,
512–​23.
Cappellini MD, Motta I (2017). New therapeutic targets in transfusion-​
dependent and -​independent thalassemia. Hematology, 2017, 278–​3.
22.6.4  Iron metabolism and its disorders
5401
Cappellini MD, et al. (2011). Iron chelation with deferasirox in adult
and pediatric patients with thalassemia major: efficacy and safety
during 5 years’ follow-​up. Blood, 118, 884–​93.
Coffey R, Ganz T (2017). Iron homeostasis: an anthropocentric per-
spective. J Biol Chem, 292, 12727–​34.
Cohen AR, et al. (2008). Effect of transfusional iron intake on response
to chelation therapy in beta-​thalassemia major. Blood, 111, 583–​7.
Cox TM (1998). Iron salts, iron–​dextran complex and iron–​sorbitol
citrate. In: Dollery CT (ed.) Therapeutic drugs: a clinical pharmaco-
poeia, 2nd edition, vol. 2, pp. 178–​83. Baillière Tindall, Edinburgh.
Dallman PR (1989). Iron deficiency: does it matter? J Intern Med, 226,
367–​72.
Di Maggio R, Maggio A (2017). The new era of chelation treatments: ef-
fectiveness and safety of 10 different regimens for controlling iron
overloading in thalassaemia major. Br J Haematol, 178, 676–​88.
Donker AE, et al. (2016). Iron refractory iron deficiency anemia: a het-
erogeneous disease that is not always iron refractory. Am J Hematol,
91, E482–​90.
Drakesmith H, Nemeth E, Ganz T (2015). Ironing out ferroportin. Cell
Metab, 22, 777–​87.
Enns CA, Zhang AS (2009). Iron homeostasis: recently identified pro-
teins provide insight into novel control mechanisms. J Biol Chem,
284, 711–​15.
Ganz T (2013). Systemic iron homeostasis. Physiol Rev, 93, 1721–​41.
Heeney MM, Finberg KE (2014). Iron-​refractory iron deficiency an-
emia (IRIDA). Hematol Oncol Clin North Am, 28, 637–​52.
High KA, Roncarolo MG (2019). Gene Therapy. N Engl J Med, 381,
455–64.
Hurrell RF (2007). Iron fortification: its efficacy and safety in relation
to infections. Food Nutr Bull, 28 Suppl 4, S585–​94.
Johnson BE (1990). Pica. In: Walker HK, Hall WD, Hurst JW (eds)
Clinical methods: the history, physical, and laboratory examinations,
3rd edition, pp. 709–​10. Butterworths, Boston, MA. https://​www.
ncbi.nlm.nih.gov/​books/​NBK255/​
Kaplan J (2019). Roxadustat and the anemia of chronic renal failure.
N Engl J Med. doi: 10.1056/NEJMe1908978.
Kautz L, et al. (2014). Identification of erythroferrone as an erythroid
regulator of iron metabolism. Nat Genet, 46, 678–​84.
Lopez A, et al. (2016). Iron deficiency anaemia. Lancet, 387, 907–​16.
Magio A (2007). Light and shadows in the chelation treatment of
haematological diseases. Br J Haematol, 138, 407–​21.
McCance RA, Widdowson EM (1937). Absorption and excretion of
iron. Lancet, ii, 680–​4.
McCann JC, Ames BN (2007). An overview of evidence for a causal
relation between iron deficiency during development and deficits in
cognitive or behavioral function. Am J Clin Nutr, 85, 931–​45.
McCulley L, et al. (2016). Reports to FDA of fatal anaphylaxis asso-
ciated with intravenous iron products. Am J Hematol, 91, E496–​7.
McGann PT, Nero AC, Ware RE (2017). Clinical features of β-​
thalassemia and sickle cell disease. Adv Exp Med Biol, 1013, 1–​26.
McCleod C, et al. (2009). Deferasirox for the treatment of iron overload
associated with regular transfusions (transfusional haemosiderosis)
in patients suffering with chronic anaemia: a systemic review and
economic evaluation. Health Technol Assess, 13, iii–​iv, ix–​xi, 1–​121.
Melis MA, et al. (2008). A mutation in the TMPRSS6 gene, encoding
a transmembrane serine protease that suppresses hepcidin pro-
duction, in familial iron deficiency anemia refractory to oral iron.
Haematologica, 93, 1473–​9.
Miceli MH, et al. (2006). Iron overload is a major risk factor for severe
infection after autologous stem cell transplantation: a study of 367
myeloma patients. Bone Marrow Transplant, 37, 857–​64.
Modell B, et al. (1982). Survival and desferrioxamine in thalassaemia
major. Br Med J, 284, 1081–​4.
Modell B, Darlison M (2008). Global epidemiology of haemoglobin
disorders and derived service indicators. Bull World Health Organ,
86, 480–​7.
Motta I, Scaramellini N, Cappellini MD (2017). Investigational drugs
in phase I and phase II clinical trials for thalassemia. Expert Opin
Investig Drugs, 26, 793–​802.
Namwanje M, Brown CW (2016). Activins and inhibins: roles in de-
velopment, physiology, and disease. Cold Spring Harb Perspect Biol,
8, a021881.
Neufeld E (2006). Oral chelators deferasirox and deferiprone for
transfusional iron overload in thalassemia major:  new data, new
questions. Blood, 107, 3436–​41.
Olivieri NF, et al. (1994). Survival in medically treated patients with
homozygous β-​thalassaemia. N Engl J Med, 331, 574–​8.
Olivieri NF, et al. (1998). Long-​term safety and effectiveness of iron-​
chelation therapy with deferiprone for thalassemia major. N Engl J
Med, 339, 417–​23.
Ottomano C, Franchini M (2012). Sports anaemia: facts or fiction?
Blood Transfus, 10, 252–​4.
Pasricha SR, McHugh K, Drakesmith H (2016). Regulation of hepcidin
by erythropoiesis: the story so far. Annu Rev Nutr, 36, 417–​34.
Pasricha SR, et al. (2013). Control of iron deficiency anemia in low-​
and middle-​income countries. Blood, 121, 2607–​17.
Percy L, Mansour D, Fraser I (2017). Iron deficiency and iron defi-
ciency anaemia in women. Best Pract Res Clin Obstet Gynaecol,
40, 55–​67.
Pennell DJ, et al. (2012). Randomized controlled trial of deferiprone or
deferoxamine in beta-​thalassemia major patients with asymptom-
atic myocardial siderosis. Blood, 107, 3738–​44.
Pippard MJ, Weatherall DJ (1984). Iron absorption in iron-​loading
anaemias. Haematologia, 17, 407–​14.
Pippard MJ, Weatherall DJ (2000). Oral iron chelation therapy for thal-
assaemia: an uncertain scene. Br J Haematol, 111, 2–​5.
Poggiali E, et al. (2015). The role of TMPRSS6 polymorphisms in iron
deficiency anemia partially responsive to oral iron treatment. Am J
Hematol, 90, 306–​9.
Porter JB (2001). Practical management of iron overload. Br J
Haematol, 115, 239–​52.
Porter JB (2006). Deferasirox: an effective once-​daily orally active iron
chelator. Drugs Today, 42, 623–​37.
Porter JB (2007). Concepts and goals in the management of
transfusional iron overload. Am J Hematol, 82 Suppl 12, 1136–​9.
Porter JB, Garbowski M (2014). The pathophysiology of transfusional
iron overload. Hematol Oncol Clin North Am, 28, 683–​70.
Rampton D, et  al. (2014). Hypersensitivity reactions to intra-
venous iron:  guidance for risk minimization and management.
Haematologica, 99, 1671–​6.
Roy A, et al. (2018). Pica is prevalent and strongly associated with iron
deficiency among Hispanic pregnant women living in the United
States. Appetite, 120, 163–​70.
Rund D (2016). Thalassemia 2016: modern medicine battles an ancient
disease. Am J Hematol, 91, 15–​21.
Shaker M, et al. (2009). An economic analysis of anemia prevention
during infancy. J Pediatr, 154, 44–​9.
Szebeni J, et al. (2015). Hypersensitivity to intravenous iron: classifica-
tion, terminology, mechanisms and management. Br J Pharmacol,
172, 5025–​36.
Taher AT, Weatherall DJ, Cappellini MD (2018). Thalassaemia. Lancet,
391, 155–​67.

22.6.5 Anaemia of inflammation 5402 Sant- Rayn Pas
22.6.5 Anaemia of inflammation 5402 Sant- Rayn Pasricha and Hal Drakesmith
section 22  Haematological disorders
5402
Taher AT, et al. (2012). Deferasirox reduces iron overload significantly
in nontransfusion-​dependent thalassemia:  1-​year results from a
prospective, randomized, double-​blind, placebo-​controlled study.
Blood, 120, 970–​7.
Tanner MA, et al. (2007). A randomized, placebo-​controlled, double-​
blind trial of the effect of combined therapy with deferoxamine and
deferiprone on myocardial iron in thalassemia major using cardio-
vascular magnetic resonance. Circulation, 115, 1876–​84.
Vlachodimitropoulou Koumoutsea E, Garbowski M, Porter J (2015).
Synergistic intracellular iron chelation combinations: mechanisms
and conditions for optimizing iron mobilization. Br J Haematol, 170,
874–​83.
Woywodt A, Kiss A (2002). Geophagia: the history of earth-​eating.
J R Soc Med, 95, 143–​6.
Zhang DL, et al. (2018). Erythrocytic ferroportin reduces intracel-
lular iron accumulation, hemolysis, and malaria risk. Science, 359,
1520–​3.
22.6.5  Anaemia of inflammation
Sant-​Rayn Pasricha and Hal Drakesmith
ESSENTIALS
The anaemia of inflammation is one of the commonest causes
of anaemia, particularly in the elderly and in those with extensive
comorbidities. Dysregulation of the hepcidin axis is central to its
pathophysiology, with upregulation of hepcidin both limiting iron
absorption from the gut and preventing effective mobilization of iron
stores for erythropoiesis. However, hepcidin-​independent mechan-
isms of the anaemia of chronic disease exist and must be considered
in assessing the underlying aetiology of otherwise unexplained an-
aemia. Interactions between the anaemia of chronic disease and
other forms of anaemia are common.
The management of the anaemia of inflammation rests in large
part on the treatment of the underlying disease(s) where pos-
sible. In cases where anaemia-​specific treatment is required, iron
supplementation—​particularly intravenous iron—​may have a role.
There is limited evidence for the use of erythropoiesis-​stimulating
agents in certain circumstances.
Introduction
Anaemia of inflammation (synonymous with anaemia of chronic
disease) is the most common cause of anaemia among hospitalized
patients and in patients with acute and chronic medical conditions.
In patients with complex morbidities, anaemia of inflammation can
be superimposed upon or complicated by other causes of anaemia.
In particular, there is an important overlap between mechanisms
involved with anaemia of inflammation and anaemia observed in
patients with renal impairment. Although generally mild in se-
verity, anaemia of inflammation may be symptomatic and delay
convalescence in clinical conditions. Anaemia is associated with
increased mortality in a range of diseases. Diagnosis of anaemia of
inflammation depends on a high degree of clinical suspicion and
exclusion of other causes of anaemia, and can usually be suggested
by a simple panel of investigations that typically reveal a normocytic
normochromic anaemia associated with reduced serum iron and
normal transferrin levels, along with evidence of inflammation. In
selected cases, more invasive tests such as bone marrow examination
may be necessary. Decisions to treat anaemia of inflammation re-
quire clinical judgement, as resolution of the underlying condition
should result in improvement to the anaemia, and specific inter-
ventions directed at ameliorating the anaemia may only be tempor-
arily effective. However, improved understanding of the molecular
pathogenesis of this condition has led to the development of dir-
ected therapies that could improve treatments in the future.
Aetiology
Anaemia is a common complication of a range of autoimmune, in-
fectious, malignant, and other medical and surgical processes which
cause systemic inflammation.
Any condition resulting in systemic inflammation sufficient to
produce elevation of hepcidin expression (see ‘Pathogenesis/​path-
ology’ for details on the role of hepcidin) can produce anaemia
of inflammation. Box 22.6.5.1 presents a list of conditions that
are associated with anaemia of inflammation. In many of these
conditions, complex systemic illness results in multiple mechan-
isms contributing to anaemia simultaneously; in particular, renal
Box 22.6.5.1  Medical conditions associated with anaemia
of inflammation
Infection
•	 Bacterial (localized, e.g. osteomyelitis, wound infections); dissemin-
ated (e.g. endocarditis, septicaemia)
•	 Parasitic (e.g. malaria)
•	 Fungal, especially if disseminated
Malignancy
•	 Solid tumours
•	 Haematological diseases (e.g. lymphoma)
Autoimmune diseases
•	 Active rheumatoid arthritis
•	 Systemic lupus erythematosus
•	 Polymyalgia rheumatica
•	 Inflammatory bowel disease
•	 Vasculitis
Renal
•	 Chronic kidney disease or acute kidney injury
Cardiac
•	 Congestive heart failure
Sterile inflammation and tissue necrosis
•	 Sterile abscess
•	 Wound healing
•	 Trauma
•	 Post surgery (e.g. cardiac surgery)
•	 Head injury
22.6.5  Anaemia of inflammation
5403
impairment may diminish the erythropoietin response, exacer-
bating the insufficient bone marrow response. Iron deficiency may
coexist with anaemia of inflammation, especially in hospitalized
patients who have undergone recent blood loss or even repeated
blood sampling.
Anaemia of inflammation and anaemia of chronic disease are
synonymous conditions, while functional iron deficiency is an over-
lapping concept that refers to a state in which, despite the presence
of stainable iron in the bone marrow and a ferritin level above the
lower threshold of normal, iron incorporation into the erythroblast
is inadequate; this situation is also seen, for example, in patients with
chronic renal impairment.
Epidemiology
Anaemia is common among patients with comorbid inflammatory
conditions, especially among hospitalized patients and the elderly.
It is probably the most common cause of anaemia among hospital-
ized patients. However, the specific proportion attributable to an-
aemia of inflammation compared with other causes of anaemia in
these patients has not always been carefully quantified. The preva-
lence and severity of anaemia tend to increase with the severity of
the underlying condition and the age of the patient group, and hence
estimates of prevalence vary widely. For example, estimates of the
prevalence of anaemia in patients with acute or chronic inflam-
mation range from 18 to 95%; in patients with cancer from 30 to
77%; and in patients with autoimmune conditions from 8 to 71%.
Anaemia of inflammation is considered the second most common
cause of anaemia in the world after iron deficiency anaemia. Thus,
overall, inadequate iron supply to the erythroblast either due to ab-
solute iron deficiency or inadequate availability accounts for the vast
majority of cases of anaemia worldwide.
Pathogenesis/​pathology
Anaemia of inflammation is mediated via both direct (hepcidin-​
independent) and indirect (hepcidin-​dependent) effects of inflam-
mation on erythropoiesis (Fig. 22.6.5.1).
Hepcidin is the master regulator of systemic iron homeostasis. It
is chiefly expressed by the liver and acts to inhibit iron export from
cells (especially duodenal enterocytes and splenic red pulp macro-
phages). Therefore, low hepcidin levels promote increased plasma
iron levels and availability of iron for erythropoiesis by facili-
tating iron export from intestinal cells and macrophages, and thus
cytokines
cytokines
Erythropoeisis
inflammation
Other
tissues
flow of iron
effects of
inflammation
iron-loaded
Transferrin
FPN
Mphage
HEPCIDIN
FPN
Fig. 22.6.5.1  Pathogenesis of anaemia of inflammation. Normally about 1 to 2 mg of iron is
absorbed from the diet every day, and is taken through the circulation bound to the dedicated
iron-​carrying protein transferrin. Most iron is delivered to the erythron for incorporation into
the haem of developing erythrocytes. The lifespan of human red blood cells is about 120 days,
after which senescent cells are phagocytosed by macrophages that degrade their contents, digest
haem, and recycle liberated iron into plasma. About 25 mg Fe is recycled per day. The transfer of
iron from duodenal enterocytes into the circulation and the recycling of iron by macrophages into
circulation are both achieved by the unique iron exporter protein ferroportin (‘FPN’). Hepcidin
inhibits ferroportin, thus controlling iron homeostasis. Inflammation is associated with release
of cytokines, some of which (especially IL-​6) induce production of hepcidin by the liver. Chronic
inflammation leads to persistently high levels of hepcidin that over time decrease the amount
of iron available for erythropoiesis. Other inflammatory cytokines impair erythropoietin (EPO)
production by the kidney and the bone marrow response to EPO. Together these hepcidin-​
dependent and hepcidin-​independent effects lead to the anaemia of inflammation.
section 22  Haematological disorders
5404
enhanced iron absorption from the diet and recycling of iron lib-
erated from senescent phagocytosed red cells. Conversely, elevated
hepcidin levels prevent iron export from enterocytes into serum,
impairing dietary iron absorption, and inhibit iron export from
macrophages, preventing iron recycling. Notably, most (>95%)
iron requirements for erythropoiesis are met by recycled iron
from erythroblasts recycled from macrophages, and hence preven-
tion of macrophage iron release rapidly results in iron-​restricted
erythropoiesis.
Hepcidin levels are directly regulated by liver iron and levels of
circulating transferrin-​bound iron; increased iron elevates hepcidin
while iron deficiency suppresses it, achieving homeostatic regula-
tion of iron. Hepcidin is also suppressed by increased erythropoi-
esis (e.g. stress erythropoiesis, administration of erythropoietin,
post-​phlebotomy, hypoxia, or in conditions such as thalassaemia).
Importantly, hepcidin expression is directly increased by inflam-
mation, mediated especially by interleukin (IL)-​6 and also by other
cytokines including IL-​22 and type I  interferon (Fig. 22.6.5.1).
Thus, any condition that produces an acute phase response may
raise hepcidin levels, resulting in restriction of iron recycling, with
reductions in serum iron and availability of iron for erythropoiesis.
Inflammation-​induced elevations in hepcidin thus also explain the
hypoferraemia (low plasma iron levels) of infection.
Anaemia of inflammation may also be mediated via hepcidin-​
independent mechanisms. The erythropoietin response to an-
aemia may be blunted in patients with anaemia of inflammation;
this may be mediated directly by inflammatory cytokines or by
impaired kidney function. Inflammation may also directly im-
pair erythropoiesis, and cytokines such as IL-​6, tumour necrosis
factor-​α, and even hepcidin itself have been shown to directly im-
pair erythropoiesis, independently of local or systemic effects on
iron handling.
Clinical features
Anaemia of inflammation typically presents with a mild to mod-
erate anaemia (haemoglobin concentration >80–​90 g/​litre)
although in some cases it may be more severe, especially if exacer-
bated by other causes of anaemia. Patients have existing or recently
diagnosed additional medical conditions. It presents with the
usual symptoms of anaemia (e.g. fatigue, lethargy, and exertional
dyspnoea) and pallor. As it complicates existing comorbid medical
conditions or hospital admissions, it may present with a deterior-
ation in well-​being, functional performance, or clinical recovery
in an already complex patient with other reasons for reduced
performance status. For example, patients may experience an ex-
acerbation in fatigue, symptoms of heart failure such as exertional
dyspnoea, and hospitalized patients may become less able to par-
ticipate in rehabilitation and physiotherapy. As such, isolation of
the clinical effects of the anaemia from the underlying condition is
difficult and requires judgement and clinical monitoring. Indeed,
even a deterioration in a patient’s clinical status associated with
worsening anaemia may not be exclusively due to the anaemia it-
self if the underlying condition has simultaneously deteriorated,
making the patient feel more unwell. A trial of attempted specific
therapy for the anaemia (as discussed later) may be the only way
to distinguish between the effects of anaemia and the effects of
the underlying condition, and to develop a rationale for anaemia-​
specific therapy.
Differential diagnosis
In patients with complex medical conditions, anaemia of inflamma-
tion may coexist with, masquerade as, or be masked by a range of
other causes of anaemia. As many of these conditions have specific
therapies available, or portend serious deteriorations in the under-
lying condition that must be recognized, the clinician must consider
all relevant causes of anaemia (discussed in the following sections)
in each case and exclude these where necessary (Box 22.6.5.2).
Iron deficiency anaemia
Reduced iron stores may complicate cases where iron stores have be-
come depleted over time, for example, in patients with inflammatory
bowel disease with systemic inflammation but chronic gastrointes-
tinal bleeding, or in patients undergoing recurrent blood sampling
for laboratory tests or blood losses following medical procedures
(e.g. intensive care inpatients). Patients may have a more microcytic
blood picture, or have a ferritin level disproportionately low for the
degree of inflammation.
Renal failure
Impaired renal function may accompany systemic inflammation
in patients with any systemic illness; alternatively, acute kidney
Box 22.6.5.2  Important causes of anaemia in the unwell patient
Reduced red cell production
•	 Anaemia of inflammation
•	 Iron deficiency
•	 Medications:
—​	 Bone marrow failure
•	 Renal failure:
—​	 Acute kidney injury
—​	 Chronic renal failure
•	 Bone marrow failure:
—​	 Replacement by carcinoma, infection
—​	 Acute leukaemia
—​	 Aplastic anaemia, paroxysmal nocturnal haemoglobinuria
—​	 Myelofibrosis
•	 Aplastic sickle cell crisis
Impaired red cell survival
•	 Acute blood loss
•	 Haemolysis:
—​	 Idiopathic
—​	 Associated with medication
—​	 Associated with infection
—​	 Delayed transfusion reaction
•	 Microangiopathic anaemia:
—​	 Haemolytic uraemic syndrome/​thrombotic thrombocytopenic
purpura
—​	 Drugs
•	 Sickle cell crisis:
—​	 Haemolytic
—​	 Sequestration
22.6.5  Anaemia of inflammation
5405
injury and chronic renal failure may occur due to a range of
noninflammatory conditions. Erythropoietin deficiency is a con-
sistent feature of renal impairment, decreasing the differentiation
of haematopoietic stem cells towards the erythroid lineage and
hence suppressing the synthesis of new erythrocytes. Renal failure
may also exacerbate anaemia of inflammation through accumu-
lation of hepcidin due to impaired renal excretion. Furthermore,
erythropoietin deficiency and its sequela of reduced erythro-
poiesis may both independently reduce the production of bone
marrow-​derived hormones that suppress hepcidin, particularly
erythroferrone, resulting in inappropriately elevated hepcidin
levels. These processes all result in functional iron deficiency due
to impaired mobilization of macrophage-​sequestered or liver-​
stored iron for erythropoiesis.
Microangiopathic anaemia
Microangiopathic anaemias may develop de novo (e.g. in thrombotic
thrombocytopenic purpura or haemolytic uraemic syndrome) and
inadequate clinical evaluation and misdiagnosis as anaemia of in-
flammation will result in an adverse outcome for the patient. Patients
with these conditions are systemically unwell, with anaemia and
renal impairment. Clinical and biochemical evidence of haemolysis,
thrombocytopenia, and identification of red cell fragmentation on
the blood film will suggest the correct diagnosis. Microangiopathic
anaemia must be considered in every patient with anaemia and
thrombocytopenia. Patients with medical (e.g. malignancy, sepsis)
or surgical disease (e.g. trauma, leaking mechanical cardiac valves,
dysfunctional arteriovenous fistulae) and obstetric complications
(e.g. pre-​eclampsia) may also develop microangiopathic anaemia,
most severely disseminated intravascular coagulation. These condi-
tions are discussed in greater detail in Chapter 22.7.5.
Bone marrow failure
Bone marrow failure may complicate metastatic malignancy or dis-
seminated infection; patients may have pancytopenia.
Drugs
Various medications used to treat the underlying conditions asso-
ciated with anaemia of inflammation may also directly cause an-
aemia through a range of mechanisms. For example, antibiotics to
treat infection can also cause haemolytic anaemia (e.g. β-​lactams,
methyldopa) or bone marrow suppression (e.g. isoniazid), while
drugs used to treat autoimmune conditions (e.g. methotrexate), and
chemotherapy to treat malignancy, may also suppress erythropoiesis.
Clinical investigations
Investigations to diagnose anaemia of inflammation must be inter-
preted in the broader clinical context. Anaemia of inflammation can
usually be diagnosed with a limited panel of simple tests but may
occasionally require additional investigations. Newer investigations
may soon assist in the diagnosis. Findings characteristic of anaemia
of inflammation are summarized in Box 22.6.5.3.
Laboratory investigations
The underlying condition responsible for the inflammation should
be investigated as appropriate. Tests specific to determining the
presence of anaemia of inflammation are described in the following
sections.
Haematology
A full blood count and blood film examination will reveal a mild to
moderate anaemia (typically, haemoglobin concentration exceed­
ing 80 g/​litre), which is normochromic and normocytic; in some
cases, mild hypochromia or microcytosis may be seen. Evidence
of infection and inflammation may also be present: for example,
granulocytosis with a left shift (evidence of immature granulocytes
such as band forms) and evidence of toxic changes. On selected
automated analysers, measurement of the reticulocyte haemoglobin
(CHr) or percentage of hypochromic red blood cells (%Hypo) may
provide information about recent and medium-​term iron supply
to the bone marrow respectively. Patients with reduced CHr
have impaired incorporation of iron into erythroblasts, indicating
functional or absolute iron deficiency. Elevated %Hypo indicates
that iron stores, erythropoietic stimulation, and/​or inflammation
are impairing functional iron availability for erythropoiesis. The
erythrocyte sedimentation rate is elevated.
Clinical chemistry
Inflammatory markers such as C-​reactive protein and α1 glyco-
protein are usually elevated. Iron indices usually demonstrate
withholding of iron from the plasma—​serum iron and transferrin
saturation are low, with a normal transferrin level. Serum ferritin
levels are likely to be elevated due to its behaviour as an acute phase
protein. The soluble transferrin receptor (sTfR) is likely to be in the
normal range. Coexistent iron deficiency can be difficult to iden-
tify in individuals with anaemia of inflammation, but should be
suspected in patients with coexistent inflammation and ferritin con-
centrations below 100 to 200 μg/​litre, although clinical judgement is
required. Iron deficiency may also be suggested by elevation of the
sTfR or the sTfR/​log10 (ferritin) ratio. In selected cases (especially
those where treatment with erythropoietin is being considered),
measurement of serum erythropoietin levels may identify an inad-
equate response to the degree of anaemia, especially in patients with
coexisting renal failure. Measurement of serum hepcidin is now
readily available in research settings and may soon be accessible
Box 22.6.5.3  Results of investigations for anaemia
of inflammation
Haematology
•	 Mild anaemia
•	 Blood film:  mild normocytic (or mildly microcytic) red cells. Toxic
changes in granulocytes
•	 Erythrocyte sedimentation rate: elevated
Biochemistry
•	 Inflammatory markers (e.g. C-​reactive protein): elevated
•	 Iron indices:
—​	 Ferritin: elevated
—​	 Serum iron: reduced
—​	 Transferrin saturation: reduced
—​	 Soluble transferrin receptor: normal
•	 Erythropoietin: elevated or normal (suggesting inadequate response to
anaemia)
•	 Hepcidin: elevated
section 22  Haematological disorders
5406
in the clinic; patients with anaemia of inflammation have elevated
serum hepcidin concentrations.
Bone marrow aspiration
In selected complex cases, especially where multiple causes of an-
aemia may be present, examination of the bone marrow should be
considered. In particular, using Perls’ stain enables definitive detec-
tion of iron deficiency, while assessment of erythropoiesis enables
identification of erythroid hypoplasia and bone marrow failure.
Treatment
Control of the underlying disease remains the most important
strategy for treatment of anaemia of inflammation. Resolution of in-
flammation and recovery of red cell production can result in a tran-
sient, erythropoiesis-​mediated suppression of hepcidin, facilitating
iron recycling and supporting recovery of anaemia. The decision to
implement specific treatment directed at anaemia of inflammation
requires clinical judgement. In the majority of cases where the an-
aemia is mild and unlikely to be independently affecting the patient’s
clinical condition or well-​being, specific therapy for the anaemia it-
self is unnecessary. It is therefore imperative to consider whether the
anaemia is likely to be significantly affecting the performance and
quality of life of the patient, exacerbating other underlying medical
conditions (e.g. producing angina in patients with underlying car-
diovascular disease), or impairing recovery and capacity to partici-
pate in rehabilitation. This may be difficult to ascertain in complex
clinical situations. National and international clinical guidelines
exist for management of anaemia associated with chronic renal
failure, cancer-​related anaemia, and to advise on transfusion prac-
tice; however, evidence-​based guidelines for treatment of anaemia of
inflammation per se are lacking.
Although anaemia is correlated with impaired survival in many
conditions (cancer, acute cardiovascular events and chronic heart
failure, and renal failure), correction of anaemia has not been shown
to reverse mortality risk, and thus the primary rationale for allevi-
ation of anaemia is improvement of morbidity. The therapeutic goal
is to achieve the minimum haemoglobin concentration at which the
symptoms of anaemia are ameliorated. A trial of anaemia-​specific
therapy to assess the value of increased haemoglobin on patient
health in difficult cases may be warranted. Treatment is more likely
to be warranted in older patients, patients with cardiovascular or
respiratory disease, or patients with an acute deterioration in their
haemoglobin concentration.
Transfusion
Blood transfusions are generally safe and are widely used when cor-
rection of anaemia is relatively urgent. Patient blood management
guidelines recommend that the decision to transfuse should not be
solely based on a patient’s haemoglobin concentration, but should
also take account of the clinical condition. These guidelines sug-
gest that transfusions are appropriate in patients with haemoglobin
concentrations less than 70 g/​litre, although well-​compensated pa-
tients can be treated conservatively; in patients with haemoglobin
concentrations of 70 to 100 g/​litre, the decision should be based on
clinical features, especially the need to urgently resolve signs and
symptoms of anaemia. There is no evidence transfusion improves
mortality in patients in this group. Guidelines do not identify evi-
dence supporting a need for a different approach among the elderly
or in patients with respiratory or cerebrovascular disease. Patients
with haemoglobin concentrations exceeding 100 g/​litre do not gen-
erally require and in some cases (e.g. cases of acute coronary syn-
drome) may be harmed by transfusion.
Iron therapy
Patients with anaemia of inflammation have functional iron deficiency
(i.e. iron stores are sufficient but iron cannot be mobilized to reach
developing red blood cells). However, they may also have concomi-
tant absolute iron deficiency (e.g. due to prolonged impairment of
intestinal iron absorption), or iron losses due to acute (including sur-
gery and upper gastrointestinal stress ulceration) or chronic bleeding
(including frequent blood sampling). Oral iron supplementation is
infrequently recommended for iron therapy in anaemia of inflamma-
tion. Elevated hepcidin levels impair the efficiency of iron absorption.
Gastrointestinal side effects (especially in patients with intestinal dis-
eases) can also limit acceptability and adherence. However, oral iron
is cheap and can be useful in selected cases with motivated patients.
Increased understanding of the pathophysiology of anaemia of
inflammation, coupled with the recent arrival in the pharmacy of a
new generation of intravenous iron formulations, has led to renewed
interest in the use of parenteral iron. New products largely overcome
the limitations of previous parenteral high molecular weight iron dex-
tran formulations (especially risk of anaphylaxis), and enable high
doses of iron to be administered in single treatments. For example,
ferric carboxymaltose can be administered in high doses (up to 1000
mg) over 15 min, and carries a very low risk of allergic reaction, and
as such, premedication with antihistamines or steroids is unnecessary.
Intravenous iron is widely used in chronic renal failure to overcome
functional iron deficiency to facilitate erythropoiesis; it is also used in
patients with inflammatory bowel disease in whom iron losses may be
high due to gastrointestinal bleeding, and avoidance of the oral route
is preferred due to concerns about exacerbation of intestinal path-
ology. Intravenous iron has been shown to improve symptoms in pa-
tients with congestive cardiac failure. Systematic reviews indicate that
hospital patients receiving intravenous iron have an increased haemo-
globin concentration and reduced requirement for transfusion com-
pared with those receiving either no iron or oral iron.
Key concerns about the safety of intravenous iron include, in patients
receiving formulations comprising iron in a carbohydrate shell (e.g. iron
sucrose, polymaltose, and carboxymaltose), hypophosphataemia; this
is mediated by elevations in intact fibroblast growth factor-​23 causing
renal phosphate wasting, which has been associated with osteomalacia
in cases of recurrent use. Monitoring of serum phosphate, parathyroid
hormone, and vitamin D levels in patients receiving frequent dosing
may therefore be advisable. Several studies and one systematic review
have indicated an increased risk of infection in patients receiving par-
enteral iron. Although data are conflicting, intravenous iron should
probably not be administered to patients with confirmed or suspected
sepsis until the infection has been addressed.
Erythropoiesis-​stimulating agents
As well as controlling anaemia through direct stimulation of erythro-
poiesis, erythropoiesis-​stimulating agents (ESAs) may indirectly
benefit red cell production by improving access to iron, through
suppression of hepcidin levels. Amelioration of elevated hepcidin in

22.6.6 Megaloblastic anaemia and miscellaneous def
22.6.6 Megaloblastic anaemia and miscellaneous deficiency anaemias 5407 A.V. Hoffbrand
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5407
these conditions may mobilize iron from macrophage stores, easing
functional iron deficiency and facilitating erythropoiesis. ESAs are
widely used to treat patients with chronic kidney disease in whom
endogenous erythropoietin production is inadequate to maintain
erythropoiesis in the context of declining haemoglobin concentra-
tions. However, use of ESAs for routine treatment of anaemia of in-
flammation is no longer recommended, although these drugs may be
valuable in selected cases where serum erythropoietin levels have been
measured and are inappropriately low. The goal should be to gradually
increase haemoglobin concentrations to a target range between 100
and 120 g/​l, as rapid increases or higher levels have been associated
with cardiovascular risk among patients with renal failure and cancer.
Prognosis and outcome
Anaemia of inflammation will persist for as long as erythroblasts are
deprived of iron by the effects of hepcidin, or via direct cytokine-​
mediated suppression of erythropoiesis. Resolution or control of
the underlying condition should relieve these processes and facili-
tate erythropoiesis. Some emerging data suggest that following re-
covery from inflammation, rebound erythropoiesis may transiently
suppress hepcidin, facilitating enhanced iron absorption and release
from macrophages, and optimizing recovery from anaemia.
Special circumstances
Anaemia of the elderly
Elderly individuals have a higher prevalence of anaemia compared
with the younger population. For example, in Australia 16% of indi-
viduals older than 75 years are anaemic compared with fewer than
5% of younger adults. The prevalence of anaemia increases with age
and among individuals in hospital or in care. Anaemia in the eld-
erly is consistently correlated with impaired physical and cognitive
performance, and increased frailty. Epidemiological studies indicate
that about one-​third of cases of anaemia of the elderly are due to
deficiencies of iron and other haematinics. A further third are due
to anaemia of inflammation due to identifiable medical comorbid-
ities and renal impairment. The causes of the final third are more
difficult to ascertain, and it is often termed ‘unexplained anaemia
of the elderly’. Some cases may also represent underlying malig-
nant haematological diseases such as myelodysplasia. The mech-
anisms of unexplained anaemia require further investigation, but
may be multifactorial, including reduced erythroid potential of the
haematopoietic stem cell compartment, hepcidin-​dependent and -​
independent mechanisms of anaemia of inflammation, reduced
erythropoietin production in response to reduced haemoglobin
concentrations, and in men, hypoandrogenism. Elderly patients
with anaemia should be appropriately investigated to identify po-
tential therapies and exclude serious underlying conditions.
Areas of uncertainty and controversy
Treatment of anaemia of inflammation remains challenging, as few
specific therapies are presently available. Further work is needed to
assess the role, timing, and optimal regimens of intravenous iron and
ESAs. The safety of and optimal targets for treatment with ESAs re-
quire investigation in nonrenal-​ or cancer-​related anaemia. The role
of measurement of hepcidin in diagnosis of anaemia of inflammation
and its role in guiding treatment need development. The epidemi-
ology and mechanisms of anaemia of the elderly requires clarification.
Future developments
The discovery that hepcidin is a key mediator of anaemia of inflam-
mation has stimulated interest in development of novel therapeutics
which either directly inhibit its action or potentially, prevent its expres-
sion. For example, hepcidin expression can be prevented by heparin,
and heparin-​derived compounds (e.g. glycol-​split or oversulphated
heparins) have been shown to reduce hepcidin levels and alleviate an-
aemia of inflammation in experimental animals. Likewise, antihepcidin
monoclonal antibodies and the antihepcidin L-​ribonucleic acid
aptamer (NOX-​H94) both neutralize circulating hepcidin, reducing
the anaemia of inflammation in animal models. Finally, recombinant
erythroferrone (the erythroid-​derived suppressor of hepcidin) may
prove to be a useful strategy for reducing hepcidin expression and
treating anaemia of inflammation in the future.
FURTHER READING
Camaschella C, Pagani A (2018). Advances in understanding iron
metabolism and its crosstalk with erythropoiesis. Br J Haematol,
182(4), 481–94.
Cullis JO (2011). Diagnosis and management of anaemia of chronic
disease: current status. Br J Haematol, 154, 289–​300.
Ganz T (2019). Anemia of inflammation. N Engl J Med, 381, 1148–57.
National Blood Authority (2012). Patient blood management guide-
lines: module 3 medical. National Blood Authority, Canberra.
Thomas DW, et al. (2013). Guideline for the laboratory diagnosis of
functional iron deficiency. Br J Haematol, 161, 639–​48.
Weiss G (2015). Anemia of chronic disorders: new diagnostic tools and
new treatment strategies. Semin Hematol, 52, 313–​20.
Weiss G, Ganz T, Goodnough LT (2019). Anemia of inflammation.
Blood, 133, 40–50.
22.6.6  Megaloblastic anaemia and
miscellaneous deficiency anaemias
A.V. Hoffbrand
ESSENTIALS
Megaloblastic anaemias are characterized by red blood cell
macrocytosis. They arise because of inhibition of DNA synthesis in
the bone marrow, usually due to deficiency of one or other of vitamin
B12 (cobalamin) or folate, but sometimes as a consequence of a drug
or a congenital or acquired biochemical defect that disturbs vitamin
B12 or folate metabolism, or affects DNA synthesis independent of
vitamin B12 or folate.
section 22  Haematological disorders
5408
Biochemical and nutritional aspects of vitamin
B12 and folate
Vitamin B12—​synthesized by bacteria; in humans the daily re-
quirement of 1 to 2 μg is acquired from secondary animal sources
including fish, eggs, milk, and meat. Processing within the body oc-
curs as follows: (1) proteolysis of food releases dietary vitamin B12 for
binding to a glycoprotein haptocorrin; (2) pancreatic trypsin degrades
the glycoprotein, releasing vitamin B12 for attachment to intrinsic
factor; (3) the vitamin B12–​intrinsic factor complex binds to a specific
receptor—​cubilin amnion—​expressed on the luminal brush border
of the mucosal cells of the ileum, and is endocytosed; (4) after lyso-
somal degradation, vitamin B12 is complexed with transcobalamin
(TC)-​II and secreted into the circulation; (5) the TCII–​B12 complex is
incorporated by cellular endocytosis in peripheral tissues and vitamin
B12 released by digestion in the lysosomal compartment.
Folate—​occurs principally in leaves and vegetables, but is des-
troyed by cooking. The daily requirement is about 100 μg, with ab-
sorption occurring through a proton-​coupled folate transporter in
the proximal small intestine and duodenum. Attached glutamate res-
idues are cleaved, releasing methyl tetrahydrofolate into the portal
plasma.
Biochemical basis of megaloblastic anaemia—​(1) folate defi-
ciency reduces the availability of the coenzyme 5,10-​methylene
tetrahydrofolate (THF) polyglutamate, thus inhibiting synthesis
of thymidylate, which is the rate-​limiting step in DNA synthesis.
(2) Vitamin B12 deficiency impairs DNA synthesis indirectly because it
is needed for conversion of methyl-​THF entering cells from plasma to
THF, the substrate of active folate coenzymes (folate polyglutamates)
including 5,10 methylene-​THF.
Causes of megaloblastic anaemia
Vitamin B12 deficiency—​(1) malabsorption—​including (a)  gastric
causes (e.g. acquired pernicious anaemia, gastrectomy); (b) intestinal
causes (e.g. bacterial overgrowth, ileal resection); and (2) nutritional
(e.g. vegans).
Folate deficiency—​(1) poor diet—​e.g. poverty, alcoholism; (2) mal-
absorption (e.g. gluten-​induced enteropathy, tropical sprue);
(3)  excessive requirements (e.g. pregnancy, haemolytic anaemia);
(4) excess excretion (e.g. chronic haemodialysis); (5) drugs (e.g. anti-
convulsants); and (6) liver disease.
Not due to vitamin B12 or folate deficiency—​(1) abnormalities of
vitamin B12 or folate metabolism—​including (a) congenital (e.g. TCII
deficiency); (b)  acquired (e.g. dihydrofolate reductase inhibitors);
(2) independent of vitamin B12 or folate—​including (a) congenital (e.g.
orotic aciduria); (b) acquired (e.g. various myeloid leukaemias); and
(c) drugs (e.g. antimetabolites, hydroxycarbamide).
Laboratory investigation
This consists of three stages: (1) recognition that megaloblastic an-
aemia is present—​the mean corpuscle volume is raised to 100 to
140 fl, and the peripheral blood shows hypersegmented neutrophils.
The bone marrow (if examined) is hypercellular, with megaloblastic
erythroblasts and giant metamyelocytes. (2)  Distinction between
vitamin B12 or folate deficiency (or rarely some other factor) as the
cause of the anaemia—​usually achieved by assay of serum vitamin
B12 and serum folate. (3) Diagnosis of the underlying disease causing
the deficiency—​depends on taking a dietary history, measurement
of parietal cell and intrinsic factor antibodies and serum gastrin,
transglutaminase antibody, and pursuing clinical clues to other pos-
sible causes. Subclinical deficiency of both vitamins is more frequent
than overt megaloblastic anaemia.
Pernicious anaemia
Antibodies in serum and gastric juice directed against parietal cells
(85–​90% of cases) and intrinsic factor (50%), and raised serum gastrin
are associated with autoimmune gastritis and failure of absorption
of vitamin B12.
Clinical features—​anaemia usually develops gradually, and symp-
toms may not occur until it is severe. Aside from pallor, other
manifestations can include (1)  mild jaundice, (2)  mild pyrexia,
(3) psychiatric disturbance, (4) glossitis and angular cheilosis, and
(5) features of an associated disorder (e.g. vitiligo, thyroid disease).
Complications include (1) peripheral sensorimotor neuropathy and
(2) subacute combined degeneration of the spinal cord—​manifest as
loss of proprioception and pyramidal weakness.
Treatment and prevention of megaloblastic anaemia
Vitamin B12 deficiency—​may be treated with intramuscular
hydroxocobalamin (1-​mg doses, six given in the first 2–​3 weeks, then
every 3 months). Oral therapy is practised by a minority and is un-
likely to be useful in pernicious anaemia. Neurological complications
are irreversible unless treated early.
Folate deficiency—​high-​dose oral folic acid (5 mg daily) over-
comes folate malabsorption, but this should not be given alone
where vitamin B12 deficiency coexists because neurological disease
may be precipitated or exacerbated (although the haematological
abnormalities improve). Where folate metabolism is disturbed by
methotrexate, oral or parenteral folinic acid is given to restore DNA
synthesis.
Prevention—​dietary folate fortification is an accepted and highly
effective public health measure in many countries (none in Europe)
for reducing the incidence of neural tube birth defects.
Introduction
The megaloblastic anaemias are a group of disorders characterized
by a macrocytic anaemia and distinctive morphological abnormal-
ities of the developing haematopoietic cells in the bone marrow. In
severe cases, the anaemia may be associated with leucopenia and
thrombocytopenia. Megaloblastic anaemia arises because of in-
hibition of DNA synthesis in the bone marrow, usually due to de-
ficiency of one or other of two water-​soluble B vitamins: vitamin
B12 (cobalamin) or folate. Vitamin B12 deficiency may also cause a
severe neuropathy. In a minority of cases, megaloblastic anaemia
arises because of a disturbance of DNA synthesis due to a drug or
a congenital or acquired biochemical defect that causes a disturb-
ance of vitamin B12 or folate metabolism or affects DNA synthesis
independent of vitamin B12 or folate. Vitamin B12 and folate are
discussed first and the other rare megaloblastic anaemias are men-
tioned later in this chapter.
Folic acid supplements in pregnancy and food fortification with
folic acid are aimed at preventing neural tube defects. Possible
relations between folate and vitamin B12, and cardiovascular or
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5409
malignant diseases and cognitive defects in older people are also
discussed.
Biochemical and nutritional aspects
of vitamin B12 and folate
Vitamin B12
Biochemistry
Four major forms of the vitamin exist in humans, all with the
same cobalamin nucleus, which consists of a planar corrin ring
(hence the term ‘corrinoids’ for vitamin B12 compounds) attached
at right angles to a nucleotide portion, 5,6-​dimethylbenzimidazole
joined to ribose-​phosphate (Fig. 22.6.6.1 and Table 22.6.6.1).
5′-​Deoxyadenosyclobalamin (adocobalamin) accounts for about
80% of vitamin B12 inside mammalian cells and is located mainly
in mitochondria; methylcobalamin is a minor cellular compo-
nent but the main form in plasma. Both are extremely light sensi-
tive and are rapidly photolysed to hydroxocobalamin by daylight;
hydroxocobalamin is present in small amounts in tissues and plasma
and is available commercially for therapeutic use. The fourth form,
cyanocobalamin, is found only in trace amounts naturally, but is
stable and used therapeutically. Hydroxo-​ and cyanocobalamins
are converted to the two biochemically active forms. The fully re-
duced compounds are termed Cob(I)alamins, and the oxidized
compounds Cob(III)alamins. Analogues of vitamin B12 (pseudo-​
vitamin B12s) exist in nature, endogenous production of which in
humans is suggested by their presence in all sera (including fetal
serum) and their fall in parallel with physiologically active vitamin
B12 in vitamin B12 deficiency.
Vitamin B12 is known to be involved in only three reac-
tions in human tissues:  as adocobalamin in the isomerization of
methylmalonyl CoA to succinyl CoA and of α-​leucine to β-​leucine,
and as methylcobalamin in the methylation of homocysteine to
methionine, a reaction that also requires methyltetrahydrofolate
(Fig. 22.6.6.2). In some bacteria, but not in humans, vitamin B12
has a direct role in DNA synthesis by virtue of its involvement in
ribonucleotide reductase.
Nutrition
Vitamin B12 is synthesized by microorganisms; animals obtain it by
consuming the flesh of other animals or their produce (milk, cheese,
eggs, etc.)—​or vegetable foods contaminated by bacteria. A healthy
mixed diet contains between 5 and 30 µg daily. In some species, but
not in humans, vitamin B12 is absorbed after synthesis by bacteria
in the large intestine. The vitamin B12 content in humans is about 3
to 5 mg; it is found mainly in the liver (c.0.7–​1.1 µg/​g). Adult daily
losses are related to body stores; to maintain normal body stores,
daily requirements are of the order of 1 to 2 µg. It takes 3 to 4 years,
on average, for deficiency to develop if supplies are totally cut off by
malabsorption. There is an enterohepatic circulation for vitamin B12,
variously estimated at 3 to 9 µg daily, which is intact in vegans, which
may partly account for their tendency to maintain low body stores
without incurring severe deficiency. The body is unable to degrade
vitamin B12 and deficiency has not been shown to be due to excess
utilization or loss.
Absorption
About 15% of dietary vitamin B12 is available for absorption. It is
released from protein binding in food by proteolytic enzymes, heat,
and acid, and combines one molecule to one molecule with a glyco-
protein R vitamin B12-​binding protein (also called haptocorrin) in
gastric juice. The glycoprotein binds dietary forms of vitamin B12
but does not facilitate its absorption. Gastric pepsin and pancre-
atic trypsin degrade this protein and so releases vitamin B12 for at-
tachment to intrinsic factor (IF) and subsequent absorption. IF is
a glycoprotein produced mainly by the gastric parietal cells (Table
22.6.6.2). The normal stomach produces a vast excess of IF, meas-
ured in units (1 unit binds 1 ng vitamin B12). Vitamin B12 in bile
is also attached to IF and reabsorbed through the ileum. At neu-
tral pH, in the presence of calcium ions, the vitamin B12–​IF com-
plex attaches passively to a complex specific IF receptor, cubilin
amnion, on the brush border of the mucosal cells of the terminal
ileum. Cubilin is a 640-​kDa peripheral membrane protein pre-
sent in the epithelium of intestine and kidney. Amnionless (AMN)
(50 kDa) binds to cubilin and is essential for production of mature
cubilin and its transport to the apical brush border. AMN directs
sublocalization and endocytosis of cubilin and the IF–​B12 complex.
Mutations of cubilin or AMN underlie hereditary malabsorption of
vitamin B12 (discussed later in this chapter).
After cubilin–​AMN-​mediated endocytosis, IF undergoes lyso-
somal degradation. After a delay of 3 to 5 h, vitamin B12 appears in
portal blood, with a peak concentration 8 h after ingestion, com-
plexed with transcobalamin II (TCII) secreted into the circulation
from the basolateral side of the intestinal cells. Ileal absorption of
vitamin B12 is limited by the number of cubilin receptors to a few
N
N
CH2
CH2
CH2
CH3
CO
NH
CH
O
O –
P
N
A
C
B
CO+
OH
C
D
N
N
N
N
C
C
O
O
O
C
CH2OH
CH3
CH3
Fig. 22.6.6.1  The structure of cyanocobalamin.
section 22  Haematological disorders
5410
micrograms daily, and although 80% of a single dose of 1 to 2 µg
may be absorbed, the proportion diminishes steeply at higher doses.
A small (<1%) trace of a large (≥1 mg) dose of vitamin B12 can be
absorbed passively and rapidly through the buccal, gastric, and duo-
denal mucosae without the involvement of the IF pathway and this
forms the basis for treating malabsorption of vitamin B12 with large
oral doses of cyanocobalamin.
Transport
Vitamin B12 in plasma is 70 to 90% attached to a glycoprotein,
transcobalamin I (TCI), and 0 to 10% to transcobalamin III (TCIII),
which do not enhance cell uptake of vitamin B12 (Table 22.6.6.2).
TCI and III belong to a group of glycoproteins, the R binders or
haptocorrins (previously mentioned), that are present in many tis-
sues and fluids; these molecules have the same amino acid compos-
ition but differ in the carbohydrate moiety. The haptocorrins may
have the role of binding analogues of vitamin B12 derived from food
or intestinal organisms and transporting them to the liver for excre-
tion in the bile. Genetic mutations of the gene TCN1, which codes
for TCI, cause subnormal serum vitamin B12 levels.
The most important plasma vitamin B12-​binding protein, TCII,
is synthesized in macrophages, the liver, the ileum, and possibly the
endothelium. TCII is loaded with vitamin B12 from the ileum and
by release of vitamin B12 from the liver and other organs. It is nor-
mally almost completely unsaturated because it actively enhances
uptake of vitamin B12 by bone marrow, placenta, and other tissues of
the body that contain TCII receptors. TCII–​vitamin B12 is internal-
ized by endocytosis; vitamin B12 is released by proteolytic cleavage
in lysosomes but TCII is not reutilized (Table 22.6.6.1). TCII has a
20% amino acid homology and greater than 50% nucleotide hom-
ology with human TCI and with rat IF. It shows at least five genetic
variants. Serum TCII is normally higher in women than men and
in black populations compared with white. The concentration of
vitamin B12 in cerebrospinal fluid is low, with a mean of 10 ng/​litre
in normal subjects. Most of this is attached to TCII. There is virtually
no vitamin B12 in normal urine.
Folate
Biochemistry
This vitamin exists in nature in over 100 forms, all of which are de-
rivatives of folic acid (pteroylglutamic acid), which has the structure
Table 22.6.6.1  Vitamin B12 and folate
Vitamin B12
Folate
Parent form
Cyanocobalamin (cyano-​B12), molecular weight 1355
Folic acid (pteroylglutamic acid), molecular weight 441.4
Crystals
Dark-​red needles
Yellow, spear-​shaped
Natural forms
Deoxyadenosylcobalamin
Reduced (di-​ or tetrahydro-​), methylated, formylated, other single carbon
additions; mono-​ and polyglutamates
Methylcobalamin
Hydroxocobalamin
Foods
Animal produce (especially liver) only
All, especially liver, kidney, yeast, greens, nuts
Adult daily
requirements
2 μg
100 μg
Adult body stores
2–​5 mg
6–​20 mg
Length of time to
deficiency
2–​4 years
4 months
Daily diet content
5–​30 μg
About 200–​250 μg
Cooking
Little effect
Easily destroyed
Absorption
Intrinsic factor (+ neutral pH + Ca2+) via ileum
Deconjugation, reduction, and methylation via duodenum and jejunum
Plasma transport
Tightly and specifically bound to transcobalamins
One-​third loosely bound albumin, other proteins;?specific protein
Enterohepatic
circulation
3–​9 μg/​day
60–​90 μg/​day
(2) S-adenosylhomocysteine
Methionine
Homocysteine
THF
Methyl THF
(3) α-leucin
B-leucine
adocobalamin
(1) Propionyl CoA
adocobalamin
Succinyl CoA
Methylmalonic Acid
Isoleucine
Odd-chain fatty acids
Thymidine
Valine
methylmalonyl
CoA mutase
Methylmalonyl CoA
S-adenosylmethionine
Fig. 22.6.6.2  Biochemical reactions of vitamin B12 (cobalamin) in
human tissues. THF, tetrahydrofolate.
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5411
of a pteridine, a para-​aminobenzoic acid moiety and L-​glutamic
acid (Fig. 22.6.6.3). Natural folates differ from folic acid by:
•	being reduced in the pteridine ring to di-​ or tetrahydo-​ forms
•	having a single carbon moiety attached at positions N5 or N10
(e.g. methyl, formyl, etc.)
•	having a chain of glutamate moieties attached by γ-​peptide bonds
to the L-​glutamate moiety
In human and other mammalian cells, the number of glutam-
ates is mainly four, five, or six. Polyglutamate forms of folate are
the active coenzymes; they show increased affinity or lowered Km
values compared to the monoglutamate equivalent compounds,
for most of the enzymes of one-​carbon metabolism. In body
fluids, however, folates are monoglutamate derivatives. In plasma,
5-​methyltetrahydrofolate (methyl-​THF) predominates.
The biochemical reactions of folates are shown in Table 22.6.6.3.
In each there is transfer of a single carbon group, methyl (–​CH3),
formyl (–​CHOH), methenyl (≡CH), methylene (=CH2), or formi­
mino (=CHNH), from one compound to another. Three of the
reactions are concerned with synthesis of DNA precursors (two
purine and one pyrimidine). During thymidylate synthesis,
oxidation of folate to the dihydro state occurs; the enzyme
dihydrofolate reductase, the principal target for the antifolates
methotrexate and pyrimethamine, returns folate to the active
tetrahydro state (Fig. 22.6.6.4). During its reactions, folate is not
completely reutilized, some degradation at the C9–​N10 bond oc-
curs to nonfolate compounds. Thus, folate utilization is increased
and folate deficiency likely when cell turnover and DNA synthesis
are increased.
Nutrition
Folate occurs in most foods, the highest concentrations (more than
30 µg/​100 g wet weight) in liver (in which it is easily destroyed by
cooking). Vitamin C protects folate from oxidative destruction. An
average Western daily intake is about 250 µg, with 50% or more in the
polyglutamate form. Body stores are about 10 to 12 mg, with a mean
liver concentration of about 7 µg/​g. Primitive or rapidly growing tis-
sues have higher folate concentrations than corresponding mature
tissues. Daily adult requirements are about 100 µg.
Absorption
Folates are absorbed rapidly, mainly through the duodenum and je-
junum. Polyglutamates are deconjugated in the intestinal lumen, at
the brush border, and possibly in lysosomes of intestinal cells by an
enzyme known as folate conjugase (γ-​glutamylcarboxypeptidase,
pteroylpolyglutamate hydrolase). They are reduced to the
tetrahydro state and methylated at the N5 position so that methyl-​
THF enters portal plasma whatever food folate is ingested
(Table 22.6.6.1). Folic (pteroylglutamic) acid itself, which is not
present in food, but is used therapeutically, enters the portal blood
largely unchanged at doses of more than 200 µg, as it is a poor sub-
strate for reduction by dihydrofolate reductase. A proton-​coupled
folate transporter (PCFT) with high affinity and a low pH op-
timum is essential for absorption of reduced folates and folic acid.
It is expressed particularly in the apical brush border of the en-
terocytes of the duodenum and jejunum. Various mutations have
been found in this transporter in patients with a specific hereditary
malabsorption of folate. The protein is expressed in other tissues
and may be involved in intracellular transportation of folates from
endocytic vesicles. As it is also active at neutral pH for methyl-​
THF, it may play a role in delivering this folate to systemic cells
(e.g. the liver). It may also transport antifolates (e.g. methotrexate)
into the acid interior of solid tumours. The small intestine has a
large capacity to absorb folate; on average 50% of natural folate
is absorbed whatever the dose. If excessive amounts are fed, the
excess is largely excreted in urine as folates or their breakdown
products after cleavage of the C9–​N10 bond. There is a substantial
enterohepatic circulation for folate, estimated at up to 90 µg folate
daily; if this is interrupted, plasma folate concentrations decrease
to about one-​third within 24 h.
Table 22.6.6.2  Vitamin B12-​binding proteins
Intrinsic factor
Transcobalamin I and IIIa
Transcobalamin II
Present in
Gastric juice
Plasma
Plasma, cerebrospinal fluid
Source
Gastric parietal cell
Granulocytes, Other organs
Macrophages, liver parenchyma, ileum
Molecular weight
45 000
60 000
45 500
Structure
Glycoprotein (15% sugar)
Glycoprotein
Polypeptide
Normal total binding capacity
30–​110 μg/​litre
700–​800 ng/​litre
900–​1000 ng/​litre
Vitamin B12 content
No vitamin B12
300–​400 ng/​litre vitamin B12
30–​60 ng/​litre vitamin B12
Function
Vitamin B12 absorption
(not itself absorbed)
? Storage of vitamin B12
Vitamin B12 delivery to marrow, placenta, brain,
and other tissues,
Vitamin B12 absorption
? Protection of vitamin B12
Binding of vitamin B12 analogues
a Related ‘R’ binders (haptocorrins) occur in other tissues and secretions, e.g. milk, gastric juice, saliva, and tears.
N
H
O
C
COOH (α)
CH
COOH (γ)
1
2
3
6
7
8
9
10
N
N
N
N
N
H
N
5
4
Fig. 22.6.6.3  The structure of pteroylglutamic (folic) acid.
section 22  Haematological disorders
5412
Transport
Folate is transported in plasma, two-​thirds unbound and about one-​
third loosely bound to albumin and possible other proteins. There
are two highly specific mammalian folate transporters. SLC19A1
is a facilitative transporter with the characteristics of an anion ex-
changer. The gene is located at chromosome 21q 22.2. The protein
has 12 transmembrane domains and both N-​ and C-​termini are dir-
ected to the cytoplasm. It is ubiquitously expressed on normal tis-
sues and tumours. Its affinity for folic acid and methotrexate is one
to two orders less than for reduced folates. The second is a group of
high-​affinity binding proteins (FRs) encoded by three genes, des-
ignated α, β, and γ, localized to chromosome 11q13.3 to 11q13.5.
FRα and FRβ are both glycosylphosphatidylinositol (GPI)-​anchored
proteins. The physiological role of the FRs is not clear. They are ex-
pressed in the apical brush border of the renal tubular epithelial cells
so may have a role in renal reabsorption of folates.
Folate taken up by membrane-​bound FRs is thought to enter
endocytic vesicles. FRα (but not FRβ) knockout mice show fatal
morphological abnormalities, suggesting a critical role in mouse
development. The FRs have enhanced expression on certain tumour
cells and this has prompted studies aimed at developing tumour-​
specific antifolates or folate-​conjugated radiopharmaceuticals or
other molecules.
Plasma folate is filtered by the glomerulus and mostly reabsorbed
unless the renal tubular maximum is exceeded. Normal urine
folate is 0 to 13 µg in 24 h. Folate is secreted into cerebrospinal fluid
(which has a mean concentration of 24 µg/​litre) and is present in
bile. Human milk has a folate concentration of 50 µg/​litre. Prostate-​
specific membrane antigen is a folate hydrolase carboxypeptidase
which can release glutamates in either α or γ linkages. The physio-
logical significance of this is unknown.
Biochemical basis of megaloblastic anaemia
All known causes of megaloblastic anaemia, whether drugs, de-
ficiencies, or inborn errors of metabolism, inhibit DNA synthesis
by reducing the activity of one of the many enzymes concerned in
Table 22.6.6.3  Biochemical reactions of folates
Reaction
Enzyme
1.
Conjugation or deconjugation
Hydrolysis of poly-​ to monoglutamates
Folate ‘conjugase’ (α-​glutamylcarboxypeptidase; pteroylpolyglutamate hydrolase)
Conjugation of monoglutamates to polyglutamates
Folate-​polyglutamate synthase
2.
Oxidation–​reduction
Oxidized or dihydrofolates converted to tetrahydrofolates
Dihydrofolate reductase
3.
Amino acid interconversions
(a) Homocysteine → methioninea
5-​Methyl-​THF methyltransferase (methionine synthase)
Methyl THF →THF
(b) 5-​Formiminoglutamic acid (Figlu) → glutamic acid
Figlu transferase
THF → formimino THF
Serine–​hydroxymethyltransferase
(c) Serine → glycine
THF → 5,10-​methylene THF
4.
DNA synthesis
Purine synthesis:
(a) GAR → formyl GAR
GAR transformylase
5,10 Methenyl THF →THF
(b) AICAR → inosinic acid
AICAR transformylase
10-​Formyl THF →THF
Pyrimidine synthesis:
Deoxyuridine monophosphate (dUMP) → thymidine
monophosphate (TMP)
Thymidylate synthase
5,10-​Methylene THF → THF
5.
Formate fixation
Formic acid + ATP + THF → 10-​formyl-​THF + ADP
THF formylase
6.
? Methylation of biogenic amines
E.g. dopamine → epinine
? Dopamine methyltransferase
Methyl THF → THF
THF, tetrahydrofolate; DHF, dihydrofolate; GAR, glycinamide ribotide; AICAR, 5-​amino-​4-​imidazolecarboxamide ribotide.
a See Figs. 22.6.6.2 and 22.6.6.4.
Reaction (6) has been demonstrated only in vitro and may not take place in vivo.
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5413
purine or pyrimidine synthesis or by inhibiting DNA polymeriza-
tion from its precursors. Folate deficiency, by reducing supply of the
active coenzyme form, 5,10-​methylene-​THF, inhibits thymidylate
synthesis, a rate-​limiting reaction in DNA synthesis. Vitamin B12
does not have a direct role in this or any other reaction in mamma-
lian DNA synthesis. Vitamin B12 deficiency inhibits DNA synthesis
indirectly because of the requirement for methylcobalamin in the
conversion of methyl-​THF that has entered cells from the plasma to
THF. Deficiency of vitamin B12 is considered to decrease the intra-
cellular supply of THF, from which the natural folate coenzymes,
folate polyglutamates, are made. Methyl-​THF cannot act as a sub-
strate for synthesis of folate polyglutamates in human cells. When
vitamin B12 is deficient there is reduced activity of all reactions re-
quiring folate coenzymes, including those involved in DNA syn-
thesis (Fig. 22.6.6.4). Misincorporation of the base uracil, because
of the accumulation of dUMP (Fig. 22.6.6.4) and hence of dUTP, has
been proposed to contribute to the DNA abnormality.
Clinical features and causes
of megaloblastic anaemia
Although pernicious anaemia (PA) is only one of the many causes
of megaloblastic anaemia (Tables 22.6.6.4–​22.6.6.6), it is convenient
to describe the general clinical features of the anaemia under this
heading; PA is the most frequent cause of megaloblastic anaemia in
Western countries. The laboratory findings and treatment of PA and
other megaloblastic anaemias are discussed later.
Pernicious anaemia (Addisonian pernicious anaemia,
Biermer’s anaemia)
Definition
An autoimmune disease in which there is atrophy of the stomach
with severely reduced or absent IF and acid secretion with conse-
quent malabsorption of vitamin B12 and vitamin B12 deficiency.
There is an autoimmune gastritis caused by pathological CD4 T cells
reacting against gastric H/​K-​ATPase.
Aetiology
PA is a disease of older people: less than 10% of patients are under
the age of 40 years. There is a female:male ratio in most (but not
all) series of about 1.6:1. There is a slightly higher prevalence (c.44
vs 40%) of blood group A in patients with PA compared with con-
trols in the United Kingdom. No overall association between PA
and HLA type has been found, but those with an endocrine disease
have a greater incidence of HLA B8, B12, and BW15. An association
between autoimmune gastritis and HLA DRB103 and DRB104 has
been reported in Finnish and Italian populations. PA occurs in all
ethnic groups including African, Indian, Native American, and
Chinese, as well as white Europeans. There is a higher incidence in
close relatives, of either sex, of an affected person.
DNA sequence variants of a gene NLRP1, located at chromosome
17p13, encoding NACHT, a leucine-​rich repeat protein which is a
regulator of the innate immune response, have been associated with
vitiligo and its associated diseases including PA.
About 55% of patients have serum thyroid antibodies and 33%
with primary myxoedema have parietal cell antibody. There is
probably no association with diabetes mellitus. Other evidence
for an immune aetiology of the gastritis of PA is the improve-
ment in mucosal appearance and function with corticosteroid
therapy, the presence of antibodies in serum and gastric juice dir-
ected against parietal cells and IF, and of cell-​mediated immunity
to IF. Parietal cell antibody is present in the serum of 85 to 90%
of patients. The autoantigens are the α-​ and β-​subunit of the gas-
trin proton pump (H+,K+ ATPase). Two antibodies to IF exist in
serum. Type I (‘blocking’) occurs in about 50% of patients and is
directed against the vitamin B12-​binding site. Type II (to the ileal
binding site) occurs in 30 to 35% but only if type I antibody is also
present. Antibodies to IF in gastric juice may neutralize the action
of remaining IF. The incidence of parietal cell and IF antibodies
in serum in PA may be different in different groups of patients,
younger patients having a lower incidence of parietal cell antibody
while black patients and Hispanic patients may have a higher inci-
dence of IF antibodies. The antibodies to IF are virtually specific for
PA but parietal cell antibodies occur in many subjects with atrophic
gastritis without PA. An autoantibody to the gastrin receptor may
also occur in serum in PA.
PA may be associated with hypogammaglobulinaemia or with se-
lective IgA deficiency when it tends to present at an early age. Serum
gastrin concentrations are raised (>200 µg/​litre) in 90% of patients
with PA, and serum pepsinogen (PG) concentrations are less than
30 µg/​litre in 92% of such patients with a low PGI/​PGII ratio.
DNA
dATP
dGTP
dTTP
dCTP
dTDP
thymidine
kinase
Thymidine
dTMP
thymidylate synthase
dUMP
DHF-polyglutamate
dihydrofolate
reductase
Methylation
of myelin, basic
proteins, lipids,
DNA, amines
S-adenosyl-
methionine
S-adenosyl
homocysteine
THF-polyglutamate
THF
Methionine
Homocysteine
Methyl THF
(synthesized by
small intestine
from dietary
folates)
5,10 methylene THF –
polyglutamate
Deoxyuridine
(dU)
Fig. 22.6.6.4  Suggested mechanisms by which vitamin B12 deficiency
affects folate metabolism and interferes with DNA synthesis. Indirect
involvement of vitamin B12, as methylcobalamin, in DNA synthesis
is suggested by the ‘methylfolate’ trap (‘tetrahydrofolate starvation’)
hypothesis. Methylcobalamin is involved in formation of intracellular
THF from plasma methyl-​THF. THF and/​or its formyl derivative, but
not methyl-​THF, are the ‘ground substances’ from which all folate
coenzymes are made by glutamate addition and single carbon unit
transfer. 5,10-​Methylene-​THF polyglutamate is involved in thymidylate
synthesis. A, adenine; C, cytosine; D, deoxyribose; DP, diphosphate; G,
guanine; T, thymine; THF, tetrahydrofolate; TP, triphosphate; U, uridine.
section 22  Haematological disorders
5414
The relationship of PA, autoimmune gastritis, and Helicobacter
pylori infection is not clear. Young subjects (<40 years) with gastritis,
hypergastrinaemia, and positive antiparietal cell antibody in serum
will usually show iron deficiency anaemia whereas older (>60 years)
with these features more frequently have macrocytic red cells and
low serum vitamin B12 levels. H. pylori infection occurs in up to
40% of such subjects less than 20 years old but in only 10% of those
older than 60 years. It has been proposed that H. pylori is an infective
trigger to autoimmune gastritis by molecular mimicry.
Pathology
There is a gastritis in which all layers of the body and fundus of the
stomach are atrophied with loss of normal gastric glands, mucosal
architecture, and absence of parietal and chief cells, but mucous cells
lining the gastric pits are well preserved. An infiltrate of plasma cells
and lymphocytes with an excess of CD8 cells occurs and intestinal
metaplasia may be present. The antral mucosa is well preserved ex-
cept in hypogammaglobulinaemia, and, like the fundus, shows an
increased number of gastrin-​secreting cells.
Clinical features
The general features of megaloblastic anaemia are similar, whatever
the underlying cause. Particular clinical features may point to the
underlying disease, whether PA or some other cause. In PA, the an-
aemia usually develops gradually, perhaps over several years, and
symptoms may not occur until it is severe. The most common com-
plaints are due to the anaemia, but loss of mental and physical drive,
numbness, or difficulty in walking suggest neurological complica-
tions. Psychiatric disturbances are common and range from mild
neurosis to severe organic dementia. They may occur in the absence
of anaemia or macrocytosis. Mild jaundice, loss of appetite and
weight, indigestion, and episodic diarrhoea are frequent. An inter-
current infection may precipitate severe anaemia and thus symp-
toms. Older patients may present with congestive heart failure. In
a few patients, bruising due to thrombocytopenia is marked. Many
symptomless patients are diagnosed because a routine blood test
is made.
Physical signs, if present, are those of anaemia, perhaps with
mild jaundice, giving the patient a so-​called lemon-​yellow tint.
A few patients with deficiency of either vitamin B12 or folate de-
velop a widespread brown pigmentation, affecting nail beds and
skin creases particularly, but not mucous membranes. This is re-
versible with the appropriate therapy. The tongue may be red,
smooth, and shiny, occasionally with ulcers. A mild pyrexia up
to 38°C is common in patients with moderate to severe anaemia.
Table 22.6.6.4  Causes of vitamin B12 deficiency and malabsorption
of vitamin B12
Causes of severe
vitamin B12
deficiency
(a) Nutritional:
Vegans
Long-​continued extremely poor diet (rarely)
(b) Malabsorption:
Gastric causes:
(Addisonian) pernicious anaemia
Congenital intrinsic-​factor deficiency or abnormality
Total and partial gastrectomy
Destructive lesions of stomach
Intestinal causes:
Gut flora (associated with jejunal diverticulosis,
ileocolic, fistula, anatomical blind loop, stricture,
Whipple’s disease, scleroderma, HIV disease)
Ileal resection and Crohn’s disease
Chronic tropical sprue
Selective malabsorption with proteinuria
Fish tapeworm
Transcobalamin II deficiency
Causes of
malabsorption
of vitamin B12
usually without
severe vitamin
B12 deficiency
Malabsorption of food vitamin B12 (due to simple
atrophic gastritis, gastric bypass, proton pump
inhibitors, H. pylori infection); severe chronic
pancreatitis,
Zollinger–​Ellison syndrome, adult gluten-​induced
enteropathy, giardiasis, HIV disease, graft-​versus-​host
disease
Drugs: p-​aminosalicylic acid, colchicine, neomycin,
slow K, ethanol, metformina, phenformin,
anticonvulsants
a Recent studies suggest metformin lowers serum vitamin B12 by reducing the level of TCI.
Table 22.6.6.5  Causes of folate deficiency
Poor diet
Especially poverty, psychiatric disturbance, alcoholism,
dietary fads, scurvy, kwashiorkor, goats’ milk anaemia
Malabsorption
Gluten-​induced enteropathy (child or adult or
associated with dermatitis herpetiformis)
Tropical sprue
Congenital specific malabsorption
Minor factors: jejunal resection, inflammatory bowel
disease, systemic infections
Drugs: cholestyramine, sulphasalazine, methotrexate, ?
others (see ‘Drugs’).
Excessive
requirements
Physiological:
Pregnancy
Prematurity and infancy
Pathological:
(a) Malignancies—​leukaemia, carcinoma, lymphoma,
myeloma, etc.
(b) Blood disorders—​haemolytic anaemia
(especially sickle-​cell anaemia, thalassaemia major),
myeloproliferative diseases
(c) Inflammatory—​tuberculosis, malaria, Crohn’s disease,
psoriasis, exfoliative dermatitis, rheumatoid arthritis, etc.
(d) Metabolic—​homocystinuria (some cases)
Excess urinary
excretion
Congestive heart failure, acute liver damage, chronic
dialysis
Drugs
Mechanism uncertain
Anticonvulsants (diphenylhydantoin, primidone,
barbiturates)
? Alcohol
Also drugs causing malabsorption of folate (see
‘Malabsorption’ )
Liver disease
Mixed causes as above, and poor storage
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5415
The liver may be enlarged while the cardiovascular system shows
changes due to anaemia. Patients with PA may also have features
of an associated disorder on presentation, most commonly myxoe-
dema. Other thyroid disorders, vitiligo, carcinoma of the stomach
(incidence three times that of controls), Addison’s disease, and
hypoparathyroidism, may precede, occur simultaneously with, or
follow the onset of the anaemia. A few cases of PA with gastric at-
rophy, achlorhydria, and IF antibodies have occurred in children.
They may show associated autoimmune conditions, for example,
myxoedema, hypoparathyroidism, Addison’s disease, or chronic
mucocutaneous candidiasis.
Neurological complications of vitamin B12 deficiency
Vitamin B12 deficiency may cause a symmetrical neuropathy, af-
fecting the lower limbs more than the upper, which usually pre-
sents with paraesthesiae or with ataxia, particularly in the dark. In
some cases, loss of cutaneous sensation, spastic paraparesis, muscle
weakness, urinary or faecal incontinence, an optic neuropathy, or
psychiatric disturbance dominates. The nervous system disease
is due to severe deficiency judged by serum vitamin B12 levels or
methylmalonic acid (MMA) excretion, but may occur with mild or
no anaemia. A similar neurological syndrome with paraparesis has
been described in dentists and others repeatedly exposed to nitrous
oxide (N2O), which inactivates methionine synthase. The biochem-
ical explanation for the neurological disease is not clear. A defect in
fatty acid metabolism in myelin tissue has been suggested. Studies in
N2O-​treated monkeys have also suggested that the neuropathy re-
sults from accumulation of S-​adenosyl homocysteine (caused by the
block in conversion of homocysteine to methionine) with inhibition
of transmethylation of biogenic amines, proteins, phospholipids,
and neurotransmitters in the spinal cord and brain. Methionine has
been shown to prevent the neurotoxicity caused by N2O in experi-
mental animals.
General tissue effects of vitamin B12 and folate deficiencies
Both deficiencies cause macrocytosis and related cytopathic effects
in proliferating epithelial cells throughout the body (e.g. bronchial,
bladder, buccal, and uterine cervix), with glossitis and angular
cheilosis, a mild malabsorption syndrome, and reduced regen-
eration of damaged liver cells. In both sexes, sterility (reversible
with vitamin B12 or folate therapy) may result from effects on the
gonads. It is possible that the deficiencies in children affect overall
body growth. Nutritional vitamin B12 deficiency in infants long term
causes failure to thrive and poor brain growth with poor intellectual
outcome.
Generalized, reversible melanin pigmentation occurs in a few
patients with vitamin B12 or folate deficiency, the cause of which is
uncertain. Defective bactericidal activity of phagocytes due to im-
paired intracellular killing has been described in vitamin B12 but not
in folate deficiency. Vitamin B12 deficiency reduces serum concen-
trations of the osteoblast-​related proteins alkaline phosphatase and
osteocalcin, but whether clinically important bone disease occurs is
unknown.
Neural tube defects
Folic acid supplements at the time of conception and in early (first
weeks) of pregnancy reduce the incidence of neural tube defects
(NTDs) (anencephaly, encephalocoele, and spina bifida) in the
first and subsequent pregnancies when such a malformation has
occurred previously.
Folic acid fortification of the diet has led to a substantial reduc-
tion of incidence of NTDs (e.g. in North America). The explanation
for the effect of folic acid on NTDs is not certain. Women carrying
affected fetuses have on average lower serum folate and vitamin B12
concentrations and higher serum homocysteine levels than matched
controls. There is a linear relationship when plotted on logarithmic
scales between the birth incidence of NTDs and maternal red cell
folate, indicating that an increase in red cell folate even within
normal range is associated with a constant, proportional decrease in
the birth frequency of NTDs. Folic acid prevention of NTDs (and in
some studies cleft lip and palate), despite apparently normal serum
and red cell folate concentrations, suggests that folic acid is over-
coming a metabolic abnormality in folate metabolism. Only one
such defect, a mutated 5,10 methylene tetrahydofolate reductase
(MTHFR) enzyme, has been identified. Periconceptional use of vita-
mins or supplements containing folic acid is also associated with a
reduced incidence of birth defects associated with maternal diabetes
mellitus.
Mutated MTHFR, a common thermolabile variant (677C→T)
(Ala225Val) is associated with lower serum and red cell folate con-
centrations and with higher plasma homocysteine than in control
subjects in the general population. The prevalence of the homozy-
gous state in the population is approximately 5% and in parents of
fetuses with NTDs the prevalence is approximately 13%. The pres-
ence of this mutation can therefore account for only a small propor-
tion of NTDs. Mutations of other genes (e.g. VANGL1) not related to
folate metabolism, have been found in NTD families. Serum vitamin
B12 levels are also lower in sera of mothers with NTD infants than
in controls. Also, TCII receptor polymorphisms are associated with
Table 22.6.6.6  Megaloblastic anaemia not due to vitamin B12 or
folate deficiency
Abnormalities
of vitamin
B12 or folate
metabolism
Congenital:
Transcobalamin II deficiency or functional abnormality
Inborn errors of folate metabolism, e.g. methylfolate
transferase deficiency
Homocystinuria and methylmalonic aciduria (some cases)
Acquired:
Nitrous oxide
Dihydrofolate reductase inhibitors: methotrexate,
pyrimethamine, trimethoprim, ?pentamidine, triamterene
Independent
of vitamin B12
or folate
Congenital:
Orotic aciduria (responds to uridine)
Lesch–​Nyhan syndrome, ? responds to adenine
Thiamine-​responsive
Some cases of congenital dyserythropoietic anaemia
Acquired:
Erythroleukaemia, other myeloid leukaemias (some cases)
Myelodysplasia
Drugs:
Antimetabolites: 6-​mercaptopurine, cytosine arabinoside,
hydroxycarbamide, 5-​fluorouracil, azathioprine, etc.
section 22  Haematological disorders
5416
increased risk for NTD. There are, however, no studies showing that
vitamin B12 therapy or dietary fortification with vitamin B12 reduces
the incidence of NTDs.
Mental deterioration
There is a more rapid decline in cognitive function in subjects with
low serum vitamin B12 levels, serum holotranscobalamin, and raised
serum MMA concentrations. Studies suggest that low serum folate
levels are not associated with cognitive loss or depression in the eld-
erly. Meta-​analysis suggests administration of folic acid and vitamin
B12 may have a mild effect on memory but does not improve or sta-
bilize cognitive function in older people with or without low serum
vitamin B12 or folate concentrations. Although low serum vitamin
B12 levels and raised serum homocysteine and methylmalonate
levels have been reported to be more frequent in subjects with de-
mentia, including Alzheimer’s disease, than in controls, trials of
vitamin B12 therapy have generally shown no benefit in treating the
dementia or slowing its progression. Vitamin B12 has been used to
treat chronic fatigue syndrome but there are no controlled trials
which validate this.
Cardiovascular disease and stroke
McCully (1969) first implicated homocysteine as a cause of ath-
erosclerosis. This was based on pathological studies of children or
young adults with congenital homocystinuria, whether due to a
defect of cystathionine synthase, methionine synthase, or MTHFR
(Fig. 22.6.6.5). In these children, plasma homocysteine concentra-
tions are raised to 10 to 100 times normal. It is now apparent that
milder rises in plasma homocysteine are associated with coronary
or peripheral arterial disease, stroke, and deep vein thrombosis.
Homocysteine can directly injure endothelial cells, activate platelets
and leucocytes, stimulate vascular smooth muscle proliferation, oxi-
dize low-​density lipoprotein (LDL), and disturb collagen and extra-
cellular matrix formation. Determinants of plasma homocysteine
include age, sex, renal function, protein intake, vitamin B6, folate,
and vitamin B12 status, the presence of the thermolabile variant
MTHFR, smoking, and alcohol consumption, as well as intake of
various drugs.
Folate deficiency assessed by serum or red cell folate or by dietary
folate intake is also associated with coronary vascular disease, myo-
cardial infarct, and peripheral vascular disease. Meta-​analysis of
prospective trials shows that a 25% lower starting homocysteine
level is associated with 11% lower coronary heart disease risk (and
19% lower stroke risk). There is also an association of the MTHFR
homozygous state TT, which is associated with a higher homocyst-
eine level in serum than the wild type CC state, and ischaemic heart
disease. Meta-​analysis of 75 studies showed an increased risk of is-
chaemic heart disease in TT compared to CC homozygotes, odds
ratio 1.16 (1.04–​1.29). However, meta-​analysis of 26 randomized
control trials enrolling 58 804 participants showed that folic acid
supplementation was not associated with a significant change in car-
diovascular disease or all-​cause mortality, although it was linked to
a decreasing trend in stroke risk. This was more marked in popu-
lations without mandatory fortification of the diet with folic acid
(Yang et al. 2012). Huo et al. (2014) in the most recent meta-​analysis
of 15 randomized trials (N = 55 764) found an overall reduction in
stroke (relative risk 0.92; P = 0.04). This was more marked in those
followed for more than 3 years, those without background fortifica-
tion of breakfast cereals with folic acid and those not taking statins.
It may be relevant that a reduction in incidence of stroke occurred in
the United States of America and Canada coinciding with the intro-
duction of dietary folic acid fortification, whereas no reduction in
incidence occurred over the same period (1998–​2002) in England
and Wales without fortification. A recent Chinese randomized con-
trolled study over 4.5 years has confirmed that folic acid 0.8 mg daily
is effective in reducing primary occurrence of ischaemic stroke, par-
ticularly in those starting with lower folate levels, with the MTHFR
TT mutation and for longer receiving folic acid (Huo et al. 2015).
Those with the TT mutation had the highest stroke risk. Wald et al.
(2011) have suggested that in trials of prevention of secondary cor-
onary disease, aspirin may be negating or reducing the effect of
lowering homocysteine and folic acid prophylaxis. On this basis,
folic acid would have a role in primary prevention of ischaemic heart
disease in those not taking aspirin.
Malignancy
Positive and negative associations between the occurrence of various
types of lymphoblastic or myeloblastic leukaemias in infancy and
childhood and polymorphisms of folate-​metabolizing enzymes have
been reported. Folic acid prophylactically in pregnancy has been
reported to reduce the incidence of a subsequent childhood acute
lymphoblastic leukaemia and of brain tumours. In Canada, food for-
tification with folic acid has been associated with a 60% reduction in
incidence of neuroblastoma.
Epidemiological studies show an inverse risk of colorectal cancer
or adenoma and folate status and a less clear-​cut relation exists with
other gastrointestinal, lung, breast, ovary, and cervical carcinomas.
Also small randomized and nonrandomized trials suggest a benefit
of supplemental folic acid on incidence of colorectal cancer. A large
randomized trial has shown no difference in overall incidence of co-
lonic adenomas in women between controls and subjects receiving
folic acid, vitamin B6, and vitamin B12 supplements.
It has also been suggested that an increased incidence of colorectal
cancer in the United States of America and Canada is associated with
the fortification of the diet with folic acid, but the temporal disasso-
ciation between fortification and risk in colorectal cancer incidence
Methionine
synthase
Methionine
Homocysteine
Cystathionine
synthase
Cystathionine
Tetrahydrofolate
5-Methyl
tetrahydrofolate
5,10-Methylene
tetrahydrofolate
reductase
5,10-Methylene
tetrahydrofolate
Fig. 22.6.6.5  The role of three enzymes (cystathione synthase,
methionine synthase, and MTHFR) and three vitamins (vitamin B12,
vitamin B6, and folate) in homocysteine metabolism.
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5417
makes this unlikely—​increased detection by screening endoscopy is
a more likely explanation. There has been no increase in mortality
rate from colorectal cancer in the United States or Canada since for-
tification. Most recent analysis has shown no significant effect of
folic acid at large doses over prolonged periods on cancer incidence.
Meta-​analysis of 13 trials carried out before 2011, 10 for cardiovas-
cular disease prevention and 3 for colonic adenoma and cancer in-
cidence, involving almost 50 000 subjects in the supplemented and
control groups and lasting for a mean of 5.2 years, did not show an
effect on cancer incidence at doses of folic acid daily of 2 to 5 mg.
This was true for individual cancers of breast, prostate, lung, or large
bowel, as well as for rarer cancers. No overall increased incidence of
cancer was found in the analysis of 14 trials of B vitamin supplemen-
tation (some included in the previous study).
Other effects
Complications of pregnancy that have been ascribed to folic acid,
including miscarriage and multiple pregnancy, have no sound basis.
Malabsorption of vitamin B12
Congenital deficiency or structural abnormality
of intrinsic factor
Fewer than 100 cases have been reported of a child being born with
absent or nonfunctioning IF due to a mutation of the IF gene. There
is an otherwise normal stomach on biopsy and normal secretion of
acid. Inheritance is autosomal recessive. In different cases, IF may
be present in the gastric juice but susceptible to acid degradation
or cannot bind vitamin B12, or binds it but cannot attach it to ileal
receptors. These children tend to present with irritability, vomiting,
diarrhoea, and loss of weight, and are found to have megaloblastic
anaemia. The usual age of diagnosis is about 2 years, although a
few have been diagnosed as early as 4 months and others only in
their teens.
Gastrectomy
All patients who have total gastrectomy will develop vitamin
B12 deficiency, which usually presents between 2 and 6  years
postoperatively. They should be treated with prophylactic vitamin
B12 injections from the time of the operation. Iron deficiency usually
accounts for the anaemia that occurs after partial gastrectomy. A mi-
nority develop megaloblastic anaemia due to vitamin B12 deficiency.
In most of these patients, malabsorption of vitamin B12 is due to an
abnormal jejunal flora. The exact incidence of vitamin B12 deficiency
depends mainly on the size of the gastric remnant. After gastric pli-
cation or Roux-​en Y surgery for obesity, vitamin B12 deficiency may
occur and oral vitamin supplementation is often used. Monitoring
for vitamin B12 deficiency is advisable.
Small-​intestinal lesions
Colonization of the upper small intestine with colonic bacteria, if
sufficiently heavy as in the stagnant-​loop syndrome, leads to mal-
absorption of vitamin B12. The most common causes are listed in
Table 22.6.6.4. It appears that the bacteria destroy IF. Infestation
with the fish tapeworm (Diphyllobothrium latum) has a similar ef-
fect but is now almost completely eradicated; infestation is only suf-
ficiently marked in Finland and Russian lake regions to suggest a
possible cause of megaloblastic anaemia.
Resection of 1 m or more of terminal ileum
This causes severe malabsorption of vitamin B12. Other diseases
that may affect ileal structure and function include tropical sprue,
in which severe vitamin B12 deficiency with anaemia or, rarely, neur-
opathy is a manifestation only in the chronic phase; gluten-​induced
enteropathy in which megaloblastic anaemia, if it occurs, is always
due to folate deficiency (and vitamin B12 deficiency, if it occurs, is
mild); and in Crohn’s disease, malabsorption of vitamin B12 is fre-
quent but severe vitamin B12 deficiency is unusual unless there is an
ileal resection, fistula, or stagnant loop.
Selective malabsorption of vitamin B12 with proteinuria
(Imerslund’s disease, Imerslund–​Gräsbeck syndrome,
recessive megaloblastic anaemia, MGA1) (OMIM 261100)
This congenital disorder with autosomal recessive inheritance is the
most common cause of megaloblastic anaemia due to vitamin B12
deficiency in nonvegan children. The child secretes IF normally but
is unable to transport vitamin B12 across the ileum to portal blood.
Most Finnish patients with MGA1 carry the disease-​specific mu-
tation P1297L (FM1) in cubilin. A second less frequent mutation
(FM2) activates a cryptic splice site with insertion of multiple stop
codons in the CUB6 domain. Other mutations in cubilin have been
described. In Norway at least six different mutations of the AMN
gene have been reported in affected families. The proteinuria, pre-
sent in over 90% of cases, is benign, nonspecific, and persists after
vitamin B12 therapy. The clinical presentation of the disease is iden-
tical to that of congenital IF deficiency.
Other causes of malabsorption of vitamin B12
The most frequent cause of subclinical vitamin B12 deficiency in the
United Kingdom and United States of America, shown by a border-
line or low serum vitamin B12 level, and normal blood count with or
without a raised serum homocysteine and methylmalonate levels,
is malabsorption of dietary vitamin B12. This is thought to be due
to failure of release of dietary vitamin B12 from its protein binding
in food. It is usually due to an atrophic gastritis resulting in reduced
pepsin activity. The gastritis may be associated with a positive par-
ietal cell antibody test or with H. pylori infection. The incidence is
slightly more common in the elderly and the deficiency rarely pro-
gresses to megaloblastic anaemia or vitamin B12 neuropathy. Several
other conditions and drugs may cause malabsorption of food
vitamin B12, for example, proton pump inhibitors which very rarely
cause deficiency of clinical severity.
Other causes of vitamin B12 malabsorption by unidentified
mechanisms include p-​aminosalicylate, colchicine, neomycin, and
‘slow’ potassium tablets. Recent studies suggest the fall in serum
vitamin B12 level with metformin is due to a reduced serum level
of TCI rather than malabsorption. In chronic pancreatitis and the
Zollinger–​Ellison syndrome, there is failure to release vitamin B12
from gastric haptocorrin due to absence or inactivation of pancre-
atic trypsin.
Serum vitamin B12 concentrations fall progressively in untreated
HIV-​infected patients and subnormal serum values occur in 10 to
35% of individuals with AIDS; increased concentrations of TCII
are usual and malabsorption of vitamin B12, not corrected by IF, has
been found in some of these patients. An abnormal small-​intestinal
flora is the most likely cause of the vitamin B12 malabsorption.
section 22  Haematological disorders
5418
Malabsorption of vitamin B12 also occurs in inherited TCII defi-
ciency and temporarily after total-​body irradiation before stem
cell transplantation. In chronic graft-​versus-​host disease affecting
the gut, malabsorption of vitamin B12 is usual, due to the abnormal
gut flora as well as to an ileal defect. Irradiation to the ileum during
radiotherapy treatment for carcinoma of the cervix has also been
reported to cause vitamin B12 malabsorption.
Dietary vitamin B12 deficiency
This occurs most commonly in vegans. The incidence of overt meg-
aloblastic anaemia is much lower than the incidence of subclinical
deficiency assessed by the serum vitamin B12 assay. These individ-
uals have low vitamin B12 stores. Babies have been born vitamin B12
deficient with megaloblastic anaemia caused by severe vitamin B12
deficiency (due to poor diet or tropical sprue) in the mother. Breast
milk may also be vitamin B12 deficient if the mother’s stores are low.
Dietary deficiency of vitamin B12 also occurs rarely in nonvegetarian
people living on inadequate diets because of poverty.
Folate deficiency
Clinical features
The main clinical features of megaloblastic anaemia due to folate
deficiency are similar to those seen when the anaemia is due to
vitamin B12 deficiency, except that a neuropathy does not occur
and the underlying aetiology tends to be different. Cognitive
changes and depression may be caused by the deficiency, and
neurological abnormalities do occur with inborn errors of folate
metabolism and may be precipitated by antifolate drugs. Folate de-
ficiency may develop rapidly in a few months, and although many
mildly deficient patients do not progress for months or years, in
some patients the deficiency may lead to a severe pancytopenia
(‘arrest of haematopoiesis’) over a short period, particularly if an
infection supervenes.
Nutritional folate deficiency
Minor degrees of nutritional folate deficiency are frequent in
most countries. Potentially millions of people in northern China,
Bangladesh, Burma, Malaysia, Africa, or India have low levels of
folate due to a poor dietary intake and nutritional folate deficiency
is the main cause of megaloblastic anaemia, often presenting in
pregnancy. In many countries—​for example, Caribbean islands, Sri
Lanka, and South-​East Asia—​tropical sprue (see Chapter 15.10.8) is
an important cause of both deficiencies and is difficult to distinguish
from ‘pure’ nutritional deficiency.
Severe folate deficiency has been estimated to account for about
17% of all cases of megaloblastic anaemia in the United Kingdom,
where it occurs mainly in the context of a poor diet and/​or alco-
holism. In some cases, barbiturates or consumption of spirits or
cough mixtures or a physical abnormality such as rheumatoid arth-
ritis, or tuberculosis may aggravate the effect of a poor diet. A few
cases have developed because a special diet is taken, such as for
phenylketonuria or for slimming. Scurvy is usually accompanied by
severe folate deficiency. Goats’ milk anaemia is a nutritional folate
deficiency due to the low (6 µg/​litre) folate content of goats’ milk.
Malabsorption
(Also see Diseases of the gastro-intestinal tract described in
Section 15.)
Gluten-​induced enteropathy
Folate deficiency due to malabsorption of folates occurs in virtually
all untreated patients, the serum folate being subnormal in virtu-
ally 100% and red cell folate subnormal in 80% or more. Anaemia
occurs in about 90% of adult cases, due to folate deficiency alone in
30 to 50%, and to mixed iron and folate deficiency in the remainder.
Mild vitamin B12 deficiency may also occur, but it is not a cause of
anaemia in uncomplicated cases. Spontaneous atrophy of the spleen
occurs in most of the patients; in about 10 to 15% of cases; the blood
film shows the presence of Howell–​Jolly bodies, and other features
of hyposplenism. A gluten-​free diet produces a spontaneous rise
in serum and red cell folate in those patients who respond. In chil-
dren with gluten-​induced enteropathy, anaemia is most often due
to combined iron and folate deficiency. Patients with dermatitis
herpetiformis almost all show some degree of gluten-​induced duo-
denal and jejunal abnormality; the severity of folate malabsorption
and deficiency correlates with the severity of the intestinal lesion.
Tropical sprue
(Also see Chapter 15.10.8). Malabsorption of folate occurs in all
severe, untreated patients in the acute phase and megaloblastic an-
aemia due to folate deficiency may develop within a few months.
Not only does the anaemia respond to folate therapy but in many pa-
tients all the clinical features, and malabsorption of fat, vitamin B12,
and other substances, improves on folate therapy alone. In the first
year, about 60% of patients appear to be cured by folic acid alone.
Long-​standing cases are more likely to be vitamin B12 deficient and
thus to require vitamin B12 as well as folate and antibiotic therapy.
Congenital specific malabsorption of folate
This is a rare, autosomal recessive abnormality. Affected children
show features of damage to the central nervous system (mental re-
tardation, fits, athetotic movements) and present with megaloblastic
anaemia responding to physiological doses of folic acid given paren-
terally but not orally. Folate levels in cerebrospinal fluid are low. It is
due to inherited mutations of the proton-​coupled folate transporter
(PCFT) affecting protein stability or its membrane trafficking.
Other causes
Absorption of folate is impaired by systemic infections. Mild de-
grees of folate malabsorption have also been reported after jejunal
resection or partial gastrectomy, with Crohn’s disease, and with
lymphoma. In the intestinal stagnant-​loop syndrome, folate levels
tend to be high due to absorption of bacterially produced folate.
Alcohol, anticonvulsants, oral contraceptives, antituberculous
drugs, nitrofurantoin, and sulfasalazine have been suggested, on
variable evidence, to cause malabsorption of folate in some subjects
but none is definitely established except sulfasalazine.
Increased folate utilization
A general mechanism of increased folate utilization in conditions of
increased cell turnover has emerged. This consists of partial degrad-
ation of folate at the C9–​N10 bond rather than complete recycling of
the folate coenzymes required in DNA synthesis.
Pregnancy
This, associated with poor nutrition, is probably the most common
cause of megaloblastic anaemia world-​wide, unless folic acid
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5419
supplements are taken. The frequency of the anaemia was about
0.5% in most Western cities and up to 50% in some areas of Asia
and Africa until the introduction of prophylactic folic acid. The
incidence increases with parity and is higher in twin pregnan-
cies. Folate requirements in a normal pregnancy are increased to
about 300 to 400 µg daily. Serum and red cell folate tend to fall
as pregnancy progresses, and to rise spontaneously about 6 weeks
after delivery. Lactation may prove an additional cause of folate
deficiency, however, which may precipitate megaloblastic anaemia
postpartum.
The cause of the deficiency in pregnancy is increased degrad-
ation of folate. Folate transfer to the fetus may play a minor part; in
a few, megaloblastic anaemia of pregnancy is the first sign of gluten-​
induced enteropathy. The statistical association of iron and folate de-
ficiencies in pregnancy is probably due to a poor quality of the diet
in certain women.
Prophylactic folic acid should now be given routinely in preg-
nancy; 400 µg/​day is recommended (mentioned previously) and
intake in women who may become pregnant should be at least
this amount daily from food or supplements. Larger doses (4–​
5 mg/​day) should be used if there has been a previous infant with
an NTD.
Prematurity
Newborn infants have higher serum and red cell folate concentra-
tions than adults. These fall to a nadir at about 6 weeks of age. In pre-
mature infants, the decline is particularly steep and megaloblastic
anaemia may develop, particularly if infections, feeding difficulties,
or haemolytic disease with exchange transfusion have occurred.
Prophylactic folic acid (e.g. 1 mg/​week for the first 3–​4 weeks of life)
may be given, particularly to those babies weighing less than 1.5 to
1.8 kg at birth.
Malignant diseases
Mild folate deficiency is frequent in patients with cancer
(Table 22.6.6.5). In general, the severity correlates with the extent
and degree of dissemination of the underlying disease. However, pa-
tients with megaloblastic anaemia due to folate deficiency are un-
usual and, supplementation tends to be avoided in the absence of a
categorical indication for its use (e.g. anaemia, leucopenia, etc.) due
to concerns regarding promoting tumour growth.
Blood disorders
Chronic haemolytic anaemia  Requirements for folate are in-
creased in patients with increased erythropoiesis, particularly when
there is ineffective erythropoiesis with a high turnover of primitive
cells. Occasional patients, presumably those with a poor folate in-
take, develop megaloblastic anaemia, particularly in sickle cell
anaemia, thalassaemia major, hereditary spherocytosis, and warm-​
type autoimmune haemolytic anaemia; prophylactic folic acid is
usually given in these disorders.
Primary myelofibrosis  Megaloblastic haematopoiesis has been
reported in as many as one-​third of patients. Circulating meg-
aloblasts, increased transfusion requirements, severe thrombo-
cytopenia, or pancytopenia may be the first indication that folate
deficiency has developed. Polycythaemia vera is not a typical cause
of folate deficiency.
Inflammatory diseases
Folate deficiency has been described in patients with tubercu-
losis, malaria, Crohn’s disease, psoriasis, widespread eczema, and
rheumatoid arthritis. The degree of deficiency is related to the extent
and severity of the underlying disorder. Increased demand for folate
probably is a factor but reduced appetite is also important in those
who develop megaloblastic anaemia.
Metabolic
Homocystinuria  Patients with the most common form of this dis-
order, due to cystathionase deficiency, may show folate deficiency,
possibly due to excess conversion of homocysteine to methionine
and thus excess utilization of the folate coenzyme concerned (see
Chapter 12.2).
Excess urinary loss of folate
Urine folate excretion of 100 µg a day or more occurs in some pa-
tients with congestive cardiac failure or active liver disease causing
necrosis of liver cells. It is presumed that losses are due to release of
folate from damaged liver cells. Haemodialysis and peritoneal dia-
lysis remove folate from plasma. Folic acid (e.g. 5 mg/​week) is now
usually given prophylactically to patients with renal failure who re-
quire long-​term dialysis.
Drugs
Dihydrofolate reductase inhibitors
Methotrexate, aminopterin, pyrimethamine, and trimethoprim
all inhibit dihydrofolate reductase (DHFR) but have different rela-
tive activities against the human, malarial, and bacterial enzymes.
Methotrexate is converted to polyglutamate forms, which increases
its activity against DHFR and also increases its retention in cells.
These methotrexate derivatives invariably impair human folate me-
tabolism. Trimethoprim, used as an antibacterial agent, may aggra-
vate pre-​existing folate or vitamin B12 deficiency but does not in
itself cause megaloblastic anaemia.
Alcohol
Folate deficiency may occur in alcoholism. The main factor is poor nu-
trition and it is likely that alcohol interrupts the enterohepatic circu-
lation for folate. It also has a direct effect on haematopoiesis, causing
vacuolation of normoblasts, impaired iron utilization, sideroblastic
changes, macrocytosis, megaloblastosis, and thrombocytopenia, even
in the absence of folate deficiency. Beer drinkers usually avoid folate de-
ficiency because of the high folate content of beer. The usual red cell
macrocytosis in nonanaemic alcoholics is not related to folate deficiency.
Anticonvulsants and barbiturates
Diphenylhydantoin, primidone, and barbiturate therapy may be as-
sociated with folate deficiency. The more severe deficiency is associ-
ated with poor dietary intake of folate and prolonged drug therapy
at high doses. The mechanism of the deficiency is unknown and
double-​blind trials have shown no effect of folic acid supplementa-
tion on the frequency of seizures.
Other drugs
Nitrofurantoin, triamterene, proguanil, and pentamidine have been
reported to cause folate deficiency.
section 22  Haematological disorders
5420
Liver disease
Folate deficiency occurs most commonly in alcoholic cirrhosis
where alcohol, poor nutrition, and release of stored folate with ex-
cess urine losses may all be important. The deficiency is less frequent
in other types of liver disease.
Laboratory investigation of megaloblastic anaemia
This consists of three stages: (1) recognition that megaloblastic an-
aemia is present; (2) distinction between vitamin B12 or folate de-
ficiency (or rarely some other factor) as the cause of the anaemia;
and (3) diagnosis of the underlying disease causing the deficiency
(Table 22.6.6.7).
Recognition of megaloblastic anaemia
Peripheral blood
The mean corpuscle volume (MCV) is raised to between 100 and
140 fl. Oval macrocytes are seen in the blood film. In mild cases,
macrocytosis is present before anaemia has developed. Cabot rings
(composed of arginine-​rich histone and nonhaemoglobin iron) and
occasional Howell–​Jolly bodies (DNA fragments) may occur due to
extramedullary haematopoiesis in the liver and spleen. The MCV
may be normal if there is associated iron deficiency, when the blood
film appears dimorphic, or if the anaemia (usually due to folate de-
ficiency or antimetabolite drug therapy) develops acutely over the
course of a few weeks. The MCV is also normal in some severely an-
aemic cases involving excess red cell fragmentation. The reticulocyte
count is low for the degree of anaemia, usually of the order of 1 to 3%.
The peripheral blood also shows hypersegmented neutrophils
(which have nuclei with more than five lobes; Fig. 22.6.6.6) and the
leucocyte count is often moderately reduced in both neutrophils and
lymphocytes, although the total leucocyte count rarely falls to less
than 1.5 × 109/​litre. The platelet count may be moderately reduced
but rarely falls below 40 × 109/​litre.
Biochemical changes
These are confined to the anaemic patient and include a mild rise in
serum bilirubin (up to 50 µmol/​litre), mainly unconjugated, a rise
in serum lactic dehydrogenase of up to 10 000 IU/​litre. The serum
iron and ferritin are also raised and fall with effective treatment. The
serum cholesterol is low and alkaline phosphatase mildly reduced.
Absence of haptoglobins is usual. In severe cases, free haemoglobin
may be present in plasma, Schumm’s test for methaemalbumin in
serum is positive, and haemosiderin and fibrin degradation prod-
ucts are present in urine. The direct antiglobulin test is weakly posi-
tive in some patients, due to complement.
Bone marrow
The bone marrow is hypercellular in moderate or severely anaemic
cases. The myeloid–​erythroid ratio is often reduced or reversed. The
erythroblasts are larger than normal and show asynchronous mat-
uration of nucleus and cytoplasm, nuclear chromatin remaining
primitive with an open, lacy, fine granular pattern despite normal
maturation and haemoglobinization of the cytoplasm. Excessive
numbers of dying cells and nuclear remnants including Howell–​
Jolly bodies, mitoses, and multinucleate cells may be present.
Because of death (by apoptosis) of later cells, there is a dispropor-
tionate accumulation of early cells. Giant and abnormally shaped
metamyelocytes and megakaryocytes with hypersegmented nuclear
lobes are also usually present (Fig. 22.6.6.7).
Table 22.6.6.7  Laboratory diagnosis of megaloblastic anaemia
General tests
Peripheral blood film and count
Bone marrow
Serum bilirubin, iron, lactate dehydrogenase
Tests for
vitamin B12 or
folate deficiency
Serum vitamin B12 and folate; red cell folate
Serum homocysteine and methylmalonic acid levels
Tests for cause
of vitamin B12 or
folate deficiency
Vitamin B12 deficiency:
Serum antibodies to parietal cell, intrinsic factor
Serum gastrin
Gastric secretion; intrinsic factor, acid
Endoscopy, gastric biopsy
Upper gastrointestinal endoscopy
Proteinuria, fish tapeworm ova, intestinal flora, etc.
Folate deficiency:
Transglutaminase, endomysial antibodies
Small-​intestinal function
Duodenal biopsy
Barium follow-​through
Tests for many underlying conditions
Fig. 22.6.6.6  Megaloblastic anaemia. Hb 40 g/​litre MCV 120 fl.
Hypersegmented neutrophil, oval macrocytes, and a small lymphocyte to
show size of macrocytes. The fragmentation of advanced megaloblastosis
is present. Thrombocytopenia is marked.
Fig. 22.6.6.7  Megaloblastic anaemia. Bone marrow aspirate showing
megaloblasts at different stages and giant metamyelocytes.
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5421
The severity of these changes parallels the degree of anaemia.
In milder cases, changes, described as ‘intermediate’, ‘transitional’,
or ‘moderate’, are principally in the size and nuclear chromatin
pattern of the erythroblasts, with giant metamyelocytes present;
hypercellularity and gross dyserythropoiesis may be absent. In very
mild cases, megaloblastic changes are difficult to recognize. In pa-
tients with severe anaemia but only mild megaloblastic changes,
some additional cause for the anaemia should be sought.
Chromosomes
Changes found in marrow and other proliferating cells include
(1) random chromatin breaks; (2) exaggeration of centromere con-
striction; and (3) thin, elongated, uncoiled chromosomes.
Ineffective haematopoiesis
The increased cellularity of the marrow with degenerate forms, and
the low reticulocyte count suggest that many developing cells are
dying in the marrow. This occurs by apoptosis, especially of late
erythroblasts. The raised unconjugated serum bilirubin, lactic de-
hydrogenase, and lysozyme are all due to ineffective haematopoiesis.
Differential diagnosis
Other causes of macrocytosis include a high reticulocytosis (e.g.
haemolytic anaemia or regeneration of blood after haemorrhage),
aplastic anaemia, red cell aplasia, liver disease, alcoholism and
myxoedema, the myelodysplastic syndromes, myeloid leukaemias,
cytotoxic drug therapy, chronic respiratory failure, plasma cell mye-
loma, and other paraproteinaemias. If a bone marrow biopsy has
been performed, the principal differentiation is from other causes of
megaloblastosis, particularly myelodysplasia. Other causes of meg-
aloblastic anaemia not due to vitamin B12 or folate deficiency are
listed in Table 22.6.6.6.
Some patients with rapidly developing megaloblastic anaemia,
particularly due to folate deficiency, may develop almost complete
aplasia of the red cell series, and the peripheral blood and bone
marrow may resemble that of acute myeloid leukaemia.
Diagnosis of vitamin B12 or folate deficiency
The peripheral blood and bone marrow appearances are identical in
folate or vitamin B12 deficiency. Special tests are, therefore, needed
to distinguish between the two deficiencies.
Vitamin B12 deficiency
The assay of serum vitamin B12 content of serum is by competi-
tive binding of intrinsic factor and immunochemiluminescence-​
based assays. The reference range, depending on the assay, is from
160 to 200 mg/​litre to 960 to 1200 ng/​litre. Some report in pmol/​
litre (1 pmol = 1.355 ng/​litre). It is difficult to determine a normal
range and each laboratory should establish this independently. The
concentrations are low in vitamin B12 deficiency, being extremely
low in patients with neurological disease. Unfortunately, using
competitive-​binding assays, false-​normal results have been reported
in some patients with untreated pernicious anaemia and intrinsic
factor antibodies in serum, which interfere with the test. Subnormal
serum vitamin B12 concentrations in the absence of tissue vitamin
B12 deficiency have been reported in pregnancy, in inherited mu-
tations of TCI (haptocorrin), in severe nutritional folate deficiency,
in subjects taking large doses of vitamin C, and occasionally in iron
deficiency. In the elderly, low serum vitamin B12 concentrations usu-
ally in the range 100 to 200 ng/​litre may occur in the absence of an-
aemia or macrocytosis.
In some research studies, serum holo-​TCII levels have been meas-
ured to diagnose vitamin B12 deficiency.
Raised serum vitamin B12 concentrations, if not due to therapy,
are most commonly caused by a rise in TCI as in a leucocytosis due
to a myeloproliferative disease. Raised haptocorrin also occur in as-
sociation with some tumours, especially hepatoma and fibrolamellar
tumour of the liver. In benign leucocytosis, the rise is mainly of
TCIII and this is often not accompanied by a high serum vitamin
B12. Raised serum TCII concentrations occur in conditions where
macrophages are stimulated, for example, autoimmune diseases
such as systemic lupus erythematosus, rheumatoid arthritis, in
Gaucher disease, and in some monocytic or monoblastic leukae-
mias, and in inflammatory bowel disease. In active liver diseases,
vitamin B12 leaks from the liver with saturation of the serum vitamin
B12 binders.
A second and less widely used test for vitamin B12 deficiency is
serum MMA. Serum MMA levels are raised in vitamin B12 defi-
ciency but not in folate deficiency. Raised levels may also occur in
renal failure. Rare cases of congenital methylmalonic aciduria have
been described, due to a variety of enzyme defects.
A sensitive method of measuring MMA in serum was introduced
and combined with serum homocysteine assay for the diagnosis of
vitamin B12 or folate deficiency. The minor increases in serum MMA
concentration found particularly in older people in the absence of
macrocytosis or anaemia, with or without borderline vitamin B12
concentrations, may suggest ‘biochemical’ vitamin B12 deficiency
which does not progress to megaloblastic anaemia, but the normal
ranges for MMA for different age groups are difficult to determine.
Randomized trials are needed to assess the value of preventing or
treating putative vitamin B12 deficiency in these subjects.
Folate deficiency
Direct tests include the serum and red cell folate assay. The serum
folate is always low (<3 µg/​litre, 7 nmol/​litre) in folate deficiency
(and is normal or raised in vitamin B12 deficiency unless folate de-
ficiency is also present). Raised levels occur after folate therapy and
also in vitamin B12 deficiency and in the stagnant-​loop syndrome.
Red cell folate is not now recommended for routine diagnostics but
may be useful in some patients with probable folate deficiency in
whom the serum folate is found to be normal. It is low in a propor-
tion of patients with megaloblastic anaemia solely due to vitamin B12
deficiency. Serum homocysteine levels are usually raised in folate
and vitamin B12 deficiency and many other situations.
Diagnosis of the cause of vitamin B12 deficiency
Although the clinical and family history and the clinical findings
may point to PA or some other cause of vitamin B12 deficiency, it is
important to establish this for certain. A brief dietary history will
rapidly establish whether or not the patient is a vegan or takes a very
inadequate diet. Endoscopy and gastric biopsy will show features
of gastric atrophy and help to exclude gastric carcinoma. Follow-​
through radiographic examination of the small intestine will help
to exclude a small-​intestinal lesion (e.g. duodenal or jejunal diver-
ticulosis). The serum gastrin concentration is raised in patients with
PA and the serum is tested for antibodies to IF, parietal cells, and
section 22  Haematological disorders
5422
thyroid; serum immunoglobulins are measured in view of the asso-
ciation with hypogammaglobulinaemia.
Diagnosis of the cause of folate deficiency
An inadequate diet is usually at least partly implicated, but an exact
estimate of dietary intake from the clinical history is impossible be-
cause of variation in folate content of foods, losses in cooking, and
size of portions. Often it is the general social circumstances that sug-
gest a poor intake. Drug intake, particularly of barbiturates, is im-
portant. Many underlying inflammatory or malignant diseases may
exaggerate the tendency to folate deficiency in patients with inad-
equate diets. The main cause of malabsorption of folate is gluten-​
induced enteropathy; in patients with severe folate deficiency, tests
for transglutaminase and endomysial antibodies and a duodenal bi-
opsy are usually necessary. In certain tropical countries, sprue may
cause a generalized malabsorption syndrome in which folate defi-
ciency commonly occurs.
Treatment of megaloblastic anaemia
Therapy is aimed at correcting the anaemia, completely replenishing
the body of whichever vitamin is deficient, treatment of the under-
lying disorder, and prevention of relapse. In most cases, it is possible
to diagnose which deficiency is present before starting therapy.
Vitamin B12 deficiency
Hydroxocobalamin 1 mg intramuscularly given six times at several
days’ interval over the first few weeks will restore normal vitamin
B12 stores. There is no evidence that patients with vitamin B12 neur-
opathy derive greater benefit from more frequent doses, although
many physicians use these for 6 months or so.
Response to therapy
The patient feels better within 24 to 48 h, and the mild fever, if not
due to infection, abates. A painful tongue and an uncooperative, dis-
orientated state may also be improved in 48 h. The white cell count
becomes normal by 3 to 7 days and the platelet count rises and may
reach levels of 500 to 1000 × 109/​litre before falling to normal at
about 10 to 14 days. The bone marrow reverts to normoblastic by
36 to 48 h, although giant metamyelocytes persist for 10 to 12 days.
The neuropathy always improves with therapy but residual deficits
remain in some patients; this applies usually to those with the longest
histories or the most severe manifestations, particularly where there
is subacute combined degeneration of the spinal cord and spastic
paraparesis.
Maintenance
Hydroxocobalamin, 1 mg intramuscularly, is given once every
3 months for life in PA and most other causes of vitamin B12 de-
ficiency to prevent relapse. The life expectancy in PA once treated
is as good as that in the general population in women, and slightly
lower in men, probably due to the increased incidence of carcinoma
of the stomach. In a few patients with vitamin B12 deficiency, the
underlying cause can be reversed; for example, expulsion of the fish
tapeworm, improvement of vegan diet, surgical correction of an in-
testinal stagnant loop. A few micrograms of vitamin B12 can be ab-
sorbed each day in PA from oral doses of 1 mg or more by passive
diffusion, but this maintenance therapy is usually reserved for those
who cannot have injections—​for example, those with a bleeding dis-
order, or who refuse them—​and for the extremely rare individual
who is allergic to all injectable forms of vitamin B12. Vegans may be
maintained on much smaller oral doses of vitamin B12 each day, such
as 50 µg as a tablet or syrup.
Prophylaxis
Vitamin B12 therapy should be given from the time of operation
after total gastrectomy or ileal resection. Patients with PA tend to
develop iron deficiency anaemia and they may also develop thyroid
disorders or carcinoma of the stomach. It is advisable that a regular
blood count be made once a year. Routine endoscopy is not war-
ranted but these diseases must be particularly borne in mind if rele-
vant symptoms or signs develop.
It is unclear whether vitamin B12 should be given orally or par-
enterally to those with biochemical (subclinical) vitamin B12 de-
ficiency without anaemia or macrocytosis or clinical symptoms.
Trials are needed to clarify this.
Folate deficiency
This is corrected by giving 5 mg folic acid by mouth daily. It is essen-
tial to first exclude vitamin B12 deficiency so that precipitation of a
neuropathy is avoided. It is usual to continue for at least 4 months
until there is a completely new set of red cells, although body stores
will theoretically be normal within a few days of therapy. In patients
with severe malabsorption of folate, larger oral doses of folic acid
(e.g. 5 mg three times a day) may be used but it is not necessary to
give parenteral folate except for those unable to swallow tablets. The
response to therapy is as described for vitamin B12. The decision
whether or not to continue folic acid beyond 4 months depends on
whether or not the cause can be corrected. In practice, long-​term
folic acid is usually needed only in patients with severe haemo-
lytic anaemias (e.g. sickle cell anaemia and thalassaemia major),
myelofibrosis, and in gluten-​induced enteropathy when a gluten-​
free diet is either unsuccessful or not feasible.
Prophylactic folic acid
This should be given to all pregnant women to prevent megaloblastic
anaemia and reduce the incidence of NTDs (doses of 300–​400 µg/​
day are used). Only 19% of women in a Northern Ireland study had
taken folic acid preconception, resulting in a lower red cell folate
and thus increased risk of NTDs, during the first trimester, than
in women who had taken preconception folic acid. Studies in the
United States of America show that black and Hispanic women have
a lower dietary intake of folate than white non-​Hispanic women and,
therefore, particularly need folic acid supplements periconception.
Doses of 5 mg/​day would have a greater effect but currently need a
medical prescription in the United Kingdom. They are given if there
has been a previous infant with an NTD. Folic acid is given to pa-
tients undergoing regular haemodialysis or peritoneal dialysis, to
premature infants weighing less than 1.5 kg at birth, and to selected
patients in intensive care units or receiving parenteral nutrition. In
young children exposed to a high risk of malaria, combined iron and
folic acid supplements may be harmful and should be avoided.
Folate therapy has been shown to improve chromosomal stability
in the fragile X syndrome, even though these patients do not have
folate deficiency or a demonstrable defect of folate metabolism.
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5423
Food fortification
Mandatory fortification of cereals and grains with folic acid (140 μg/​
100 g cereal grain) aimed at reducing the incidence of NTDs began
in the United States of America in 1998 and is now also practised in
Canada, Chile, and 70 other countries amounting to about 20% of
the world’s population. Median serum folate in clinical specimens
in United States of America rose from 12.6 to 18.7 μg/​litre between
1997 and 1998. There was also a fall in serum homocysteine levels.
The theoretical side effects of fortification are largely in patients
with unsuspected vitamin B12 deficiency who, it has been suggested,
might present with neuropathy if the extra folate consumed prevents
the development of anaemia due to vitamin B12 deficiency. However,
the small amounts of folic acid with each meal would be largely if
not entirely converted to methyl-​THF by the small intestine and this
folate is not able to affect haematopoiesis in vitamin B12 deficiency.
In keeping with this, there is no evidence for an increased incidence
of nonanaemic subjects with low serum vitamin B12 levels in the
United States of America since fortification. In the United Kingdom,
fortification of flour with folic acid (240 μg/​100 g flour) has been re-
commended but not implemented. As discussed previously, there
is no evidence that fortification would affect the incidence in the
population of any type of cancer. Fortification of grain with vitamin
B12 has also been suggested to reduce the incidence of NTDs, but
this has not been implemented in any country.
Folinic acid (5-​formyl-​THF)
This reduced folate is used to prevent or treat toxicity due to metho-
trexate or other dihydrofolate reductase inhibitors.
Severely ill patients
Some patients, usually elderly, are admitted to hospital severely ill
with megaloblastic anaemia, perhaps in congestive heart failure or
with pneumonia. In this case, it is necessary to commence therapy
immediately after obtaining blood for vitamin B12 and folate assay,
before it is known which deficiency is present. Both vitamins should
be given simultaneously in large doses. Heart failure and infection
should be treated in conventional fashion but blood transfusion
should be avoided, except in cases of extreme anaemia, when 1 to 2
units of packed cells may be given slowly.
Other therapy
Hypokalaemia has been reported to occur during initial therapy
but is, rarely, if ever, clinically important. An attack of gout has
been reported on the days 6 to 7 of therapy. Most patients develop
hyperuricaemia at this stage but the clinical disease probably only
occurs in those with a strong gouty tendency. Iron deficiency com-
monly develops in the first few weeks of therapy and this should be
treated initially with oral ferrous sulphate in the usual way.
Megaloblastic anaemia due to inborn errors
of folate or vitamin B12 metabolism
Folate
A number of babies have been described with congenital de-
ficiency of one or other enzyme concerned in folate metab-
olism: 
5-​methyltetrahydro-​folate,
methylene
THF-​reductase,
FIGLU-​transferase, methenyl-​THF cyclohydrolase. Some of the ba-
bies had multiple congenital defects including the heart and cerebral
ventricles and nearly all showed impaired mental development. In
the methylfolate transferase deficiency, megaloblastic anaemia was
present.
Vitamin B12
Congenital deficiency of TCII was first reported as an autosomally
recessive disease in 1971 in two siblings who developed megalo-
blastic anaemia requiring therapy with large daily doses of vitamin
B12 at 3 and 5 weeks of age. Similarly affected families have been de-
scribed. A spectrum of mutations in the gene for TCII have been de-
tected. In some cases, TCII is undetected; in others, often presenting
later in life, functionally inactive TCII has been detected. The serum
vitamin B12 concentration is normal, vitamin B12 being bound to
TCI. Absorption of vitamin B12 is impaired. Treatment is with mas-
sive doses of vitamin B12 (e.g. 1 mg intramuscularly, three times each
week). Delay in treatment may allow a neuropathy to occur. In con-
trast, in subjects with rare, inherited, mutations of TCI, low serum
vitamin B12 concentrations occur, but haematopoiesis is normal.
Children with one form of congenital methylmalonic aciduria,
which responds to vitamin B12 therapy in large doses, have been
shown to have a defect in conversion of hydroxocobalamin to
adocobalamin. They do not show megaloblastic anaemia. In a
few, this defect has been associated with a defect of formation of
methylcobalamin and with homocystinuria, but some of the chil-
dren have also surprisingly not shown megaloblastic anaemia.
Neurological abnormalities are usual. Homocystinuria and meg-
aloblastic anaemia without methylmalonic aciduria have also been
reported. In some cases, the defect appears to be in maintaining
vitamin B12 bound to methionine synthase in the reduced state.
Megaloblastic anaemia due to acquired
disturbances of folate or vitamin B12 metabolism
Folate
Therapy with dihydrofolate reductase inhibitors may cause megalo-
blastic anaemia. This is usual with methotrexate and less likely with
pyrimethamine unless high doses are used or the patient is already
folate deficient. Trimethoprim and triamterene are very weak folate
antagonists in humans, but may precipitate megaloblastic anaemia
in patients already B12 or folate deficient (mentioned earlier).
Vitamin B12
Nitrous oxide (N2O)
This anaesthetic gas oxidizes vitamin B12 from the active fully reduce
cob(I)alamin form to the inactive cob(II)alamin and cob(III)alamin
forms, inactivating methylcobalamin and hence methionine syn-
thase. Megaloblastosis develops within several hours in humans. This
recovers over several days when exposure to N2O is discontinued.
After many weeks of exposure to N2O, monkeys develop a neuropathy
resembling vitamin B12 deficiency neuropathy in humans; peripheral
neuropathies and more severe neurological disease have also been
described in humans (e.g. dentists and anaesthetists) repeatedly ex-
posed to the gas. When N2O is used as anaesthetic for patients with
low vitamin B12 stores, megaloblastic anaemia or neuropathy may
section 22  Haematological disorders
5424
be precipitated months later, due to failure to replenish vitamin B12
stores by absorption. Recovery from N2O exposure needs new co-
balamin and also synthesis of new apoenzyme (methionine synthase)
because this protein is also damaged by active oxygen derived from
the N2O–​cobalamin reaction. Methylmalonic aciduria has not been
found in animals or humans exposed for short periods to N2O, as
methylmalonic CoA mutase does not need reduced B12.
Megaloblastic anaemia not due to folate or
vitamin B12 deficiency or metabolic defect
Congenital
Orotic aciduria
This is a very rare, recessive disorder involving two consecutive en-
zymes (orotidsylic pyrophosphatase and orotidylic decarboxylase)
in pyrimidine synthesis and presents with megaloblastic anaemia in
the first few months of life. The diagnosis is made if needle-​shaped,
colourless crystals of orotic acid are found in the urine, daily ex-
cretion ranging from 0.5 to 1.5 g. Heterozygotes excrete slightly
raised amounts of orotic acid but show no haematological disorder.
Treatment with uridine (1–​1.5 g/​day) leads to a haematological re-
sponse, restoration of normal haematopoiesis and growth, and re-
duction in orotic acid excretion.
Lesch–​Nyhan syndrome
A few patients with this rare disorder of purine synthesis have shown
megaloblastic change but whether this was due to associated folate
deficiency or a direct result of reduced purine synthesis is not certain
(see Chapter 12.4).
Vitamin E deficiency
This has been reported to cause megaloblastosis in a group of chil-
dren with kwashiorkor. However, many were also folate deficient.
Vitamin C deficiency
Megaloblastic appears to be due to associated folate deficiency.
Thiamine responsive
About 12 cases have been well documented. They have also shown
sideroblastic change and a defect in phosphorylation of thiamine
has been implicated. Diabetes mellitus and sensorineural deafness
are additional features. There is a fault in thiamine phosphorylation
due to a genetic defect of the phosphorylase enzyme.
Responding to large doses of vitamin B12 and folate
A single patient has been reported who needed both vitamins in
large doses, but the site of the defect was not elucidated.
Congenital dyserythropoietic anaemia
Some cases of congenital dyserythropoietic anaemia show megalo-
blastic changes not due to vitamin B12 or folate deficiency.
Acquired
Megaloblastic changes are often marked in acute myeloid leukaemia
and less commonly in other forms of acute myeloid leukaemia.
They also occur in the myelodysplastic syndromes.
Drugs that directly inhibit purine or pyrimidine synthesis
(e.g. cytosine arabinoside, 5-​fluorouracil, hydroxycarbamide,
6-​mercaptopurine, or azathioprine and azidothymidine (AZT))
may cause megaloblastic anaemia. Alcohol has also been found to
have a direct effect on the bone marrow, causing megaloblastosis in
some cases even in the absence of vitamin B12 or folate deficiency.
On the other hand, drugs that inhibit mitosis (e.g. colchicine or
daunorubicin) or alkylate preformed DNA (e.g. cyclophosphamide,
chlorambucil, or busulfan) do not cause megaloblastosis.
Other deficiency anaemias
Vitamin C
Anaemia is usual in scurvy but the pathogenesis is complicated. It is
likely that vitamin C has a direct effect on erythropoiesis but folate
and iron deficiencies, haemorrhage, or haemolysis often complicate
the picture.
Biochemical and nutritional aspects
Vitamin C is needed for collagen synthesis by its involvement in
the hydroxylation of protein and for maintenance of intercellular
substance of skin, cartilage, periosteum, and bone. It may also have
a general role in oxidation–​reduction systems, for example, gluta-
thione, cytochromes, pyridine, and flavin nucleotides. Although
vitamin C is also thought to be needed for maintaining body fol-
ates in the reduced active state, the exact reactions involved are
unclear. Vitamin C has a particular role in iron metabolism, iron
excess causing increased utilization of vitamin C and in extreme
cases clinical scurvy, whereas iron deficiency is associated with a
raised leucocyte ascorbate concentration. Vitamin C is needed for
incorporation of iron from transferrin into ferritin and for iron
mobilization from ferritin. Vitamin C therapy increases iron excre-
tion in patients receiving subcutaneous desferrioxamine infusions
and also, at least in experimental animals, affects iron distribution
by increasing parenchymal relative to reticuloendothelial iron.
Minimum adult daily requirements for vitamin C are about 10 mg
but 30 to 70 mg is recommended; utilization, and therefore require-
ments, are relatively higher in infants, children, and pregnant and
lactating women. Vitamin C may be excreted as such but is also
broken down to oxalate.
Vitamin C is present in food as its reduced (ascorbic acid) and
oxidized (dehydroascorbic acid) forms, the highest concentrations
occurring in green vegetables, fruits, liver, and kidney. Potatoes are
not a rich source but provide a substantial proportion of normal
dietary intake. Cooking, particularly in alkaline conditions with large
volumes of water, destroys the vitamin, which is also lost on storage
with exposure to the air. Absorption occurs through the length of the
small intestine and deficiency is never solely due to malabsorption.
The anaemia of scurvy is typically normochromic, normocytic
with a slightly raised reticulocyte count (to 5–​10%) and a
normoblastic marrow with erythroid hyperplasia. This suggests a
direct role for vitamin C in erythropoiesis but not all patients with
clinical scurvy are anaemic. Extravascular haemolysis with mild
jaundice and increased urobilinogen excretion occurs in many of
the patients. Moreover, in many the anaemia is complicated by folate
deficiency (due to inadequate folate intake) with a megaloblastic
22.6.6  Megaloblastic anaemia and miscellaneous deficiency anaemias
5425
marrow, or in a few by iron deficiency due to external haemorrhage,
reduced diet intake, and possibly reduced iron absorption. In a
few patients placed on a low-​folate diet, response of megaloblastic
haematopoiesis to vitamin C alone has been described. In others,
response of the megaloblastic anaemia to folic acid alone on a diet
low in vitamin C has occurred; but in most such cases, both vitamin
C and folic acid have been found necessary.
Vitamin B6
This, as its coenzyme form pyridoxal-​5-​phosphate, is involved
in many reactions of the body, especially transaminases and
decarboxylases. It is also a cofactor in the important rate-​limiting
reaction in haem synthesis, δ-​aminolaevulinic acid (ALA)-​
synthase. It occurs in natural tissues in three major forms: pyri-
doxine, pyridoxamine, and pyridoxal phosphate. Red cells are
capable of interconverting them. Anaemia due purely to vitamin
B6 deficiency has been produced in animals. It is hypochromic
and microcytic with a raised serum iron and increased iron in
erythroblasts, with some partial or complete ring sideroblasts.
A similar anaemia has occurred in humans with malabsorption,
pregnancy, or haemolysis but has not been fully documented
to respond to physiological doses of vitamin B6 alone. Vitamin
B6-​responsive anaemia is, however, well documented among
patients with sideroblastic anaemia of all types. Pyridoxine re-
sponses occur particularly in the inherited form (when it is as-
sumed that a fault in one or other enzyme of haem synthesis,
e.g. ALA-​synthase, increases the need for pyridoxal phosphate
as cofactor) and when sideroblastic anaemia occurs in patients
receiving pyridoxine antagonists, such as antituberculous drugs.
The value of pyridoxine dietary supplements in lowering serum
homocysteine and reducing the incidence of cardiovascular dis-
ease has yet to be proven.
Riboflavin
On the basis of studies in experimental animals and humans fed
a deficient diet together with a riboflavin antagonist, deficiency of
this vitamin is known to cause a normochromic, normocytic an-
aemia associated with a low reticulocyte count and red cell aplasia
in the marrow, sometimes with vacuolated normoblasts. The exact
biochemical basis is undecided. Clinically, a similar anaemia may
occur in pure form but is usually associated with the anaemia due
to protein deficiency, as in kwashiorkor or marasmus. Other clinical
features of riboflavin deficiency—​dermatitis, angular cheilosis, and
glossitis for example—​may be present.
Thiamine
For discussion, see under megaloblastic anaemia not due to folate or
vitamin B12 deficiency or metabolic defect.
Nicotinic acid, pantothenic acid, and niacin
Deficiencies of these vitamins cause anaemia in experimental ani-
mals, but anaemia purely due to one or other of these deficiencies
has not been established to occur in humans.
Vitamin E
This vitamin is needed for preventing peroxidation of cell mem-
branes. A haemolytic anaemia responding to vitamin E has been
reported in premature infants. Less well documented is a macrocytic
anaemia due to vitamin E deficiency in protein–​calorie-​deficient
infants and aggravation of anaemia in patients with thalassaemia
major because of vitamin E deficiency.
Protein deficiency
Anaemia is usual in both ‘pure’ protein deficiency (kwashiorkor)
and in protein–​calorie malnutrition (marasmus). It has been re-
ported in many parts of the world where malnutrition, especially
in children and pregnant women, is common. The anaemia also
occurs in patients with gastrointestinal disease and severe malab-
sorption. The anaemia is typically normochromic, normocytic,
and of the order of 80 to 90 g/​litre. The reticulocyte count is usu-
ally reduced and the marrow may show a selective reduction in
erythropoiesis. Experimental studies in animals suggest that the
anaemia is largely due to reduced serum erythropoietin levels con-
sequent on a lack of stimulus for erythropoietin secretion. Lack
of amino acids for synthesis of erythropoietin or globin is not the
cause. In many patients, the anaemia is complicated by infection,
folate or iron deficiency, and possibly other vitamin deficiencies
(e.g. riboflavin, vitamin E) and then it may be more severe and
show additional morphological abnormalities in the blood and
marrow.
FURTHER READING
General
Devalia V, Hamilton MS, Molloy AM (2014). Guidelines for the diag-
nosis and treatment of cobalamin and folate disorders. Br J Haem,
166, 496–​513.
Longo DL (2015). Drug-​induced megaloblastic anemia. N Engl J Med,
373, 1649–​58.
McLean E, de Benoist B, Allen LH (2008). Review of the magnitude of
folate and vitamin B12 deficiencies worldwide. Food Nutr Bull, 29
Suppl 2, 538–​51.
Vitamin B12
Carmel R (2011). Biomarkers of cobalamin (vitamin B12-​12) status in
the epidemiologic setting: A critical overview of context, applica-
tions, and performance characteristics of cobalamin, methylmalonic
acid and holotranscobalamin 11. Am J Clin Nutr, 97, 541–​56.
Carmel R (2012). Subclinical cobalamin deficiency. Curr Opin
Gastroenterol, 28, 151–​8.
Green R (2017). Vitamin B12 deficiency from the perspective of the
practising hematologist. Blood, 129, 2603–11.
Green R, et al. (2017). Vitamin B12 deficiency. Nat Rev Dis Primers,
3, 17040.
Hershko C, et al. (2006). Variable hematological presentation of auto-
immune gastritis: age related progression from iron deficiency to
cobalamin depletion. Blood, 107, 1673–​9.
Lewerin C (2008). Serum biomarkers for atrophic gastritis and anti-
bodies against Helicobacter pyloric in the elderly: Implications for
vitamin B12, folic acid and iron status and response to oral vitamin
therapy. Scand J Gast, 143, 1502–​8.
Mills JL (2011). Do high blood folate concentrations exacerbate meta-
bolic abnormalities in people with low vitamin B12 status. Am J Clin
Nutr, 94, 495–​500.

22.6.7 Disorders of the synthesis or function of h
22.6.7 Disorders of the synthesis or function of haemoglobin 5426 Deborah Hay and David J. Weatherall†
section 22  Haematological disorders
5426
Stabler SP (2013). Clinical practice. Vitamin B12 deficiency. N Eng J
Med, 368, 149–60.
Toh B-​H, et al. (2012). Cutting edge issues in autoimmune gastritis.
Clin Rev Allerg Immunol, 42, 269–​78.
Weir DG, Scott JM (1997). Brain function in the elderly: role of B12 and
folate. Br Med Bull, 55, 669–​82.
Weng TC, et al. (2015). Anaemia and related nutrient deficiencies after
Roux-​en-​Y by-​pass surgery: a systemic review and meta-​analysis.
BMJ Open, 5, e006964.
Folate
Bergen NE, et al. (2012). Homocysteine and folate concentrations in
early pregnancy and the risk of adverse pregnancy outcomes: the
Generation R Study. Br J Obstet Gynaecol, 119, 739–​51.
Kibar Z, et al. (2007). Mutations in VANGLI associated with neural
tube defects. N Engl J Med, 356, 1432–​7.
Mills JH, et al. (2003). Low B12 concentrations in patients without an-
emia: the effect of folic acid fortification of grain. Am J Clin Nutr,
77, 1474–​7.
Qui A, et al. (2006). Identification of an intestinal folate transporter
and the molecular basis for hereditary folate malabsorption. Cell,
127, 917–​28.
Neural tube defects
Correa A, et al. (2012). Lack of periconceptional vitamins or supple-
ments that contain folic acid and diabetes mellitus-​associated birth
defects. Am J Obstet Gynecol, 206, 218.e1–​13.
Hoffbrand AV (2014). Professor John Scott, folate and neural tube de-
fects. Br J Haematol, 164, 496–​502.
McNulty B, et al. (2011). Women’s compliance with current folic acid
recommendations and achievement of optimal vitamin status for
preventing neural tube defects. Hum Reprod, 26, 1530–​6.
Molloy AM, et  al. (2009). Maternal vitamin B 12 status and risk
of neural tube defects in a population with high neural tube de-
fect prevalence and no folic acid fortification. Pediatrics, 123,
917–​23.
Morris JK, et al. (2016). Prevention of neural tube defects in the UK:
a missed opportunity. Arch Dis Child, 101, 604–7.
Mental disorders
Li MM, et al. (2014). Efficacy of vitamin B supplementation in mild
cognitive impairment and Alzheimer’s disease: a systematic review
and meta-​analysis. Curr Alzheimer Res, 11, 844–​52.
Moore EM, et al. (2014). Among vitamin B12 deficient older people,
high folate levels are associated with worse cognitive func-
tion: combined data from three cohorts. J Alzheimers Dis, 39,
661–​8.
Morris MS, Selhub J, Jacques PF (2012). Vitamin B-​12 and folate
status in relation to decline in scores on the mini-​mental state
examination in the Framingham heart study. J Am Geriatr Soc, 60,
1457–​64.
Wald DS, et  al. (2011). Serum homocysteine and dementia:  meta-​
analysis of eight cohort studies involving 8669 participants.
Alzheimers Dement, 7, 412–​17.
Cardiovascular disease
Huo Y, et al. (2015). Efficacy of folic acid therapy in primary preven-
tion of stroke among adults with hypertension in China. The CSPPT
randomized clinical trial. JAMA, 313, 1325–​35.
Huo Y, et al. (2012). Efficacy of folic acid supplementation in stroke
prevention: new insight from a meta-​analysis. Int J Clin Pract, 66,
544–​51.
Ji Y, et al. (2013). Vitamin B supplementation, homocysteine levels and
the risk of cerebrovascular disease. A meta-​analysis. Neurology, 81,
1298–​307.
Stampfer M, Willett W (2015). Folate supplements is stroke preven-
tion. Targeted trial trumps the rest. JAMA, 313, 1321–​2.
Wald DS, et  al. (2011). Reconciling the evidence on serum homo-
cysteine and ischaemic heart disease: a meta-​analysis. PLoS One,
6, e1643.
Yang H-​T, et al. (2012). Efficacy of folic acid supplementation in car-
diovascular disease prevention. An updated meta-​analysis of ran-
domized controlled trials. Eur J Int Med, 23, 745–​54.
Cancer
Vollset SE, et  al. (2013). Effects of folic acid supplementation on
overall and site -​ specific cancer incidence during randomized
trials:  meta-​analysis of data on 50,000 individuals. Lancet, 381,
1029–​36.
Zhang D, Guo Y, Cui W (2015). Elevated homocysteine level and folate
deficiency associated with an increased overall risk of carcinogen-
esis: meta-​analysis of 83 case control studies involving 35,758 indi-
viduals. PLos One, 10e0123423.
Miscellaneous
Adams EB (1970). Anemia associated with protein deficiency. Semin
Haematol, 7, 55–​66.
Cox EV (1968). The anaemia of scurvy. Vitam Horm, 26, 635–​52.
Rindi G, et al. (1994). Further studies of erythrocyte thiamin trans-
port and phosphorylation in seven patients with thiamin-​responsive
megaloblastic anaemia. J Inher Metabol Dis, 17, 667–​77.
22.6.7  Disorders of the synthesis or
function of haemoglobin
Deborah Hay and David J. Weatherall†
ESSENTIALS
The inherited disorders of haemoglobin are the commonest single-​
gene disorders in the world. They cause significant morbidity and
mortality in those individuals who are severely affected and place
a major burden on health services in some places, in particular the
Mediterranean region, sub-​Saharan Africa, and South-​East Asia,
when economic conditions improve and infant and childhood death
rates fall. Migrations of populations from high-​incidence areas for
the haemoglobin disorders, together with the general ease of inter-
national travel, means that patients with these conditions are now
seen in all regions of the world.
† It is with great regret that we report that David J. Weatherall died on 8
December, 2018.
22.6.7  Disorders of the synthesis or function of haemoglobin
5427
Disorders of haemoglobin can be genetic or acquired and may be
caused by disordered production of one or more globin chains or
structural change in the globin chain. The most important disorders
are the genetic conditions thalassaemia and sickle cell disease.
Thalassaemias
A heterogeneous group of genetic disorders, all resulting from a re-
duced rate of production of one or more of the globin chains of
haemoglobin and inherited in a simple Mendelian fashion. They
are clinically classified according to their severity into major (a se-
vere transfusion-​dependent disorder), intermediate (characterized
by anaemia and splenomegaly), and minor (a symptomless carrier
state) forms.
The β thalassaemias, which occur in patients with ethnic origin
from a broad belt ranging from the Mediterranean and parts of
North and West Africa through the Middle East and Indian sub-
continent to South-​East Asia, are the most important types of
thalassaemia because they are very common and produce severe
anaemia in their homozygous and compound heterozygous states.
Most countries in which the disease is common are putting a major
effort into programmes for its prevention (population screening
and prenatal diagnosis). Symptomatic management of severe dis-
ease requires regular blood transfusion, judicious use of splenec-
tomy if hypersplenism develops, and chelating agents to reduce iron
overload.
Sickle cell disease
Haemoglobin S differs from haemoglobin A  by the substitution
of valine for glutamic acid at position 6 in the β globin chain, and
homozygosity for haemoglobin S produces the state of sickle cell
disease. This occurs very frequently in African populations and
sporadically throughout the Mediterranean region and the Middle
East, with extensive pockets in India. Typical presentation is in in-
fancy with symptoms related to anaemia or infection, but clinical
manifestations are very variable, ranging from an almost incidental
finding on routine haematological examination to severe haemolytic
anaemia interspersed with frequent exacerbations or crises, which
can take various forms and may be life-​threatening. Management of
both acute and chronic complications remains largely supportive,
with hydroxycarbamide being the only clinically proven effective
treatment to date in routine clinical use. However, investigational
agents targeting the complex pathophysiology of sickle cell anaemia
are in clinical trials and promise to improve outcomes for patients
with this disease.
Introduction
Disorders of the synthesis or structure of haemoglobin may be
either inherited or acquired. The inherited disorders of haemo-
globin are the commonest single-​gene disorders with hundreds
of millions of carriers worldwide, and at least 300 000 severely af-
fected homozygotes or compound heterozygotes born annually. The
greatest prevalence is in low and middle income countries of the
tropical belt. The main reason for the extremely high frequency of
these conditions in these regions is that heterozygotes show a vari-
able degree of resistance to infection with Plasmodium falciparum
malaria. There has therefore been intense selection for these muta-
tions in countries where malaria is common. As economic condi-
tions improve in these countries, and infant and childhood death
rates fall, the genetic disorders of haemoglobin place a major
burden on health services.
As a result of migration of populations from high-​incidence
areas for the haemoglobin disorders, these conditions are now seen
with increasing frequency in all parts of the world. Some of them,
particularly sickle cell anaemia and the more severe forms of thal-
assaemia, can produce life-​threatening medical emergencies. It is
important for all clinicians to have a working knowledge of their
clinical features, management, and prevention.
Haemoglobin disorders are also of particular interest because
they were the first group of diseases to be analysed genetically. More
is known about their molecular pathology than any other genetic
disorders and their study has given us insight into the wide reper-
toire of mutations that underlie inherited diseases in humans.
Before describing the haemoglobin disorders, it is necessary to
discuss briefly the structure, function, and synthesis of haemoglobin
and the way that it is genetically determined.
The structure, function, genetic control, and
synthesis of haemoglobin
Structure
Human haemoglobin is heterogeneous at all stages of develop-
ment; different haemoglobins are synthesized in the embryo, fetus,
and adult, each adapted to the particular oxygen requirements.
Each human haemoglobin has a tetrameric structure made up of
two different pairs of globin chains, each attached to one haem mol-
ecule (Fig. 22.6.7.1). At each stage of development, the tetramer is
made from two alpha (α)-​like chains and two beta (β)-​like chains.
The α-​like chains are α and zeta (ζ) globins, encoded by adjacent
genes on the telomeric tip of chromosome 16. Of these, ζ globin is
transcribed in embryonic life, and α globin in fetal and adult life. The
β globin locus on chromosome 11 encodes β-​like chains expressed
at different stages of maturation—​epsilon (ε) globin expressed em-
bryonically, gamma (γ) globin in the fetus, and delta (δ) globin and β
globin in the adult. Thus, adult and fetal haemoglobins have α chains
combined with β chains (Hb A, α2β2), δ chains (Hb A2, α2δ2), or γ
chains (Hb F, α2γ2). In embryos, ζ chains combine with γ chains to
produce Hb Portland (ζ2γ2), or with ε chains to make Hb Gower 1
(ζ2ε2), and α and ε chains combine to form Hb Gower 2 (α2ε2). Fetal
haemoglobin (Hb F α2γ2) is itself heterogeneous; there are two kinds
of γ chains which differ in their amino acid composition at position
136, where they have either glycine (Gγ) or alanine (Aγ). The Gγ
and Aγ chains are the products of separate (Gγ and Aγ) loci in the
β globin cluster.
Function
The sigmoid shape of the oxygen dissociation curve, which reflects
the allosteric properties of haemoglobin, ensures that oxygen is rap-
idly taken up at high oxygen tensions in the lungs, and that it is re-
leased readily at the lower tensions encountered in the tissues. The
shape of the curve is due to cooperativity between the four haem
molecules. When one takes on oxygen, the affinity for oxygen of the
section 22  Haematological disorders
5428
remaining haems of the tetramer increases dramatically. This is be-
cause haemoglobin can exist in two configurations, deoxy (T) and
oxy (R) (T and R stand for tight and relaxed states, respectively).
The T form has a lower affinity than the R form for ligands such as
oxygen. During the sequential addition of oxygen to the four haems,
transition from the T to R configuration occurs and the oxygen af-
finity of the partially liganded molecule increases rapidly.
The position of the oxygen dissociation curve can be modi-
fied in many ways. First, oxygen affinity decreases as CO2 tension
rises, the Bohr effect. This facilitates oxygen delivery to the tissues,
where the pH falls due to CO2 generation. The opposite effect oc-
curs in the lungs. Oxygen affinity is also modified by the level of
2,3-bisphosphoglycerate (2,3-​BPG) in the red cell. Increasing con-
centrations move the curve to the right, reducing oxygen affinity.
Diminishing concentrations have the opposite effect. The 2,3-​BPG
mechanism plays an important role in the response to hypoxia (see
Chapter 22.6.2).
Genetic control
The arrangement of the two main families of globin genes is illus-
trated in Fig. 22.6.7.2. The β-​like globin genes form a linked cluster
on chromosome 11 that spans about 60 kb; they are arranged in the
order 5′-​ε-​Gγ-​Aγ-​ψβ-​δ-​β-​3′. The α-​like globin genes form a linked
cluster on chromosome 16, in the order 5′-​ζ-​ψζ-​ψα-​α2-​α1-​3′. The
ψβ, ψζ, and ψα genes are pseudogenes; their sequences resemble
the β, ζ, or α genes but contain mutations which prevent them from
functioning as structural genes. They may be ‘burnt out’ remnants of
genes which were functional at an earlier stage of evolution.
The molecular machinery required for gene expression has been
defined comprehensively for the globin genes, in part through the
study of patients with unusual forms of thalassaemia. The promoters,
regulatory elements, 5′ and 3′ untranslated regions and splice sites
are all well defined for both α globin and β globin. Individual globin
chains combine with haem, which is synthesized through a separate
pathway to form definitive tetrameric haemoglobin molecules.
Classification of the disorders of haemoglobin
The main groups of disorders of haemoglobin are shown in
Box 22.6.7.1. The genetic disorders are divided into those in which
there is a reduced rate of production of one or more of the globin
chains, the thalassaemias, and those in which a structural change in
a globin chain leads to instability or to abnormal oxygen transport.
In addition, there is a harmless group of mutations, known collect-
ively as hereditary persistence of fetal haemoglobin, that interfere
with the normal switching of fetal to adult haemoglobin production.
Tyr HC2
Val E11
FG2
F9
G1
F1
G5
V
M
F8
C3
C7
M
CD1
CD2
CD
E7
C
P
D1
E1
D7
B5
A16
B1
AB1
G19
GH4
A1
H5
E20
EF1
G15
NA2
EF3
H16
V
M
NA1
NH3
+
HI
C1
E5
Fig. 22.6.7.1  The α chain subunit of human haemoglobin showing
the position of the haem molecule in a cleft formed by the globin chain.
The helical parts of the chain are given letters of the alphabet and each
amino acid residue in each helical region has a specific number, for
example, Val E11 is the 11th amino acid in the E helical region. The
nonhelical regions of the N-​ and C-​terminal ends of the chains are
labelled NA and HC respectively.
Reproduced by permission of Dr M F Perutz and the editors of the Cold Spring
Harbor Symposia for Quantitative Biology.
1 Kb
11
Embryo
Fetus
Hb Portland
Hb Gower 1
Hb Gower 2
HbF
HbA
HbA2
16
105
104
31
31
30
32 99 100
Adult
ζ2
ψζ1
ψα 2
ψβ
α2
α1
ε
Gγ
Aγ
β
δ
Fig. 22.6.7.2  The genetic control of human haemoglobin. Two of the genes are enlarged
to show the introns (unshaded) and exons (purple) plus 3ʹ and 5ʹ untranslated regions (lilac).
1 kb = 1000 nucleotide bases.
22.6.7  Disorders of the synthesis or function of haemoglobin
5429
The acquired disorders of haemoglobin can also be subdivided
into those characterized by defective synthesis of the globin chains
and those in which the structure of the haem molecules is altered,
leading to inefficient oxygen transport.
Like all biological classifications, this way of classifying the
haemoglobin disorders is not entirely satisfactory. For example,
some structural variants are synthesized in reduced amounts and
hence produce the clinical picture of thalassaemia.
The thalassaemias
Historical introduction
The thalassaemias are the commonest of the inherited haemato-
logical disorders and, indeed, are the commonest single-​gene dis-
orders in the world population. The condition was first recognized
in 1925 by Thomas B. Cooley, who described infants who became
profoundly anaemic and developed splenomegaly over the first year
of life. A  milder form was described independently in the same
year by Fernando Rietti. As further cases were identified the dis-
order was variously called von Jaksch’s anaemia, splenic anaemia,
erythroblastosis, Mediterranean anaemia, or Cooley’s anaemia. In
1936, George Whipple and Lesley Bradford recognized that many
of their patients came from the Mediterranean region and hence
they invented the word ‘thalassaemia’ from the Greek word meaning
‘the sea’. Although it was realized later that the disorder occurs
throughout the world and is not localized to the Mediterranean re-
gion, the name has stuck.
Thalassaemia is extremely heterogeneous. Its clinical picture can
result from the interaction of many different genetic defects. This
chapter concentrates mainly on the clinical and haematological
aspects; readers who wish to learn more about the molecular
pathology and population genetics of thalassaemia are referred to
reviews and monographs listed at the end of this chapter.
Definition and classification
The thalassaemias are a heterogeneous group of genetic disorders
of haemoglobin synthesis, all of which result from a reduced rate
of production of one or more of the globin chains of haemoglobin.
They are divided into the α, β, δβ, or εγδβ thalassaemias, according to
which globin chain is produced in reduced amounts (Box 22.6.7.2).
In some thalassaemias, no globin chain is synthesized at all; these are
called α° or β° thalassaemias. In others, the α+ or β+ thalassaemias,
globin chain is produced but at a reduced rate. Thalassaemia occurs
in populations in which structural haemoglobin variants are also
common and an individual may inherit a thalassaemia gene from
one parent and a gene for a structural haemoglobin variant from the
other. Both α and β thalassaemia occur commonly in some countries
and hence individuals may carry genetic changes causing both types.
These different interactions produce an extremely complex and clin-
ically diverse series of genetic disorders which range in severity
from death in utero to extremely mild, symptomless hypochromic
anaemias.
The thalassaemias are inherited in a simple Mendelian fashion.
Heterozygotes are usually symptomless, although they can be
easily recognized haematologically. More severely affected patients
are either homozygotes for α or β thalassaemia, compound het-
erozygotes for different molecular forms of α or β thalassaemia,
or compound heterozygotes for thalassaemia and a structural
haemoglobin variant. Clinically, the thalassaemias are classified
according to their severity into major, intermediate, and minor
forms. Thalassaemia major is a severe transfusion-​dependent dis-
order. Thalassaemia intermedia is characterized by anaemia and
splenomegaly though not of such severity as to require regular
transfusion. Thalassaemia minor is the symptomless carrier
state. While these descriptive terms do not have a precise genetic
meaning, they remain useful in clinical practice and may be sim-
plified further into transfusion-dependent and non transfusion-
dependent thalassaemia.
β Thalassaemias
The β thalassaemias are the most important types of thalassaemia
because they are very common and produce severe anaemia in their
homozygous and compound heterozygous states (Table 22.6.7.1).
Distribution
Patients with the β thalassaemias have an ethnic origin that re-
lates to a broad belt ranging from the Mediterranean and parts of
North and West Africa through the Middle East and Indian subcon-
tinent to South-​East Asia (Fig. 22.6.7.3). The high incidence zone
stretches north through the Balkans and the southern parts of Russia
and includes the southern regions of China. The disease is particu-
larly common in South-​East Asia where it occurs from southern
Box 22.6.7.1  Disorders of haemoglobin
Genetic
•	 Thalassaemia
•	 Structural variants
•	 Hereditary persistence of fetal haemoglobin
•	 α Thalassaemia/​mental retardation syndromes
Acquired
•	 Methaemoglobin
•	 Carbonmonoxyhaemoglobin
•	 Sulphaemoglobin
•	 Glycosylated haemoglobin
•	 Acquired HbH disease
•	 Disorders associated with raised levels of haemoglobin F
Box 22.6.7.2  The thalassaemias
α Thalassaemia
•	 α°
•	 α+
β Thalassaemia
•	 β°
•	 β+
δβ Thalassaemia
•	 (δβ)°
•	 Haemoglobin Lepore (δβ)+
•	 (εγδβ)° Thalassaemia
•	 δ Thalassaemia
section 22  Haematological disorders
5430
China, through Thailand, the Malay peninsula and Indonesia, to
some of the Pacific islands. In these populations, and in some of the
Mediterranean islands and mainland countries, gene frequencies
for the various forms of β thalassaemia range between 2 and 20%.
It should be remembered that β thalassaemia is not entirely con-
fined to these high-​incidence regions; it occurs sporadically in every
racial group.
Molecular pathology
The precise molecular lesions responsible for the defective synthesis
of the β globin chains have been determined for many patients with
β thalassaemia. The disease is extremely heterogeneous with nearly
400 different mutations found to date which result in the clinical
phenotype of β thalassaemia.
With the exception of a deletion of about 600 bases at the 3´
end of the β globin gene, which is only found in certain popula-
tions of northern India, deletions are an uncommon cause of β
thalassaemia. Most of the mutations are single base changes or
small deletions and insertions of one or two bases. These occur
in both introns and exons, and also outside the coding regions.
Nonsense, frameshift, and splice site mutations have all been de-
scribed. Mutations activating cryptic splice sites have been ob-
served, causing a β+ thalassaemia with severity dependent on
the relative usage of the normal and abnormal splice sites. Many
single base substitutions have also been found in the flanking re-
gions of the β globin genes. They alter either the proximal pro-
moter regions or adjacent transcriptional regulatory machinery
(e.g. enhancers).
Because there are so many different β thalassaemia mutations it
follows that many patients who are apparently homozygous for β
thalassaemia are, in fact, compound heterozygotes for two different
molecular lesions.
Pathophysiology
The mutations that cause β thalassaemia result in absent or re-
duced β chain production. The synthesis of α chains proceeds at a
normal rate and hence there is imbalanced globin chain synthesis
(Fig. 22.6.7.4). In the absence of their partner chains the excess
α chains are unstable and precipitate in the red cell precursors,
forming large intracellular inclusions. These interfere with red cell
maturation, and hence there is a variable degree of intramedullary
destruction of red cell precursors, (ineffective erythropoiesis).
Those red cells which mature and enter the circulation contain α
chain inclusions which interfere with their passage through the
microcirculation, particularly in the spleen. These cells are prema-
turely destroyed. However, the mechanisms of the destruction of
red cell precursors and their progeny are extremely complex and
are not simply a reflection of mechanical damage to the red cells.
Free α chains and their degradation products, particularly haem
and iron, cause severe oxidative damage to the red cell membrane
proteins. The end result is a dehydrated, rigid erythrocyte with a
markedly shortened survival.
Table 22.6.7.1  The β, δβ, and γδβ thalassaemias
Type of thalassaemia
Findings in homozygote
Findings in heterozygote
β°
Thalassaemia majora, b
Thalassaemia minor
Hbs F and A2
Raised Hb A2
β+
Thalassaemia majora, b
Thalassaemia minor
Hbs F, A, and A2
Raised Hb A2
δβ
Thalassaemia intermedia
Thalassaemia minor
Hb F only
Hb F 5–​15%; Hb A2 normal
(δβ)+
Thalassaemia major or intermedia
Thalassaemia minor
(Lepore)
Hbs F and Lepore
Hb Lepore 5–​15%; Hb A2 normal
εγδβ
Not viable
Neonatal haemolysis
Thalassaemia minor in adults, with normal Hbs F and A2
a Occasionally have thalassaemia intermedia phenotype.
b Many patients with thalassaemia are compound heterozygotes for different molecular forms of β° or β+ thalassaemia.
CODON 6 – 1 bp
IVS 1 – 1G   A
IVS 2 – 1G   A
IVS 2 – 745 C   G
CODON 39 CAG   TAG
IVS 1 – 6T   C
IVS 1 – 110 G   A
IVS 1 – 5 G   C
IVS 1 – 1 G   T
CODONS 41– 42.bp DEL.
CODONS 26 GAG   AAG (HbE)
IVS 1 – 5 G   C
IVS 2 – 654 C   T
CODONS 41 –  42.4bp DEL.
CODON 17 AAG   TAG
CODON 26 GAG   AAG(HbE)
–28A   G
–29A   G
–29 A   G
–88 C   T
CODON 24 T   A
POLY-A T   C
IVS 1 – 5 G   C
619 bp DELETION
CODON 8/9 + G
IVS 1 – 1 G   T
CODONS 41– 42.4 bp DEL.
IVS 1 – 110 G   A
IVS 1 – 5 G   C
IVS 1 – 6 T   C
CODON 39 CAG   TAG
CODON 8 2bp DEL
Fig. 22.6.7.3  World map showing the distribution of the different β
thalassaemia mutations.
22.6.7  Disorders of the synthesis or function of haemoglobin
5431
If untreated, the anaemia acts as a stimulus to increase erythro-
poietin production, causing massive expansion of the bone
marrow which may lead to serious deformities of the skull and
long bones. Because the spleen is being constantly bombarded
with abnormal red cells, it hypertrophies. The resulting spleno-
megaly and bone marrow expansion gives rise to an increase in
the plasma volume which, together with pooling of the red cells in
the enlarged spleen, causes an exacerbation of an already severe
degree of anaemia.
As mentioned previously, fetal haemoglobin production largely
ceases after birth. However, some adult red cell precursors (F cells)
retain the ability to produce a small number of γ chains. Because
the latter can combine with excess α chains to form haemoglobin F,
cells which make relatively more γ chains in the bone marrow of β
thalassaemics are partly protected against the deleterious effect of
α chain precipitation. Red cell precursors which produce haemo-
globin F are selected in the marrow and peripheral blood of these
patients. Thus, they have relatively large amounts of haemoglobin
F in their red cells. Furthermore, because δ-​chain synthesis is un-
affected, the disorder is characterized by a relative or absolute in-
crease in haemoglobin A2 (α2δ2) production.
If the anaemia is corrected with blood transfusion the erythro-
poietic drive is reduced, growth and development are improved,
and bone deformities do not occur. However, as each unit of
blood contains 200 mg of iron, regular transfusion results in the
steady accumulation of iron in the liver, endocrine glands, and
myocardium. Even though well-​transfused thalassaemic chil-
dren grow and develop normally, they die of iron overload unless
steps are taken to remove iron (see iron chelation, in ‘Symptomatic
treatment’).
The severe homozygous or compound heterozygous forms
of β thalassaemia
These are the commonest and most important forms of thalas-
saemia and give rise to a major public health problem in many
parts of the world.
Clinical features
Most severe forms of β thalassaemia present within the first year
of life, as fetal haemoglobin production declines, with failure
to thrive, poor feeding, intermittent bouts of fever, or failure
to improve after an intercurrent infection. At this stage, the af-
fected infant is pale and splenomegaly may already be present.
Diagnosis depends on the haematological changes outlined in the
following paragraphs. The clinical manifestations of the severe
forms of β thalassaemia have to be described in two contexts: (1)
the well-​transfused child and (2) the child with chronic anaemia
throughout early life.
In the well-​transfused thalassaemic child, early growth and de-
velopment is normal. Splenomegaly is minimal. However, there is
a gradual accumulation of iron and the effects of tissue siderosis
start to appear by the end of the first decade in the unchelated
patient. The normal adolescent growth spurt fails to occur.
Hepatic, endocrine, and cardiac complications of iron overloading
produce a variety of problems including diabetes, hypopara-
thyroidism, adrenal insufficiency, and progressive liver failure.
Secondary sexual development is delayed or does not occur at all.
Short stature and lack of sexual development may lead to serious
psychological problems. By far the commonest cause of death,
which usually occurs toward the end of the second or early in the
third decade in patients who do not receive iron chelation, is pro-
gressive cardiac damage. Ultimately these patients die either as a
result of protracted cardiac failure or suddenly, as the result of an
acute arrhythmia.
Children who have been both adequately transfused and che-
lated may grow and develop normally, pass through a normal
puberty, and survive to adult life in good health. However, even
children who have been well managed in this way may still suffer
from complications as they get older, particularly delayed sexual
maturation, growth disturbances, and osteoporosis. It seems
likely that many of these problems are due to subtle damage to
the hypothalamic–​pituitary axis with secondary hypogonadism.
The clinical picture in children who are inadequately trans-
fused is quite different. Rates of growth and development are
Excess
Precipitation
Haemolysis
Anaemia
Transfusion
Tissue hypoxia
Erythropoietin
Marrow expansion
Iron loading
Destruction of
red blood
cell precursors
Splenomegaly
(pooling, plasma
volume
expansion)
Ineffective
erythropoiesis
High oxygen
affinity of red cells
Bone deformity
Increased metabolic rate
Wasting
Gout
Folate deficiency
Endocrine deficiencies
Cirrhosis
Cardiac failure
Death
Selective survival of
HbF-containing cells
HbF
Increased iron
absorption
α
α2γ2
β
γ
Fig. 22.6.7.4  The pathophysiology of β thalassaemia.
section 22  Haematological disorders
5432
markedly slowed. There is progressive splenomegaly; hyper­
splenism may cause a worsening of the anaemia. Because of
the bone marrow expansion there may be deformities of the
skull with marked bossing and overgrowth of the zygomata
giving rise to the classical facial appearance of β thalassaemia
(Fig. 22.6.7.5). These findings are reflected by radiological
changes which include a lacy, trabecular pattern of the long
bones and phalanges and a typical ‘hair-​on-​end’ appearance of
the skull (Fig. 22.6.7.7). These bone changes may be associated
with recurrent fractures. There is increased susceptibility to in-
fection which may cause a catastrophic drop in the haemoglobin
level. Because of the massive marrow expansion, these chil-
dren are hypermetabolic, run intermittent fevers, lose weight
(Fig. 22.6.7.6), have increased requirements for folic acid, and
may become acutely folate depleted with worsening of their
anaemia. Increased turnover of red cell precursors occasion-
ally gives rise to hyperuricaemia and secondary gout. There
is a bleeding tendency which, partly due to thrombocyto-
penia secondary to hypersplenism, may be exacerbated by
liver damage associated with iron loading and extramedullary
haemopoiesis. There is also an increased risk of thrombotic
complications, reflecting procoagulant properties of the ab-
normal red cell membranes. The bone deformities of the
skull can cause distressing dental complications with poorly
formed teeth and malocclusion, and inadequate drainage of
the sinuses and middle ear which may lead to chronic sinus
infection and deafness. If these children survive to puberty,
they develop the same complications of iron loading as the
well-​transfused patients. In this case, some of the iron accu-
mulation results from an increased rate of gastrointestinal
absorption (due to decreased hepcidin production—​see also
Chapter  22.6.4) as well as that derived from the inadequate
transfusion regimen.
Laboratory features
There is always a severe anaemia. The haemoglobin values on
presentation range from 20 to 80 g/​litre. The appearance of the
stained peripheral blood film is grossly abnormal (Fig. 22.6.7.8).
The red cells show marked hypochromia and variation in shape
Fig. 22.6.7.5  Homozygous β thalassaemia: skull and facial deformity
due to bone marrow expansion.
Fig. 22.6.7.6  Gross wasting of the limbs and hepatomegaly in an
undertransfused child.
Fig. 22.6.7.7  Radiological changes of the skull in homozygous β
thalassaemia.
22.6.7  Disorders of the synthesis or function of haemoglobin
5433
and size. There are many hypochromic macrocytes and misshapen
microcytes, some of which are mere fragments of cells. There is
anisochromia, basophilic stippling, and some nucleated red cells in
the peripheral blood. After splenectomy, nucleated cells are found
in large numbers. In the postsplenectomy film, many of the nucle-
ated cells and mature erythrocytes show ragged inclusions after
incubation of the blood with methyl violet, representing free ex-
cess α globin. There is usually a slight elevation in the reticulocyte
count. The white cell and platelet counts are normal unless there is
hypersplenism in which case they are reduced. The bone marrow
shows marked erythroid hyperplasia.
The haemoglobin F level is always elevated. In β° thalassaemia
there is no haemoglobin A and the haemoglobin consists of F and
A2 only. In β+ thalassaemia the level of haemoglobin F ranges from
30 to 90% of the total haemoglobin. The haemoglobin A2 level is
usually normal and is of no diagnostic value.
There are biochemical changes of increased haemolysis and
progressive iron loading. The bilirubin level is usually elevated
and haptoglobins are absent. The serum iron and serum ferritin
rises progressively. Most transfusion-​dependent children have
a totally saturated iron-​binding capacity. Liver biopsies show a
marked increase in hepatic iron, which may be distributed both in
the reticuloendothelial and parenchymal cells, and magnetic res-
onance imaging (MRI) of the heart and liver manifest markedly
shortened relaxation times, consistent with iron loading (see later
in this chapter: Fig. 22.6.7.18).
As well as folic acid deficiency, vitamin E and ascorbate defi-
ciency is common in thalassaemic children. The endocrine com-
plications of iron loading include diabetes, and parathyroid or
adrenal insufficiency. Growth hormone levels are usually normal.
Heterozygous β thalassaemia
Carriers for β thalassaemia, apart from symptoms of mild an-
aemia, are usually well except in periods of stress such as preg-
nancy, when they may become more anaemic. Splenomegaly is
rarely present.
There is a mild degree of anaemia with haemoglobin values of
90 to 110 g/​litre. The red cells are hypochromic and microcytic.
The reticulocyte count is usually normal. The bone marrow shows
moderate erythroid hyperplasia.
Haemoglobin analysis shows an elevated haemoglobin A2 level
of 4 to 6%, and there may also be a slight elevation of haemoglobin
F to approximately 1 to 3%. A less common form occurs in which
the haemoglobin A2 is not elevated (see ‘Other β thalassaemia
variants’).
β Thalassaemia in association with haemoglobin variants
Although numerous interactions between thalassaemia and struc-
tural haemoglobin variants have been described, in clinical practice
only three are of importance: sickle cell β thalassaemia, haemo-
globin C β thalassaemia, and haemoglobin E β thalassaemia.
Sickle cell β thalassaemia
The clinical manifestations which result from the interaction of
the β thalassaemia and sickle cell genes vary considerably from
race to race, and depend on the severity of the thalassaemia de-
terminant. In African populations, there are mild forms of β+
thalassaemia which, when they interact with the sickle cell gene,
produce a condition characterized by mild anaemia and few
sickling crises. By contrast, in Mediterranean populations, the
combination of a β° or severe β+ thalassaemia determinant from
one parent with a sickle cell gene from the other may give a clin-
ical picture which is indistinguishable from sickle cell anaemia
(see later).
The diagnosis of sickle cell thalassaemia rests on the clinical
features of a sickling disorder found in association with a per-
ipheral blood picture with typical thalassaemic red cell changes,
that is, a low mean cell haemoglobin and mean cell volume. In
the more severe forms of sickle cell β° thalassaemia, there may
be an elevated reticulocyte count and sickled red cells are found
on the peripheral blood film. The diagnosis can be confirmed
by high-​performance liquid chromatography (HPLC) or haemo-
globin electrophoresis, which in sickle cell β+ thalassaemia shows
haemoglobin S together with 10 to 30% haemoglobin A  and
an elevated haemoglobin A2. In sickle cell β° thalassaemia, the
haemoglobin consists mainly of haemoglobin S with an elevated
level of haemoglobins F and A2 and is therefore indistinguishable
from homozygous sickle cell disease. To confirm the diagnosis,
DNA analysis is required.
Haemoglobin C thalassaemia
This disorder is restricted to patients of West African ethni-
city and to some North African and southern Mediterranean
populations. It is characterized by a mild haemolytic anaemia
associated with splenomegaly. The peripheral blood film shows
numerous target cells and thalassaemic red cell changes with a
moderately elevated reticulocyte count. Haemoglobin HPLC
shows a preponderance of haemoglobin C. The diagnosis is con-
firmed by finding the haemoglobin C trait in one parent and the
β thalassaemia trait in the other, or again by recourse to DNA
sequencing.
Haemoglobin E β thalassaemia
This is the commonest form of severe thalassaemia in many Asian
countries where it causes a serious public health burden. The
Fig. 22.6.7.8  Peripheral blood film in homozygous β thalassaemia
(×630, Leishman stain).
section 22  Haematological disorders
5434
mutation that underlies haemoglobin E produces an alternative
splice site in the β globin gene. This mutation results in the pro-
duction of a β globin variant which is produced at a much lower
rate than normal β globin. Thus, when a haemoglobin E gene is
inherited together with a severe β thalassaemia mutation there
is a marked inefficiency of β chain production. However, one of
the major characteristics of haemoglobin E β thalassaemia, which
causes particular difficulties for its management, is its extraor-
dinary clinical heterogeneity. At one end of the spectrum it is
indistinguishable from β thalassaemia major, while at the other
end there are patients who grow and develop normally without
the need for transfusion. While some of this phenotypic vari-
ability can be ascribed to the inheritance of β thalassaemia alleles
of varying severity, some is also due to coinheritance of various
modifier genes, including those for α thalassaemia or increased
haemoglobin F production. The explanation for much of this
phenotypic variation remains unclear.
In the more severe forms of this condition the findings are
very similar to those in severe β thalassaemia (Fig. 22.6.7.9),
while in the milder forms they resemble those of β thalassaemia
intermedia, as described later in this chapter. Complications in-
clude susceptibility to infection, hypersplenism, iron loading,
neurological lesions due to extramedullary erythropoietic
masses extending inwards from the inner tables of the skull or
vertebrae, folate deficiency, and recurrent pathological frac-
tures. From the limited data that are available, it seems that
patients at the milder end of the clinical spectrum, though
often quite anaemic, survive in good health well into adult life.
They do not appear to develop cardiac complications unless they
have become particularly iron loaded from increased intestinal
absorption.
The diagnosis of haemoglobin E thalassaemia is confirmed by
finding haemoglobins E and F and little or no haemoglobin A on
HPLC and by demonstrating the haemoglobin E trait in one parent
and the β thalassaemia trait in the other.
Other β thalassaemia variants
It is not uncommon to encounter patients with the clinical and
haematological features of heterozygous β thalassaemia who do
not have an elevated haemoglobin A2 level. Many of these in-
dividuals are heterozygotes for both β and δ thalassaemia. It is
important to recognize this interaction because, if it is inherited
together with a typical β thalassaemia gene, it can produce a severe
transfusion-​dependent disorder. Hence this variant is important
in antenatal screening programmes. It can only be identified for
certain by genetic analysis of the affected locus. Families are oc-
casionally encountered in which there is a more severe form of
heterozygous β thalassaemia associated with anaemia, jaundice,
and splenomegaly. In some of these families it is apparent that
the affected individuals are in fact compound heterozygotes for
β thalassaemia and the so-​called silent β thalassaemia gene, that
is, a determinant which cannot be identified haematologically in
heterozygotes. In other families, a severe form of β thalassaemia
behaves as a single gene disorder with full expression in heterozy-
gotes, that is, it follows a dominant form of inheritance. In most of
these families, the disorder results from the synthesis of a highly
unstable β globin chain.
The δβ thalassaemias
See Table 22.6.7.1.
Molecular genetics and classification
Disorders due to reduced β and δ chain synthesis are much less
common than those due to defective β chain production alone.
They are remarkably heterogeneous at the molecular level and
may result from deletions of the β and δ globin genes (the (δβ)°
thalassaemias).
Unequal crossing over between the δ and β globin gene loci
may also occur, with the production of δβ fusion genes. These
produce δβ fusion chains which combine with α chains to form
haemoglobin variants called the Lepore haemoglobins (Lepore
was the family name of the first patient to be recognized with this
disorder).
Clinical and haematological changes
The (δβ)° thalassaemias have been reported in many populations,
although there are no high-​frequency areas. In the homozygous
state there is a mild degree of anaemia with haemoglobin values of
80 to 100 g/​litre. There is often a moderate degree of splenomegaly
but these patients are usually symptomless except during periods of
stress such as infection or pregnancy. Haemoglobin analysis shows
100% haemoglobin F.  Heterozygous carriers have thalassaemic
blood pictures, elevated levels of haemoglobin F of 5 to 20%, and
normal levels of haemoglobin A2.
The homozygous state for haemoglobin Lepore is character-
ized by a clinical picture which is usually similar to that of homo-
zygous β thalassaemia although in some cases it may be milder
and nontransfusion dependent. The haematological findings are
similar to those of β thalassaemia. The haemoglobin consists
of F and Lepore only. Heterozygous carriers have thalassaemic
Fig. 22.6.7.9  Bossing of the skull in haemoglobin E thalassaemia.
22.6.7  Disorders of the synthesis or function of haemoglobin
5435
blood pictures associated with about 5 to 15% haemoglobin
Lepore.
The (εγδβ)0 thalassaemias
There are several rare forms of thalassaemia which result from
long deletions of the β globin gene cluster which, as well as re-
moving or inactivating the β genes, involve the δ, γ, and embryonic
ε genes. They also involve the main regulatory sequence upstream
of the β globin gene cluster, the locus control region. This means
that there is no output of globin chains from this gene cluster at
all. Clearly, the homozygous state for these disorders would not be
compatible with survival. Heterozygotes often have severe haemo-
lytic disease as neonates with anaemia and hyperbilirubinaemia.
If they survive the neonatal period they grow and develop nor-
mally; in adult life they have the haematological picture of het-
erozygous β thalassaemia with mild anaemia, hypochromic
microcytic red cells, and a haemoglobin pattern consisting of
haemoglobin A, no elevation of haemoglobin F, and a normal level
of haemoglobin A2.
Hereditary persistence of fetal haemoglobin
There is a complex family of conditions characterized by per-
sistent fetal haemoglobin synthesis into adult life associated
with no major haematological abnormalities. In some cases
they result from long deletions of the β globin gene cluster,
similar to those that cause δβ thalassaemia. Indeed, they form
a continuum with this condition; homozygotes have 100% fetal
haemoglobin, elevated haemoglobin levels, and no clinical find-
ings. Other forms result from point mutations in the promoter
regions of the γ globin genes. In this case there is increased γ
chain production together with reduced β chain production on
the affected chromosome. Hence, homozygotes have markedly
elevated levels of haemoglobin F but also produce some haemo-
globin A. Finally, there is a group in which persistent low levels
of haemoglobin F, in the 3 to 10% range, are observed. They re-
sult from mutations either within the β globin gene cluster or on
other chromosomes.
The only clinical importance of this complex group of condi-
tions is that they may interact with the thalassaemias or struc-
tural haemoglobin variants and reduce the severity of different
phenotypes by increasing the amount of haemoglobin F that is
produced.
The α thalassaemias
Although the α thalassaemias are commoner on a global basis than
the β thalassaemias, they pose less of a public health problem be-
cause their severe forms only occur in a few regions.
Distribution
The α thalassaemias occur in patients whose ethnic origin re-
lates to the Mediterranean region, parts of West Africa, the
Middle East, parts of the Indian subcontinent, and throughout
South-​East Asia from southern China through Thailand, the
Malay peninsula, and Indonesia to the Pacific island popula-
tions (Fig. 22.6.7.10). The serious forms of α thalassaemia are
restricted mainly to patients of Mediterranean and South-​East
Asian ethnicity.
Inheritance and molecular pathology
As both haemoglobins A and F have α chains, genetic disorders of
α chain synthesis result in defective fetal and adult haemoglobin
production. In the fetus, deficiency of α chains means there is
a relative excess of γ chains which form γ4 tetramers also known
as haemoglobin Bart’s (Fig. 22.6.7.11). In adults, a deficiency of α
chains leads to a relative excess of β chains which form β4 tetramers,
or haemoglobin H, the adult counterpart of haemoglobin Bart’s.
However, a critical level of globin chain imbalance is required be-
fore detectable amounts of haemoglobins Bart’s or H appear in the
red cells, and in individuals with mild forms of α thalassaemia this
level is not reached; significant amounts of these haemoglobin vari-
ants occur only in the red cells of patients who have a severe degree
of α chain deficiency.
As normal individuals receive two α globin genes from each of
their parents, αα/​αα, the genetic basis of the α thalassaemias is
5–40%
1–15%
60%
40–80%
5–80%
5–15%
α + Thalassaemia
α ° Thalassaemia
Fig. 22.6.7.10  World map showing the distribution of the α
thalassaemias.
Adult
Fetus
Normal
Excess
Excess
HbF
HbA
Hb Bart’s
[High oxygen affinity]
HbH
High oxygen affinity
Unstable  Inclusions
Hypochromia
Haemolysis
Hypoxia
α Thalassaemia
γ2
α2
α2
α2γ2
α2β2
β4
γ4
β2
Fig. 22.6.7.11  The pathophysiology of α thalassaemia.
section 22  Haematological disorders
5436
more complicated than that of the β thalassaemias. It is useful to
define these conditions in heterozygotes. First, there is a more se-
vere form, α° thalassaemia, which results from loss of both of the
linked α globin genes, − −/​αα. The second type, α+ thalassaemia,
arises due to the deletion –​α/​αα, so there is still some output of α
globin from the affected chromosome. This is almost completely
silent in carriers; their red cells are normal or are only slightly
hypochromic.
In clinical practice we encounter two symptomatic types of α thal-
assaemia, the haemoglobin Bart’s hydrops syndrome and haemo-
globin H disease (Table 22.6.7.2). The former results from the
homozygous inheritance of α° thalassaemia. Haemoglobin H dis-
ease by contrast usually results from the coinheritance of both α°
and α+ thalassaemia. These genetic interactions are summarized in
Fig. 22.6.7.12.
Like the β thalassaemias, the α thalassaemias are extremely het-
erogeneous at the molecular level. Various deletions can remove
either both the α globin genes or the main regulatory regions of
the α globin gene cluster and cause α° thalassaemia, but there are
only two that are common. One is found in patients of South-​East
Asian ethnicity. The other occurs mainly in Mediterranean popu-
lations. Similarly, there are several different-​sized deletions that re-
move a single α globin gene to produce the deletion forms of α+
thalassaemia; the commonest are those that remove either 3.7 or
4.2 kb of the α gene cluster (Fig. 22.6.7.13). Nondeletion forms
of α+ thalassaemia are also seen, and many of them are similar to
those that produce β thalassaemia. A particularly common form of
nondeletion α+ thalassaemia, found in up to 5% or more of some
South-​East Asian populations, results from a single base change in
the α globin chain termination codon UAA, which changes to CAA.
The latter is the code for the amino acid glutamine. When the ribo-
somes reach this point, instead of the chain terminating, they read
through mRNA that is not normally translated until another stop
codon is reached. An elongated α chain variant is synthesized, but
the mRNA is destabilized by read-​through of sequences which are
not normally translated and so the variant is also produced at a re-
duced rate. It is called haemoglobin Constant Spring after the name
of the town in Jamaica in which it was discovered.
Genotype–​phenotype relationships
Molecular studies explain much of the clinical variability of α thal-
assaemia in different populations. Since the haemoglobin Bart’s
hydrops syndrome requires the homozygous inheritance of α°
thalassaemia (–​ –​/​–​ –​), this condition only occurs in populations
in which α° thalassaemia is common. Most forms of haemoglobin
H disease are due to the inheritance of α° thalassaemia from one
parent and α+ thalassaemia from the other (–​α/​–​ –​ or –​αT/​–​ –​). Thus,
Table 22.6.7.2  The α thalassaemias
Type
Homozygotes
Heterozygotes
α°
Hb Bart’s hydrops
Thalassaemia minor
α+ (deletion)
Thalassaemia minor
Thalassaemia minorb
αT (nondeletion)
Hb H diseasea
Thalassaemia minoret
a Haemoglobin H disease more commonly results from the compound heterozygous
inheritance of α° and either variety of α+ thalassaemia.
b Heterozygotes for the α+ determinant typically have reduced mean cell volume and
mean cell haemoglobin; in a minority of cases, the red cells indices fall within the
normal range.
Normal
Hb Bart’s
hydrops
Hb H
disease
Normal
α° Thal. trait
α° Thal. trait
α° Thal. trait
α° Thal. trait
α+ Thal. trait
α+ Thal. trait
α° Thal. trait
α° Thal. trait
Fig. 22.6.7.12  The genetics of α thalassaemia. The purple α genes
represent gene deletions or otherwise inactivated genes. The open α
genes represent normal genes. α° Thalassaemia and α+ thalassaemia are
defined in the text.
–
–
–
3′HVR
30
20
10
0
−10
−50
--MC
--CAL
--THAI
--FIL
--CL
--BRIT
--SA
–(α)20.5
--MED
--SEA
--SPAN
–(α)5.2
θ1
ψζ1
inter-
ζHVR
ζ2
ψα2
α2
α1
α3.7
α3.5
α4.2
yα1
Fig. 22.6.7.13  The different-​sized deletions responsible for some
forms of α° or α+ thalassaemia. The α globin gene cluster is shown at
the top of the figure. Two highly variable regions (HVR) are shown. The
abbreviations on the right-​hand side indicate the source of origin of
patients with the deletions: MED, Mediterranean; SEA, South-​East Asia.
The three smaller deletions at the bottom of the figure show some of the
main classes of α+ thalassaemia. The superscripts 3.7, 4.2, and 3.5 indicate
the size of the deletions in kb.
22.6.7  Disorders of the synthesis or function of haemoglobin
5437
haemoglobin H disease is also restricted mainly to Mediterranean
and Asian populations. On the other hand, α+ thalassaemia occurs
very commonly throughout the whole of the tropical belt; α° thalas-
saemia does not occur commonly in many of these regions, so that
the haemoglobin Bart’s hydrops syndrome and haemoglobin H dis-
ease are not seen. The homozygous state for α+ thalassaemia (–​α/​
–​α) is characterized by a mild hypochromic anaemia, very similar
to the heterozygous state for α° thalassaemia. To complicate matters,
sometimes the homozygous state for the nondeletion forms of α+
thalassaemia, αTα/​αTα, are more severe and cause haemoglobin H
disease.
Pathophysiology
The pathophysiology of α thalassaemia is different from that of β
thalassaemia. A deficiency of α chains leads to a relative excess of γ
chains or β chains which form haemoglobins Bart’s and H respect-
ively. These more soluble tetramers do not precipitate significantly
in the bone marrow, and erythropoiesis is thus more effective than
in β thalassaemia. However, haemoglobin H is unstable and precipi-
tates in red cells as they age. The large inclusion bodies produced in
this way are trapped in the spleen and other parts of the microcir-
culation leading to a shortened red cell survival. Both haemoglobins
Bart’s and H have a very high oxygen affinity; because they have no
α chains there is no haem–​haem interaction and their oxygen dis-
sociation curves resemble that of myoglobin, making them physio-
logically useless.
Haemoglobin Bart’s hydrops syndrome
This condition is a cause of fetal loss throughout South-​East Asia
and in Greece and Cyprus. Affected infants produce no α chains and
hence can make neither fetal nor adult haemoglobin.
The clinical picture is very characteristic (Fig. 22.6.7.14). Infants
are usually stillborn between 28 and 40 weeks. Live-​born infants
take a few gasping respirations and then expire within the first hour
after birth. They show the typical picture of hydrops fetalis with ex-
treme pallor, generalized oedema, and massive hepatosplenomegaly.
There is a high frequency of other congenital abnormalities, and a
very large, friable placenta, all due to severe intrauterine anaemia.
The haemoglobin values are in the 60 to 80 g/​litre range and there
are gross thalassaemic changes of the peripheral blood film. The
haemoglobin consists of approximately 80% haemoglobin Bart’s and
20% of the embryonic haemoglobin Portland (ζ2γ2). It is believed
that these infants survive to term because they continue to produce
embryonic haemoglobin at this level; haemoglobin Bart’s is, as men-
tioned previously, useless as an oxygen carrier.
This syndrome is also characterized by a high incidence of ma-
ternal pre-eclampsia and considerable obstetric difficulties due to
the presence of the large, abnormal placenta.
Haemoglobin H disease
Haemoglobin H disease usually results from the inheritance of
α° thalassaemia from one parent and α+ from the other. It may
also result from the inheritance of α° thalassaemia and haemo-
globin Constant Spring or from the homozygous state for a severe,
nondeletion form of α thalassaemia. The latter form of inheritance
is particularly common in Saudi Arabia. Recent evidence suggests
that, overall, this condition is more severe in those who have in-
herited α° thalassaemia together with haemoglobin Constant Spring
or other nondeletion forms of the disease compared with those who
have inherited three α gene deletions.
There is a variable degree of anaemia and splenomegaly but it is
most unusual to see severe thalassaemic bone changes or the growth
retardation characteristic of homozygous β thalassaemia. Patients
usually survive into adult life although the course may be inter-
spersed with severe episodes of haemolysis associated with infec-
tion, or worsening of the anaemia due to progressive hypersplenism.
Oxidant drugs such as sulphonamides may increase the rate of pre-
cipitation of haemoglobin H and therefore exacerbate the anaemia.
(a)
(b)
Fig. 22.6.7.14  The haemoglobin Bart’s hydrops syndrome: (a) a
hydropic infant with massively enlarged placenta; (b) autopsy findings
with an enlarged liver.
By permission of Professor P. Wasi.
section 22  Haematological disorders
5438
Haemoglobin values range from 70 to 100 g/​litre. The blood
film shows typical thalassaemic changes. There is a moderate
reticulocytosis. Incubation of the red cells with brilliant cresyl
blue generates numerous inclusion bodies by precipitation of the
haemoglobin H under the redox action of the dye (Fig. 22.6.7.15)
The haemoglobin comprises 5 to 40% haemoglobin H together with
haemoglobin A and a normal or reduced level of haemoglobin A2.
The haematological findings in the α° and α+ thalassaemia traits
are summarized in Table 22.6.7.2. They can only be identified with
certainty by analysis of the α globin genes.
α Thalassaemia and intellectual disability or myelodysplasia
There is an increasingly important and heterogeneous group of α
thalassaemias which are not restricted to individuals from tropical
backgrounds. They are observed in all racial groups and have been
best characterized in those of northern European origin. These con-
ditions are characterized by variable degrees of intellectual disability,
dysmorphic features, and α thalassaemic blood pictures. They follow
a completely different form of inheritance from the commoner gen-
etic forms of α thalassaemia. There are two major varieties of this
condition. The first is due to lesions that involve the α globin gene
cluster on chromosome 16, ATR-​16. There is another group re-
sulting from mutations on the X chromosome, ATR-​X.
The ATR-​16 disorders are characterized by a variable degree of
intellectual disability and dysmorphic features. The blood film shows
mild α thalassaemic changes and some cells which contain typical
haemoglobin H inclusion bodies. In some cases the condition re-
sults from long deletions which remove the end of the short arm of
chromosome 16 and extend for 1 to 2 Mb. In other cases, the loss of
the end of the short arm of chromosome 16 is the result of an in-
herited cytogenetic abnormality, including translocations and other
rearrangements.
The ATR-​X syndrome is characterized by a much more consistent
series of dysmorphic features including typical facial features and
genital abnormalities, and more severe intellectual disability. This is
accompanied by a very mild form of haemoglobin H disease. This
condition is inherited as a typical sex-​linked disorder affecting males
and results from mutations of the ATR-​X gene which regulates tran-
scription via an effect on chromatin structure. Female carriers may
show a very small proportion of red cells containing haemoglobin H
bodies. Acquired mutations of ATR-​X are sometimes found in older
patients who have a mild form of haemoglobin H disease associated
with myelodysplasia. The relationship between the mutations and
the disease of the bone marrow is still not clear.
Thalassaemia intermedia
Definition and pathogenesis
The term ‘thalassaemia intermedia’ is used to describe patients with
the clinical picture of thalassaemia which, although not transfusion
dependent, is associated with a much more severe degree of anaemia
than that found in carriers for α or β thalassaemia. Many of the con-
ditions which have been described previously in this section follow
this clinical course, for example, haemoglobin C or E thalassaemia,
the various δβ thalassaemias and haemoglobin Lepore disorders,
haemoglobin H disease, and the wide variety of conditions which
can result from the interactions of the different β and δβ thalassaemia
determinants. However, some children with this condition have
parents with typical heterozygous β thalassaemia blood pictures and
elevated haemoglobin A2 levels. These patients appear to be homo-
zygous for β thalassaemia, yet they run a much milder course than
is usually the case with this condition. Some of them have inherited
an α thalassaemia determinant as well as being homozygous for β
thalassaemia. This reduces the overall degree of globin chain imbal-
ance and consequently the severity of the dyserythropoiesis which
usually accompanies homozygous β thalassaemia; hence these chil-
dren run a milder clinical course. In other cases, particularly in indi-
viduals of African ethnicity, relatively mild forms of homozygous β
thalassaemia seem to reflect the action of less severe β thalassaemia
mutations. Finally, some intermediate forms of β thalassaemia seem
to result from the coinheritance of a gene for unusually effective
haemoglobin F production.
Clinical and haematological changes
The clinical features of the intermediate forms of thalassaemia are
extremely variable. At one end of the spectrum are patients who are
virtually symptom free except for moderate anaemia. At the other
end there are patients who have haemoglobin values of 50 to 70 g/​
litre and who develop marked splenomegaly, skeletal deformities
due to expansion of bone marrow, and, as they get older, become iron
loaded because of increased intestinal iron absorption. Recurrent
leg ulceration, folate deficiency, symptoms due to extramedullary
haemopoietic tumour masses in the chest and skull (Figs. 22.6.7.16
and 22.6.7.17), gallstones, and a tendency to infection are character-
istic of this group of thalassaemias.
Due to the heterogeneity of these disorders, it is only possible to
determine the course that is likely to evolve in any individual patient
by following the disorder very carefully from early childhood.
Differential diagnosis of the thalassaemias
There are few conditions that are likely to be confused with the more
severe forms of homozygous β thalassaemia or haemoglobin H dis-
ease. The ethnic background of the patient, the presence of anaemia
from early life, and the characteristic haematological changes make
the diagnosis relatively easy. Once thalassaemia is suspected, the
parents and near relatives should be examined for the carrier states
for α or β thalassaemia. Both disorders can be distinguished from
simple iron deficiency by the finding of a normal ferritin level and
by the associated changes in the haemoglobin pattern on HPLC.
Fig. 22.6.7.15  Supravital staining with brilliant cresyl blue highlights
prominent red cell inclusion bodies in Hb H disease.
22.6.7  Disorders of the synthesis or function of haemoglobin
5439
It should be remembered, however, that in some groups iron defi-
ciency and heterozygous thalassaemia frequently occur together in
the same person, particularly during pregnancy. The sideroblastic
anaemias can be easily distinguished from thalassaemia by the mor-
phological appearances of the red cells and the presence of ring
sideroblasts in the bone marrow. It should be remembered that there
are some rare forms of acquired haemoglobin H disease in elderly
patients with myelodysplasia.
Laboratory diagnosis of thalassaemia
The homozygous states for the severe forms of β thalassaemia are
easily recognized by the haematological changes associated with
very high levels of haemoglobin F; haemoglobin A2 values vary so
much that they are of no diagnostic help. The heterozygous states
are recognized by microcytic hypochromic red cells, a high red cell
count and an elevated level of haemoglobin A2. The δβ thalassaemias
are characterized by the finding of 100% haemoglobin F in homo-
zygotes and 5 to 15% haemoglobin F together with a normal level of
haemoglobin A2 in heterozygotes (Table 22.6.7.1).
When β thalassaemia is diagnosed, quantitative HPLC or haemo-
globin electrophoresis will exclude the presence of an abnormal
haemoglobin variant such as haemoglobin E or Lepore. The precise
nature of the genetic lesion may need be determined by sequencing
the affected loci, or by multiplex-​ligation dependent probe analysis
for deletions.
The haemoglobin Bart’s hydrops syndrome is recognized by the
finding of a hydropic infant with a severe anaemia, a thalassaemic
blood picture, and 80% or more haemoglobin Bart’s on HPLC.
Haemoglobin H disease is identified by the finding of a typical
thalassaemic blood picture with an elevated reticulocyte count, and
variable amounts of haemoglobin H on HPLC. There are no really
useful, simple diagnostic tests for the different α thalassaemic carrier
states although α° thalassaemia heterozygotes usually have typical
thalassaemic red cell changes with a normal haemoglobin A2 value.
It is essential for counselling purposes to diagnose the different car-
rier states for α thalassaemia; blood samples should be referred to a
laboratory for DNA analysis of the globin genes.
Prevention and treatment
Thalassaemia produces a severe public health problem and a serious
challenge for medical resources in many populations. Since there is no
definitive treatment, most countries in which the disease is common
are putting a major effort into programmes for its prevention.
Prevention
Since the carrier states for the β thalassaemias can be easily rec-
ognized, it is possible to screen populations and provide ante-
natal genetic counselling. When heterozygous carrier mothers are
found, their partners are tested; if they are also carriers, the couple
are offered the possibility of prenatal diagnosis and the option to
discuss termination of pregnancies where fetuses are affected by
severe forms of thalassaemia.
Fig. 22.6.7.16  (a) Chest radiograph with a right paravertebral mass of extramedullary haemopoietic tissue in β thalassaemia
intermedia. (b) Transverse thoracic image from computed tomography scan of the same patient, showing the extent of the
extramedullary haemopoietic mass complicated by haemothorax.
Fig. 22.6.7.17  Cranial MRI of a patient with thalassaemia intermedia,
showing significantly thickened skull vault consistent with extramedullary
haematopoiesis.
section 22  Haematological disorders
5440
Prenatal diagnosis
Prenatal diagnosis can be offered to couples at risk for having chil-
dren with severe forms of β thalassaemia and haemoglobin Bart’s
hydrops. Prenatal diagnosis of thalassaemia is typically carried out
by genetic analysis of fetal tissue obtained by chorionic villus sam-
pling between the 11th and 14th week of gestation. As prenatal diag-
nosis of thalassaemia is now well established in many countries, it is
important to discuss the genetic implications of the condition when
carriers are detected by chance, even in low-​prevalence areas.
Symptomatic treatment
The symptomatic management of severe β thalassaemia requires
regular blood transfusion, the judicious use of splenectomy if
hypersplenism develops, and the administration of chelating agents
to prevent iron overload. When the diagnosis of severe β thalassaemia
is suspected during the first year of life, the infant should be followed
for several weeks to make sure that the haemoglobin has fallen to
a level at which regular transfusion will be necessary. It is difficult
to be dogmatic about exactly when transfusions should be started.
A severely anaemic infant who is feeding poorly, inactive, or other-
wise failing to thrive, will almost certainly need to be transfused. The
object is to maintain the pretransfusion haemoglobin level at about
95 g/​litre. This usually requires transfusion of 10 to 15 mg/​kg red cells
every 4 weeks, with extended red cell phenotyping to reduce the risk
of alloimmunization. The rate of transfusion should not exceed 4 to
5 ml/​kg per h. In patients who are profoundly anaemic or show evi-
dence of cardiac insufficiency, the rate should be no more than 2 ml/​
kg per h. It is important to calculate the annual blood consumption
by dividing the total volume of blood transfused over 12 months
by the patient’s weight in the middle of the year. If it is higher than
200 ml/​kg body weight, splenectomy may be considered.
Hypersplenism is becoming much less common where children
are maintained on an adequate transfusion regimen. Increasing
blood requirements, or other evidence of hypersplenism, such as
pancytopenia, should prompt one to consider splenectomy. It should
be avoided before the age of 6 years because of the particularly high
incidence of infection in asplenic children. Two to three weeks
before splenectomy the child should be given (1)  pneumococcal
vaccine, (2) Haemophilus influenzae type B vaccine, and (3) menin-
gococcal A and C vaccine. After the operation the children should
be maintained on oral penicillin V, 125 mg twice daily, increasing to
250 mg twice daily for older children. For those who are allergic to
penicillin, erythromycin should be given.
For patients given adequate transfusion support, iron overload be-
comes a critical factor in the determining the morbidity and mortality
of thalassaemia. Meticulous attention to chelation is required if pa-
tients are to avoid significant iron loading in the heart, liver, and endo-
crine organs, with clinical manifestations including diabetes mellitus,
hypogonadotropic hypogonadism, hypothyroidism, and hypopara-
thyroidism. The anterior pituitary appears to be especially sensitive
to the effects of iron overload, and delayed sexual maturation and
subfertility may be observed even in the context of good iron chelation.
Regular assessment of endocrine function therefore forms a
key part of the long-​term management of transfusion-​dependent
thalassaemic patients. The secondary effects of endocrine dys-
function, such as osteoporosis (which may have contributions
from reduced growth hormone and sex hormone secretion, hypo-
parathyroidism and vitamin D deficiency, as well as collagen gene
polymorphisms), must also be sought and treated where possible.
Assessment of iron loading has historically relied on serum fer-
ritin assays, which often reflect total body iron stores only imper-
fectly, and on formal measurement of the iron concentration in tissue
obtained at liver biopsy. Liver iron concentration can now be accur-
ately determined using MRI (e.g. Ferriscan®—​Fig. 22.6.7.18) which
is a favoured modality for monitoring iron overload and the response
to chelation. All patients with transfusion-​dependent thalassaemia
should undergo regular MRI assessment of liver iron concentration,
with a target of 3–​7 mg/​g dry weight. Cardiac T2* MRI should also be
undertaken in patients with evidence of iron loading, since this will
give a reproducible estimate of the cardiac iron burden. Relaxation
times of greater than 20 ms suggest effective iron chelation, while a
T2* MRI less than 15 ms suggests a need for intensified chelation.
Three iron chelating agents are available for clinical use: desferri­
oxamine (deferoxamine), deferasirox, and deferiprone. Randomized
clinical trials comparing all three agents are still needed, and the
choice of first-​line agent is therefore based on consideration of its
side effect profile and tolerability, along with patient preference.
0
50
100
150
200
250
317
R2(/s)
Voxels
Transverse Relaxation Rate (R2) Image
Transverse Relaxation Rate (R2) Distribution
Transverse Relaxation Rate R2 (/s)
Distribution Mean ± SD: 204.1 ± 43.2
264
211
158
105
52
0
0
80
160
240
320
400
Fig. 22.6.7.18  MRI assessment of hepatic iron loading. The increased transverse relaxation rate is consistent with
significant iron deposition.
22.6.7  Disorders of the synthesis or function of haemoglobin
5441
There is greatest experience with desferrioxamine, which has
been used for iron chelation in patients with thalassaemia for over
40  years. Although an effective chelator, its route of administra-
tion remains a significant disadvantage: desferrioxamine must be
delivered parenterally, typically by subcutaneous infusion over 12
hours for five nights out of seven. Difficulties with compliance are
therefore the main limitation to its usefulness in the clinic.
The initial dose of desferrioxamine should not exceed 25 to 35 mg/​
kg body weight per 24 h, and iron excretion may be potentiated if
patients also receive 100 mg vitamin C by mouth on the days of the
infusion. A careful titration of ferritin level against dosage prevents
over-​treatment, and monitoring must include regular audiometry
and ophthalmic assessment to assess for the known complications
of this treatment.
Randomized phase III trials comparing desferrioxamine with
deferasirox have shown that both agents can effect similar reductions
in liver iron concentrations. However, its oral bioavailability makes
deferasirox an increasingly popular first-​line agent. Complications
include transient gastrointestinal upset, typically reversible renal
dysfunction with proteinuria, and rashes. It is unsuitable for use in
patients with renal dysfunction.
Although deferiprone may be less effective in reducing total body
iron in some thalassaemic patients, preliminary data suggest it may
have a specific role in reducing myocardial iron deposition, particu-
larly in conjunction with desferrioxamine. This, plus the recognized
risk of marrow suppression and agranulocytosis with deferiprone
(such that patients are advised to have a weekly full blood count once
starting this agent) has meant that deferiprone has been less widely
adopted as a first-​line choice for chelation.
Decompensated cardiac failure as a consequence of cardiac
siderosis remains a major cause of mortality in patients with
transfusion-​dependent thalassaemia. The development of features
suggestive of cardiac failure in the context of significant cardiac iron
loading should prompt immediate treatment with a continuous in-
fusion of desferrioxamine pending stabilization. The addition of
deferiprone may also be of use in this setting.
Management of patients with thalassaemia intermedia
The intermediate forms of thalassaemia should be managed by
careful observation, folic acid supplementation, and, in the face of
a falling haemoglobin and increasing spleen size, the judicious use
of splenectomy. The increased risk of venous thrombosis in patients
with thalassaemia intermedia may be exacerbated be splenectomy,
and the risks and benefits of the procedure must be considered
on an individual patient basis. It is important to monitor the iron
status regularly because some of these patients become iron loaded
as a consequence of increased intestinal absorption and chelation
therapy may be necessary later in life.
Toward a cure for thalassaemia major
Currently, haematopoietic stem cell transplantation (see
Chapter 22.8.2) is the only cure for thalassaemia major. To min-
imize transplant related mortality, the procedure should ideally
be undertaken before patients develop end-​organ damage due to
iron deposition—​typically in childhood. Although few prospective
studies and fewer controlled trials have been performed in this field,
disease-​free survival is now approximately 80%, with a transplant-​
related mortality of approximately 5% in young patients with a
matched sibling donor. The transplant-​related mortality in adult
patients is significantly higher, and haemopoietic stem cell trans-
plantation is therefore limited to patients who have had excellent
iron chelation with limited end-​organ damage.
There has been process toward the development of gene therapy
for the thalassaemias. This process aims to genetically modify the
patient’s own haematopoietic stem cells, whether by lentiviral gene
transfer of intact β globin genes, or by CRISPR-​mediated gene
editing of the globin genes themselves or other loci implicated in
the silencing of the fetal γ globin expression. The modified haem-
atopoietic cells are then returned to the patient in an autologous
transplant. In July 2019 Bluebird, a gene therapy company, has an-
nounced positive clinical results from the use of third-generation
lentiviral vectors in transfusion-dependent beta-thalassaemia—
about 80% of recipients able to be free of blood transfusions. Similar
studies are underway using CRISPR-Cas9 editing technology to
correct the sickle haemoglobin defect in human autologous haem-
atopoietic stem cells. The long-awaited hope of a definitive therapy
by gene correction for these important  inherited disorders of
haemoglobin synthesis appears to becoming a reality. However,
ensuring realistic access to these costly and labour-intensive stra-
tagems for the innumerable patients who are affected, will be a
formidabe challenge.
Structural haemoglobin variants
Over 400 structural haemoglobin variants have been described,
most of which result from single amino acid substitutions. Many
of them are harmless and have been discovered during surveys of
the electrophoretic patterns of human haemoglobin. Of course,
this approach underestimates the number of variants because it
only identifies those in which the amino acid substitution alters the
charge of the haemoglobin molecule.
Single amino acid substitutions cause clinical disorders only
if they alter the stability or functional properties of the haemo-
globin molecule. A  classification of these diseases is shown in
Table 22.6.7.3. They include the sickling disorders, chronic or drug-​
induced haemolytic anaemia associated with unstable haemo-
globins, and polycythaemia or congenital cyanosis, associated
with high-​ and low ​oxygen ​affinity haemoglobin variants, respect-
ively. There is a rare group of haemoglobin variants that produce
methaemoglobinaemia.
Table 22.6.7.3  Clinical disorders due to structural haemoglobin
variants
Disorder
Variants
Haemolysis and tissue damage
Haemoglobin S
Drug-​induced haemolysis
Haemoglobin Zürich and other unstable
haemoglobins
Chronic haemolysis
Unstable haemoglobin variants
Haemoglobin C
Congenital polycythaemia
High-​affinity variants
Congenital cyanosis
Haemoglobin(s) M
Low-​affinity variants
Hypochromia: thalassaemic
phenotype
Haemoglobin E
Haemoglobin Constant Spring
section 22  Haematological disorders
5442
Nomenclature
The structural haemoglobin variants are named by letters of the al-
phabet or by the place of origin of the first patient in whom they were
characterized. The heterozygous carrier state is termed the ‘trait’ and
the homozygous condition the ‘disease’.
The sickling disorders
Sickling disorders (Table 22.6.7.4) consist of the homozygous
state of sickle cell disease (SS), and the compound heterozygous
state for haemoglobin S together with haemoglobins C, D, E, or
other structural variants. Several disorders result from the inher-
itance of the sickle cell mutation together with different forms of
thalassaemia (described previously).
Pathogenesis
Haemoglobin S differs from haemoglobin A by the substitution
of valine for glutamic acid at position 6 in the β globin chain.
Although this has been known for well over half a century, it is still
not absolutely clear how it gives rise to the sickling phenomenon.
The latter appears to be due to the unusual solubility characteristics
of haemoglobin S which undergoes liquid crystal (tactoid) forma-
tion as it becomes deoxygenated. In this state, aggregates of sickled
haemoglobin molecules arrange themselves in parallel, rod-​like
fibres, made up of a complex solid core about 21 nm in diameter,
composed of 14 filaments arranged as 7 pairs of double filaments.
Much is now known about the complex interactions whereby the
β6 valine substitution stabilizes the molecular stacks in the deoxy
configuration of haemoglobin. There is considerable variation in
the extent to which different haemoglobins are able to participate
with haemoglobin S in the sickling process. This accounts for some
of the clinical variability of the different sickling conditions. For
example, haemoglobin F is almost completely excluded from the
sickling process; increasing concentrations in the red cell reduce
the rate of sickling.
The pathophysiology of sickling is a dynamic process. Red cells
containing sickle haemoglobin at a high concentration endure a
series of cycles of sickling (prompted by deoxygenation or inflam-
matory stimuli) and desickling, with progressive membrane damage
and loss of plasticity. Finally these dry, rigid cells become irrevers-
ibly sickled (Fig. 22.6.7.19). Sickling of this type has two main ef-
fects. First, sickled erythrocytes have a shortened survival, leading
to a chronic haemolytic anaemia. This in turn results in anaemia,
cholelithiasis, and free haemoglobin mediated changes in nitric
oxide availability with endothelial dysfunction and increased resting
vascular tone. Second, the abnormal red cells tend to adhere to vas-
cular and intercellular adhesion molecules on endothelial cells, with
the production of aggregates, blockage of the vessels, vascular stasis,
subsequent reperfusion damage and, ultimately, oxidant and inflam-
matory damage to tissues.
Distribution
The sickling disorders occur very frequently in African popula-
tions and, sporadically, throughout the Mediterranean region and
the Middle East There are extensive pockets in India. The high fre-
quency of the sickle cell gene occurs because carriers are more re-
sistant than normal individuals to P. falciparum malaria.
Clinical features
Except in conditions of extreme hypoxia, such as flying in an un-
pressurized aircraft, the sickle cell trait causes no clinical disability.
However, it is possible for individuals to suffer vaso-​occlusive epi-
sodes if they become oxygen deprived under anaesthesia. Therefore
all individuals of the appropriate racial background should have a
sickling test (see ‘Laboratory diagnosis’ under ‘Haemoglobin SC dis-
ease’) before receiving an anaesthetic. If the test is positive, homozy-
gous sickle cell disease should be excluded first; but even in patients
with sickle trait alone anaesthetics should be given with special
attention to oxygenation and care should be taken to avoid post-
operative dehydration.
Sickle cell anaemia runs an extremely variable clinical course.
At one end of the spectrum it is characterized by a severe haemo-
lytic anaemia interspersed with frequent exacerbations, or crises.
Other cases may be extremely mild and only found by chance on
routine haematological examination. The reason for these remark-
able differences in phenotypic expression, which are only partly
understood, include the level of haemoglobin F, coinheritance of α
thalassaemia, climate, and socioeconomic factors. Since reactivation
of haemoglobin F would be a potential therapeutic intervention in
the β globin disorders (both sickling diseases and β thalassaemia),
the genetic control of γ globin expression has been subject to intense
investigation. Polymorphisms in the γ G promoter, at the HMIP
Table 22.6.7.4  The major sickling disorders
Disorder
Genotype
(Normal = αα/​αα β/​β)
SS disease
αα/​αα
βS/​βS
SC disease
αα/​αα
βS/​βC
SD disease
αα/​αα
βS/​βD
S–​β thalassaemia
αα/​αα
βS/​β° or βS/​β+
S–​hereditary persistence of fetal Hb
αα/​αα
βS/​–​a
S–​α thalassaemia
α–​/​αα or α–​/​α–​
βS/​βS
SS, sickle cell anaemia. See text for details of other conditions.
a Indicates β gene deletion.
Fig. 22.6.7.19  Sickled red cells in homozygous HbS.
22.6.7  Disorders of the synthesis or function of haemoglobin
5443
locus and BCL11a locus (see ‘Towards a cure for the sickling dis-
orders’) have been shown to account for much of the variation in
fetal haemoglobin expression, but many more modifiers are likely
to be found.
Typically, sickle cell anaemia presents in infancy with symptoms
related to anaemia or infection. Dactylitis is also commonly seen.
Infants begin to develop anaemia from about the third month of
life. During early development they often have significant spleno-
megaly that gradually resolves due to repeated infarction resulting
in functional hyposplenism, though splenic sequestration crises (see
following ‘Complications’ section) can result in significant splenic
enlargement. The haemoglobin is typically between 60 to 80 g/​litre
with a reticulocyte count of 10 to 20%. There is chronic, mild icterus
with an elevated bilirubin level. Examination of the peripheral blood
film shows anisochromia and poikilocytosis with a variable number
of sickled erythrocytes. As the children grow older, the haemato-
logical changes of hyposplenism develop with the appearance of pits
on the surface of the red cells, Howell–​Jolly bodies, and distorted
red cells. The white cell and platelet counts are usually normal or
slightly elevated.
Complications
The chronic haemolysis of sickle cell disease is interspersed
with acute exacerbations of the illness called sickling crises.
Furthermore, there are a series of serious and life-​threatening
long-​term complications which develop in many patients with
sickle cell anaemia.
The different forms of sickle cell crises are summarized in
Box 22.6.7.3. The commonest is the painful crisis. This is some-
times precipitated by infection, dehydration, or exposure to cold,
although quite often no underlying cause can be found. The epi-
sode starts with vague pain, often in the back or bones of the
limbs, which worsens gradually. The pain is almost certainly due
to blockage of small vessels with sickled erythrocytes; aspiration
over areas of bone tenderness has shown infarction of marrow
tissue. Occasionally, abdominal pain is the major symptom and
this may be associated with distension and rigidity, a picture very
similar to an acute abdominal emergency. The diagnostic difficul-
ties in distinguishing between an abdominal crisis and a surgical
abdomen are compounded by the fact that the bowel sounds are
often diminished during abdominal crises.
The acute chest syndrome is the second commonest cause of hos-
pitalization for patients with sickle cell disease. In this particularly
serious form of crisis, sickling within the pulmonary vasculature ini-
tiates a vicious circle of hypoxia, microvascular occlusion, and fur-
ther downstream hypoxia, manifest as acute dyspnoea and pleuritic
pain together with infiltrates on the chest radiograph. It is some-
times accompanied by a fall in the packed cell volume and platelet
count which also may reflect sequestration of sickled cells in the
pulmonary vessels. More than 1 in 10 patients will need ventila-
tory assistance, and there is a mortality rate of approximately 3%.
Patients are treated with supportive therapy including high-​flow
oxygen, top-​up transfusions where possible, and broad-​spectrum
antibiotics. However, many patients will need an exchange transfu-
sion to lower the haemoglobin S percentage in order to break the
downward spiral of sickling and hypoxia.
Neurological complications may present in a variety of ways.
Stroke is particularly common and 11% of patients with sickle cell
disease will have had a stroke by the age of 20. Although the exact
mechanism by which stroke arises in sickle cell disease is unclear,
it is likely to be multifactorial with contributions from endothelial
dysfunction, leucocytosis, anaemia, and nitric oxide dysregulation.
Transcranial Doppler to identify children with increased middle
cerebral arterial blood velocity has been shown to identify those
children who will benefit from prophylactic transfusion to minimize
the risk of stroke. Haemorrhagic strokes are also seen, typically in
older patients, and are thought to be caused by rupture of aneurysms
or collaterals akin to those seen in moyamoya disease. MRI studies
also show a high frequency of silent infarcts, even within the first
few years of life and neurocognitive impairment is not unusual as
a result.
Sequestration crises occur mainly in babies and young children,
and are characterized by a rapid enlargement of the spleen which
becomes engorged with sickled erythrocytes. As the crisis pro-
gresses a large proportion of the total red cell mass may be trapped
in the spleen. Untreated, death may occur due to profound an-
aemia, while caution must be exercised with transfusion to ensure
that a reversal of the sequestration does not result in an excessively
high haemoglobin level. Parents of children with sickling dis-
orders should be taught how to detect splenic enlargement in their
children, and immediate medical attention should be sought for
this serious complication. Hepatic sequestration may also occur,
including in adults, and is easily overlooked if the liver size is not
monitored carefully.
Priapism is another common and distressing acute complication,
which, if recurrent, can result in fibrosis of the corpus cavernosa and
subsequent sexual dysfunction.
During painful crises, there may be a marked increase in the
rate of haemolysis with a fall in the haemoglobin level. Such acute
haemolytic episodes are uncommon. More serious are periods of
transient red cell aplasia called aplastic crises, which result from
intercurrent infection with parvovirus (erythrovirus) B19. Infection
with this erythrovirus temporarily blocks the maturation of red cell
precursors and results in a sharp drop in haemoglobin in the context
of haemolytic anaemias. The combination of worsening anaemia
with a reticulocytopenia suggests this diagnosis, and transfusional
support is needed until the virus is cleared.
Pregnancy in women with sickling disorders may be uneventful,
but there is an increased risk of fetal loss, intrauterine growth retard-
ation, premature labour, and an increased incidence of painful crises
for the expectant mother.
Box 22.6.7.3  Acute exacerbations (‘crises’) in sickle cell disease
•	 Thrombotic:
—	 Generalized or localized bone pain
—	 Abdominal
—	 Pulmonary
—	 Neurological
•	 Aplastic
•	 Haemolytic
•	 Sequestration:
—	 Spleen
—	 Liver
—	 ?Lung
•	 Various combinations of above
section 22  Haematological disorders
5444
Chronic complications
Many of the chronic complications of sickle cell anaemia result from
infarcts following repeated episodes of vascular occlusion. Almost
any organ can be involved. Those at particular risk are areas which
rely largely on small vessels for their blood supply. The bones are
particularly prone to infarction, and avascular necrosis of the hu-
meral or femoral heads may lead to deformity of the shoulder and
hip joints (Fig. 22.6.7.20). Bone infarcts may result in chronic se-
questra formation which may become secondarily infected with the
production of osteomyelitis. Chronic leg ulcers are also commonly
seen, and may prove very difficult to treat effectively.
Another organ at particular risk is the kidney. During early childhood,
renal function may be impaired but this can be corrected by blood trans-
fusion, suggesting that it is due to reversible changes in the renal vascula-
ture. However, alterations in renal function are not reversible in later life.
Chronic renal failure is one of the commonest causes of death in adults
with sickle cell anaemia, and nephrotic syndrome may be seen.
Pulmonary disease is seen, with repeated episodes leading to se-
vere pulmonary hypertension and right heart failure. Irrespective
of the presence of pulmonary hypertension, there is usually some
degree of cardiomegaly. A variety of flow murmurs may be heard,
many of which are the result of chronic anaemia. Myocardial infarc-
tion or fibrosis is not a typical feature of the disease.
Ocular manifestations are also relatively common in sickle cell
anaemia although they tend to be more serious in haemoglobin SC
disease; they will be considered in this context below. Other im-
portant chronic complications include a greatly increased suscepti-
bility to pigment gallstone formation and gallbladder disease.
Course and prognosis
There are still large gaps in our knowledge about the natural history
of sickle cell anaemia, with socioeconomic and ill-​defined genetic
factors being important factors in determining prognosis. In the
developing world, the disease still has a high mortality in the first year
or two of life with infection being a major cause of death. Data from
the United States Cooperative Study of Sickle Cell Disease (1994) sug-
gest that the median age at death for males is 42 years and for females
48 years; more recent data are lacking. In Saudi Arabia and India, a
particularly mild form of the condition occurs; mortality is extremely
low in childhood and a normal survival seems to be common.
Other sickling disorders
The other sickling disorders include the interaction of haemoglobin
S with haemoglobins C, D, and some of the rarer haemoglobin vari-
ants. The interactions with the different forms of β thalassaemia were
described earlier. In many of these conditions, the clinical manifest-
ations are little different from the sickle cell trait, but haemoglobin
SC disease and SD disease more closely resemble sickle cell anaemia.
Haemoglobin SC disease
This disease is found in West Africa and less frequently in North
Africa. Characterized by a milder anaemia than sickle cell disease,
it may go unrecognized until adult life. It may present with a com-
plication resulting from damage to the microvasculature, probably
because of the relatively high haemoglobin level and the combined
effects of sickling and red cell rigidity caused by haemoglobin C
(see ‘Haemolysis due to common haemoglobin variants other than
haemoglobin S’). Aseptic necrosis of the femoral or humeral heads
and unexplained haematuria are common complications, and re-
peated blockage of the retinal vessels may lead to retinitis proliferans,
retinal detachment, and vitreous haemorrhage.
Haemoglobin SC disease is diagnosed by finding a mild anaemia,
sometimes with splenomegaly, and characteristic morphological
changes of the red cells including many target forms, intracellular
crystals, and sickle cells. The sickling test is positive and haemo-
globin HPLC shows haemoglobins S and C in about equal propor-
tions (Fig. 22.6.7.21).
Laboratory diagnosis
The presence of haemoglobin S can be determined by the sickle solu-
bility test. A variety of such tests are available but each is based on the
insolubility of reduced sickle haemoglobin in a phosphate buffer. If
red cells containing haemoglobin S are lysed in a phosphate buffer,
the addition of a reducing agent such as hydrosulphite will result in
the formation of a turbid suspension—​a positive sickle solubility test.
Sickle cell trait causes no haematological changes and is diagnosed
by the finding of a positive sickling test together with haemoglobins
A and S on electrophoresis or HPLC (Fig. 22.6.7.22). Sickle cell an-
aemia is diagnosed by the finding of a variable degree of anaemia, an
elevated reticulocyte count, sickled erythrocytes on the peripheral
blood film, a positive sickling test, and a haemoglobin electrophor-
esis or HPLC pattern characterized by the absence of haemoglobin
A and a preponderance of haemoglobin S with a variable amount of
haemoglobin F (Figs. 22.6.7.21 and 22.6.7.22).
Management
Prospective genetic counselling for couples with sickling disorders
and sickle trait is available in developed countries. Although prenatal
diagnosis of sickle cell disease can be carried out by DNA analysis
following chorionic villus sampling, it has not been taken up as ex-
tensively as it has for the thalassaemias, not least because the pheno-
type of affected children cannot be so accurately predicted. Universal
screening programmes for all neonates help identify affected infants
Fig. 22.6.7.20  Aseptic necrosis of the left femoral head in sickle cell
disease.
22.6.7  Disorders of the synthesis or function of haemoglobin
5445
0.96
1.07
F
1.33
1.24
1.73
2.40
A2
3.62
0
0.0
7.5
%
15.0
22.5
30.0
37.5
45.0
(a)
1
2
Time (min.)
3
4
5
6
45.0
(b)
37.5
30.0
22.5
15.0
%
7.5
0.0
0
1
2
3
Time (min.)
F
1.09
1.25
2.14
2.28
A2
3.62
4.31
4
5
6
45.0
(c)
37.5
30.0
22.5
15.0
F
1.10
1.27
1.80
2.34
A2
3.64
4.49
5.17
%
7.5
0.0
0
1
2
3
Time (min.)
4
5
6
1
Fig. 22.6.7.21  (a) Normal HPLC trace showing dominant peak for HbA, with a smaller peak
for HbA2, and no variant haemoglobins. (b) HPLC trace for homozygous sickle cell disease
(HbSS). Note the increased HbF peak. (c) HPLC trace showing HbSC disease; the rightmost
peak corresponds to HbC.
section 22  Haematological disorders
5446
at the earliest opportunity; this helps to minimize the risk of early
deaths due to infection through the administration of prophylactic
antibiotics and immunization. Affected infants should be given
oral penicillin at a dosage of 62.5 mg three times a day, up to 1 year
of age, 125 mg twice a day from the age of 1 to 3 years, and 250 mg
twice a day thereafter. It is also standard practice for these babies to
receive pneumococcal vaccine, and vaccines against meningococcus
and H. influenzae. While it used to be believed that the high death
rate among infants with sickle cell disease in sub-​Saharan Africa and
similar environments was due to malaria infection, studies have dem-
onstrated that many of these deaths are due to infection with the same
organisms that occur in nonmalarious parts of the world. Appropriate
prophylactic programmes are therefore critically important.
Patients with sickle cell anaemia adapt well to their low haemo-
globin levels and regular blood transfusion is not required. Regular
folate supplements should be given. Patients should be given access
to a centre that has expertise in the management of this disorder and
advised to present at the first sign of a painful crisis. They should also
be given a card to carry which states their haemoglobin genotype.
Painful crises not responding to simple analgesia, oral hydration
and rest should be managed in hospital. Patients should be examined
for evidence of underlying infection and given adequate rehydration,
oxygen, antibiotics where appropriate, and, in particular, analgesia.
The use of patient-​controlled analgesia pumps can result in the rapid
resolution of pain, but must be accompanied by careful monitoring of
respiratory function to avoid oversedation. The haemoglobin level and
reticulocyte count should be estimated at frequent intervals to antici-
pate an aplastic crisis or sequestration episode. It is important to be
alert to the possibility of developing acute chest syndrome, the majority
of which arise in the context of a pre-​existing painful crisis.
The acute chest syndrome may be managed initially with oxygen
and top-​up transfusion (with extended phenotyped, sickle-​free blood)
where the baseline haemoglobin is low enough to permit it; any deteri-
oration warrants red cell exchange transfusion and a low threshold for
involvement of the intensive care team. Similarly, cerebral complica-
tions should be treated by exchange or top-​up transfusion. Exchange
transfusion should also be used to cover major surgical interventions,
such as total hip replacement for avascular necrosis of the femoral
head, or for patients who are having recurrent crises.
Ocular manifestations, particularly proliferative retinopathy, re-
quire expert ophthalmological treatment, likely to involve laser
photocoagulation. Pre-​emptive ophthalmic assessment is advised
annually for patients with sickling disorders to detect proliferative
retinopathy prior to complications such as vitreous haemorrhage.
Haematuria is common, and usually resolves without treatment,
but it is important to be aware of the possibility of renal medullary
carcinoma which is seen almost exclusively in this patient popula-
tion. Proteinuria is also a common manifestation of sickle nephrop-
athy, and treatment with angiotensin-​converting enzyme inhibitors
may slow the rate of development of renal impairment. Endstage
renal failure should be managed as for any other form of renal in-
sufficiency; renal transplantation has been shown to be successful
in several studies though regular exchange transfusions are subse-
quently needed to maintain the health of the graft.
Recurrent priapism may be a problem. Nearly two-​thirds of major
episodes are preceded by stuttering attacks and therefore it has been
suggested that effective therapy at this stage may reduce the risk of
sustaining a major attack, with danger of permanent deformity of the
penis. Several forms of management have been suggested although
none has been studied in sufficient detail. One approach has been to
commence etilefrine, an α-​adrenergic agonist during the stuttering
phase. Acute or fulminant cases may require intracavernosal irriga-
tion with epinephrine. Centres with experience of this complication
suggest that conservative treatment should be restricted to 24 h at
the most. If there is no improvement, surgical correction is recom-
mended, with a cavernosum–​spongiosum shunt.
The management of leg ulcers is unsatisfactory. They may heal
with bed rest and debridement but often relapse. Skin grafting does
not always give good results and controlled trials have shown that
transfusion does not appear to increase the rate of healing.
Increasingly, efforts are being made to emphasize a preventa-
tive rather than reactive approach to sickle crises, with the recog-
nition that long-​term, subclinical sickling will result in end-​organ
damage however effective the treatment of acute crises. For patients
with more than three painful crises per year, treatment with oral
hydroxycarbamide has been shown to improve quality of life and
reduce the overall mortality of sickle cell disease. Variable increases
in fetal haemoglobin production are seen in patients treated with
hydroxycarbamide, and this may underlie its beneficial effect. Initial
concerns about the possible leukaemogenicity of hydroxycarbamide
have not been borne out by long-​term studies of its safety. If
hydroxycarbamide is not tolerated or ineffective, long-​term elective
red cell exchange programmes may be used, with good effect.
Although this may be a significant burden for the patient and pose a
risk of red cell alloimmunization, it may free patients with especially
severe clinical phenotypes from frequent and disabling crises.
Regular top-​up transfusions are avoided where possible to avoid iron
overload (with the exception of transfusion in the light of Doppler studies
suggesting an increased risk of stroke in children—​mentioned previ-
ously). The increased viscosity of the blood in patients with sickle cell dis-
ease means that over-​transfusion (>100 g/​litre) should also be avoided.
Hb A
Hb S
5
4
3
2
1
Origin
−
+
Fig. 22.6.7.22  Haemoglobin electrophoresis showing the haemoglobin
pattern in the sickling disorders (starch gel electrophoresis, protein stain,
pH 8.5). The following are shown (left to right): (1 and 2) the sickle cell
trait; (3) normal; (4) sickle cell anaemia; (5) normal.
22.6.7  Disorders of the synthesis or function of haemoglobin
5447
There has been significant progress in recent years in developing
new treatments for sickle cell disease which target the underlying
pathophysiology of this condition. Crizanlizumab, monoconal anti-
body targeted against the adhesion molecule P-selectin, has been
shown in randomized studies to reduce the rate of sickle-related
painful crises relative to placebo, by disrupting the cell-cell inter-
actions which are thought to be central to development of such ex-
acerbations. An alternative strategy of reducing sickle haemoglobin
polymerization has also shown some benefit in phase III random-
ized trials; voxelotor, a drug which reversibly binds to, and stabilizes,
the oxygenated form of haemoglobin has been shown to improve
parameters associated with haemolysis. L-glutamine has been ap-
proved as a new agent for the management of patients with sickle cell
anaemia, again reducing the frequency of painful crises in a phase
III randomized controlled study, presumably through anti-oxidative
mechanisms. Whether these agents will have an impact on the long
term outcomes of patients with sickling disorders remains to be seen.
Towards a cure for the sickling disorders
The management of sickle cell anaemia remains largely supportive,
with hydroxycarbamide being the only widely available effective treat-
ment to date. As with thalassaemia major, allogeneic bone marrow
transplant programmes exist which offer the possibility of cure to
patients who have sustained only minimal end-​organ damage and
who can therefore tolerate the procedure. Currently this means that
bone marrow transplantation is undertaken mostly in children, and
requires a careful discussion of the risks and benefits of such a major
procedure if the patient’s true clinical phenotype is still unclear. Efforts
to understand the normal control of γ globin transcription, and thus
to reverse its silencing in adult life, remain the focus of many transla-
tional research groups. The discovery of the critical role of the tran-
scription factor BCL11a in the silencing of fetal haemoglobin has been
an important step forward in this process.
Haemolysis due to common haemoglobin variants other
than haemoglobin S
After haemoglobin S, the second commonest variant in West Africa
is haemoglobin C. Because of its relatively low solubility haemoglobin
C appears to exist in a precrystalline state in red cells, causing their
rigidity and premature destruction in the microcirculation. The homo-
zygous state, haemoglobin C disease, is characterized by a mild haemo-
lytic anaemia with splenomegaly, and 100% target cells on the blood
film. Haemoglobin analysis shows haemoglobin C with small amounts
of haemoglobin F.  By contrast with haemoglobin SC, homozygous
haemoglobin C is a mild disorder and no specific treatment is required.
The commonest haemoglobin variant throughout South-​East
Asia and the Indian subcontinent is haemoglobin E. The homozy-
gous state for this variant, haemoglobin E disease, is characterized
by a very mild degree of anaemia with a slight reticulocytosis. The
blood film shows mild morphological changes of the red cells which
are hypochromic and microcytic, resembling the changes seen in β
thalassaemia. No treatment is required.
Haemoglobin variants which migrate in the position of haemoglobin
S on electrophoresis but which do not sickle have been given the gen-
eral title of haemoglobin D. There are several different molecular var-
ieties of this variant; the commonest is haemoglobin D Los Angeles. The
homozygous state is associated with moderate anaemia, splenomegaly,
and a mild degree of haemolysis. The compound heterozygous state with
haemoglobin S produces a disorder very similar to sickle cell anaemia.
The unstable haemoglobin disorders
The unstable haemoglobin disorders are a rare group of inherited
haemolytic anaemias which result from structural changes in the
haemoglobin molecule that cause intracellular precipitation with
the formation of Heinz bodies. Their true incidence is not known.
There have been several well-​documented families in which patients
with one of these haemoglobin variants have had no affected rela-
tives, suggesting that the condition has arisen by a new mutation.
Aetiology and pathogenesis
Most of the unstable haemoglobin variants result from single amino
acid substitutions at critical areas of the molecule. For example,
substitutions in or around the haem pocket can disrupt the normal
structure and allow in water, with subsequent oxidative damage to
haem which leads to precipitation of the haemoglobin. Some sub-
stitutions, such as those involving proline residues, cause a marked
disruption of the secondary structure of a globin chain. A few of
these variants result from deletions of either single or several amino
acid residues. For example, in haemoglobin Gun Hill, five amino
acids are missing including the haem binding site. As the unstable
haemoglobins precipitate in the red cells or their precursors, they
produce intracellular inclusions, or Heinz bodies, which make the
cells more rigid causing their premature destruction in the microcir-
culation (Fig. 22.6.7.23). The degradation products of the precipi-
tated haemoglobin, notably haem and iron, cause oxidative damage
to the red cell membrane proteins in much the same way as the ex-
cess α and β chains produced in the thalassaemias.
Clinical features
All these conditions are characterized by a haemolytic anaemia of
varying severity and splenomegaly. There may be a history of the pas-
sage of dark urine, particularly during episodes of infection. As in
all chronic haemolytic anaemias, there is an increased incidence of
pigment gallstones. The condition may become worse during periods
of intercurrent infection. In the more severe forms, such episodes
are associated with life-​threatening anaemia. Patients with unstable
haemoglobins are at particular risk of haemolytic episodes following
Fig. 22.6.7.23  The peripheral blood film of a patient with an
unstable haemoglobin disorder, haemoglobin Hammersmith. This is a
postsplenectomy film, which shows small inclusions in many of the red
cells (×1000, Leishman stain).
section 22  Haematological disorders
5448
the administration of oxidant drugs. Apart from intermittent icterus
and splenomegaly there are no characteristic physical findings.
Laboratory diagnosis
This condition should be considered in any familial haemolytic an-
aemia, particularly if a red cell enzyme deficiency cannot be demon-
strated. The peripheral blood film shows the features of haemolysis but
the red cell morphology may be relatively normal. Occasionally there
is a mild degree of hypochromia and microcytosis. Unless splenectomy
has been carried out, Heinz bodies are not seen in the peripheral blood.
The most characteristic feature of the unstable haemoglobins is
their heat instability. If a dilute haemoglobin solution is heated at
50°C for 15 min, most of the unstable haemoglobins precipitate as a
dense cloud. A similar phenomenon can be induced by isopropanol.
Sequencing of the globin genes allows a precise molecular diagnosis,
and over 140 unstable variants have been identified to date.
Treatment
Because these conditions are so rare, there has been very little experience
of the effects of splenectomy. From the information that is available, and
from the senior author’s personal experience, it appears that if a child has
had several life-​threatening episodes of anaemia or is running a steady-​
state haemoglobin level which is impairing development or well-​being,
splenectomy should be undertaken. It is interesting to note that some of
these haemoglobin variants produce a ‘right shift’ in the oxygen dissoci-
ation curve, and a measurement of the P50 as part of the pre-​splenectomy
assessment may help to decide whether to proceed to surgery; a marked
right shift, that is, an increased P50, indicates that the anaemia should be
more easily tolerated than if the oxygen dissociation curve is moved in
the opposite direction with a low P50. An accurate history from the child
or its parents is probably more helpful, however.
Haemoglobin variants which cause abnormal
oxygen binding
The first high-​affinity haemoglobin identified was haemoglobin
Chesapeake, detected as an abnormal haemoglobin band in a pa-
tient with otherwise unexplained polycythaemia. Since then, over 90
haemoglobin variants of this type have been defined, all associated
with familial polycythaemia.
Aetiology
The high ​oxygen ​affinity haemoglobin variant may result from single
amino acid substitutions in either the α or β globin chains, in critical
parts of the haemoglobin molecule which are involved in the config-
uration changes that underlie haem–​haem interaction and the pro-
duction of a sigmoid oxygen dissociation curve. Many occur at the
junctions between the α and β subunits. Others involve the amino
acids which are involved with the binding of 2,3-​bisphosphoglycerate
(2,3-​BPG) to haemoglobin. As mentioned earlier, increasing con-
centrations of 2,3-​BPG tend to push the oxygen dissociation curve
to the right; fetal haemoglobin has a high oxygen affinity (left-​shifted
curve) because it cannot interact with 2,3-​BPG; mutations of the
BPG binding sites have a similar effect.
Pathophysiology
The high ​oxygen ​affinity variants have a left-​shifted oxygen dis-
sociation curve with a reduced P50, which may be detected using a
standard blood gas analyser. The variant haemoglobin holds on to
oxygen more avidly than normal haemoglobin. This leads to tissue
hypoxia. This in turn causes an increased output of erythropoietin and
an elevated red cell mass.
Clinical features
Many patients with high ​oxygen ​affinity variants are completely
healthy and are only found to carry the variant when a routine haem-
atological examination shows an unusually high haemoglobin level
or packed cell volume. There have been one or two reports of arterial
or venous occlusive disease in these patients. However, this is un-
common. Most patients are asymptomatic. There is no splenomegaly
and no other associated haematological findings. Although it might
be expected that a high ​oxygen ​affinity haemoglobin would cause de-
fective oxygenation of the fetus, this has not been observed clinically.
Diagnosis
The condition should be suspected in any patient with polycy-
thaemia associated with a left-​shifted oxygen dissociation curve.
A raised or inappropriately normal serum erythropoietin will be
seen. The diagnosis can be confirmed by haemoglobin analysis.
Treatment
In asymptomatic patients with high ​oxygen ​affinity haemoglobin
variants no treatment is necessary. The difficulty arises if the patient
has associated vascular disease with symptoms of coronary or cere-
bral artery insufficiency. These patients require a high haemoglobin
level for oxygen transport; half their haemoglobin is physiologically
useless. Venesection is therefore not usually recommended for these
patients, though there is insufficient evidence to support categorical
statements how these patients should be managed.
Low ​oxygen ​affinity variants
At least 60 haemoglobin variants with reduced oxygen affinity have
been reported. The first to be described, haemoglobin Kansas, was
found in a mother and son with unexplained cyanosis. The subjects
were asymptomatic and had normal haemoglobin levels without
any evidence of haemolysis. Like many of the high affinity vari-
ants, the amino acid substitution in this variant was at the interface
between the α and β globin chains. This condition should be thought
of in any patient with an unexplained congenital cyanosis; the differ-
ential diagnosis is considered later in this chapter.
Methaemoglobinaemia,
carboxyhaemoglobinaemia,
and sulphaemoglobinaemia
Methaemoglobinaemia is a condition characterized by in-
creased quantities of haemoglobin in which the iron of haem is
oxidized to the ferric (Fe3+) form. Carboxyhaemoglobinaemia
(carbonmonoxyhaemoglobinaemia) results from the binding of
carbon monoxide to the haem molecules. Sulphaemoglobinaemia
is a rare condition in which there is a mixture of haemoglobin de-
rivatives whose structure is poorly characterized but which can be
defined by their specific spectral characteristics.
Pathogenesis
As mentioned earlier, each haemoglobin molecule has four haem moi-
eties. At first sight it is not clear why the oxidation of a proportion of
the iron atoms, or the fact that they are liganded to carbon monoxide,
22.6.7  Disorders of the synthesis or function of haemoglobin
5449
should cause such profound changes in oxygen transport. However,
oxidation of 30% of the haem molecules has a much more serious ef-
fect on tissue oxygenation than a reduction of the haemoglobin level
by the same amount. This is because, if a single haem is oxidized, it so
alters the conformation of the haemoglobin molecule that the oxygen
affinity of the other three haems is increased. Thus methaemoglobin,
carboxyhaemoglobin, and cyanmethaemoglobin all have very high
oxygen affinities with left-​shifted oxygen dissociation curves, and
hence are associated with impaired unloading of oxygen to the tissues.
Methaemoglobinaemia
Methaemoglobin causes a variable degree of cyanosis. It should
be suspected in any patient with significant central cyanosis in
whom there is no evidence of cardiorespiratory disease. The de-
gree of cyanosis produced by 50 g/​litre of deoxygenated haemo-
globin can be produced by 15 g/​litre methaemoglobin and 5 g/​litre
of sulphaemoglobin. Methaemoglobin concentrations of 10 to 20%
are tolerated quite well. It is useless as an oxygen carrier; levels above
this are thus often associated with dyspnoea and headache. Much
depends on the rapidity at which it is formed. Many patients with
lifelong methaemoglobinaemia are asymptomatic, while individuals
who have accumulated a similar level of methaemoglobin acutely
may be acutely dyspnoeic. For reasons that are not clear, it is un-
usual for patients with chronic methaemoglobinaemia to have an
increased haemoglobin level or red cell count.
Methaemoglobinaemia may arise as a result of a genetic defect in
red cell metabolism or haemoglobin structure, or may be acquired
following the ingestion of various oxidant drugs and toxic agents.
Genetic methaemoglobinaemia
There are two forms of inherited methaemoglobinaemia. The first,
and less common, results from a deficiency of red cell NADH-​
cytochrome b5 reductase, the second from a structural alteration in
either the α or β globin chains of haemoglobin.
NADH-​diaphorase (NADH methaemoglobin reductase) cata-
lyses a step in the major pathway for methaemoglobin reduction. The
enzyme reduces cytochrome b5 using NADH as a hydrogen donor.
The reduced cytochrome b5, in turn, reduces methaemoglobin
to haemoglobin. There are several different molecular forms of
NADH-​cytochrome b5 reductase deficiency which have been iden-
tified by electrophoretic analysis of NADH-​cytochrome b5 reductase
in the red cells of affected patients. The condition is inherited in an
autosomal recessive manner. Homozygotes have elevated levels of
methaemoglobin and are cyanotic from birth. Heterozygotes do not
have elevated levels of methaemoglobin but seem to be unusually
susceptible to the oxidant action of drugs. For example, severe cyan-
osis has been precipitated by the use of antimalarial drugs.
There are several abnormal haemoglobin variants which are
associated with genetic methaemoglobinaemia, all of which
are designated haemoglobin M, and further identified by their
place of discovery (e.g. haemoglobin M Boston, haemoglobin M
Milwaukee). These variants may affect either the α or β chain, but
usually result from amino acid substitutions near the haem pocket.
Normally, haem lies between two histidine residues, one called the
proximal histidine to which it is attached, and the other called
the distal histidine. Oxygen is bound to haem at a site opposite to
the distal histidine. If the latter is substituted by tyrosine, as occurs
in the α chain variant haemoglobin M Boston and in the β chain
variant M Saskatoon, a stable bond is formed between the haem
iron and the phenolic ring of the tyrosine. The iron atom is ‘fixed’
in the Fe3+ state. These variants are associated with cyanosis which
is present from early life. In the case of the α chain variants it is
present from birth, while the β chain haemoglobin variants only
produce cyanosis after the first few months of life as adult haemo-
globin synthesis becomes established. Unlike NADH diaphorase
deficiency, which is inherited as a recessive trait, the haemoglobin
Ms have a dominant form of inheritance. Thus the diagnosis of gen-
etic methaemoglobinaemia and even the affected globin chain can
be ascertained by a good clinical history.
The diagnosis is confirmed by spectroscopic examination of the
blood and by determination of methaemoglobin levels. The precise
cause can be established by combination of HPLC and globin gene
sequencing, or by an assay of NADH-​diaphorase.
Genetic methaemoglobinaemia due to NADH-​diaphorase defi-
ciency is readily treated by the administration of ascorbic acid, 300
to 600 mg daily by mouth in divided doses, or by the administration
of methylene blue, either intravenously (1 mg/​kg body weight) or by
mouth 60 mg three to four times daily. On the other hand, the gen-
etic methaemoglobinaemias due to structural haemoglobin variants
do not respond to ascorbic acid, methylene blue, or any other treat-
ment. Most affected individuals go through life asymptomatic and
require no treatment.
Acquired methaemoglobinaemia
Acquired methaemoglobinaemia usually results from the adminis-
tration of drugs or exposure to chemicals which cause oxidation of
haemoglobin. There are many agents which are capable of exceeding
the red cells’ ability to reduce methaemoglobin. They include ferri-
cyanide, bivalent copper, chromate, chlorate, quinones, and certain
dyes with a high oxidation–​reduction potential. Nitrite, often used
as a preservative, is one of the most common methaemoglobin-​
forming agents. Nitrates, after conversion to nitrites in the gut,
may cause serious methaemoglobinaemia in infants. Other agents
which commonly cause methaemoglobinaemia include phenacetin,
primaquine, sulfonamides, and various analine dye derivatives.
If any of the agents listed previously is given in low dose over a
long period of time it may lead to chronic methaemoglobinaemia
with or without a haemolytic anaemia. However, after exposure to a
large amount of these agents, and the development of in excess of 50
to 60% methaemoglobin, the symptoms of acute anaemia develop
because methaemoglobin lacks the capacity to transport oxygen.
Thus the clinical picture may be characterized by vascular collapse,
coma, and death.
Methaemoglobinaemia with haemolytic anaemia
The haemolytic action of oxidant drugs is described elsewhere (see
also Chapter 22.6.11). Chronic methaemoglobinaemia with haemo-
lytic anaemia, characterized by Heinz body formation and frag-
mented red cells, occurs commonly in patients receiving dapsone,
salazopyrine, or phenacetin. This condition is usually innocuous
and can be modified by adjusting the dose of the drug.
Occasionally, acute intravascular haemolysis associated with
methaemoglobinaemia and disseminated intravascular coagulation
occurs. It usually follows the ingestion or infusion of a strong oxi-
dizing agent such as chlorate or arsine. There is gross intravascular
haemolysis and methaemoglobinaemia together with evidence of
disseminated intravascular coagulation. The haemoglobin level may
fall very rapidly and may be complicated by renal failure.

22.6.8 Anaemias resulting from defective maturatio
22.6.8 Anaemias resulting from defective maturation of red cells 5450 Stephen J. Fuller and James S. Wiley
section 22  Haematological disorders
5450
Treatment
In cases of chronic acquired methaemoglobinaemia, the drug or
chemical agent should be removed where possible. If continued
therapy is required, it should be administered at a lower dose.
Acute toxic methaemoglobinaemia presents a serious medical
emergency. Methylene blue should be administered in a dose of 1
to 2 mg/​kg intravenously over a 5-​min period. Repeated doses may
be needed. Toxicity is uncommon, although doses of over 15 mg/​kg
may cause haemolysis in young infants. The drug should not be used
if the methaemoglobinaemia is due to chlorate poisoning, as it may
convert the chlorate to hypochlorite which is an even more toxic
compound. In cases of acute methaemoglobinaemia with intravas-
cular haemolysis, haemodialysis with exchange transfusion is the
treatment of choice.
Carboxyhaemoglobinaemia
Carbon monoxide has an affinity for haemoglobin approximately
210 times that of oxygen. Following acute exposure, it is so tightly
bound that it takes about 4 h for an individual with normal ventila-
tion to expel half of it. At levels of 5 to 10% there may be no symp-
toms, but above 20% there is usually headache and weakness. Levels
of 40 to 60% or more lead to unconsciousness and death.
Carbon monoxide poisoning is discussed in Chapters 10.1 and
10.2 and secondary polycythaemia due to chronic exposure is con-
sidered elsewhere in this section (see Chapter 22.3.5).
Sulphaemoglobinaemia
This poorly defined condition derives its name from the fact that
it can be produced in vitro by the action of hydrogen sulphide
on haemoglobin. It has not been reported as a genetic disorder.
It is usually associated with the administration of drugs, particu-
larly sulfonamides or phenacetin. It has also been reported in
patients with chronic constipation or malabsorption syndromes
(enterogenous cyanosis) although its relationship to these dis-
orders is far from clear.
Other acquired abnormalities of the structure
or synthesis of haemoglobin
Glycosylated haemoglobin, haemoglobin A1c
Haemoglobin may undergo post-​translational modification in pa-
tients with diabetes. The abnormal haemoglobin, haemoglobin
A1c, is formed by the nonenzymic combination of glucose with the
N-​terminus of the β chain, first forming a Schiff base which then
undergoes a rearrangement to form a stable ketoamine. The level of
haemoglobin A1c is raised in diabetics and is related to the blood
sugar level over the previous weeks. The value of the estimation of
haemoglobin A1c as an index of the control of diabetes is considered
elsewhere.
Fetal haemoglobin production in adult life
A number of haematological disorders are associated with a re-
version to fetal haemoglobin production after the neonatal period.
These include juvenile myelomonocytic leukaemia and congenital
hypoplastic anaemias. Haemoglobin F may also appear transitorily
during rapid regeneration of the bone marrow after drug-​induced
hypoplasia, viral infection, or bone marrow transplantation,
and the level is also slightly elevated during the mid trimester of
pregnancy.
FURTHER READING
Forget BG (2011). Progress in understanding the haemoglobin switch.
N Engl J Med, 365, 852–​4.
Higgs DR, Gibbons RJ (2010). The molecular basis of alpha-​thalassaemia:
a model for understanding human molecular genetics. Hematol Oncol
Cin North Am, 24, 1033–​54.
Houwing ME, et al. (2019). Sickle cell disease: Clinical presentation
and management of a global health challenge; SCORE Consortium.
Blood Rev, 37, 100580.
Orkin SH, et al. (eds) (2014). Nathan and Oski’s hematology of infancy
and childhood, 9th edition. Saunders and Elsevier, Philadelphia.
Piel FB, Steinberg MH, Rees DC (2017). Sickle Cell Disease. N Engl J
Med, 376, 1561–73.
Serjeant GR, Serjeant BE (2001). Sickle cell disease, 3rd edition. Oxford
University Press, Oxford.
Steinberg MH, et al. (Eds) (2009). Disorders of hemoglobin, 2nd edi-
tion. Cambridge University Press, New York.
Telen MJ, Malik P, Vercellotti GM (2019). Therapeutic strategies for
sickle cell disease: towards a multi-agent approach. Nat Rev Drug
Discov, 18(2), 139–58.
Weatherall DJ, Clegg JB (2001). The thalassaemia syndromes, 4th edi-
tion. Blackwell Science, Oxford.
Weatherall DJ (2013). The role of the inherited disorders of haemo-
globin, the first ‘molecular diseases’ in the future of human genetics.
Annu Rev Genomics Hum Genet, 14, 1–​24.
Weatherall DJ, Schechter AN, Nathan DG (eds) (2013). Hemoglobin
and its diseases. Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, New York.
22.6.8  Anaemias resulting from
defective maturation of red cells
Stephen J. Fuller and James S. Wiley
ESSENTIALS
Defective maturation of red cells leads to premature destruction of
nucleated red cell precursors before they leave the bone marrow,
which results in expansion of the marrow, haemolytic jaundice,
peripheral signs of increased erythroid turnover on blood films,
and (in long-​standing disorders) iron overload due to enhanced
absorption.
Causes of ineffective erythropoiesis
These include (1) inhibition of erythroid DNA synthesis (e.g. meg-
aloblastic anaemias most commonly caused by vitamin B12 or
folate deficiency, and drugs blocking DNA synthesis); (2) clonal dis-
orders of erythropoiesis (e.g. myelodysplastic syndromes with ring
22.6.8  Anaemias resulting from defective maturation of red cells
5451
sideroblasts, and acute erythroid leukaemia); (3) genetic disorders of
erythropoiesis (e.g. thalassaemia syndromes, hereditary sideroblastic
anaemia, and congenital dyserythropoietic anaemia); and (4) other
causes (e.g. alcohol).
Sideroblastic anaemias
These result from defects in haem biosynthesis, with most cases
being acquired as a clonal disorder of erythropoiesis, with varying
degrees of myelodysplasia. Other causes are (1) hereditary (e.g. in-
herited defects of the erythroid-​specific 5-​aminolevulinic acid syn-
thase 2 gene on the X-​chromosome causes congenital sideroblastic
anaemia); (2)  ­acquired non-clonal (e.g. drugs or toxins such as
ethanol, isoniazid, or lead; following chemotherapy or irradiation;
or of unknown cause).
Diagnosis is confirmed by finding ring sideroblasts (erythroblasts
containing five or more iron-​positive granules arranged in a peri-
nuclear location around one-​third or more of the nucleus) on a
bone marrow aspirate stained with Prussian blue iron reagent. Aside
from supportive care with blood transfusion and iron chelation, a
trial of pyridoxine is generally indicated (25% of hereditary cases—​
but few acquired cases—​show some response). Acquired clonal
sideroblastic anaemia has a median survival of 70–100  months,
with 3 to 12% progressing to acute leukaemia.
Introduction
Erythroid cell maturation is specialized towards the coordinated
synthesis of large amounts of haem and globin necessary to attain
the high concentration of haemoglobin found in the mature red
cell. Hereditary or acquired defects in the production of either of
these cause a maturation block, which leads to ineffective erythro-
poiesis in which many of the developing nucleated erythroblasts
are destroyed in the marrow before they can reach the circulation.
Thus in thalassaemia, defective synthesis of either α-​ or β-​globin
leads to unbalanced production of the other chain, which pre-
cipitates and leads to destruction of the precursor erythroblast.
Defective haem synthesis in the sideroblastic anaemias also leads
to an anaemia which is characterized by ineffective erythropoiesis
(Box 22.6.8.1). Abnormalities of DNA synthesis in the developing
erythroid cells, produced, for example, by vitamin B12 or folic acid
deficiency, blocks cell division required for erythroid maturation
and produces morphological and biochemical evidence of inef-
fective erythropoiesis.
Ineffective erythropoiesis may be recognized by the character-
istic erythroid hyperplasia of the bone marrow with normal or
only slightly increased reticulocyte numbers. Some other fea-
tures of ineffective erythropoiesis may be variably present: a mild
increase in bilirubin, a decrease in haptoglobin, and increased
serum lactic dehydrogenase activity. As a result, iron absorp-
tion is increased, serum iron and ferritin become elevated, and,
after many years, iron overload develops which is indistinguish-
able from idiopathic haemochromatosis. However, the degree of
iron overload does not depend on either the severity of the an-
aemia or the presence of the characteristic mutation (Cys282Tyr,
His63Asp, and Ser65Cys) of the HFE gene associated with genetic
haemochromatosis.
Sideroblastic anaemias
The sideroblastic anaemias are a group of hereditary or acquired
anaemias of varying severity diagnosed by the finding of ring
sideroblasts in the bone marrow aspirate. The peripheral blood film
shows hypochromic red cells which are microcytic in the heredi-
tary form (Fig. 22.6.8.1), but are often macrocytic in the acquired
forms of the disease. Normochromic and normocytic red cells are
also present which gives the film a dimorphic distribution of red
cell sizes. The diagnostic procedure is bone marrow aspirate fol-
lowed by staining of the smear with Prussian blue iron reagent.
Ring sideroblasts are diagnostic (Fig. 22.6.8.2) and are defined as
erythroblasts containing five or more iron-​positive granules ar-
ranged in a perinuclear location around one-​third or more of the
Box 22.6.8.1  Anaemias with defective red cell maturation and
ineffective erythropoiesis
•	 Inhibition of erythroid DNA synthesis:
—	 Megaloblastic anaemias (most commonly due to vitamin B12 or
folate deficiency)
—	 Drugs blocking DNA synthesis (e.g. hydroxycarbamide,
6-​mercaptopurine)
•	 Clonal disorders of erythropoiesis:
—	 Refractory anaemia
—	 Acquired clonal sideroblastic anaemias (refractory anaemia with
ring sideroblasts, with and without thrombocytosis)
—	 Acute erythroid leukaemia (WHO classification subtypes: erythro-
leukaemia and pure erythroid leukaemia)
•	 Genetic disorders of erythropoiesis:
—	 Thalassaemia syndromes
—	 Hereditary sideroblastic anaemias
—	 Congenital dyserythropoietic anaemias (CDAs)
•	 Miscellaneous:
—	 Alcohol
—	 Drugs
—	 Heavy metal poisoning (e.g. arsenic)
—	 Falciparum malaria
Fig. 22.6.8.1  Peripheral blood smear in hereditary sideroblastic
anaemia showing a population of hypochromic and microcytic
erythrocytes.
section 22  Haematological disorders
5452
nucleus. Electron microscopy reveals that the iron-​containing gran-
ules are mitochondria containing precipitated ferric phosphate and
ferric hydroxide. The sideroblastic anaemias have diverse aetiologies
(Box 22.6.8.2) but have in common an impaired biosynthesis of haem
in the erythroid cells of the marrow. Most sideroblastic anaemias are
acquired as a clonal disorder of erythropoiesis, classified as a subtype of
myelodysplasia. The hereditary forms are uncommon. Most are found
in males with an X-​linked pattern of inheritance. A number of drugs
have been associated with reversible sideroblastic anaemia, chiefly in
patients with alcohol abuse (Box 22.6.8.2).
Hereditary sideroblastic anaemias
Aetiology and pathogenesis
Haem biosynthesis occurs by a cascade of eight enzymes
(Fig. 22.6.8.3). In humans, mutations affecting the first enzyme
of this pathway produce hereditary sideroblastic anaemia. Inborn
errors that occur in later enzymes in this pathway result in meta-
bolic disorders known as the porphyrias (Fig. 22.6.8.3). The
pathway begins with the condensation of glycine with succinyl
CoA to form 5-​aminolaevulinic acid (ALA), a step which is under
the control of the mitochondrial enzyme ALA synthase. This en-
zyme requires pyridoxal phosphate as a cofactor. Two isoenzymes
of ALA synthase have been identified. One is found in liver and
other tissues (ALAS1); the other is confined to erythroid cells
of the bone marrow (ALAS2). In most families with inherited
sideroblastic anaemia, males are affected with an X-​linked pattern
of inheritance consistent with a mutation on the X chromosome
(Fig. 22.6.8.4). The gene for the erythroid-​specific ALAS2 isoen-
zyme resides on the X chromosome and is now known to be the site
of most mutations giving rise to X-​linked hereditary sideroblastic
anaemia. Over 100 different mutations have been described in
different families and nearly all result from a single amino acid
alteration arising from a point mutation in the ALAS2 coding re-
gion of DNA. However, in nearly half the families with hereditary
sideroblastic anaemia, the structure of the ALAS2 gene is normal,
and a number of inherited variations in other genes involved in
haem synthesis have recently been identified. These hereditary
sideroblastic anaemias consist of two nonsyndromic forms, which
have a similar phenotype to X-​linked sideroblastic anaemia and
five rare syndromic forms where haem synthesis is affected in
nonhaematopoietic tissues in addition to red cells (Box 22.6.8.2).
Of the nonsyndromic forms, inherited mutations in the SLC25A38
and GLRX5 genes cause autosomal recessive pyridoxine-​refractory
sideroblastic anaemia.
Clinical and laboratory features
Typically the anaemia presents in infancy or childhood, but when
the condition is mild the diagnosis may not be made until adult
life. Occasionally, such patients may present with features of iron
overload such as diabetes or cardiac failure. Others may be found
in family surveys, which should be undertaken when this anaemia
is diagnosed. Slight enlargement of the liver or spleen may occur.
The degree of anaemia is variable, ranging from severe (haemo-
globin <80 g/​litre) to mild (>100 g/​litre) but even with mild or no
anaemia the mean corpuscular volume (MCV) is below the normal
range. The blood film shows a population of cells with hypochromic,
microcytic morphology. In X-​linked sideroblastic anaemia, female
carriers may show the characteristic red cell dimorphism. White
cell counts are normal, while platelet counts are normal or slightly
elevated. Serum iron and ferritin concentrations are invariably in-
creased and transferrin shows an increased percentage saturation
with iron. The differential diagnosis includes idiopathic haemo-
chromatosis, since both diseases have evidence of iron overload.
Examination of the blood film, the MCV, and the bone marrow
should establish the diagnosis.
Fig. 22.6.8.2  Bone marrow smear stained with Prussian blue, showing
ring sideroblasts.
Box 22.6.8.2  Classification of sideroblastic anaemias
Hereditary
•	 Nonsyndromic:
—​	 X-​linkeda
—​	 Autosomal inheritancea (includes pyridoxine-​refractory autosomal
recessive sideroblastic anaemia)
•	 Syndromic:
—​	 X-​linked sideroblastic anaemia with ring sideroblasts and cere-
bellar ataxia
—​	 Myopathy, lactic acidosis, and sideroblastic anaemia
—​	 Pearson syndrome
—​	 Thiamine-​responsive megaloblastic anaemia
—​	 Sideroblastic anaemia with immunodeficiency, fevers, and devel-
opmental delay
Acquired
•	 Refractory anaemia with ring sideroblasts (RARS), now classified as
myelodysplastic syndrome with RS (MDS-RS)a
•	 Associated with previous chemotherapy, irradiation
•	 RARS with thrombocytosis (RARS-T), now classified as myelodysplastic/
myeloproliferative neoplasm with RS and thrombocytosis (MDS/
MPN-RS-T)
Drugs
•	 Alcohol
•	 Isoniazid, cycloserine, pyrazinamide
•	 Chloramphenicol
Rare causes
•	 Erythropoietic protoporphyria
•	 Copper deficiency or zinc overload
•	 Hypothermia
a Trial of pyridoxine indicated.
22.6.8  Anaemias resulting from defective maturation of red cells
5453
The syndromic forms of hereditary sideroblastic anaemia present
with anaemia in combination with either muscle, neurological, or
pancreatic tissue involvement (Box 22.6.8.2). These disorders show
both X-​linked and autosomal patterns of inheritance. The labora-
tory features of these sideroblastic anaemias are similar to X-​linked
sideroblastic anaemia; however, mutations in the high-​affinity
thiamine transporter gene, SLC19A2, cause the unusual feature of
megaloblastic erythroid maturation with ring sideroblasts.
Treatment and prognosis
A trial of pyridoxine, 100 to 200 mg/​day taken orally, is indicated
for 3 months in all patients with proven or suspected hereditary
sideroblastic anaemia. About 25% of patients experience a full or
partial correction. This vitamin should be continued lifelong in re-
sponders but at a lower maintenance dosage. Regular transfusions
of packed red cells are the mainstay of treatment of severe anaemia.
These should be given to relieve symptoms and allow normal child-
hood development. Splenectomy is contraindicated in this con-
dition. Iron overload progresses rapidly once transfusions begin.
Chelation therapy with desferrioxamine or oral deferasirox should
thus be commenced after the first 10 to 20 transfusions. Iron re-
moval may greatly benefit patients with mild or moderate anaemia
and evidence of iron overload. Furthermore, patients should avoid
alcohol and ascorbic acid supplements, both of which enhance iron
absorption.
Acquired clonal sideroblastic anaemias
Acquired clonal sideroblastic anaemias may either be idiopathic or
develop following chemotherapy or irradiation (Box 22.6.8.2). Since
nearly all cases also show evidence of dyserythropoiesis, this an-
aemia was classified as one of the myelodysplastic syndromes and
Hereditary
coproporphyria
Protoporphyrinogen III
Protoporphyrin IX
Haem
Glycine +
succinyl CoA
ALA
ALA
Porphobilinogen
Acute intermittent
porphyria
Hydroxymethylbilane
Uroporphyrinogen III
Coproporphyrinogen III
Porphyria cutanea
tarda
Congenital porphyria
Plumboporphyria
X-linked hereditary
sideroblastic anaemia
Erythropoietic
protoporphyria
Variegate porphyria
Fig. 22.6.8.3  Pathway of haem biosynthesis in mammalian cells. The first step in the pathway
is catalysed by ALAS and occurs within the mitochondrion using pyridoxal 5′-​phosphate as
a cofactor. ALA then leaves the mitochondrion and is converted by ALA dehydratase to give
a monopyrrole, porphobilinogen. Four molecules of this are converted by porphobilinogen
deaminase to a linear tetrapyrrole, hydroxymethylbilane. This molecule is then cyclized by
uroporphyrinogen III synthase to uroporphyrinogen III, which is then decarboxylated to
coproporphyrinogen III. This molecule enters the mitochondrion and is oxidized in succession
by coproporphyrinogen III oxidase and protoporphyrinogen III oxidase. The product is
protoporphyrin IX, a substrate for ferrochelatase, which catalyses the insertion of Fe2+ to form
haem. A mitochondrial haem exporter has been identified as feline leukaemia virus subgroup
C receptor 1b (FLVCR1b). The defective steps associated with specific porphyrias and X-​linked
hereditary sideroblastic anaemias are shown.
From Hoffman R, et al. (eds) (2013). Hematology: basic principles and practice, 6th edition. WB Saunders,
Philadelphia, with permission.
1
2
3
4
?
?
?
?
?
?
?
Fig. 22.6.8.4  Pedigree of a family with pyridoxine-​responsive
sideroblastic anaemia showing X-​linked recessive inheritance. Filled
square, affected; filled circle, carrier; ?, unknown status. Diagonal lines
indicate deceased members. The pedigree has been abbreviated to show
only the affected branches of the family. The arrow indicates the proband.
Copyright 1994 Massachusetts Medical Society. All rights reserved. Reproduced from
Cox et al. (1994).
section 22  Haematological disorders
5454
termed ‘refractory anaemia with ring sideroblasts (RARS)’. The World
Health Organization (WHO) classification system now includes 2
myelodysplastic syndromes with ring sideroblasts: (1) refractory an-
emia with ring sideroblasts (RARS), now classified as myelodysplastic
syndromes with RS (MDS-RS); and (2) RARS with thrombocytosis
(RARS-T), now classified as myelodysplastic/myeloproliferative
neoplasm with RS and thrombocytosis (MDS/MPN-RS-T). Clonal
defective haematopoiesis has also been shown in the acute eryth-
roid leukaemias, in which bizarre dysplastic changes are seen in the
developing erythroblasts. These comprise a majority (>50%) of all nu-
cleated marrow cells in erythroleukaemia, a subtype of acute erythroid
leukaemia. The fact that more than 20% of the myeloid cells are blasts
distinguishes acute erythroleukaemia from one of the myelodysplastic
syndromes. A second form of acute erythroid leukaemia, pure eryth-
roid leukaemia, is defined within the WHO classification as a neo-
plastic proliferation of immature erythroid cells comprising 80% or
more of bone marrow cells with no significant myeloblastic element.
Ring sideroblasts are seen in the erythroleukaemia subtype of acute
erythroid leukaemia, but are uncommon in pure erythroid leukaemia.
Aetiology and pathogenesis
The cause of the defective haem synthesis in acquired sideroblastic
anaemia is unclear. Recent reports indicate that levels of ALAS in
bone marrow are normal. Indirect evidence points to an acquired
defect in the mitochondrial respiratory chain that impairs the re-
duction of Fe3+ to Fe2+ since ferrous iron is essential for the terminal
ferrochelatase reaction (Fig. 22.6.8.3). Clonal haematopoiesis has
been demonstrated in this anaemia by both molecular and karyo-
typic analysis. Thus a single glucose-​6-​phosphate dehydrogenase
(G6PD) isoenzyme was found in erythrocytes of a woman het-
erozygous for G6PD who expressed two isoenzymes in her som-
atic tissues. Clonal chromosome changes are also found in bone
marrow cells in many patients with acquired sideroblastic anaemia.
Characteristic changes include monosomy 7, trisomy 8, deletions
involving chromosomes 5, 7, 11, 13, or 20, and a number of trans-
locations. When sideroblastic anaemia is acquired secondary to
chemotherapy or irradiation, chromosomal changes are nearly al-
ways found and tend to be multiple. However, they are probably a
late event in the course of this anaemia and may be preceded by the
expansion of a clone of genetically unstable stem cells. Recurrent
mutations in the SF3B1 gene have been described in 85% of ac-
quired sideroblastic anaemia cases. The product of SF3B1 is associ-
ated with mRNA splicing, and mutations in this gene may influence
a number of mitochondrial gene networks, including changes in
the expression of the iron transporter ABCB7, which results in the
accumulation of iron-​laden mitochondria during erythroid devel-
opment. Patients with MDS/MPN-RS-T have features of MDS-RS
and thrombocytosis; with the bone marrow showing proliferation
of large atypical megakaryocytes. Mutations in SF3B1 and JAK2
(V617F) are commonly found in these patients.
Clinical and laboratory features
Acquired sideroblastic anaemia typically has an insidious onset.
Most patients are middle aged or older, but young adults can be
affected. Mild splenomegaly may be present. White cell and platelet
counts are usually normal; some patients may have thrombocytosis.
The bone marrow shows erythroid hyperplasia with varying degrees
of dyserythropoiesis, including irregular nuclear contour, nuclear
fragmentation (karyorrhexis), bi-​ or trilobed nuclei, and inter-
nuclear bridges. Iron staining of the aspirate shows ring sideroblasts
which should total 15% or more of the nucleated erythroid cells to
make the diagnosis. However, the prognostic significance of this
level of ring sideroblasts has recently been questioned. Dysplasia
of myeloid precursors or megakaryocytes may be present; however,
when 10% or more of nonerythroid cells are dysplastic then cases
are classified as ‘refractory cytopenia with multilineage dysplasia’. If
the overall blast count in the peripheral blood is 2% or greater, or 5%
or greater in the bone marrow, or the peripheral blood monocyte
count exceeds 1.0 × 109/​litre, the condition falls within a different
category of the myelodysplastic syndromes. Thus, ring sideroblasts
may be seen in other myelodysplastic conditions such as refrac-
tory anaemia with excess blasts. Distinguishing acquired idiopathic
sideroblastic anaemia from a mild hereditary sideroblastic anaemia
presenting in adult life can be difficult. However, careful examin-
ation of the marrow for dysplastic changes, the MCV, possible re-
sponse to pyridoxine, and a family survey all help to distinguish
these two entities.
Treatment and prognosis
Transfusions of packed red cells should be given for relief of symp-
tomatic anaemia. A  trial of pyridoxine, 100 to 200 mg/​day for
3 months, is worthwhile but few patients respond to this vitamin.
Myelodysplastic syndromes with ring sideroblasts and refractory an-
aemia have the most favourable outlook among the myelodysplastic
syndromes, with a median survival of 70 to 100 months and a 3
to 12% incidence of progression to acute leukaemia. The Revised
International Prognostic Scoring System (IPSS-​R) uses the bone
marrow blast percentage, karyotypic analysis, and the presence of
peripheral blood cytopenias to group newly diagnosed cases into
one of five prognostic groups (Table 22.6.8.1).
Table 22.6.8.1  Calculation of IPSS-​R score for myelodysplastic syndromes
Variable
0
0.5
1
1.5
2
3
4
Karyotypea
Very good
Good
Intermediate
Poor
Very poor
Bone marrow blasts (%)
≤2
2–​<5
5–​10
10
Haemoglobin (g/​litre)
≥100
80–​<100
<80
Platelet count (× 109/​litre)
≥100
50–​<100
<50
Absolute neutrophil count (× 109/​litre)
≥0.8
<0.8
a Very good: −Y, del(11q). Good: normal, del(5q), del(12p), del(20q), double including del(5q). Intermediate: del(7q), +8, +19, i(17q), other single or double independent clones.
Poor: −7, inv(3)/​t(3q)/​del(3q), double including −7/​del(7q), complex (3 abnormalities). Very poor: complex (>3 abnormalities)
22.6.8  Anaemias resulting from defective maturation of red cells
5455
Patients with very low risk disease, scores of 1.5 or less, have a
median survival of 8.8 years, low-​risk patients with scores of greater
than 1.5 to 3 have a median survival of 5.3 years, intermediate-​risk
patients with scores of more than 3 to 4.5 have a median survival of
3 years, high-​risk patients with scores of greater than 4.5 to 6 have a
median survival of 1.6 years, and very high-​risk patients with a score
of more than 6 have a median survival of only 0.8 years.
A number of agents, including erythropoietin, 5-​azacytidine,
decitabine, and lenalidomide, have been studied in therapeutic
trials for myelodysplastic syndromes, which have included ac-
quired sideroblastic anaemia cases; however, overall outcomes have
been poor. It is unclear if iron chelation therapy with the oral iron-​
chelator deferasirox is of benefit, but this question may be answered
by ongoing clinical trials. The value of cytoreductive therapy in
MDS/MPN-RS-T is uncertain, as anaemia may be worsened; how-
ever, hydroxyurea, lenalidomide, interferon alpha and busulfan have
all been used in the setting of very high platelet counts.
Defective red cell maturation secondary to alcohol
and drugs
Alcohol has a direct toxic effect on erythropoiesis, manifested by
the macrocytosis that characterizes red cells of subjects chronic-
ally ingesting alcohol in excess. Malnourished and anaemic alco-
holics may exhibit ring sideroblasts in the bone marrow as well as
vacuolation of erythroblasts. These manifestations gradually dis-
appear over 4 to 12 days when alcohol is withdrawn, although the
macrocytosis may take several months to normalize. The antibiotic
chloramphenicol when given in dosages greater than 2 g/​day pro-
duces a reversible inhibition of erythropoiesis associated with ring
sideroblasts and vacuolation of erythroblasts. This effect, due to in-
hibition of mitochondrial protein synthesis, is quite separate from
the rare idiosyncratic side effect of aplastic anaemia. Protracted ex-
posure to the antituberculous drug isoniazid has been occasionally
associated with development of a sideroblastic anaemia.
Defective red cell maturation secondary to lead, arsenic,
or zinc ingestion, or copper deficiency
Patients suffering lead poisoning show clinical and laboratory evi-
dence of reduced haem biosynthesis. Basophilic stippling of red cells
is prominent. Mild hypochromic, microcytic anaemia may develop.
Red cell protoporphyrin, increased due to inhibition of the terminal
step in the haem pathway, provides a sensitive measure of lead ex-
posure. The peripheral neuropathy of lead poisoning may be a result
of reduced haem biosynthesis, as in the porphyrias. Acute or chronic
arsenic ingestion can cause anaemia with marked dyserythropoiesis.
Arsenic trioxide (As2O3) is now used to treat patients with acute
promyelocytic leukaemia at doses far less than is required to cause
sideroblastic anaemia. However, sufficiently high levels may be en-
countered in patients with renal failure if the dose of As2O3 has not
been appropriately adjusted. Basophilic stippling of red cells is char-
acteristic while neutropenia and thrombocytopenia may be present.
Copper deficiency has been described only in malnourished pre-
mature infants or in patients receiving long-​term parenteral hyper-
alimentation. This syndrome consists of anaemia and neutropenia
associated with marrow findings of ring sideroblasts and vacuolated
erythroid and myeloid precursors. Large quantities of ingested zinc
interfere with copper absorption and reproduce the sideroblastic an-
aemia and neutropenia characteristic of copper deficiency.
Congenital dyserythropoietic anaemias
This rare group of inherited refractory anaemias, covered in more
detail in Chapter 22.5.1, is characterized by gross multinuclearity of
erythroid precursors in the marrow, ineffective erythropoiesis, and
associated iron overload. Four types have been described based on
morphology of the bone marrow and serological features. The most
common, type II (OMIM 224 100), is also known as HEMPAS (her-
editary erythroblast multinuclearity with positive acidified serum
test) since red cells are lysed by acidified (pH 6.8) serum from about
30% of normal subjects. In congenital dyserythropoietic anaemia
(CDA) type II, a defect in glycosylation of erythroblast membrane
proteins has been identified. Most patients are diagnosed in late child-
hood or adolescence with mild to moderate anaemia, with intermit-
tent jaundice or in older patients with manifestations of iron overload.
Splenomegaly or hepatomegaly may be variably present. CDA carries
a good prognosis with few patients requiring transfusions. The degree
of iron overload should be monitored and treated when appropriate.
FURTHER READING
Aivado M, et al. (2006). X-​linked sideroblastic anemia associated with
a novel ALAS2’ mutation and unfortunate skewed X-​chromosome
inactivation patterns. Blood Cells Mol Dis, 37, 40–​5.
Bergmann, AK, et al. (2010). Systematic molecular genetic analysis
of congenital sideroblastic anemia: evidence for genetic heterogen-
eity and identification of novel mutations. Pediatr Blood Cancer,
54, 273–​8.
Bottomley SS, Fleming MD (2014). Sideroblastic anemia: diagnosis
and management. Hematol Oncol Clin North Am, 28, 653–​70.
Campagna DR, et  al. (2014). X-​linked sideroblastic anemia due to
ALAS2 intron 1 enhancer element GATA-​binding site mutations.
Am J Hematol, 89, 315–​9.
Chakraborty PK, et al. (2014). Mutations in TRNT1 cause congenital
sideroblastic anemia with immunodeficiency, fevers, and develop-
mental delay (SIFD). Blood, 124, 2867.
Cotter PD, et  al. (1995). Late-​onset X-​linked sideroblastic anemia.
Missense mutations in the erythroid δ-​aminolevulinate synthase
(ALAS2) gene in two pyridoxine-​responsive patients initially diag-
nosed with acquired refractory anemia and ringed sideroblasts.
J Clin Invest, 96, 2090–​6.
Cox TC, et al. (1994). X-​linked pyridoxine-​responsive sideroblastic
anemia due to a THR388-​ to -​SER substitution in erythroid 5-​
aminolevulinate synthase. N Engl J Med, 330, 675–​9.
Donker AE, et al. (2014). Practice guidelines for the diagnosis and
management of microcytic anemias due to genetic disorders of iron
metabolism or heme synthesis. Blood, 123, 3873.
Ducamp S, Fleming MD (2019). The molecular genetics of sideroblastic
anemia. Blood, 133(1), 59–69.
Fuller SJ, Wiley JS (2013). Heme biosynthesis and it disorders:
porphyrias and sideroblastic anemias. In: Hoffman R, et al. (eds)
Hematology: basic principles and practice, 6th edition, pp. 466–​72.
Churchill Livingstone/​Elsevier, Philadelphia.
Greenberg P, et al. (2012). Revised international prognostic scoring
system (IPSS-​R) for myelodysplastic syndromes. Blood, 120,
2454–​65.
Guernsey, DL, et al. (2009). Mutations in mitochondrial carrier family
gene SLC25A38 cause nonsyndromic autosomal recessive con-
genital sideroblastic anemia. Nat Genet, 41, 651–​3.

22.6.9 Disorders of the red cell membrane 5456 Pat
22.6.9 Disorders of the red cell membrane 5456 Patrick G. Gallagher
section 22  Haematological disorders
5456
Haas D, et al. (2007). New insights into the prognostic impact of the
karyotype in MDS and correlation with subtypes. Evidence from a
core dataset of 2124 patients. Blood, 110, 4385–​95.
Hasserjian RP, et al. (2008). Refractory anemia with ring sideroblasts.
In: WHO classification of tumours of haemopoietic and lymphoid tis-
sues, 4th edition. pp. 96–​7. IARC Press, Lyon.
Labay V, et  al. (1999). Mutations in SLC19A2 cause thiamine-​
responsive megaloblastic anaemia associated with diabetes mellitus
and deafness. Nat Genet, 22, 300–​4.
Papaemmanuil E, et  al. (2011). Somatic SF3B1 mutation in
myelodysplasia with ring sideroblasts. N Engl J Med, 365, 1384–​95.
Patnaik MM, Tefferi A (2017). Refractory anemia with ring sideroblasts
(RARS) and RARS with thrombocytosis (RARS-T): 2017 update on
diagnosis, risk-stratification, and management. Am J Hematol, 92(3),
297–310.
Patnaik MM, et al. (2012). Prognostic irrelevance of ring sideroblast
percentage in World Health Organization-​defined myelodysplastic
syndromes without excess blasts. Blood, 119, 5674–​7.
Raskin WH, et al. (1984). Evidence for a multistep pathogenesis of a
myelodysplastic syndrome. Blood, 63, 1318–​23.
Renalla R, Wood R (2009). The congenital dyserythropoietic anemias.
Hematol Oncol Clin North Am, 23, 283–​306.
Savage D, Lindenbaum J (1986). Anemia in alcoholics. Medicine, 65,
322–​38.
Szpurka H, et al. (2006). Refractory anemia with ringed sideroblasts
associated with marked thrombocytosis (RARS-​T) another
myeloproliferative condition characterized by JAK2 V617F muta-
tion. Blood, 108, 2173–​81.
Ye H, et al. (2010). Glutaredoxin 5 deficiency causes sideroblastic an-
emia by specifically impairing heme biosynthesis and depleting
cytosolic iron in human erythroblasts. J Clin Invest, 120, 1749–​61.
22.6.9  Disorders of the red
cell membrane
Patrick G. Gallagher
ESSENTIALS
The integrity of the red cell membrane depends on molecular inter-
actions between proteins and the phospholipid membrane: vertical
interactions stabilize the membrane lipid bilayer; horizontal inter-
actions provide resistance against shear stress.
Hereditary spherocytosis
This disorder affects 1 in 2000–5000 individuals of northern European
descent. There is typically a dominant family history, but the con-
dition is genetically heterogeneous: combined spectrin and ankyrin
deficiency is the most common defect observed, followed by band
3 deficiency, isolated spectrin deficiency, and protein 4.2 deficiency.
These affect vertical membrane interactions with loss of surface area
relative to red cell volume.
Clinical features and diagnosis—​the key clinical manifestations are
anaemia and signs of persistent haemolysis, with jaundice and a
marked propensity to gallstones. Diagnostic tests include the incu-
bated osmotic fragility test (in which spherocytes burst at higher sa-
line concentrations than normal); the eosin-​5-​maleimide binding
test in which binding of a fluorescent dye to key red cell membrane
components is assessed by flow cytometry; and next-​generation
targeted sequencing of candidate genes.
Complications and treatment—​parvovirus B19 infection of erythro-
poietic precursors may cause acute aplastic crises. Megaloblastic
anaemia due to folate deficiency occurs in response to increased
requirements during growth and pregnancy, but is preventable with
supplementation. Splenectomy can alleviate the anaemia in many
patients and reduces the risk of gallstones.
Hereditary elliptocytosis
This disorder occurs with a frequency of 1 in 2000 to 1 in 4000
worldwide, and is more frequent in parts of Africa. The inheritance
is usually dominant, with defects in red cell proteins such as α-​ and
β-​spectrin causing disturbances in horizontal interactions in the
erythrocyte membrane.
Clinical features, diagnosis, and treatment—​most patients are
asymptomatic and are typically diagnosed incidentally during testing
for unrelated conditions, but about 10% experience haemolysis,
anaemia, splenomegaly, and intermittent jaundice. Diagnosis is
based on the presence of elliptocytes on a peripheral blood smear.
Treatment is rarely required.
Other conditions
These include (1) hereditary pyropoikilocytosis—​a rare cause of se-
vere haemolytic anaemia, usually seen in patients of African descent;
(2) South-​East Asian (or Melanesian) ovalocytosis—​an asymptom-
atic autosomal dominant condition due to band 3 protein ab-
normalities that confer resistance to invasion by malaria parasites;
(3) stomatocytosis—​characterized by red cells with a characteristic-
ally shaped slit-​like area of central pallor—​a heterogeneous group
of disorders that are often asymptomatic but may cause haemolysis
and anaemia, and which may be hereditary (e.g. missense muta-
tions in band 3) or acquired (e.g. cholestatic liver disease, alcoholism,
and vinca alkaloids); and (4) acanthocytosis—​characterized by con-
tracted red cells with spiky projections, again with both hereditary
(e.g. neuroacanthocytosis syndromes, abetalipoproteinaemia) and
acquired (e.g. severe hepatic disease) aetiologies.
The red cell membrane
Composition and function
Although the primary structure and a number of the important func-
tions of the red cell membrane have been known for many years, their
study continues to yield important insights into our understanding
of membrane structure and function. The red cell membrane is com-
posed of three major structural elements: a lipid bilayer primarily
comprising phospholipids and cholesterol; integral proteins em-
bedded in the lipid bilayer that span the membrane; and a membrane
skeleton on the internal side of the red cell membrane.
The membrane and its skeleton provide the erythrocyte with the
ability to undergo significant deformation without fragmentation or
loss of integrity during its travel through the microcirculation. The
membrane also assembles and organizes the proteins of the lipid
22.6.9  Disorders of the red cell membrane
5457
bilayer and the membrane skeleton, allowing the red cell to partici-
pate in a wide range of functions. These include influencing cellular
metabolism by selectively and reversibly binding and inactivating
glycolytic enzymes, retaining organic phosphates and other vital com-
pounds, removing metabolic waste, and sequestering the reductants
required to prevent corrosion by oxygen. During erythropoiesis, the
membrane responds to erythropoietin and imports the iron required
for the synthesis of haemoglobin. The lipid bilayer provides an imper-
meable barrier between the cytoplasm and the external environment
and helps avoid red cell adherence to endothelial cells or aggregation in
the microcirculation. The membrane also participates in erythrocyte
biogenesis and ageing. Finally, it is involved in the maintenance of pH
homeostasis by participating in chloride–​bicarbonate exchange.
Interactions of membrane proteins and disorders
of red cell shape
Membrane protein–​protein and protein–​lipid interactions have been
classified into two categories, vertical and horizontal interactions
(Fig. 22.6.9.1). Vertical interactions stabilize the lipid bilayer mem-
brane while horizontal interactions support the structural integrity
of erythrocytes after their exposure to shear stress. The interactions
between proteins and lipids of the erythrocyte membrane are more
complex than this simplistic model, but it serves as a useful starting
point for understanding red cell membrane interactions, particularly
in membrane-​related disorders. According to this model, hereditary
spherocytosis (HS) is a disorder of vertical interactions. Although the
primary molecular defects in HS are heterogeneous (see ‘Hereditary
spherocytosis’), one common feature of HS erythrocytes is a weak-
ening of the vertical contacts between the skeleton and the lipid bi-
layer. As a result, the lipid bilayer membrane is destabilized, leading
to release of lipids in the form of skeleton-​free lipid vesicles, which in
turn results in membrane surface area deficiency and spherocytosis.
By contrast, in this model, hereditary elliptocytosis is a defect of hori-
zontal interactions, primarily those involving spectrin dimer self-​
association. Defects of horizontal interactions disrupt the membrane
skeletal lattice leading to elliptocytic shape in mild cases and skeletal
instability and cell fragmentation in severe cases.
Hereditary spherocytosis
This group of inherited disorders is characterized by the presence of
spheroidal erythrocytes on a peripheral blood smear. HS occurs in all
racial and ethnic groups. It is the most common inherited anaemia
in individuals of northern European descent, affecting approximately
1 in 2500 individuals in the United States of America and the United
Kingdom. It is much more common in Caucasians than in individuals
of African ethnicity. Clinical, laboratory, biochemical, and genetic
heterogeneity characterize the spherocytosis syndromes.
Aetiology and pathogenesis
The primary defect in HS is loss of membrane surface area relative
to intracellular volume, accounting for the spheroidal shape and de-
creased deformability of the red cell. This loss of surface area results
from increased membrane fragility due to defects in erythrocyte
membrane proteins. Increased fragility leads to membrane vesicula-
tion and membrane loss. Splenic destruction of these nondeformable
erythrocytes is the primary cause of haemolysis experienced by HS
patients. Physical entrapment of erythrocytes in the splenic micro-
circulation and ingestion by phagocytes have been proposed as
mechanisms of destruction. Furthermore, the splenic environment
is hostile to erythrocytes. Low pH, glucose, and ATP concentrations,
and high local concentrations of toxic free radicals produced by ad-
jacent phagocytes, all contribute to membrane damage.
Membrane loss is due to defects in several membrane proteins,
including ankyrin, band 3, α-​spectrin, β-​spectrin, and protein 4.2.
Combined spectrin and ankyrin deficiency is the most common de-
fect observed, followed by band 3 deficiency, isolated spectrin defi-
ciency, and protein 4.2 deficiency. As might be expected, therefore,
the genetic defects underlying HS are heterogeneous. Multiple gen-
etic loci are implicated and various abnormalities, including point
mutations, defects in mRNA processing, and gene deletions, have
all been described. Except for a few rare exceptions, HS mutations
are private, that is, each individual kindred has a unique mutation.
Clinical features
The clinical manifestations of the spherocytosis syndromes vary
widely. The typical picture of HS combines evidence of haemolysis
(anaemia, jaundice, reticulocytosis, gallstones, and splenomegaly)
with spherocytosis (spherocytes on a peripheral blood smear) and
a positive family history (Box 22.6.9.1). Mild, moderate, and severe
forms of HS have been defined according to differences in haemo-
globin, bilirubin, and reticulocyte counts correlated with the degree
of compensation for the haemolysis (Table 22.6.9.1). Initial assess-
ment of a patient with suspected HS should include a family history
and questions about history of anaemia, jaundice, gallstones, and
splenectomy. Physical examination should seek signs such as scleral
icterus, jaundice, and splenomegaly. After diagnosing a patient with
HS, family members should be examined for the presence of HS.
HS typically presents in childhood, but may present at any age.
In children, anaemia is the most frequent presenting complaint
(50%), followed by splenomegaly, jaundice, or a positive family his-
tory. Two-​thirds to three-​quarters of HS patients have incompletely
compensated haemolysis and mild to moderate anaemia. The an-
aemia is often asymptomatic except for fatigue and mild pallor.
Jaundice is seen at some time in about 50% of patients, usually in
association with viral infections. When present, it is acholuric, that
Vertical interaction
Horizontal interaction
Band 3
4.2
Ankyrin
Glycophorin A
Adducin
Actin
p55
Tropomyosin
Tropomodulin
Glycophorin C
4.9
Band 3
Band 3 Band 3
α Spectrin
β Spectrin
Fig. 22.6.9.1  Schematic diagram of the red cell membrane (not to scale).
Membrane–​protein and membrane–​lipid interactions can be divided into
two categories: (1) vertical interactions, which are perpendicular to the
plane of the membrane and involve spectrin–​ankyrin–​band 3 interactions,
spectrin–​protein 4.1–​glycophorin C interactions, and weak interactions
between spectrin and the lipid bilayer; and (2) horizontal interactions,
which are parallel to the plane of the membrane.
From Tse WT, Lux SE (1999). Red blood cell membrane disorders. Br J Haematol, 104,
2–​13, with permission.
section 22  Haematological disorders
5458
is, there is unconjugated hyperbilirubinaemia without detectable
bilirubinuria. Palpable splenomegaly is detectable in most (75–​95%)
older children and adults. Typically, the spleen is modestly enlarged
but it may be massive.
About 20 to 30% of HS patients have ‘compensated haemolysis,’
that is, erythrocyte production and destruction are balanced.
Although the erythrocyte lifespan may only be about 20–​30 days,
these patients adequately compensate for their haemolysis with in-
creased marrow erythropoiesis. They are not anaemic and are usu-
ally asymptomatic. Many of these individuals escape detection until
adulthood, when they are being evaluated for unrelated disorders or
when complications related to anaemia or chronic haemolysis occur.
Haemolysis may become severe with illnesses that cause spleno-
megaly, such as infectious mononucleosis, or may be exacerbated
by other factors such as pregnancy. Because of the asymptomatic
course of HS in these patients, diagnosis of HS should be considered
during evaluation of splenomegaly, gallstones at a young age, or an-
aemia from viral infection.
Approximately 5 to 10% of HS patients have moderate to severe
anaemia. Patients with moderately severe disease typically have a
haemoglobin of 60 to 80 g/​litre, reticulocytes about 10%, bilirubin
2 to 3 mg/​dl, and 40 to 80% of the normal red cell spectrin content.
This category includes patients with both dominant and recessive
HS and a variety of molecular defects. Patients with severe disease,
by definition, have life-​threatening anaemia and are transfusion
dependent. They almost always have recessive HS. Most have iso-
lated severe spectrin deficiency. In addition to the risks of recurrent
transfusions, these patients often suffer from haemolytic and aplastic
crises and may develop complications of severe uncompensated an-
aemia including growth retardation, delayed sexual maturation, or
aspects of thalassaemic facies.
The parents of patients with recessive HS are clinically asymptom-
atic and do not have anaemia, splenomegaly, hyperbilirubinaemia,
or spherocytosis on peripheral blood smears (‘Trait’, Table 22.6.9.1).
Most have subtle laboratory signs of HS including a slight reticulo­
cytosis. It has been estimated that at least 1.4% of the population are
silent carriers.
Inheritance
The genes responsible for HS include ankyrin, β-​spectrin, band 3
protein, α-​spectrin, and protein 4.2. In approximately two-​thirds
to three-​quarters of HS patients, inheritance is autosomal dom-
inant. In the remaining patients, inheritance is nondominant
due to autosomal recessive inheritance or a de novo mutation.
Cases with autosomal recessive inheritance are due to defects in
either α-​spectrin or protein 4.2. A surprising number of de novo
mutations have been reported in the HS genes. A  few cases of
‘double-​dominant’ HS due to defects in band 3 or spectrin that
result in fetal death or severe haemolytic anaemia presenting in
the neonatal period have been reported. In general, affected in-
dividuals of the same kindred experience similar degrees of
haemolysis.
Complications
Gallbladder disease
Chronic haemolysis leads to the formation of bilirubinate gall-
stones, the most frequently reported complication in HS patients.
Although gallstones have been detected in infancy, most occur be-
tween 10 and 30 years of age. Management should include interval
ultrasonography to detect gallstones, as many patients with chole-
lithiasis and HS are asymptomatic. Timely diagnosis and treatment
will help prevent complications of symptomatic biliary tract dis-
ease including biliary obstruction, cholecystitis, and cholangitis.
Box 22.6.9.1  Characteristics of hereditary spherocytosis
Clinical manifestations
•	 Anaemia
•	 Splenomegaly
•	 Intermittent jaundice:
—	 From haemolysis
—	 From biliary obstruction
•	 Haemolytic, aplastic, and megaloblastic crises
•	 Inheritance:
—	 Dominant (c.75%)
—	 Nondominant (c.25% de novo or recessive)
•	 Rare manifestations:
—	 Leg ulcers, gout, chronic dermatitis
—	 Extramedullary haematopoietic tumours
—	 Thrombosis
—	 Neuromuscular disorders
—	 Cardiomyopathy
—	 Spinocerebellar abnormalities
•	 Excellent response to splenectomy
Laboratory characteristics
•	 Reticulocytosis
•	 Spherocytosis
•	 Elevated MCHC
•	 Reduced eosin-​5-​maleimide binding
•	 Normal direct antiglobulin test
Table 22.6.9.1  Clinical classification of hereditary spherocytosis
Trait
Mild spherocytosis
Moderate spherocytosisa
Severe spherocytosis
Haemoglobin (g/​litre)
Normal
110–​150
80–​120
≤80
Reticulocytes (%)
1–​3
3–​8
≥8
≥10
Bilirubin (mg/​dl)
0–​1
1–​2
2
3
Spectrin contentb (% of normal)
100
80–​100
50–​80
20–​80
Peripheral smear
Normal
Mild spherocytosis
Spherocytosis
Spherocytosis and poikilocytosis
a Values in untransfused patients.
b In most patients, ankyrin content is decreased to a comparable degree. A minority of hereditary spherocytosis patients lack band 3 or protein 4.2 and may have mild to moderate
spherocytosis with normal amounts of spectrin and ankyrin.
22.6.9  Disorders of the red cell membrane
5459
Haemolytic, aplastic, and megaloblastic crises
Haemolytic crises are usually associated with viral illnesses and typ-
ically occur in childhood. They are generally mild and are charac-
terized by jaundice, increased splenomegaly, decreased haematocrit,
and reticulocytosis. Intervention is rarely necessary. When severe
haemolytic crises occur, there is marked jaundice, anaemia, lethargy,
abdominal pain, and tender splenomegaly. Hospitalization and red
cell transfusion may be required.
Aplastic crises following virally induced bone marrow suppres-
sion are uncommon, but may result in severe anaemia with serious
complications including congestive heart failure. The most common
aetiological agent in these cases is parvovirus (erythrovirus) B19.
Parvovirus selectively infects erythropoietic progenitor cells
and inhibits their maturation. In addition to causing a profound
reticulocytopenia, parvovirus infections may also be associated
with mild neutropenia, thrombocytopenia, or even pancytopenia.
During the aplastic phase, the haemoglobin and the production of
new red cells fall, the cells that remain age, and microspherocytosis
increases. Aplastic crises usually last 10 to 14 days (about half the
lifespan of HS red cells), and the haemoglobin typically falls to half
its usual level before recovery occurs. In patients with severe HS,
the anaemia may be profound, requiring hospitalization and trans-
fusion. As the marrow recovers, granulocytes, platelets, and, finally,
reticulocytes return to the peripheral blood. Aplastic crisis brings
many patients to medical attention, particularly asymptomatic HS
patients with normally compensated haemolysis. Because parvo-
virus may infect several members of a family simultaneously, leading
to aplastic crises, there have been reports of ‘outbreaks’ of HS.
Megaloblastic crisis occurs in HS patients with increased folate
demands, for example, the pregnant patient, growing children, or
patients recovering from an aplastic crisis. With appropriate folate
supplementation, this complication is preventable.
Diagnosis
The laboratory findings in HS are heterogeneous. Initial laboratory
investigation should include a complete blood count with periph-
eral smear, reticulocyte count, Coombs’ test, and serum bilirubin.
Although specialized diagnostic tests may not be required in the con-
text of a proven family history of HS, index cases and those in whom
there is diagnostic uncertainty should undergo additional testing. The
most widely used current tests are the eosin-​5-​maleimide binding test
(described in more detail later) and the incubated osmotic fragility test.
Peripheral blood smear
Erythrocyte morphology is quite variable. Typical HS patients
have blood smears with obvious spherocytes lacking central pallor
(Fig. 22.6.9.2a). Less commonly, patients present with only a few
spherocytes on peripheral smear or, at the other end of the spectrum,
with numerous small, dense spherocytes and bizarre erythrocyte
morphology with anisocytosis and poikilocytosis (Fig. 22.6.9.2b).
(b)
(a)
(c)
(d)
Fig. 22.6.9.2  Peripheral blood smears: (a) typical hereditary spherocytosis; (b) severe, recessively inherited spherocytosis; (c) hereditary elliptocytosis;
(d) hereditary pyropoikilocytosis.
section 22  Haematological disorders
5460
Specific morphological findings have been identified in patients
with certain membrane protein defects such as pincered erythro-
cytes (band 3) or spherocytic acanthocytes (β-​spectrin).
Erythrocyte indices
Most patients have mild to moderate anaemia. The mean cell volume
(MCV) is normal except in severe HS cases, when it is slightly de-
creased despite reticulocytosis, reflecting membrane loss and
cellular dehydration. The mean cell haemoglobin concentration
(MCHC) is increased (≥350 g/​litre) due to relative cellular dehydra-
tion in around 50% of patients. Strategies using erythrocyte indices
have combined MCHC and red cell distribution width (>354 g/​litre
and >14, respectively) or utilized histograms of hyperdense erythro-
cytes (MCHC >400 g/​litre) obtained from laser-​based cell counters,
sometimes combined with elevated MCHC, in attempts to rapidly
identify HS patients.
Osmotic fragility
In the normal erythrocyte, membrane redundancy gives the cell its
characteristic discoid shape and provides it with abundant surface
area. In spherocytes, there is a decrease in surface area relative to cell
volume, resulting in their abnormal shape. This change is reflected in
the increased osmotic fragility found in these cells (Fig. 22.6.9.3), and
this feature has been exploited in diagnostic testing for HS. Osmotic
fragility is tested by adding increasingly hypotonic concentrations
of saline to red cells. The normal erythrocyte is able to increase its
volume by swelling, but spherocytes, which are already at maximum
volume for surface area, burst at higher saline concentrations than
normal. Approximately 25% of HS individuals will have a normal os-
motic fragility on freshly drawn red cells, with the osmotic fragility
curve approximating the number of spherocytes seen on periph-
eral smear. However, after incubation at 37°C for 24 h, HS red cells
lose membrane surface area more readily than normal because their
membranes are leaky and unstable. Thus incubation accentuates the
defect in HS erythrocytes and brings out the defect in osmotic fra-
gility, making incubated osmotic fragility the standard test for diag-
nosing HS. When the spleen is present, a subpopulation of very fragile
erythrocytes, which have been conditioned by the spleen, form the
‘tail’ of the osmotic fragility curve; this disappears after splenectomy
(Fig. 22.6.9.3).
Osmotic fragility testing suffers from poor sensitivity as about
20% of mild cases of HS are missed after incubation. It is also unreli-
able in patients with small numbers of spherocytes, including those
who have been recently transfused, and will give a positive result
for patients who have spherocytosis for reasons other than HS. This,
along with the labour-​intensive nature of the test, means that it is
now been largely outmoded in the diagnosis of HS.
Eosin-​5-​maleimide binding test
Flow cytometry can be used to examine the degree of binding of the
fluorescent dye eosin-​5-​maleimide to the red cell surface as a poten-
tial diagnostic test of HS. There is an interaction between eosin-​5-​
maleimide and the band 3 protein, with individuals affected by HS
having a lower binding of the dye to the red cell surface. A binding
ratio of 0.8 is typically used as a cut-​off to distinguish normal controls
for HS patients, with sensitivity and specificity of greater than 90%.
Positive EMA binding test results can also be seen in some cases of
congenital dyserythropoietic anaemia, and abnormalities of erythro-
cyte hydration and in some HS cases with spectrin deficiency, EMA
binding may be normal—​but to date this rapid and readily available
test remains the best laboratory method of screening for HS.
Molecular genetics and targeted next-​generation resequencing
panels may be of use in the diagnosis of HS, though the large number
of private mutations makes this strategy challenging.
Other laboratory manifestations in HS are the markers of ongoing
haemolysis. Reticulocytosis, increased bilirubin, increased lactate
dehydrogenase, increased urinary and faecal urobilinogen, and de-
creased haptoglobin reflect increased erythrocyte production or
destruction.
Differential diagnosis
HS should be able to be distinguished from other haemolytic
anaemias by additional diagnostic testing—​with as autoimmune
haemolytic anaemia, distinguishable by a positive a Coombs’ test,
being a key morphological differential. Other causes of haemo-
lytic anaemia with spherocytes on peripheral smear (Box 22.6.9.2)
should be considered in the appropriate clinical context. Occasional
spherocytes are also seen in patients with a large spleen (such as in
cirrhosis and myelofibrosis) or in patients with microangiopathic
anaemias, but the differentiation of these conditions from HS is not
usually difficult.
Lysis (%)
100
Saline concentration (%)
0.5
0.4
0.3
0.6
0.7
0.8
Severe HS
Typical HS
Normal
range
Tail
80
60
40
20
Fig. 22.6.9.3  Osmotic fragility curves in hereditary spherocytosis. The
shaded region is the normal range. Results representative of both typical
and severe spherocytosis are shown. A tail, representing very fragile
erythrocytes that have been conditioned by the spleen, is common in
many spherocytosis patients prior to splenectomy.
Box 22.6.9.2  Conditions with spherocytes on peripheral
blood smear
•	 Hereditary spherocytosis
•	 Autoimmune haemolytic anaemia
•	 Liver disease
•	 Thermal injury
•	 Microangiopathic and macroangiopathic haemolytic anaemias
•	 Transfusion reaction with haemolysis
•	 Clostridial sepsis
•	 Severe hypophosphataemia
•	 Poisoning from certain snake, spider, bee, and wasp venoms
•	 Heinz body anaemias
•	 Hypersplenism
•	 ABO incompatibility (neonates)
22.6.9  Disorders of the red cell membrane
5461
Treatment
Splenectomy
Splenic sequestration is the primary determinant of erythro-
cyte survival in HS patients. Thus splenectomy alleviates the an-
aemia in the majority of patients, reducing or eliminating the
need for transfusions and decreasing the incidence of cholelith-
iasis. Postsplenectomy, spherocytosis and altered osmotic fra-
gility persist, but erythrocyte lifespan nearly normalizes, and
reticulocyte counts fall to normal or near normal levels. Typical
postsplenectomy changes, including Howell–​Jolly bodies, target
cells, and acanthocytes, become evident on peripheral smear.
Postsplenectomy, patients with the most severe forms of HS still
suffer from shortened erythrocyte survival and haemolysis, but
their clinical improvement is striking.
Early complications of splenectomy include local infection,
bleeding, and pancreatitis due to injury to the tail of the pancreas
incurred during surgery. Overwhelming postsplenectomy infection,
typically from encapsulated organisms, is an uncommon but sig-
nificant late complication of splenectomy, especially in the first few
years of life; pneumococcal vaccination preoperatively is needed,
and antibiotic prophylaxis may be recommended, especially for chil-
dren. For all splenectomized patients, early antibiotic therapy in the
context of possible infection is advised. Another postsplenectomy
complication is the increased risk of cardiovascular disease, particu-
larly thrombosis and pulmonary hypertension.
Indications for splenectomy
In the past, splenectomy was considered routine in HS patients.
However, the risk of overwhelming postsplenectomy infec-
tion with penicillin-​resistant pneumococci, and other potential
complications have led to a re-​evaluation of the role of splen-
ectomy in the treatment of HS. The risks and benefits of splen-
ectomy should be reviewed and discussed between healthcare
providers, patient, and family when splenectomy is considered.
Considering the risks and benefits, a reasonable approach would
be to splenectomize all patients with severe spherocytosis and
all patients who suffer from significant signs or symptoms of
anaemia including growth failure, skeletal changes, leg ulcers,
and extramedullary haematopoietic tumours. Other candidates
for splenectomy are older HS patients who suffer vascular com-
promise of vital organs.
Whether patients with moderate HS should undergo splenectomy
remains controversial. Patients with mild HS and compensated
haemolysis can be followed and referred for splenectomy if clinically
indicated.
When splenectomy is warranted, laparoscopic splenectomy is
the method of choice as it results in less postoperative discomfort,
shorter hospitalization, and decreased costs. Partial splenectomy
has been advocated for infants and young children with significant
anaemia associated with HS and it may be of benefit in typical HS
patients. The goals of this procedure are to allow for the palliation of
haemolysis and anaemia while maintaining some residual splenic
immune function. Long-​term follow-​up data for this procedure are
lacking.
Before splenectomy, preferably several weeks preoperatively,
patients should be immunized with vaccines against pneumo-
coccus, Haemophilus influenzae type b, and meningococcus. The
use and duration of prophylactic antibiotics postsplenectomy
is controversial. Prior to splenectomy, and in severe cases,
postsplenectomy, HS patients should take folic acid to prevent
folate deficiency.
Elliptocytosis, pyropoikilocytosis, and
related disorders
Hereditary elliptocytosis (HE) is characterized by the presence of
elliptical or cigar-​shaped erythrocytes on peripheral blood smears
of affected individuals. The worldwide incidence of HE has been
estimated to be 1 in 2000 to 1 in 4000 individuals. The true inci-
dence of HE is unknown because most patients are asymptomatic.
It is common in individuals of African and Mediterranean ancestry,
presumably because elliptocytes confer some resistance to malaria.
In parts of Africa, the incidence of HE approaches 1 in 100. HE is
typically inherited in an autosomal dominant pattern. Rare cases of
de novo mutations have been described.
Hereditary pyropoikilocytosis (HPP) is a rare cause of severe
haemolytic anaemia with erythrocyte morphology reminiscent of
that seen in severe burns. Initial studies of erythrocytes from these
patients revealed abnormal thermal sensitivity compared to normal
erythrocytes. HPP occurs predominantly in patients of African
descent. There is a strong relationship between HPP and HE.
Approximately one-​third of parents or siblings of patients with HPP
have typical HE. Many patients with HPP experience severe haem-
olysis and anaemia in infancy that gradually improves, evolving to-
ward typical HE later in life.
Aetiology and pathogenesis
The principal defect in HE/​HPP erythrocytes is an intrinsic mech-
anical weakness or fragility of the erythrocyte membrane skeleton
due to a defect of horizontal interactions (discussed previously).
This is due to defects in the red cell membrane proteins α-​spectrin,
β-​spectrin, protein 4.1, band 3, or glycophorin C. The majority of de-
fects occur in spectrin, the principal structural protein of the mem-
brane skeleton. A variety of mutations in the genes encoding these
proteins have been described, with several mutations identified in a
number of individuals on the same genetic background, suggesting
a ‘founder effect’ for these mutations.
Clinical features
The clinical presentation of HE is heterogeneous, ranging from
asymptomatic carriers to patients with severe, transfusion-​
dependent anaemia. Most patients with HE are asymptomatic and
are typically diagnosed incidentally during testing for unrelated
conditions. The erythrocyte lifespan is normal in most patients. The
10% of patients with decreased red cell lifespan are the ones who
experience haemolysis, anaemia, splenomegaly, and intermittent
jaundice. Many of these symptomatic patients have parents with
typical HE and thus are homozygotes or compound heterozygotes
for defects inherited from each of the parents. Symptomatology
may vary between members of the same family; indeed, it may
vary in the same individual at different times. To explain these ob-
servations, modifier alleles have been hypothesized to influence
spectrin expression and clinical severity. One such allele, αLELY
(low-​expression Lyon), has been identified and characterized.
section 22  Haematological disorders
5462
Diagnosis
The hallmark of HE is the presence of elliptocytes on peripheral
blood smears (Fig. 22.6.9.2c). These normochromic, normocytic
elliptocytes number from a few to 100%. The degree of haem-
olysis and anaemia do not correlate with the number of elliptocytes
present. A  few ovalocytes, spherocytes, stomatocytes, and frag-
mented cells may also be seen. Elliptocytes may be seen in asso-
ciation with several disorders including megaloblastic anaemias,
hypochromic microcytic anaemias (iron deficiency anaemia and
thalassaemia), myleodysplasic syndromes, and myelofibrosis; how-
ever, elliptocytes generally make up less than one-​third of red cells
in these conditions. History and additional laboratory testing usu-
ally clarify the diagnosis of these disorders. In addition to the per-
ipheral blood smear findings found in HE, HPP erythrocytes are
bizarre-​shaped with fragmentation and budding (Fig. 22.6.9.2d).
Microspherocytosis is common and the MCV is frequently de-
creased (50–​65 mm3).
The osmotic fragility is increased in severe HE and HPP. Other
laboratory findings in HE are similar to those found in other haemo-
lytic anaemias and are nonspecific markers of increased erythrocyte
production and destruction. When indicated, specialized testing,
such as membrane protein quantitation and genetic studies can be
performed.
Treatment
Therapy is rarely necessary. In rare cases, occasional red blood
cell transfusions may be required. In cases of severe HE and HPP,
splenectomy has been useful. The same indications for splenec-
tomy in HS can be applied to patients with symptomatic HE or HPP.
Postsplenectomy, patients with HE or HPP experience increased
haemoglobin, decreased haemolysis, and improvement in clinical
symptoms.
During acute illnesses, patients should be followed for signs of
haematological decompensation. Ultrasonography at regular inter-
vals to detect gallstones should be performed. In patients with sig-
nificant haemolysis, folate should be administered daily.
South-​East Asian ovalocytosis
South-​East Asian ovalocytosis (SAO) is characterized by the pres-
ence of oval erythrocytes with a central longitudinal slit or trans-
verse bar on peripheral blood smears of affected individuals. It is
common in parts of the Philippines, Indonesia, Malaysia, and
New Guinea and is inherited in an autosomal dominant fashion.
Incredibly rigid, SAO erythrocytes are resistant to invasion by mal-
aria parasites. The underlying defect is a mutation in a critical region
of band 3. Haematologically, patients with SAO are asymptomatic,
with little or no evidence of haemolysis or anaemia. The finding of
characteristic ovalocytes in the peripheral blood of an asymptomatic
individual from one of the earlier-​mentioned ethnic backgrounds is
highly suggestive of the diagnosis. Biochemical and DNA diagnostic
techniques are available to detect this condition.
Stomatocytosis
The hereditary stomatocytosis syndromes are a heteroge-
neous group of rare disorders characterized by mouth-​shaped
(stomatocytic) erythrocyte morphology on peripheral blood
smear (Fig. 22.6.9.4). The clinical severity of stomatocytosis pa-
tients is variable; some patients experience haemolysis and an-
aemia, while others are asymptomatic. An unusual feature of the
stomatocytosis syndromes is a dramatically increased predispos-
ition to thrombosis or pulmonary hypertension postsplenectomy.
Fortunately, anaemia is well compensated in most patients and
splenectomy is not required.
The red blood cell membranes of stomatocytosis patients usu-
ally exhibit abnormal permeability to the cations sodium and
potassium, with consequent modification of intracellular water
content, ranging from dehydrated (xerocytosis) to overhydrated
(hydrocytosis) erythrocytes. The variable clinical, laboratory, and
pathophysiological findings associated with the stomatocytosis
syndromes suggest these are a complex collection of syndromes
caused by various molecular defects.
Dehydrated
stomatocytosis
(also
termed
stomatocytic
xerocytosis) has been shown to be associated with mutations
in the Gardos and Piezo channels in the red cell membrane.
Overhydrated stomatocytosis, by contrast, is caused by defects in
the Rh-​antigen associated glycoprotein (encoded by the RHAG
(a)
(b)
Fig. 22.6.9.4  Peripheral blood smears: (a) dehydrated stomatocytosis;
(b) overhydrated stomatocytosis.
From Lande WM, Mentzer WC (1985). Haemolytic anaemia associated with
increased cation permeability. Clin Haematol, 14, 89–​103, with permission.

22.7 Haemostasis 5490 22.7.1 The biology of haemos
22.7 Haemostasis 5490 22.7.1 The biology of haemostasis and thrombosis 5490 Gilbert C. White, II, Harold R. Roberts, and Nigel S. Key
CONTENTS
22.7.1	 The biology of haemostasis and thrombosis  5490
Gilbert C. White, II, Harold R. Roberts, and Nigel S. Key
22.7.2	 Evaluation of the patient with a bleeding tendency  5509
Trevor Baglin
22.7.3	 Thrombocytopenia and disorders of platelet
function  5520
Nicola Curry and Susie Shapiro
22.7.4	 Genetic disorders of coagulation  5532
Eleanor S. Pollak and Katherine A. High
22.7.5	 Acquired coagulation disorders  5546
T.E. Warkentin
22.7.1  The biology of haemostasis
and thrombosis
Gilbert C. White, II, Harold R. Roberts,
and Nigel S. Key
ESSENTIALS
Haemostasis—​a component of the wound defence mechanism—​is a
process by which vessel wall components and platelets act in concert
with procoagulant and anticoagulant proteins to form a plug of cells
and cross-​linked fibrin. The plug is later remodelled and replaced by
new tissue as part of wound healing. These processes are very com-
plex and involve highly controlled pathways of interaction between
cells, glycans, and membrane-​bound and soluble proteins of coagu-
lation and fibrinolysis, as well as their cognate inhibitors.
Thrombosis—​this is an abnormal state leading to formation
of a clot obstructing blood vessel flow; dislodgement leads to
thromboembolism.
Blood vessel wall
Vascular endothelial cells—​these make many contributions to
haemostasis by (1)  regulating vascular tone—​through production
of (a) vasodilators, most notably nitric oxide and prostacyclin (PGI2),
and (b) vasoconstrictors, particularly endothelin and angiotensin 2;
(2) exerting anticoagulant effects—​through production of PGI2, nitric
oxide, thrombomodulin, tissue factor (TF) pathway inhibitor, glycosa-
minoglycans, CD39, and tissue plasminogen activator (tPA); (3) pro-
moting procoagulant effects—​the dominant effect of endothelial
cells is anticoagulant, but they store/​produce von Willebrand factor
and TF; and (4)  up-​regulating expression of receptors—​including
thrombin receptors, thrombomodulin, and endothelial cell protein
C receptor, and a number of adhesive receptors that are important
for the interaction of leucocytes and the vessel wall.
Other elements—​these include (1) extracellular matrix—​promotes
platelet adhesion, cellular migration, cell proliferation, and endothe-
lial and smooth muscle cell interactions; (2) smooth muscle cells; and
(3) adventitia.
Platelets
Platelets are key components of the haemostatic plug. They adhere
to damaged vessels where subendothelial matrix is exposed, aggre-
gating to form an initial plug that prevents blood loss by occluding
the breach in the vessel wall. Their involvement in haemostasis
can broadly be divided into the following processes:  (1) platelet
adhesion—​accomplished by a number of glycoproteins and other
adhesion receptors on the platelet surface; (2) platelet activation—​
following adhesion and in response to soluble agonists, platelets
undergo reactions (including changes in metabolism of membrane
inositol phospholipids) that lead to generation of platelet coagulant
activity, thrombin, and release of ADP, which lead to activation of add-
itional platelets; and (3) platelet aggregation—​mediated by binding
of activated platelet surface glycoprotein αIIb–​β3 to fibrinogen or
fibrin, which by virtue of its dimeric structure can bind to more than
one platelet and thereby facilitate their aggregation, which serves to
localize the haemostatic plug at the site of injury.
Blood coagulation
Blood coagulation depends on the presence of serial proenzymes
that are sequentially activated in the presence of activators and cofac-
tors, with key elements being (1) TF—​this is constitutively produced
in several extravascular cell types such as fibroblasts and smooth
muscle cells, but not in cells exposed to the circulating blood; it
functions as a receptor for factor VII and initiates the blood coagu-
lation pathway after it binds to and activates factor VII; (2) TF–​VIIa
complex—​this activates factors IX and X which, in the presence of
their respective cofactors (VIII and V), rapidly convert prothrombin
22.7
Haemostasis
22.7.1  The biology of haemostasis and thrombosis
5491
(factor II) to thrombin; (3) thrombin converts soluble fibrinogen to
fibrin; and (4) fibrin undergoes cross-​linking by activated factor XIII
to form the stable haemostatic plug.
Important aspects of the system include (1) platelets which pro-
vide the surface for activated clotting factors, leading to the explosive
generation of thrombin and subsequent clot formation; and (2) the
initial generation of relatively small amounts of thrombin is essen-
tial for feedback activation of factors V, VIII, XI, and XIII, as well as of
platelets.
Inhibitors of the coagulation reactions—​there are numerous in-
hibitors of the reactions involved in blood coagulation, which
are essential for the temporal control and safety of the process.
These include (1)  TF pathway inhibitor—​occurs in forms free
within the circulation and anchored to platelets and endothelial
cells; inhibits the VIIa–​TF–​Xa complex; (2) antithrombin—​a serpin
inhibitor of thrombin, factor X, and other proteases; (3)  other
inhibitors—​these include α1-​antitrypsin, C-​1 esterase inhibitor, and
protein Z-​dependent protease inhibitor.
The fibrinolytic system
The fibrinolytic system depends on the activation of plasminogen
adsorbed on the fibrin surface by tPA to form plasmin, which de-
grades fibrin to form specific fibrin degradation products, and when
generated in excess also degrades fibrinogen, factors VIII and V, and
von Willebrand factor.
Important aspects of the system include (1) free plasmin in the
circulation is rapidly inhibited by α2-​antiplasmin; (2) plasminogen
and tPA associate in the circulation with fibrinogen, hence when fi-
brinogen is converted to fibrin, the clot is rich in both of these pro-
teins, which are protected from the inhibitory action of antiplasmin,
hence clots can be lysed without interference from inhibitors;
(3) many other regulatory mechanisms exist, including plasminogen
activator inhibitor 1, urokinase plasminogen activator, and thrombin-​
activatable fibrinolytic inhibitor.
The balance of fibrinolysis and coagulation
Fibrinolysis and coagulation are interrelated: fibrin clots are nor-
mally lysed by plasmin locally released from plasminogen by the
action of tPA, and this process can be enhanced by some procoagu-
lant factors (e.g. activated factor XII, and protein C). This system, so
delicately controlled and normally maintained in a dynamic equi-
librium, is strongly influenced by components involved in inflam-
matory and other defence mechanisms in the host. An integrated
understanding of these processes offers the potential for improved
means to predict the adverse complications of many diseases and
ultimately to prevent their occurrence.
Introduction
Fluid blood is contained within the vascular tree, but as a result
of minor trauma that occurs during the wear and tear of everyday
living, leaks occur in the blood vessel wall that must be sealed
by a solid impermeable fibrin clot in order to prevent significant
blood loss. The clot is formed from clotting factors in flowing
blood and is located and restricted to the site of the leak without
dissemination throughout the vascular tree. This is the process
of haemostasis, an exquisitely controlled mechanism that re-
quires components of the vessel wall, blood platelets, and soluble
procoagulant and anticoagulant proteins. The haemostatic plug
consists of a mass of platelets, red blood cells, and leucocytes en-
meshed in interlocking strands of insoluble and cross-​linked fi-
brin fibres that plug the leak.
Once formed, the haemostatic plug is gradually replaced by new
tissue as a part of wound healing. This process requires lysis of the
blood clot by the fibrinolytic system and subsequent ingrowth of
new cells. Thus, haemostasis is not an isolated phenomenon, but
is one component of the defence mechanisms that lead to eventual
wound healing.
Thrombosis, as opposed to haemostasis, is a pathological state in
which a clot is formed that partially or completely obstructs the flow
of blood within the blood vessel and sometimes dislodges to become
an embolus.
To understand the biology of haemostasis and thrombosis, it is
necessary to know the roles of the vessel wall, the platelets, the co-
agulation and fibrinolytic systems, and their respective inhibitors.
Blood vessel wall
The anatomy of the wall of both an artery and a vein is shown sche-
matically in Fig. 22.7.1.1. All blood vessels are lined by an intima
consisting of a monolayer of endothelial cells that rest upon a loose
network of tissue called the extracellular matrix. In addition to the
intima, larger and intermediate sized arteries contain two other
layers: the media, composed mostly of smooth muscle cells, and the
adventitia, consisting largely of connective tissue, nerves, and nu-
trient vessels. These three layers also exist in veins, but the media and
adventitia are much less distinct and are not visible in the smaller
arterioles, venules, and capillaries.
Endothelium
Internal
elastic
lamina
Intima
External
elastic
lamina
Media
Adventitia
Fig. 22.7.1.1  Schematic diagram of a vessel wall consisting of the
intima, the media (smooth muscle cells), and the adventitia. The intima
consists of a layer of endothelium that is exposed to the circulating
blood. The subendothelial matrix lies below the endothelium and is
separated from the media by the internal elastic membrane. See text for
detailed description of each layer.
section 22  Haematological disorders
5492
Endothelial cells
Endothelial cells form the basis of vascular development and are
derived from embryonic mesoderm. Embryonic endothelial cells
(angioblasts) develop under the influence of growth hormones
including basic fibroblast growth factor (b-​FGF) and vascular endo-
thelial growth factor (VEGF), both of which interact with recep-
tors on the cell membrane termed receptor tyrosine kinases. These
early blood vessels expand into a vascular tree under the influence
of two major hormones, angiopoietin 1 and 2, that bind to a family
of tyrosine kinase receptors called tie-​1 and tie-​2 (tyrosine kinase
plus Ig and epidermal growth factor-​like domains) on endothelial
cells. To fully develop into an intact vascular tree, endothelial cells
must interact with the extracellular matrix and other cells, a pro-
cess that requires cell–​cell adhesion that is dependent upon cell sur-
face cytoadhesive molecules (CAMs) such as platelet-​endothelial
cell adhesion molecule-​1 (PECAM-​1), and vascular endothelial
cell cadherin (VE-​cadherin). Endothelial cell structure is also de-
pendent upon the integrin family of molecules and interactions with
the extracellular matrix.
Endothelial cells are heterogeneous in appearance, function, and
genetic regulation. In the brain, endothelial cells form very tight
junctions with one another to preserve the blood–​brain barrier; in
the spleen and liver, the interendothelial gaps are wide, permitting
soluble and cellular trafficking between blood and the extravascular
space. Not all endothelial cells synthesize the same proteins. Tissue
plasminogen activator (tPA) is synthesized by only about 3% of
cells. Even von Willebrand factor (VWF), often regarded as a spe-
cific marker for endothelial cells, is not expressed in all cells. The
microenvironment also plays an important role in regulating endo-
thelial cell function. Haemodynamic forces, including hydrostatic
pressure, and shear stresses and strains can influence endothelial
cell structure and function. Haemodynamic forces can even regulate
endothelial cell gene expression. For example, there is a shear-​stress
response element in the gene governing the synthesis of the β chain
of the platelet-​derived growth factor (PDGF). Other endothelial
cell genes responsive to shear forces include those coding for tPA,
intercellular adhesion molecule (ICAM), and vascular cell adhesion
molecule-​1 (VCAM-​1).
Endothelial cells contribute to haemostasis by their contributions
to vascular tone and procoagulant, anticoagulant, fibrinolytic, and
antifibrinolytic activities.
Vascular tone
Vasoregulatory substances produced by endothelial cells are shown
in Table 22.7.1.1. The most important vasoregulators are nitric
oxide, previously known as endothelial cell-​derived relaxation factor
(EDRF), and prostacyclin. Nitric oxide and prostacyclin are also im-
portant antiplatelet agents. On the other hand, the most important
vasoconstrictors are endothelin and angiotensin 2.  Endothelin is
also a mitogen for smooth muscle cells.
Anticoagulant properties
The anticoagulant properties of the endothelial cells are shown in
Table 22.7.1.2. Prostacyclin not only causes vasodilation, but it is a po-
tent inhibitor of platelet aggregation. Nitric oxide has a similar effect.
An important anticoagulant function of endothelial cells is the expres-
sion of thrombomodulin, a transmembrane-​bound protein that acts
as a receptor for thrombin. The thrombomodulin–​thrombin complex
is the physiological activator of protein C. Activated protein C, in turn,
inactivates clotting factors Va and VIIIa to turn off coagulation. The
Table 22.7.1.1  Vasoregulatory substances produced
by endothelial cells
Vasoregulatory substance
Action
Vasodilators
Nitric oxide (NO)
↑cGMP in SMC
Prostacyclin (PGI2)
↑Cyclic AMP in platelets
Monoamine oxidase (MAO)
↓Catecholamines
Vasoconstrictors
Endothelin
Activates Ca2+ channels in SMC
Angiotensin 2
Converts angiotensin 1 to 2 by ACE
Prostaglandin G2,H2(PGG2, PGH2)
Acts on SMC
ACE, angiotensin-​converting enzyme on endothelial cells; AMP, adenosine-​
monophosphate (converted to adenosine); SMC, smooth muscle cells.
Table 22.7.1.2  Procoagulant and anticoagulant properties of endothelial cells
Procoagulant and anticoagulant
Synthesis
Action
Procoagulants
von Willebrand factor (VWF)
Constitutive
Carrier of Factor VIII; platelet adhesion to vessel wall
Tissue factor (TF)
Inducible
Receptor for Factor VII
Anticoagulants
Prostacyclin (PGI2)
Constitutive
Inhibits platelet aggregation
Nitric oxide (NO)
Constitutive
Vasodilation
Thrombomodulin (TM)
Constitutive
TM/​thrombin complex, activates protein C
Tissue factor pathway inhibitor (TFPI)
Constitutive
Inhibits TF–​VIIa–​Xa complex
Glycosaminoglycans (GAGs)
Constitutive
Antithrombins
CD 39 (ectonucleotidase)
Constitutive
Degrades ATP and ADP and inhibits platelet aggregation
Tissue plasminogen activator (tPA)
Constitutive
Converts plasminogen to plasmin
22.7.1  The biology of haemostasis and thrombosis
5493
action of the thrombin–​thrombomodulin complex is enhanced when
protein C occupies the endothelial cell protein C receptor (EPCR),
also located on the endothelial surface (see ‘Receptors’).
Endothelial cells contribute to the control of coagulation by syn-
thesizing tissue factor pathway inhibitor, which inhibits the tissue
factor-​mediated initiation of the clotting reactions. They also syn-
thesize glycosaminoglycans such as heparan sulphate and other
proteoglycans that inhibit thrombin via their interaction with
antithrombin. In addition, they express vascular ectonucleoside tri-
phosphate diphosphohydrolase, otherwise known as CD39, on their
surface. CD39 acts to convert ATP/​ADP to AMP and then to ad-
enosine, which inhibits platelet aggregation. Endothelial cells also
secrete fibrinolytic factors including prostacyclin and tissue plas-
minogen activator, among others.
Procoagulant properties
Although the overall effect of endothelial cells is anticoagulant,
these cells do participate in coagulation by storing proteins such as
VWF and by synthesizing tissue factor (TF) under certain condi-
tions. Procoagulant properties of the endothelial cell are also listed
in Table 22.7.1.2. VWF is synthesized constitutively by endothelial
cells and is essential for platelet adhesion to the vessel wall and as
a carrier for blood clotting factor VIII. VWF is stored in Weibel–​
Palade bodies as depicted in Fig. 22.7.1.2. It is released into the cir-
culation in multimers of heterogeneous molecular mass ranging
from 1000 kDa to about 20 000 kDa. Endothelial cells also secrete
very large VWF multimers abluminally into the extracellular matrix.
TF acts as a binding protein for factor VII and is essential for the
initiation of coagulation. It is not constitutively produced by endo-
thelial cells, but it can be induced by tissue necrosis factor (TNF),
endotoxin, and other inflammatory substances.
Receptors
The receptor function of endothelial cells plays an important role in
haemostasis and thrombosis (Table 22.7.1.3). They express thrombin
receptors such as protease-​activated receptor (PAR)-​1, -​2, and -​4.
Thrombin cleaves the C-​terminal end of the receptor, which then
binds to the remaining cell-​associated protein (a so-​called tethered
ligand) and triggers intracellular signalling through G proteins, re-
sulting in activation of endothelial cells. PARs influence vascular tone
but do so through different intracellular signalling mechanisms.
The thrombin–​thrombomodulin complex not only activates pro-
tein C, but also activates a protein known as the thrombin-​activatable
fibrinolytic inhibitor (TAFI), a procarboxypeptidase that functions
to inhibit fibrinolysis. Endothelial cells also express EPCR that acts
to modulate the activity of activated protein C. EPCR resides on the
endothelial cell and enhances protein C activation by about 20-​fold
in vivo. EPCR binds protein C or activated protein C and presents
it to the thrombin–​thrombomodulin complex. Binding of APC to
EPCR is also important in the inflammatory process in that through
cell signalling mechanisms it decreases inflammatory cytokines and
other molecules involved in inflammation. In addition, EPCR acts as
a receptor for coagulation factor VII which binds to EPCR with an
affinity similar to that of protein C.
Urokinase plasminogen activator receptors are not found on
resting endothelial cells, but are found on those involved in angiogen-
esis. There are a number of adhesive receptors on the surface of endo-
thelial cells, as shown in Table 22.7.1.3. The adhesion of neutrophils
is dependent upon the expression of P-​selectin. P-​selectin is rapidly
internalized by the endothelial cell, but this is followed by expression
of another cytoadhesive molecule, E-​selectin, which is necessary for
continued adherence and rolling of neutrophils along the endothelial
cell surface. ICAM and VCAM are receptors for leucocytes and are
important for the interaction of leucocytes and the vessel wall.
Extracellular matrix
The extracellular matrix is a complex, heterogeneous structure be-
neath the endothelium with a number of constituents that contribute
to haemostasis and thrombosis. The matrix consists of a network
of collagens, elastins, proteoglycans, and glycoproteins, including
fibronectin, vitronectin, laminin, tenascin, thrombospondin, VWF,
and osteopontin, as shown in Table 22.7.1.4. The matrix proteins
promote platelet adhesion, cellular migration, cell proliferation, and
endothelial and smooth muscle cell interactions.
Collagens are the most abundant proteins in subendothelial con-
nective tissue. Collagen types I, II, III, IV, V, VI, and VII have been
identified in various matrix tissues. The collagens are synthesized
by endothelial cells, smooth muscle cells, and adventitial fibroblasts.
The various collagens contribute to the integrity of the vessel wall,
Fig. 22.7.1.2  Electron micrograph of an endothelial cell. Weibel–​Palade
bodies containing immunogold-​labelled multimers of VWF are depicted
by the arrows.
Table 22.7.1.3  Receptor function of endothelial cells
Receptor
Ligand
Protease activated receptor 1 (PAR-​1)
Thrombin
Thrombomodulin
Thrombin
Protein C receptor
Protein C
Urokinase plasminogen activator receptor (u-​PAR)
Urokinase
Adhesive receptors:
Intercellular cytoadhesive molecule (ICAM)-​1, -​2
Integrins α1β2; αmβ2
Vascular cytoadhesive molecule (VCAM)
α1β2; α4β1
P-​selectin
PSGL-​1
E-​selectin
ESL-​1; PSGL-​1; LAMP,
Mac2-​BP
section 22  Haematological disorders
5494
but they also play a role in platelet activation and, in some instances,
coagulation. For example, collagen IV has been shown to be a spe-
cific high-​affinity binding protein for blood coagulation factor IX,
although the function of this complex is not yet known.
The proteoglycans constitute a heterogeneous group of mol-
ecules composed of a core protein attached to a glycosaminoglycan.
These include decorin, biglycan, heparan sulphate, dermatan sul-
phate, and others. Heparan sulphate, for example, can combine with
antithrombin (AT) to inhibit thrombin. The precise role of all of the
proteoglycans is not known, but some attach to collagen and are ne-
cessary for maintaining the structure of the vessel wall.
The matrix also contains elastin, which is secreted by endothelial
and smooth muscle cells as tropoelastin that is converted to mature
elastin in the matrix where it is assembled into fibres. One function
of elastin is simply to maintain the elastic structure of the vessel wall.
This substance is found interspersed between smooth muscle cells
as well as the matrix. It may also function in cell migration from the
vessel wall to the extravascular space.
Fibronectin, vitronectin, and laminins are also components of the
extracellular matrix which function in fibrinolysis and platelet adhesion.
Within the extracellular matrix there are a number of matrix
metalloproteinases (MMPs), a group of enzymes useful in matrix
degradation and repair. They are secreted as proenzymes and con-
verted to active enzymes that require zinc or calcium as cofactors.
They have several functions as listed in Table 22.7.1.5. Their activ-
ities in matrix degradation and repair are controlled by tissue inhibi-
tors of metalloproteinases.
Smooth muscle cells
The smooth muscle cell layer, found in medium-​ and larger-​sized
vessels and more prominently in arteries, has several functions re-
lated to the biology of haemostasis and thrombosis. Smooth muscle
cells possess contractile, biosynthetic, and proliferative functions.
Contractile properties governed by such substances as nitric oxide,
prostacyclin, and endothelin play important roles in vasodilation
and vasoconstriction, respectively. Smooth muscle cells, like endo-
thelial cells, synthesize growth factors such as VEGF, insulin-​like
growth factors (IGF), epidermal growth factors (EGF), activins, and
others that are important in smooth muscle cell generation. Smooth
muscle cell proliferation is a hallmark of the atherosclerotic lesion.
Biosynthetic products of the smooth muscle cells include various
types of collagens, elastin, glycoproteins, and proteoglycans. When
exposed to injury, smooth muscle cells can also express functionally
active TF, contributing to the initiation of blood coagulation.
Adventitia
The adventitia is composed of a loose network of cells consisting
of fibroblasts, adipocytes, and mast cells. Collagens I and III, glyco-
proteins, and elastin are synthesized by fibroblasts. Fibroblasts also
contain large amounts of TF. Adipocytes secrete collagen I and III
and synthesize lipids.
Platelets
Platelets are the smallest of the circulating blood cells, about 2 µm
in diameter. They are essential components of the haemostatic plug
and are derived from bone marrow megakaryocytes. Although plate-
lets are anucleate and appear to be simple cells composed of cyto-
plasm, a surface connected canalicular system (SCCS), and storage
granules (δ or dense granules and α-​granules); they are, nevertheless,
complicated cells with a variety of very important functions essen-
tial for normal haemostasis. These can be broadly divided into the
following: (1) platelet adhesion, defined as platelets adhering to the
damaged area of the vessel wall where subendothelial matrix tissue is
exposed; (2) platelet activation, both by agents within the matrix as
well as by soluble agonists; (3) platelet secretion of granule contents;
(4) platelet aggregation, defined as platelets sticking to one another in
an aggregated mass, forming a platelet plug. The following sections
describe each of these broad areas of platelet function in more detail.
Platelet adhesion
The initial platelet response to vascular injury is adhesion to the
vessel wall. Resting, nonactivated platelets are not attracted to the
vessel wall. However, following vascular damage, platelets rolling
Table 22.7.1.5  Matrix metalloproteinases
MMP number
Activity
Substrate
MMP-​1
Collagenase
Collagen I, II, III, VII, VIII, X
MMP-​2
Gelatinase
Collagen IV, V, VII, X
MMP-​3
Stromelysin
Microglycans
MMP-​8
Collagenase
–​
MMP-​7
Matrilysin
Fibronectin, laminin, collagen IV
MMP-​9
Gelatinase
Elastin, fibronectin
MMP-​10
Stromelysin
Fibronectin, laminin, elastin, various
collagens
MMP-​11
Stromelysin
Fibrinogen, fibrin
MMP-​12
Elastase
Elastin
MMP-​14
–​
Collagen IV, progelatinase A
MMP-​15
–​
Gelatin
MMP, matrix metalloproteinase.
Modified from Plow EF, Ugarova T, Miles LA (1998). In: Localzo J, Shafer AI, eds.
Thrombosis and hemorrhage, 2nd edn, ch. 18, p. 381. Williams and Wilkins, Baltimore.
Table 22.7.1.4  The extracellular matrix
Structural proteins
Collagens I, III, IV, V, VI, VII
Elastin
Adhesive proteins
Fibronectin
Vitronectin
Laminin
Von Willebrand factor
Antiadhesive
Tenascin
Thrombospondin
Ground substance
Hyaluronic acid
Proteoglycans
Chondroitin sulphate
Dermatan sulphate
Heparan sulphate
Degradation and repair
Matrix metalloproteinases
22.7.1  The biology of haemostasis and thrombosis
5495
along the endothelium rapidly adhere to the subendothelial inter-
stitial matrix that is exposed by injury. A number of matrix proteins,
such as VWF, fibronectin, fibrinogen, and thrombospondin, are also
present in platelet granules as well as in the circulating blood.
Platelets possess numerous mechanisms for adhering to the
subendothelial matrix (Fig. 22.7.1.3). Adhesion is accomplished by
a number of protein receptors on the surface of platelets as described
in the following sections.
Glycoprotein Ib–​IX–​V (CD42a–​d)
The main function of the platelet membrane glycoprotein (GP) Ib–​
IX–​V complex is to act as a receptor that mediates VWF-​dependent
binding of platelets to collagen, resulting in adhesion of platelets to
the vessel wall. Glycoproteins Ibα, Ibβ, IX, and V are members of
the leucine-​rich glycoprotein family and are characterized by the
presence of a common structural motif in the extracellular domain
composed of a leucine-​rich sequence. GPIbα contains seven leucine-​
rich repeats; GPV contains 15 leucine-​rich repeats; while GPIbβ and
GPIX each contain a single leucine-​rich repeat, all in the extracellular
domains. GPIbα, GPIbβ, GPIX, and GPV are synthesized as separate
gene products which coassociate in a ratio of 2:2:2:1 during transit
through the endoplasmic reticulum. Coassociation of GPIbα, GPIbβ,
and GPIX, but not GPV, is required for the complex to be expressed
on the surface of cells. The role of GPV, a substrate for thrombin, in
the function of the complex is uncertain, and mice deficient in GPV
bind VWF normally and have normal platelet adhesion.
Adhesion to VWF-​coated surfaces through GPIb–​IX–​V is in-
creased by shear, which is thought to induce a structural change in
the receptor that enhances the interaction with VWF. The A1 domain
of VWF forms the principal site that interacts with GPIb. Binding oc-
curs in the N-​terminal 45-​kDa tryptic fragment from GPIbα. Within
this region of GPIbα, an anionic site, 276YDYYPEE282, containing two
sulphated tyrosine residues at tyrosines 278 and 279, has been fur-
ther implicated in VWF binding. The A3 domain of VWF mediates
the interaction with collagens type I and III. A model derived from
the crystal structure of the VWF A3 domain suggests that the VWF–​
collagen interaction is primarily between negatively charged residues
in the A3 domain and positively charged residues in collagen. Plasma
VWF does not interact with unstimulated circulating platelets. For
binding to occur, platelets have to be activated and plasma VWF must
undergo a conformational change. After secretion by endothelial cells,
VWF binds to underlying connective tissue matrix, providing an ac-
tive surface for platelet attachment after the vessel wall is damaged.
Glycoprotein IIb–​IIIa (αIIb–​β3)
Under conditions of low shear, platelets can adhere to matrix-​
bound VWF through a mechanism that involves platelet glycopro-
tein GPIIb–​IIIa (αIIb–​β3). αIIb and β3 are members of the integrin
superfamily, a conserved family of heterodimeric surface recep-
tors, each composed of a larger two-​chain α subunit and a smaller
β subunit, bound noncovalently. Integrins were initially identified
by an ability to bind adhesive glycoproteins containing a tripeptide
sequence, arginine–​glycine–​aspartic acid (RGD), although sub-
sequent work has identified other ligand sequences recognized by
integrins. The interaction of VWF with αIIb–​β3 is mediated by an
RGD sequence in the C4 domain of VWF. αIIb–​β3 is also able to
bind fibronectin, thrombospondin, and vitronectin and may there-
fore represent an adhesion receptor with broad specificity.
Glycoprotein Ia–​IIa (α2β1)
GPIa–​IIa is a receptor for types I  and IV collagen and mediates
platelet adhesion to the vessel wall independent of VWF. The in-
tegrin sequences that mediate the interaction with collagen reside in
a broad sequence called the I domain in the extracellular portion of
the molecule. GPIa–​IIa is constitutively active and does not require
activation to interact with collagen.
Glycoprotein VI–​Fc receptor γ-​chain complex
GPVI–​Fc receptor γ-​chain is the major platelet receptor mediating
collagen-​induced activation of platelets. GPVI is a member of the im-
munoglobulin superfamily and is characterized by immunoglobulin
domains, a transmembrane domain, and a short cytoplasmic tail
that lacks known signalling components. GPVI is associated on
the platelet surface with Fc receptor γ-​chain, apparently as a dimer
in a 1:1 stoichiometry. The complex binds collagen and mediates
collagen-​generated signals through the immunoglobulin receptor
tyrosine-​based activation motif (ITAM) of the Fc receptor γ-​chain.
Cross-​linking of the GPVI–​Fc receptor γ-​chain leads to tyrosine
phosphorylation of the ITAM sequence by Src family kinases, Fyn
and Lyn. Syk, another tyrosine kinase, binds to the phosphorylated
ITAM sequence through its SH2 domains, initiating a signal that
eventually leads to tyrosine phosphorylation of phospholipase Cγ2
and the generation of inositol phospholipids.
Glycoprotein IV (CD36)
CD36 is a highly glycosylated transmembrane protein present on
platelets, monocytes, endothelial cells, and nucleated erythrocytes
which binds thrombospondin. The thrombospondin-​binding site
has been mapped to amino acids 90 to 110 in a single disulphide
loop in the extracellular domain of GPIV. Platelets deficient in GPIV
have mild disturbances in platelet function and poor responses to
pathological agonists such as oxidized LDL and MRP8/​14.
Integrins
Fibrinogen,
Fibronectin,
Vitronectin, vWf,
Thrombospondin
LRG
IIb/IIIa
Ia/IIa
Ic/IIa
VnR
Ib/IX/V
Ig
GPVI
GPVI
CD36
Collagen
Fibronectin
Vitronectin
Laminin
Collagen
Collagen
vWf
Ie/IIa
Fig. 22.7.1.3  Receptors mediating the interaction of platelets with
subendothelial matrix proteins. Adhesion receptors on platelets include
members of the integrin family, leucine-​rich glycoproteins (LRG), members
of the immunoglobulin (Ig) family, and others. Integrins on the surface of
platelets are glycoproteins (GP) IIb–​IIIa, which binds multiple ligands; GPIa–​
IIa, a collagen receptor, GPIc–​IIa, which binds fibronectin; VnR, which is a
receptor for vitronectin; and GPIc′–​IIa, a laminin binding site. Glycoproteins
Ib/​IX/​V are leucine-​rich glycoproteins. Glycoprotein VI (GPVI)/​Fc receptor
γ-​chain is a member of the immunoglobulin family and a collagen receptor.
Glycoprotein IV (GPIV, CD36) is also a collagen receptor.
section 22  Haematological disorders
5496
Other adhesion receptors
Platelets can also adhere to subendothelial matrix through their
fibronectin receptor (α5β1, glycoprotein Ic–​IIa, VLA-​5), lam-
inin receptor (α6β1, glycoprotein Ic′–​IIa, VLA-​6), or vitronectin
receptor (αvβ3, VnR). α5β1 is a constitutively active receptor for
fibronectin that does not require cell activation. There are two
sequences in fibronectin which interact with GPIc–​IIa:  an RGD
sequence in the tenth type III repeat which interacts primarily with
the β1 subunit and a synergy sequence in the adjacent ninth type III
repeat which interacts primarily with the α5 subunit. α6β1 is a lam-
inin receptor which is expressed on platelets. Immunoprecipitation
studies suggest that α6β1 may exist on the cell surface in a com-
plex with proteins with four transmembrane domains, so-​called
tetraspanins, such as CD9, CD81, and NAG-​2. The nature of these
interactions is presently unclear. α6β1 recognizes a sequence in the
long-​arm E8 fragment of laminin obtained after elastin digestion.
The binding requires the presence of divalent cations which bind to
specific sites on the integrin α subunit. Small numbers of αvβ3 are
expressed on platelets.
Current evidence indicates that all of these adhesion mechan-
isms may be important. The redundancy in adhesion receptors
may (1) provide backup mechanisms to protect against blood loss;
(2) generate different signals in response to interaction with dif-
ferent matrix proteins; or (3) represent different systems at work
in different parts of the vascular tree. An example of the latter
might be the relative roles of GPIb–​IX–​V and GPIIb–​IIIa in the
VWF-​mediated adhesion of platelets to collagen. Under high-​
shear conditions, such as those found in capillaries and small
arterioles, GPIb–​IX–​V may be the predominant mechanism
mediating platelet adhesion to collagen and VWF-​dependent ad-
herence; whereas under low shear conditions, like those found in
large veins and in arteries, GPIb–​IX–​V may be less effective and
other mechanisms that require a shorter residence time of plate-
lets on the subendothelial matrix, including GPIc–​IIa interaction
with fibronectin and GPIa–​IIa interaction with collagen, may be
important. The presence of multiple receptors for collagen on the
platelet surface, including GPIb–​IX–​V, GPIIb–​IIIa, GPIa–​IIa,
GPIV, and GPVI is interesting and raises the possibility of different
collagen responses. Vitronectin also appears to be important for ad-
hesion at high shear, and can bind to both GPIIb–​IIIa and specific
vitronectin receptors. Recent evidence suggests that platelet adhe-
sion to collagen types I and III in flowing blood is dependent on
both VWF and fibronectin. Collagen types I, II, and III have been
shown to bind VWF.
Platelet activation
Following adhesion and in response to soluble agonists such as
thrombin, platelets undergo a series of complex biochemical reac-
tions leading to cell activation. As a result, platelets undergo changes
in shape, alterations in surface lipid composition leading to the ex-
pression of platelet coagulant activity (see later) and thrombin gen-
eration, as well as secretion of the contents of intracellular granules
leading to the release of ADP. The thrombin generated at the platelet
surface and ADP secreted from platelet granules lead to activation of
additional platelets. These reactions involve the metabolism of mem-
brane inositol phospholipids, changes in cellular levels of calcium,
activation of contractile proteins, stimulation of heterotrimeric
and low molecular weight GTP-​binding proteins, and tyrosine and
serine–​threonine phosphorylation of proteins, among other events.
These biochemical reactions initiate second messenger signals that
drive the functional changes that occur in platelets which transform
them from the resting state to an activated one, and which play a
crucial role in haemostasis. Some of these signalling pathways are
described in the following sections (Fig. 22.7.1.4).
Phospholipid metabolism
Metabolism of membrane phospholipids is one of the first signal-
ling pathways identified in platelets and remains one of the most
important. Platelet stimulation by a variety of agonists results in
activation of membrane-​associated phospholipases, including
phospholipases C, A2, and D, which cleave fatty acids from the
phospholipid. The lipid products generated by these pathways are
signalling compounds which are important for changes in cyto-
plasmic calcium and activation of kinases and phosphatases.
Fig. 22.7.1.4  Signalling pathways involved in platelet activation.
Following the interaction of agonist with receptor, there is G protein
(Gs) coupled activation of phospholipid metabolic pathways
through phospholipase A2 (PL-​A2), phospholipase Cγ (PL-​Cγ), and
phospholipase D (PL-​D) leading to generation of thromboxane A2
(TxA2), inositol trisphosphate (IP3), diacylglycerol, and phosphatidic acid
(PA). Arachidonic acid generated by the action of phospholipase A2 is
converted by cyclooxygenase-​1 (COX-​1) to prostaglandin endoperoxides
G2 (PGG2) and H2 (PGH2) which are, in turn, converted to thromboxane
A2 through the action of thromboxane synthase (TxS). Thromboxane
generated through arachidonate metabolism plays a key role in secretion,
perhaps through membrane fusion. Granule contents, including
adenosine diphosphate, are emptied into the surface connected
canalicular system (SCCS) and make their way to the outside of the cell.
Diacylglyerol stimulates activation of protein kinase C (PKC), resulting in
serine-​threonine phosphorylation of proteins such as pleckstrin. Inositol
trisphosphate (IP3) stimulates calcium release from storage sites in the
dense tubular system (dts). The release of calcium from the dense tubular
system is antagonized by cyclic AMP, generated through G protein (Gαs)
coupled inhibitory receptor activation of adenylate cyclase. Calcium,
released in response to IP3, activates gelsolin, an actin-​capping and -​
severing protein, which generates actin monomers that then serve as
nucleation sites for formation of actin filaments and assembly of the
activation-​dependent cytoskeleton. Assembly of the cytoskeleton and
interaction of the cytoskeletal proteins with surface integrins such as
αIIb–​β3 (glycoproteins IIb and IIIa) are involved in integrin activation.
Calcium also activates myosin light chain kinase which phosphorylates
myosin light chain, generating actinomyosin contraction, important for
changes in platelet shape and the secretion process.
22.7.1  The biology of haemostasis and thrombosis
5497
The most intensively studied of these pathways is the metabolism
of inositol phospholipids through phospholipase C.  Membrane
phosphatidylinositol (PI) exists in multiple phosphorylation
states: PI, PI–​P, PI–​P2 which is phosphorylated in the 3,4 or 4,5
positions, and PI–​P3 which is phosphorylated in the 3,4,5 posi-
tions. Phosphatidylinositol-​specific kinases and phosphatases
maintain pools of phosphorylated phosphoinositides in a proper
concentration range. Platelets contain several isoforms of phospho-
lipase C which are activated by different mechanisms. All cleave
phosphatidylinositol 4,5-​bisphosphate (PI 4,5–​P2) and, later,
phosphatidylinositol, as well as phosphatidylinositol 4-​phosphate
(PI 4–​P), to yield diglyceride and inositol trisphosphate (IP3).
Phospholipase Cα and Cβ are coupled to heterotrimeric G proteins
where phospholipase Cα is coupled to growth factor receptors.
IP3 generated by phospholipase C cleavage of inositol phospho-
lipids has been implicated in the release of calcium from intracel-
lular storage sites in the platelet-​dense tubular system. The other
product of phospholipase C cleavage, diacylglycerol, activates pro-
tein kinase C, which phosphorylates pleckstrin, a 47-​kDa protein,
and other proteins.
Phospholipase A2 is linked to G-​protein coupled receptors and
cleaves fatty acids in the sn-​2 position in membrane phospho-
lipids, primarily phosphatidylcholine. In most individuals in devel-
oped countries, the fatty acid in this position is arachidonic acid.
Arachidonic acid, liberated by the action of phospholipase A2, is
converted to a variety of possible products by the microsomal en-
zymes, cyclooxygenase and lipoxygenase. Cyclooxygenase converts
arachidonic acid to prostaglandin endoperoxides, prostaglandins
F2, E2, and D2, whose main fate in platelets is rapid conversion to
thromboxane A2 by thromboxane synthase. Thromboxane A2 is be-
lieved to play an important role in the release of intracellular gran-
ules by acting as a membrane fusogen, fusing granule membranes
with the membrane of the surface connected canalicular system and
permitting secretion of the granule contents to the outside of the
cell. Thromboxane A2 is also an exceptionally potent constrictor of
vascular smooth muscle and a strong platelet-​aggregating agent.
Inhibition of the arachidonate pathway has been a primary target
for platelet inhibition. Cyclooxygenase is irreversibly inhibited by
aspirin, which acetylates serine 340 of cyclooxygenase, and revers-
ibly inhibited by nonsteroidal anti-​inflammatory agents. Inhibition
of cyclooxygenase inhibits thromboxane formation and results in
inhibition of the release of intracellular granules. One of the mech-
anisms by which aspirin is thought to act as an antiatherosclerosis
agent is by inhibition of the release of PDGF from platelet granules.
Phospholipase D acts primarily on phosphatidylcholine to pro-
duce choline and phosphatic acid. Protein kinase C and PI–​P2 play
an important role in activation of phospholipase D. Phosphatidic
acid is an intracellular messenger which is proposed to play a role in
platelet activation. In addition, phosphatidic acid can be converted
to lysophosphatidic acid through the action of phospholipase A2.
Like phosphatidic acid, lysophosphatidic acid is an intracellular mes-
senger which is involved in phospholipase activation, signalling by
low molecular weight G proteins, and cytoskeleton reorganization.
Calcium metabolism
Calcium ions are extremely important in platelet function, as de-
scribed in subsequent discussions. In resting platelets, the cyto-
plasmic concentration of calcium is maintained at a low level by
active transport of calcium both inside and outside the cell and into
the dense tubular system (DTS), a sarcoplasmic reticulum-​like frac-
tion in platelets. Calcium transport in the platelet is accomplished
by a plasma membrane sarcoplasmic–​endoplasmic reticulum-​like
calcium ATPase (SERCA2-​b), a dense tubular system SERCA3,
a sodium–​calcium exchange pump in the plasma membrane, and
passive calcium fluxes. During platelet activation, IP3, generated
by metabolism of membrane inositol phospholipids, induces the
rapid release of calcium stored in the dense tubular system. This in-
crease in cytoplasmic calcium is essential for platelet activation, and
agents that cause decreases in cytoplasmic calcium inhibit platelet
activation while agents that increase cytoplasmic calcium stimulate
platelet activation.
Calcium functions as a major intracellular messenger in plate-
lets, mediating calcium-​dependent reactions important in almost
all phases of platelet activation. An increase in the concentration of
cytoplasmic free calcium activates gelsolin, the calcium-​dependent
actin capping and severing protein, which plays an important role
in reorganization of the cytoskeleton. Calcium also activates the
calcium and calmodulin-​dependent myosin light chain kinase,
leading to phosphorylation of myosin light chains, activation of
actin-​stimulated myosin ATPase activity, and the development
of contractile forces. The contraction generated by actin and my-
osin mediates changes in platelet shape and is important for events
leading to platelet secretion. In the absence of calcium ions, tropo-
myosin inhibits the interaction of myosin with actin, and this may
be an additional regulatory role of calcium in platelets. Calpain, a
calcium-​dependent thiol protease, hydrolyses numerous proteins
involved in platelet signalling. Activation of calpain is believed to
be important both for regulation of cytoskeletal events and integrin-​
mediated signalling.
Cytoskeletal reorganization
Resting platelets are discoid in shape and feature a cellular cyto-
skeleton that consists of a network of actin filaments that fill and
shape the cytoplasm of the cell and a single microtubule coil at the
margin of the disc. Upon activation, platelets undergo remarkable
morphological changes (Fig. 22.7.1.5). There is an initial change
from the normal discoid shape of the resting platelet to a sphere as
calcium levels in the cell increase. Filamentous actin appears in the
form of stress fibres, and the cellular content of filamentous actin
increases. Membrane ruffles form as long cellular projections called
pseudopodia, processes that also involve the low molecular weight
GTPases rac, rho, and cdc 42. Actin cables are present in these
pseudopodia, extending to the end of the projections. Also during
activation, microtubules contract and ‘squeeze’ granules toward the
centre of the cell.
The energy for contraction is provided by a magnesium ion-​
dependent ATPase present in myosin and stimulated by actin.
Contraction occurs by actin filaments and myosin rods sliding over
one another. Myosin light-​chain phosphatase dephosphorylates and
inhibits myosin. Membrane glycoproteins GPIIb–​IIIa, GPIb–​IX–​V,
and other membrane proteins are associated with the cytoskeleton
and provide direction for the contractile process. This activation-​
dependent cytoskeleton is more than just a structural scaffold for
platelet shape changes. Numerous signalling proteins are incorpor-
ated into the cytoskeleton and may function in specialized compart-
ments by virtue of their association with the cytoskeleton.
section 22  Haematological disorders
5498
Platelet coagulant activity (platelet factor 3)
Platelet membranes have an asymmetrical distribution of phospho-
lipids, with almost all of the acidic (negatively charged) phospho-
lipids such as phosphatidylserine and phosphatidylinositol located
in the inner leaflet of the plasma membrane in resting platelets. After
platelet activation, the acidic phospholipids are translocated to the
outer half of the membrane, while phosphatidylcholine moves to the
inner half, through the action of TMEM16F, a membrane-​bound,
calcium-​dependent lipid scramblase.
The exposed phosphatidylserine and other negatively charged
phospholipids account for some of the activity traditionally known
as platelet factor 3 by contributing to surface properties for binding
of factor X and prothrombin activation complexes. This interaction
with platelet phospholipids increases the rate of factor X activation
and prothrombin activation nearly a thousand fold. In addition to
phospholipids on the platelet membrane, there appear to be other
specific binding proteins for blood clotting factors.
cAMP pathway
A major mechanism for down-​regulation of platelet function is the
stimulation of adenylate cyclase, which increases cAMP concentra-
tions. Adenylate cyclase is mainly localized in microsomal fractions
and is stimulated by adenosine, prostacyclin, and prostaglandin E1
through activation of Gs, a heterotrimeric GTPase associated with
the prostaglandin receptor on the platelet surface. cAMP inhibits
platelet aggregation, platelet secretion, and platelet adhesion to the
vessel wall. These effects are probably exerted by inhibiting calcium
flux and/​or promoting calcium reuptake.
Activation by soluble agonists
In addition to activation through interaction with subendothelial
connective tissues, platelets may also be activated by soluble agonists.
These include ADP, adrenaline, and thrombin. In general, this acti-
vation occurs through the interaction between soluble agonist and
specific receptors on the platelet surface.
Thrombin is one of the most powerful of platelet agonists.
Generated during blood coagulation, thrombin activation of
platelets occurs through a novel family of receptors called PARs.
These are G protein-​coupled, seven-​membrane-​spanning mol-
ecules which are activated by proteolysis. Thrombin cleaves the
N-​terminal exodomain, unmasking a new N-​terminal, which
functions as a tethered peptide agonist. The tethered peptide binds
intramolecularly to the remainder of the receptor to trigger acti-
vation. Four members of the PAR family of receptors have been
identified, but only PAR-​1 and PAR-​4 mediate activation of human
platelets by thrombin.
Thrombin interacts with other proteins on the surface of platelets,
but the nature of these interactions is uncertain. Glycoprotein V, part
of the GPIb–​IX–​V complex, is a substrate for thrombin although the
absence of GPV does not appear to inhibit thrombin activation of
platelets. GPIb is an equilibrium binding site for thrombin. Patients
with a deficiency of GPIb have been reported to have changes in
the rate of activation of platelets by thrombin which is overcome at
higher concentrations of agonist.
There are at least three receptors for ADP on platelets, all members
of the seven-​transmembrane-​spanning members of the purinergic
(P2) receptor family, either P2Y (G-​protein-​coupled purinergic re-
ceptors) or P2X (ligand-​gated channel receptors). One receptor, des-
ignated P2Y1, is coupled to phospholipase Cβ, probably through Gq.
A second receptor, P2Y12, is coupled to adenylate cyclase through
Gi2. The third receptor, P2X1, is coupled to rapid calcium influx
and is a member of the intrinsic ion channel family. Full platelet
activation by ADP likely involves an interaction of ADP with all
three receptors. ADP-​induced activation of GPIIb–​IIIa on platelets
Microtubules
Surface-connected
canalicular system
Alpha granule
Dense granule
Glycogen
Mitochondrion
(b)
(a)
Fig. 22.7.1.5  Platelet morphology. Platelets are small, anucleate cells. In the resting state (a), platelets
are discoid shaped and contain a marginal rim of microtubules. After activation (b), platelets undergo
changes in shape, becoming more rounded, and extend cytoplasmic projections, called pseudopods,
outward.
22.7.1  The biology of haemostasis and thrombosis
5499
requires both P2Y1 and P2Y12 and concomitant signalling through
the GTP-​binding proteins Gq and Gi2.
Platelet secretion
A primary endpoint of platelet activation is the secretion of platelet
granule contents to the outside of the cell. During platelet activa-
tion, the granules are ‘squeezed’ to the centre of the cell where the
granules fuse with the surface-​connected canalicular system, a
series of intracellular canals that are connected to the cell surface.
The contents of the granules make their way to the outside of the
cell. Secretion requires prostaglandin metabolism and is dependent
on contractile events and members of the soluble N-​ethylmaleimide
sensitive factor attachment protein receptors (SNAREs) which me-
diate granule tethering and docking with the plasma membrane.
Products of prostaglandin metabolism, primarily thromboxane
A2, may act in the fusion of the granule membrane with that of the
surface-​connected canalicular system.
Platelets possess two types of storage granules (Table 22.7.1.6),
both of which are involved in secretion of active ingredients that
modulate platelet function. One type is the dense granule, so called
because it is dense when viewed by electron microscopy. The other
type is the α-​granule.
Dense granules contain adenine nucleotides, calcium, and sero-
tonin. Adenine nucleotides are sequestered in the dense granules
mainly as ADP and ATP in a complex with calcium ions and pyro-
phosphate, and are not interchangeable with the nucleotides involved
in general cell metabolism. ADP released from platelet-​dense gran-
ules activates additional platelets and recruits them to the growing
platelet thrombus. Serotonin, a potent modulator of vascular tone
and integrity, is also a constituent of dense granules.
α-​Granules contain PDGF, β-​thromboglobulin, PF4, fibrinogen,
factor V, VWF, and thrombospondin. PDGF is mitogenic for smooth
muscle cells and when released from platelets at a site where the
vessel wall is damaged, it stimulates proliferation and migration of
smooth muscle cells in the intima, contributing to the atheroscler-
otic process. β-​Thromboglobulin and PF4 are basic, lysine-​rich pro-
teins that interact with glycosaminoglycans such as heparan sulphate,
dermatan sulphate, and chondroitin sulphate, which are components
of the endothelial cell surface. PF4 has a strong heparin-​neutralizing
activity and has been implicated in the aetiology of heparin-​induced
thrombocytopenia. Thrombospondin is a major α-​granule glycopro-
tein, but it is also secreted by fibroblasts, endothelial and smooth-​
muscle cells. Thrombospondin is a high molecular weight adhesive
protein which binds to glycosaminoglycans, fibrinogen, plasminogen,
histidine-​rich glycoprotein, type V collagen, and calcium ions. It as-
sociates with cell surfaces and extracellular matrices and facilitates
cell–​cell and cell–​matrix interactions.
Platelet aggregation
Platelet aggregation, the interaction of one platelet with another, is a
major function of platelets and is very important in the haemostatic
process. The formation of an aggregated platelet mass at the site of
injury provides a physical plug that occludes the defect in the vessel
wall and prevents blood loss.
Aggregation is mediated by two glycoproteins on the platelet
surface, αIIb–​β3, which constitute a receptor for fibrinogen/​fibrin.
Thus, αIIb–​β3 on one platelet binds fibrinogen or fibrin which, by
virtue of its dimeric structure, interacts with αIIb–​β3 on another
platelet. On resting platelets, αIIb–​β3 is in an inactive state and is
unable to bind fibrinogen. Following platelet activation, αIIb–​β3
becomes activated through a process that involves calcium, protein
kinase C, heterotrimetric G proteins, and talin. Activation of αIIb–​
β3 requires energy and is a multistep process. Fibrinogen binding
to αIIb–​β3 occurs through a C-​terminal dodecapeptide sequence,
HHLGGAKQAGDV (His, His, Leu, Gly, Gly, Ala, Lys, Glu, Ala, Gly,
Asp, Val), in the α chain of fibrinogen where the AGDV sequence
has been suggested to have structural similarity to the RGD (Arg,
Gly, Asp) sequence, a common binding motif for integrins.
Blood coagulation
The blood coagulation system consists of a number of zymogens
(proenzymes) that are proteolytically converted to active enzymes
in a series of steps involving activators and cofactors. The coagula-
tion reactions are initiated by TF in complex with activated factor
VII (VIIa). The TF–​VIIa complex then activates both factor IX and
factor X, which, in the presence of their respective cofactors (the
activated forms of factors VIII and V), lead to the rapid conver-
sion of prothrombin to thrombin. Thrombin converts fibrinogen
into a solid fibrin clot that finally undergoes cross-​linking by acti-
vated factor XIII to become a stable haemostatic plug. Platelets are
Table 22.7.1.6  Platelet granule contents
α Granules
α2-​Antiplasmin
Immunoglobulin
Albumin
Multimerin
β-​Amyloid precursor
Plasminogen activator
β-​Thromboglobulin (β-​TG)
Platelet-​derived growth factor (PDGF)
Clusterin
Platelet factor 4 (PF4)
Endothelial cell growth factor
(ECGF)
P-​selectin (GMP-​140)
Factor V
Transforming growth factor (TGF)-​α
Fibrinogen
Transforming growth factor (TGF)-​β1
Fibronectin
Vitronectin
Granule membrane protein
(GMP) 33
von Willebrand factor (VWF)
Dense (δ) granules
Adenosine diphosphate (ADP)
Granulophysin
Adenosine triphosphate (ATP)
Polyphosphate (PPi)
Calcium
Magnesium
Guanosine diphosphate (GDP)
Serotonin (5-​hydroxytryptamine)
Guanosine triphosphate (GTP)
Lysosomal (γ) granules
β-​Galactosidase
Elastase
β-​Glucuronidase
Endoglucosidase
β-​Glycerophosphatase
LAMP-​1
β-​Hexosaminidase
LAMP-​2
Cathepsins
LIMP-​CD63
Collagenase
N-​acetylglucosaminidase
section 22  Haematological disorders
5500
essential in several steps of the clotting mechanism and form the
surface for activated clotting factors, which lead to the explosive
generation of thrombin and subsequent clot formation. Activated
platelets aggregate and localize the haemostatic plug at the site
of injury. The initial generation of relatively small amounts of
thrombin is essential for feedback activation of factors V, VIII, XI,
and XIII as well as platelets.
Understanding the modern concept of the clotting reactions
requires a detailed knowledge of each of the clotting factors.
Table 22.7.1.7 depicts the clotting factors and their inhibitors,
including the vitamin K-​dependent clotting proenzymes, the
nonvitamin K-​dependent zymogens, the cofactors, the inhibitors of
the clotting factors, and the structural proteins.
Vitamin K-​dependent zymogens
The vitamin K-​dependent blood clotting zymogens include pro-
thrombin, factor VII, factor IX, factor X, protein C, and protein S;
their characteristics are listed in Table 22.7.1.7 and their schematic
structures in Fig. 22.7.1.6. A common feature of all these clotting
factors is the presence of γ-​carboxyglutamic acid (Gla) domains in
the N-​terminal region of the molecules. Glutamic acid residues in
these proteins undergo carboxylation, a post-​translational event
Table 22.7.1.7  Characteristics of coagulation proteins
Protein
Plasma concentration (µg/​ml)
Biological half-​life (h)
Chromosome
Vitamin K-​dependent zymogens
Prothrombin
100–​150
60–​70
11p11–​q12
Factor VII
0.5
3–​6
13q34
Factor IX
4–​5
18–​24
Xq27.1–​q27.2
Factor X
8–​10
30–​40
13q34
Protein C
4–​5
6
2q13–​q14
Nonvitamin K-​dependent zymogens
Factor XI
5
72
4q32–​q35
Factor XII
30
60
5q33
Prekallikrein
50
35
4q35
Factor XIII-​A chaina,b
10
240
6p24–​p25
Soluble cofactors
Factor Vb
5–​10
12
1q21–​q25
Factor VIII
0.1–​0.2
8–​12
Xq28
Von Willebrand factor
10
12
12p13.2
Protein Sc
25
42
3p11.1–​q11.2
Protein Z
2.9
?
13q34
High molecular weight kininogen
70
150
3q26
Factor XIII-​B chaina
1q31–​q32.1
Cellular cofactors
Tissue factor
–​
–​
1p21–​p22
Thrombomodulin
–​
–​
20p12–​cen
Structural protein
Fibrinogen
2000–​4000
72–​120
  Aα chain
4q23–​32
  Bβ chain
4q23–​q32
  γ chain
4q23–​q32
Inhibitors
Antithrombin
150–​400
72
1q23–​q25
Tissue factor pathway inhibitor
0.1
2q31–​q32.1
Protein Z-​dependent protease inhibitor (ZPI)
a All of the plasma factor XIII-​A chain is in complex with factor XIII-​B chain; only half of factor XIII-​B chain is in complex with factor XIII-​A chain, the rest is free in
plasma.
b Platelets carry significant amounts of factor XIIIA (roughly half of the total factor XIII activity) and factor V (20% of circulating factor V). The B chain of factor XIII
is not in platelets.
c Some protein S is in complex with C4b binding protein.
Reprinted by permission of McGraw-​Hill Companies from Roberts HR et al. (2001). Molecular biology and biochemistry of the coagulation factors. Williams
Hematology, 6th edn, p.1460.
22.7.1  The biology of haemostasis and thrombosis
5501
that is affected by hepatic carboxylase that requires the reduced
form of vitamin K as a cofactor. The vitamin K-​dependent factors
are highly homologous in terms of amino acid sequence. Factors
VII, IX, X, and protein C have a similar domain structure with a Gla
domain, two EGF domains, and a catalytic domain (Fig. 22.7.1.6).
Prothrombin differs from other vitamin K-​dependent factors in that
it has two kringle domains (Fig. 22.7.1.6). Both factor X and protein
C are secreted as two-​chain zymogens while the others are secreted
as single-​chain proteins.
The Gla domains of these factors are necessary for binding to
phospholipid membranes, such as the surface of activated platelets.
Calcium ions occupy the Gla domain to result in a conformational
change in the protein that favours binding to platelet membrane
surfaces. Phosphatidylserine is the major phospholipid in these
reactions.
The vitamin K zymogens are all serine proteases with the typ-
ical active site: a serine/​histidine/​aspartic acid triad. Exposure of
the active site requires that the zymogen be activated by cleavage
of specific arginyl residues. As a result, all the activated vitamin
K-​dependent zymogens become two-​chain enzymes linked by di-
sulphide bonds as depicted in Fig. 22.7.1.6. Despite the high degree
of sequence homology of these proteins, they are highly specific in
their interaction with their cofactors and substrates.
Prothrombin/​thrombin
Prothrombin is synthesized in the liver and has a molecular mass
of about 72 kDa. The molecule has 10 Gla residues that play a role
in the binding of prothrombin to the surface of activated platelets
where it is converted to the active enzyme, thrombin, by the so-​
called prothrombinase complex consisting of factors Xa/​Va/​Ca2+ on
the platelet surface. Thrombin is a potent enzyme with a molecular
mass of about 38 kDa that rapidly converts fibrinogen to a fibrin clot.
Thrombin also has many other actions including its role as a potent
activator of platelets; an activator of smooth muscle cells; an activator
of factor V, VIII, and XIII; an activator of protein C in the presence of
its cofactor thrombomodulin; an activator of procarboxypeptidase
to form thrombin-​activatable fibrinolysis inhibitor (TAFI); and as a
growth factor. The primary inhibitor of thrombin is AT.
Factor VII
Factor VII is synthesized in the liver and has a molecular mass
of about 50 kDa. It has a very short half-​life of 3.5 h. The specific
Prothrombin
Pre-pro
leader
GLA domain
Catalytic domain
B chain
Factor VII
Growth
factor
domains
Factor IX
Protein C
Factor X
Arg169-Leu
Growth
factor
domains
Activation
peptide
Activation
peptide
Pre-pro
leader
GLA domain
Growth
factor
domains
Growth
factor
domains
Kringle
domains
Catalytic domain
Catalytic domain
Pre-pro
leader
Pre-pro
GLA domain
leader
GLA domain
Pre-pro
leader
GLA domain
Catalytic domain
Activation
peptide
Catalytic domain
Arg180-Val
Arg145-Ala
Arg152-Ile
Arg 320-Ile
Arg271-Thr
Arg194-Ile
Fig. 22.7.1.6  Schematic diagram of the vitamin K-​dependent factors, prothrombin and factors VII, IX, X, and protein C. •, γ-​carboxyglutamic acid
residues; ♦, active site triad of serine, histidine, and aspartic acid; arrows denote cleavage site.
section 22  Haematological disorders
5502
receptor (and cofactor) for factor VIIa is TF, found on the surface of
many cells such as pericytes that surround small vessels, fibroblasts,
activated monocytes, and many other cell types. The EPCR is also a
specific receptor for factor VII.
Once bound to TF, factor VII must be activated for the complex to
be functional. The physiological activator of factor VII is unknown,
although it has been suggested that it might be activated factor
X. The factor VIIa–​TF complex activates both factors IX and X. The
factor VIIa–​TF–​Xa complex is inhibited by tissue factor pathway in-
hibitor (TFPI). Factor VIIa is not appreciably inhibited by AT except
in the presence of heparin.
Factor IX
Factor IX is synthesized by hepatocytes and has a molecular mass
of about 57 kDa. Its plasma half-​life is 18 to 24 h. The molecule
has 12 Gla residues. About 40% of the factor IX molecules carry a
β-​hydroxyaspartic acid at position 64 of the molecule. Factor IX
is activated by factor VIIa–​TF and by activated factor XI, both of
which cleave an arginyl bond at position 145 and 180 of the molecule
to release an activation peptide of about 10 kDa. Factor IXa, in com-
plex with its cofactor (activated factor VIII), cleaves factor X to Xa.
AT will inhibit factor IXa, but the inhibition is not as rapid as the AT
inhibition of thrombin or factor Xa.
Factor X
Factor X is also synthesized by hepatocytes and has a molecular mass
of 59 kDa. It is secreted as a two-​chain molecule linked by disul-
phide bonds and has 11 Gla residues. When activated by factor IXa
or factor VIIa–​TF, an activation peptide is cleaved from the heavy
chain to expose the active site serine on the heavy chain. Factor
Xa, in the presence of its cofactor (factor Va), rapidly converts pro-
thrombin to thrombin on the activated platelet surface. The primary
inhibitor of factor Xa is AT. TFPI also inhibits factor Xa.
Protein C
Unlike the other vitamin K-​dependent zymogens, protein C is not a
procoagulant, but, when activated by the thrombin–​thrombomodulin
complex on the surface of endothelial cells, it becomes an anticoagu-
lant by proteolysis of factors Va and VIIIa, thus inhibiting coagula-
tion. To function in this way as an anticoagulant, activated protein C
(APC) requires a nonenzymatic cofactor, protein S, which also con-
tains vitamin K-​dependent Gla residues. Protein C is synthesized in
the liver and has a very short half-​life of about 6 h. It contains nine Gla
residues and has a molecular mass of 59 kDa. The primary inhibitor
of APC is the protein C inhibitor (PCI).
Nonvitamin-​K-​dependent zymogens
Factor XI
Factor XI is synthesized in the liver as a dimeric protein composed of
identical subunits. It has a molecular mass of 160 kDa and a plasma
half-​life of about 72 h (Table 22.7.1.7). In plasma, factor XI circu-
lates in complex with high molecular weight kininogen (HK), a
nonenzymatic cofactor. The physiological activator of factor XI is
thought to be thrombin, although in vitro, it can also be activated
by factor XIIa. The main function of factor XIa is to boost thrombin
generation by activating factor IX on the surface of platelets, over
and above the factor IX activated by the VIIa–​TF complex. A few
patients with factor XI deficiency have virtually no bleeding ten-
dency, and those who do usually exhibit mild bleeding when com-
pared to severely affected haemophilic patients.
Factor XII and prekallikrein
These factors have been collectively referred to as contact factors
since it appears that activation of factor XII is enhanced by con-
tact with a surface. Factor XII and prekallikrein (PK) are zymo-
gens, which, when activated, expose an active site serine. HK is a
nonenzymatic protein cofactor that circulates in complex with
factor XI and PK. All of these factors are synthesized in the liver.
Unlike the vitamin K-​dependent proteins, factors XI, XII, and
prekallikrein all possess so-​called ‘apple domains’ that have specific
functional characteristics. Deficiencies of factor XII and PK are not
associated with bleeding tendencies in patients, even with complete
deficiency. However, deficiency of each factor is associated with a
marked prolongation of the partial thromboplastin time. In this test
and in the presence of glass, ellagic acid, or some inert earth ma-
terial, factor XII is activated and in the active conformation it can
activate factor XI. Factor XII, PK, and HK may not play a major
physiological role in haemostasis, but there is evidence that they
participate in inflammatory responses that involve blood coagula-
tion, fibrinolysis, and kinin generation. The precise role of factor XII
in coagulation reactions in vivo is not known. Despite the fact that
patients with factor XII deficiency do not exhibit bleeding symp-
toms, the factor is considered to be part of the ‘intrinsic system’ of
coagulation and in some instances it may contribute to haemostasis
by virtue of exposure to collagen or other surfaces. Studies in ani-
mals lacking factor XII suggest that the protease may play a role in
formation of pathological thrombi.
Factor XIII
Factor XIII is a proenzyme that circulates in the plasma as a
heterotetramer composed of two A chains and two B chains. Factor
XIII has a molecular mass of 320 kDa and a half-​life of about 10 days.
It circulates in plasma in association with fibrinogen. The A chain
contains the active site cysteine, while the B chain is enzymatically
inactive and serves as a carrier for the A chain. The A chain is found
in platelets where it is not associated with the B chain. Upon acti-
vation by thrombin, the A and B chains are separated. In addition,
thrombin cleaves the A chain so as to expose the active site cysteine.
The activated A chain then cross-​links the α and γ chains of fibrin to
form a stable, impermeable fibrin clot that is more resistant to lysis
by plasmin than noncross-​linked fibrin.
Cofactors
Some of the cofactors are soluble and exist in circulation, namely
protein S, protein Z, factors V and VIII, HK, and VWF. Others are
cell bound, such as TF and thrombomodulin (Table 22.7.1.7).
Protein S
Protein S is synthesized in the liver and endothelial cells and is de-
pendent on vitamin K for complete synthesis. It circulates in plasma
and is also found in platelets. It has a molecular mass of 75 kDa
and a plasma half-​life of about 42 h. It contains 11 Gla residues in
the N-​terminal region. In structure, protein S differs dramatically
from the other vitamin K clotting factors in that the C-​terminal
end is homologous to growth hormone. Protein S acts as a cofactor
22.7.1  The biology of haemostasis and thrombosis
5503
for activated protein C. Protein S exists in two forms: one form is
bound to C4b-​binding protein and the other exists as a free form in
the circulation and is in equilibrium with the bound form. It is the
free form of protein S that acts as a cofactor for protein C.
Protein Z
Protein Z is synthesized in the liver and has a molecular mass of
62 kDa. Protein Z functions as an inhibitor of factor Xa. When pro-
tein Z is incubated with factor Xa, the activity of the latter is re-
duced. The inhibition of factor Xa activity is due to the presence of
a protease inhibitor that requires protein Z as a cofactor. Whether
protein Z has other functions is unknown.
Factor V
Factor V is synthesized in the liver and has a biological half-​
life of between 12 and 36 h. It is a large glycoprotein with a mo-
lecular mass of 330 kDa. Factor V is highly homologous to factor
VIII, and is composed of A, B, and C domains. A schematic dia-
gram of the structure is shown in Fig. 22.7.1.7. The A domains
are homologous to the copper-​binding protein caeruloplasmin,
so it is not surprising that this domain of factor V is involved in
binding to calcium and copper. The C domains are homologous
to fat-​globule proteins and are involved in the binding of factor V
to phospholipid-​rich platelet membranes. The A and C domains
are homologous to similar domains in factor VIII, but the B do-
main is completely different from that of factor VIII. For factor V
to act as a cofactor for factor Xa, it must be activated by thrombin
with cleavage of arginyl bonds at positions 708, 1018, and 1545
as shown in Fig. 22.7.1.7. It is inactivated by APC, which cleaves
bonds at 306 (slow) and 506 (fast).
Factor VIII
Like factor V, factor VIII is synthesized in the liver. Factor VIII
mRNA is found largely in sinusoidal endothelial cells and Kupffer
cells, although there is significant extrahepatic synthesis of factor
VIII in endothelial cells. It is a large glycoprotein, similar in size
to factor V. Again, like factor V, factor VIII has A, B, and C do-
mains with the A domains homologous to caeruloplasmin and the
C domains homologous to fat globule proteins (Fig. 22.7.1.8). The
C domains of factor VIII are essential for binding to phospholipid
membranes. The B domain of factor VIII is proteolytically removed
during activation. To act as a cofactor for factor IXa, factor VIII
must be activated by thrombin or factor Xa. Unlike activated factor
V, activated factor VIII exists as a heterotrimer composed of A1,
A2, and A3–​C1–​C2 domains linked by calcium ions. Factor VIII
circulates in a noncovalent complex with VWF and has a biological
half-​life of 8 to 12 h. In the complete absence of VWF, such as oc-
curs with type III von Willebrand disease, the half-​life of factor VIII
is less than 1 h. When activated factor VIII is released from VWF,
it binds to the surface of activated platelets where it interacts with
factor IXa.
von Willebrand factor
VWF is synthesized by endothelial cells and stored as large and
ultra-​large multimer in Weibel–​Palade bodies. It also circulates in
plasma bound to factor VIII. It binds to glycoprotein Ib on platelets
and is required for normal platelet adhesion to components of the
vessel wall such as collagen. A schematic diagram of VWF is shown
in Fig. 22.7.1.9. Although synthesized as a prepolypeptide with A,
B, C, and D domains, it is secreted into the plasma in multimeric
form with molecular mass ranging from 1000 kDa to 15  000 to
20 000 kDa. Higher molecular mass forms of VWF are secreted to
the abluminal surface of the endothelial cell as one component of the
extracellular matrix. The higher molecular mass VWF multimers
are very effective in promoting platelet adhesion. VWF is also im-
portant in platelet aggregation. A major function of VWF is to act as
a carrier protein for factor VIII. Factor VIII is associated with VWF
multimers of all sizes.
Inactivation
Me
iVa
C2
2196
V
Va
C1
C2
C1
A3
A3
C1
C2
3
A
A2
A1
Activation
709
306
506
1018
1545
B
A1
A2
A2
A2
A1
Me
Fig. 22.7.1.7  Schematic diagram of factor V. Factor V is activated by
thrombin to factor Va. Factor Va is inactivated (iVa) by activated protein
C. Activation of factor V by thrombin results in loss of the B chain and
formation of a heterodimeric molecule covalently linked by metal ions
(Me). Inactivation is by activated protein C that cleaves arginyl bonds at
positions 306 and 506.
A1
B
A2
372
740
C2 C1
A3
A2
A1
Thrombin
A3
C1
1689
C2
ME++
Factor VIIIa
(activated)
Factor VIII
(unactivated)
Factor VIII
Fig. 22.7.1.8  Schematic representation of factor VIII. Activation by
thrombin (or factor Xa) results in a heterotrimer noncovalently linked
by metal ions (Me). Like factor Va, factor VIIIa is inactivated by activated
protein C.
section 22  Haematological disorders
5504
High molecular weight kininogen
HK circulates in plasma, and part is bound to factor XI and
prekallikrein. HK is a cofactor for both of these zymogens. Deficiency
of HK is not associated with a bleeding tendency, although the par-
tial thromboplastin times of affected subjects are prolonged.
Tissue factor
TF is a transmembrane cell surface protein. Soluble TF circu-
lating in plasma has been described but does not appear to be
functional. TF may also be found in microparticles but again its
functional significance has not been well established. It is com-
posed of 263 amino acids and with a 219-​amino acid extracel-
lular domain, a 23-​amino acid transmembrane domain, and a
21-​amino acid intracytoplasmic domain. The characteristics of TF
are shown in Table 22.7.1.7. It has a molecular mass of about 46 kDa
and is constitutively expressed on several extravascular tissues such
as fibroblasts and smooth muscle cells. It is not constitutively ex-
pressed on cells exposed to the circulating blood, but can be induced
in endothelial cells by certain inflammatory cytokines and certain
bacterial products such as endotoxin. It can also be induced in blood
leucocytes. TF functions as a receptor for factor VII. When factor
VII binds to TF, it is rapidly converted to factor VIIa, although the
precise mechanism for its activation is not clear. The VIIa–​TF com-
plex is now thought to be the main physiological initiator of blood
coagulation by activating both factor IX and factor X, each of which
plays a distinct role in subsequent coagulation reactions as described
below. On some cells TF exists in a ‘latent’ form, sometimes referred
to as ‘encrypted TF’, as suggested by the fact that TF antigen levels
on cells may be higher than TF functional activity. De-​encryption
can be accomplished by exposure of cells to agents such as calcium
ionophores and various cytokines, but the physiological mechanism
by which this process takes place is not known.
Thrombomodulin
Thrombomodulin is a transmembrane protein synthesized by and
localized to endothelial cells although it has also been found on
mesothelial cells, monocytes, and squamous epithelial cells. It has a
molecular mass of about 78 kDa. A chondroitin sulphate moiety is
attached to thrombomodulin via a serine residue. The major char-
acteristics of thrombomodulin are depicted in Table 22.7.1.7. It
serves as a receptor on endothelial cells for thrombin. Thrombin
bound to thrombomodulin undergoes a structural transform-
ation such that it no longer activates platelets or clots fibrinogen,
but rather activates protein C.  The principal function of the
D3
D4
D3
D4
D1
D2
D4
C1 C2 C3 C4 C5 C6 CK
D3
D’
D3
D4
D3
D4
D3
D4
D3
D4
D3
D4
D3
D4
D3
D4
D3
D4
D3
D4
D3
D4
A1
A2
A3
propeptide
propeptide
Propeptide
Mature VWF
Furin cleavage
Dimerization
Multimerization
Flow at pH 7.4
In plasma
Intracellular
Fig. 22.7.1.9  Schematic diagram of von Willebrand factor (VWF). Following synthesis, VWF
is cleaved by furin into a propeptide and mature VWF protein. These remain noncovalently
associated as VWF dimerizes and multimerizes in the cell. In the acidic environment of the late
Golgi, factor VIII associates with VWF and the propeptide dissociates. Dimerization occurs through
disulphide links in the CK domain. The formation of multimeric forms of VWF occurs through links
of dimeric VWF via D3 domains. In plasma, VWF may form elongated ‘strings’ after attachment to
the vessel wall.
22.7.1  The biology of haemostasis and thrombosis
5505
thrombomodulin–​thrombin complex is to prevent the extension of
the haemostatic clot past the site of a break or leak in the vessel wall
and as such represents an important control mechanism to restrict
the haemostatic plug precisely to the point of injury. Thus, under
normal conditions, clot formation does not occur on the endothe-
lial cell surfaces.
Fibrinogen
Fibrinogen is synthesized in the liver and has a molecular mass of
340 kDa. It is a dimeric glycoprotein consisting of two sets of iden-
tical chains, the α, β, and γ chains. The synthesis of each fibrinogen
chain is governed by a separate gene, as depicted in Table 22.7.1.7.
The normal plasma half-​life of fibrinogen is about 3 to 5 days. It is
also found in the α granules of platelets as a result of endocytosis.
Fibrinogen is the soluble plasma precursor of the solid fibrin clot
that is necessary for haemostasis and normal wound healing.
A schematic diagram of fibrinogen is shown in Fig. 22.7.1.10. It
is a trinodular structure with a central E domain that includes the
disulphide-​linked N-​termini of all six polypeptide chains. The E
domain is linked to the C-​terminal domains referred to as the D
domains. Fibrinogen conversion to fibrin is accomplished by
thrombin cleavage of two fibrinopeptides (fibrinopeptide A  and
fibrinopeptide B) from each of the two α and β chains, respectively,
leading to the formation of the fibrin monomer. The molecular mass
of each fibrinopeptide A and B is about 2500 Da. The soluble fibrin
monomer then undergoes spontaneous polymerization by forming
side-​to-​side and end-​to-​end interactions, resulting in protofibrils
that aggregate into a visible fibrin clot composed of thicker, branched
fibres. During fibrin clot formation, other proteins are occluded in
the clot, including plasminogen, fibronectin, thrombospondin,
and VWF. Fibrin polymerization is enhanced by calcium ions, but
the polymerization process alone does not lead to a stable and im-
permeable fibrin clot since the fibres are held together weakly by
hydrogen bonds and electrostatic forces. A stable fibrin clot requires
cross-​linking of the α and γ chains of fibrin by the action of activated
factor XIII.
Inhibitors of the coagulation reactions
(See Table 22.7.1.7.)
Tissue factor pathway inhibitor
TFPI is synthesized by endothelial cells. It has a molecular mass
of about 34 to 40 kDa and serves to inhibit the initiation of coagu-
lation. TFPI can inhibit factor Xa in a slow reaction and also in-
hibits the VIIa–​TF–​Xa complex. TFPI exists in two forms, TPFIα
and TFPIβ. The latter is directly anchored to cell surfaces via
glycosylphosphatidylinositol links. It exists in the circulation in at
least three pools. One is bound to plasma lipoproteins; one pool is
bound to proteoglycans on the vessel wall; and one exists in platelets.
The TFPI bound to proteoglycans can be released by heparin. TFPI
is a Kunitz-​type inhibitor that is essential for control of coagulation
at the initiation phase.
Antithrombin
AT belongs to a family of protease inhibitors known as serpins that
inhibit many proteases with a serine-​active site. It is synthesized in
the liver and has a plasma half-​life of approximately 65 h. Its major
function is to inhibit thrombin and factor Xa, although it will also
inhibit the other coagulation serine proteases less well. The inhibi-
tory action of AT is greatly enhanced by heparin, which accelerates
the rate of inhibition of the serine proteases.
Protein Z-​dependent protease inhibitor
This inhibitor inhibits factor Xa in the presence of calcium, phospho-
lipids, and protein Z.  It has a molecular mass of about 72 kDa.
Like AT, it is also a member of the serpin family of serine protease
inhibitors.
Other inhibitors of clotting factors
The major inhibitor of factor XIa is thought to be α1-​antitrypsin
since it has the highest affinity for the enzyme. However, other in-
hibitors, namely C-​1 esterase inhibitor, will also inhibit factor XIa.
The other inhibitors that are of some importance in coagulation are
also listed in Table 22.7.1.7.
Coagulation pathways
The coagulation reactions have been viewed as a sequential series
of steps in which a enzymatic precursor (zymogen) clotting factor
is converted to an active enzyme that in turn activates another pre-
cursor, finally ending in the rapid conversion of prothrombin to
thrombin. Early models of the coagulation reactions are shown in
Fig. 22.7.1.11. As can be seen, when viewed in this manner, the
clotting reactions appear as a waterfall or cascade, hence the terms
waterfall or cascade hypotheses. Since TF was extrinsic to the blood
stream, the activation of factor X by the VIIa–​TF complex was
termed the extrinsic pathway. The intrinsic pathway consisted en-
tirely of clotting factors within the circulation and, upon conver-
sion of factor IX to IXa by factor XIa, the factor IXa–​VIIIa complex
could also convert factor X to Xa in the presence of phospholipids.
Although this concept of coagulation was essentially correct, it did
not explain why patients with factor XII deficiency had no bleeding
tendency, nor why factor XI-​deficient patients had only a mild
bleeding tendency. It was also pointed out that defects in the in-
trinsic system could lead to haemorrhage in affected patients even
though the extrinsic system was intact and vice versa. The dem-
onstration that the VIIa–​TF complex could activate both factor IX
and factor X led several groups to conclude that the clotting reac-
tions were, in fact, initiated by factor VIIa–​TF and that separate
intrinsic and extrinsic systems did not exist in vivo. Further work
demonstrated that the clotting reactions leading to a haemostatic
plug were controlled in large part by cell surfaces which modulated
the reactions.
D
D
E
β
γ
α
α
β
γ
Fig. 22.7.1.10  Diagram of the structure of fibrinogen. The three chains,
α, β, γ, are shown. The E domain occurs at the N-​termini while the D
domains are found at the C-​termini. Arrows represent cleavage sites for
fibrinopeptide A from the α chain and fibrinopeptide B from the β chain.
section 22  Haematological disorders
5506
Role of the TF-​expressing cell
When a blood vessel is injured or ruptured, flowing blood is exposed
to TF, which is bound through its transmembrane domain and cyto-
plasmic tail to cells exposed as the result of injury, e.g. pericytes,
fibroblasts, and other connective tissue cells. Factor VII binds to the
TF-​bearing cell and is activated. In fact, it appears that the TF found
in pericytes surrounding small vessels is already saturated with
factor VII. As a result, the VIIa–​TF complex on the TF-​bearing cell
activates both factor IX and factor X as shown in Fig. 22.7.1.12a. The
factors Xa and IXa formed in the milieu of the TF-​expressing cell
play very different and distinct roles in subsequent reactions.
The role of factor Xa in the milieu of the TF cell
Factor Xa, in concert with its cofactor Va (which is found in the vicinity
of TF cells) converts prothrombin to very small amounts of thrombin,
as shown in Fig. 22.7.1.12b. This amount of thrombin, though insuf-
ficient to clot fibrinogen, can, however, act as a ‘primer’ of subsequent
coagulation reactions to accomplish the following: activate platelets;
activate more factor V; dissociate factor VIII from VWF and activate
factor VIII; and activate factor XI as shown in Fig. 22.7.1.12b. Factor
Xa alone and in complex with VIIa–​TF is then inhibited by TFPI. The
activated co-​factors resulting from the priming amount of thrombin
in the milieu of the TF cell then occupy binding sites on the activated
platelet as shown in Fig. 22.7.1.12b. Thus, the main function of factor
Xa formed as the result of the VIIa–​TF complex is to furnish a priming
amount of thrombin sufficient to initiate further subsequent reactions
which take place on the activated platelet surface.
Role of factor IXa activated by the factor VIIa–​TF complex
Factor IXa formed by the VIIa–​TF complex on the TF-​bearing cell
diffuses away from the TF cell and occupies a site on the activated
platelet adjacent to its co-​factor VIIIa (Fig. 22.7.1.12c). This factor
IXa then plays a primary role in the subsequent burst of thrombin
generation on platelet surfaces as noted in the next sections.
Role of the activated platelet
The activated platelet mass is the primary site of thrombin gener-
ation, which is highly dependent upon the amount of factor IXa
formed both by the VIIa–​TF cell and factor XIa, which also occu-
pies sites on the platelet. Factor IXa in the presence of its cofactor
VIIIa then recruits more factor X from solution and activates it
Intrinsic pathway
XII
XI
IX
IXa
XIa
PK
HK
XIIa
HK
VIIIa
VIIa
TF
Extrinsic pathway
X
Xa
Prothrombin
Fibrin
Fibrinogen
Thrombin
XIII
XIIIa
Cross-linked fibrin
X
Va
Fig. 22.7.1.11  An earlier model of blood coagulation reactions: the
cascade or waterfall hypothesis of coagulation.
X
Vlla
Xa
ll
lla
Va
TF
(a)
(b)
(c)
TF
Vlla
lX
lXa
TFPI
Xa
Xla
X
lXa
lXa
Xa
Va
Vllla
Xla
lX
Activated platelet
ll
lla
Vllla
Vlla
Vllla
Activated platelet
Platelet
Xa
lla
Vlll/VWF
Vlll + free VWF
Xl
Xla
V
ll
Va
Va
Va
V
TF
TF
Tissue factor-
bearing cell
Tissue factor-
bearing cell
Fig. 22.7.1.12  (a) Tissue factor (TF), a transmembrane protein
expressed on TF-​bearing cells, acts as a receptor for factor VII, which is
rapidly converted to factor VIIa. The TF–​VIIa complex then accomplishes
two major functions: (1) activation of factor X to Xa and (2) activation of
factor IX to IXa. Factor Xa activates factor V on the TF-​bearing cell and
the resulting Xa–​Va complex converts small amounts of prothrombin
to thrombin. (b) The small amount of thrombin formed in the vicinity
of the TF cell acts as a ‘primer’ for coagulation by (1) activating
platelets; (2) dissociating factor VIII from VWF and activating factor VIII;
(3) activating factor V; and (4) activating factor XI. The activated platelets
then adhere to the site of vascular injury and bind the cofactors, factors
VIIIa and Va. Factor XIa also binds to platelets. The TF–​VIIa–​Xa complex
is then inhibited by TFPI. (c) Factor IXa formed by the TF–​VIIa complex
associates with VIIIa on the platelet surface and recruits additional factor
X from plasma to form factor Xa. The factor Xa then associates with its
cofactor, factor Va, on the platelet surface to rapidly convert prothrombin
to large amounts of thrombin sufficient to clot fibrinogen.
22.7.1  The biology of haemostasis and thrombosis
5507
on the activated platelet surface. This factor Xa in the presence
of its cofactor Va then converts large amounts of prothrombin to
thrombin sufficient to clot fibrinogen. All of these reactions are
summarized in Fig. 22.7.1.13. The mass of aggregated platelets
upon which these reactions take place is localized to the damaged
area of the vessel wall.
Role of the endothelial cells, vessel wall, and inhibitors
The mass of platelets interspersed with fibrin forms a plug at the
site of a leak in the vessel wall where the endothelial cell mono-
layer is disrupted. The question arises as how the haemostatic plug
is confined to the damaged area of the vessel wall. A  schematic
diagram of these events is shown in Fig. 22.7.1.14. The endothe-
lial cells express thrombomodulin, which traps thrombin to form
a thrombomodulin–​thrombin complex that controls the procoagu-
lant stimulus by activating the protein C system, resulting in inacti-
vation of both factors Va and VIIIa, all localized essentially on the
endothelial cell surface. This series of events is enhanced by acti-
vated protein C on the EPCR. In addition, endothelial cells contain
glycosaminoglycans, some of which inhibit thrombin via AT. AT
also circulates in solution to inhibit any thrombin that escapes from
the haemostatic plug. In this way the fibrin clot sealing a leak in a
blood vessel wall is confined precisely to that site such that extension
of the clot does not occur under normal circumstances.
Ongoing coagulation in vivo
It is well known that products of the coagulation reactions are found
in the circulation under normal (basal) conditions. Small but def-
inite levels of fibrinopeptides A and B can be measured in plasma.
Fragment 1+2 derived from the N-​terminal portion of prothrombin
can also be detected after thrombin is formed. Activation peptides
from several of the coagulation factors as well as complexes of acti-
vated factors with their inhibitors can also be found in the circula-
tion. These observations strongly suggest that small leaks in blood
vessels that occur during the stress and strain of everyday living are
repaired by the ongoing formation of haemostatic fibrin clots. This
has been termed ‘basal’ coagulation, a process that allows the blood
to remain fluid within the vascular tree and at the same time per-
mits small, exquisitely controlled and confined fibrin clots to plug
small leaks in the vasculature without dissemination. The fibrin plug
is then removed by the fibrinolytic system following the formation
of new tissue.
Role of TF in microparticles
It has been shown that microparticles shed from leucocytes and
other cells possess procoagulant activity at least in vitro and perhaps
in vivo. These particles have also been shown to possess TF but the
physiological role of this TF in vivo has not been definitively dem-
onstrated. Platelets contain pre-​mRNA for TF and under certain
circumstances may express TF activity but again, the physiological
significance of these interesting observations has not been clearly
delineated.
Fibrinolytic system
The fibrinolytic system is shown schematically in Fig. 22.7.1.15. The
components of the system and their characteristics are depicted in
Fibrinogen
Fibrin
VIIa
VIII/VWF
Platelet
VIIIa
Activated platelet
VIIIa + free VWF
XI
XIa
Va
V
IIa
XIa
Va
Xa
II
TF
TFPI
Xa
Tissue factor-bearing cell
TF
VIIIa
Va
V
TF VIIIa
X
II
X
IX
IX
IXa
Xa
IIa
D
D
E
D
D
E
D
D
E
Fig. 22.7.1.13  The clotting reactions summarized. After thrombin
formation, fibrinogen is converted to fibrin.
Blood flow
IIa
iV
TF
Platelet plug
TM
Va
Endothelial cell
ATIII
PC
G
A
G
APC
PS
Endothelial cell
Subendothelium
TF
IIa
IIa
PC
IIa
G
A
G
ATIII
TM
TF
TF
VIIIa
APC
PS
iVIII
Fig. 22.7.1.14  Diagram of the haemostatic plug and the control
mechanisms that restrict this plug to the site of injury and prevent
extension of the clot to normal endothelium. APC, activated protein C;
GAG, glycosaminoglycans; IIa, thrombin; iVa, inactivated Va; PC, protein
C; platelet plug = haemostatic plug iVIIIa, inactivated VIIIa; PS, protein S;
TF, tissue factor; TM, thrombomodulin.
Plasminogen
FDP
Fibrin
T-PA
PAI
pro-MMP
ECM
degradation
U-PA:u-PAR
Plasmin
MMP
TIMP
PAI
α2-Antiplasmin
Fig. 22.7.1.15  The fibrinolytic system. Plasminogen is converted
to plasmin by activators, including plasminogen activators (tPA) and
urokinase (U-​PA). The activators are inhibited mainly by plasminogen
activator inhibitor-​1 (PAI-​1). Plasmin degrades fibrin and activates matrix
metalloproteinases (MMPs), which degrades extracellular matrix (ECM).
Plasmin is inhibited by antiplasmin. FDP, fibrin degradation product;
TIMP, tissue inhibitors of metalloproteinases; U-​PAR, urokinase protease-​
activated receptor.
section 22  Haematological disorders
5508
Table 22.7.1.8. The active enzyme in the fibrinolytic system is plasmin,
which is derived from its precursor, plasminogen. Plasminogen is ac-
tivated to plasmin by activators. The physiological activator is single-​
chain tPA, which cleaves plasminogen into two-​chain plasmin.
Another activator of plasminogen in vivo is single-​chain urokinase,
but this appears to be more important for degradation of matrix pro-
teins. The physiological inhibitor of plasmin is α2-​antiplasmin.
Plasminogen and tPA associate in the circulation with fibrinogen.
When fibrinogen is converted to fibrin, free lysine residues in fibrin
promote the binding and conversion of plasminogen to plasmin by
tPA. Fibrin-​bound plasmin is protected from the inhibitory action
of antiplasmin. Thus, clots can be lysed without interference from
inhibitors, yet free plasmin in the circulation will be rapidly in-
hibited by its inhibitor.
Plasminogen
Plasminogen is synthesized in the liver and has a molecular mass of
about 92 kDa. It is composed of a single chain and exists in two forms
in the circulation:  Glu-​plasminogen and Lys-​plasminogen. Glu-​
plasminogen has an N-​terminal glutamic acid and is larger than Lys-​
plasminogen, which is formed in the circulation by plasmin cleavage
of an arginyl bond at position 78 of the Glu form, leaving lysine as the
N-​terminal residue. Lys-​plasminogen rapidly binds to fibrin via lysine
binding sites. Thus, Lys-​plasminogen is in close proximity to fibrin and
protected from the action of antiplasmin. When activated, plasminogen
is converted to active two-​chain plasmin with a serine-​active site on the
heavy chain that is connected to the light chain by disulphide bonds.
The proteolytic action of plasmin is usually characterized by
the proteolysis of fibrinogen and fibrin, but it can also degrade
several other proteins including factor VIII, factor V, VWF, and
others. The cleavage of fibrinogen and fibrin leads to the formation
of fibrin(ogen) degradation products (FDP). Fibrin(ogen) frag-
ments resulting from plasmin cleavage are shown in Fig. 22.7.1.16.
Fragment X is the first and largest fragment of plasmin digestion
of fibrinogen. It is still clottable by thrombin, although much more
slowly than native fibrinogen. Fragment X gives rise to fragments Y
and D, and fragment Y is further proteolysed to give rise to a second
fragment D plus fragment E. These fragments can be detected in a
simple laboratory test using antifibrinogen antibodies coated to latex
particles. However, the test is nonspecific and does not distinguish
between the fibrinogen or fibrin degradation products which are
quite similar, since the only difference between fibrin and fibrinogen
is the absence of the small fibrinopeptides A and B in fibrin. A better
test for detection of fibrin fragments is the so-​called D-​dimer test,
which detects D-​dimers resulting from the cross-​linking of fibrin
by factor XIIIa.
tPA
tPA is considered to be the physiological activator of plasminogen.
It is synthesized in endothelial cells and has a molecular mass of
about 68 kDa. It has high affinity for plasminogen. tPA circulates
for the most part in complex with its inhibitor, plasminogen acti-
vator inhibitor-​1 (PAI-​1). The tPA–​PAI-​1 complex can be dissoci-
ated during the process of coagulation, and free tPA associates with
fibrin, which enhances tPA activity. Single-​chain tPA has catalytic
activity, but when activated to the two-​chain form by plasmin, the
activity is increased by threefold.
Plasminogen activator inhibitor-​1
PAI-​1 is the physiological inhibitor of tPA. It belongs to the serpin
family of inhibitors. It is synthesized in endothelial cells and has a
molecular mass of 52 kDa. Elevated levels of this inhibitor have been
associated with arterial and venous thromboses. PAI-​2, found in the
placenta, also inhibits tPA, but not as efficiently as PAI-​1. PAI-​3 is
also known as the protein C inhibitor and inhibits plasminogen ac-
tivators less efficiently than PAI-​1.
Urokinase plasminogen activator
Urokinase plasminogen activator (U-​PA) exists as a single-​chain
zymogen and is found in the kidney, the urine, and fibroblast-​
like cells. It activates plasminogen by proteolysis of an arginyl
residue at position 561. Its main function is in wound healing and
vasculogenesis, and it is active in proteolysis of the extracellular
Table 22.7.1.8  Characteristics of the fibrinolytic system
Plasma molecular
mass (kDa)
Concentration
(mg/​litre)
Chromosomal
location
Plasminogen
92
20
6
Plasmin
85
–​
tPA
68
0.005
8
U-​PA
54
0.008
10
α2-​Antiplasmin
70
200
18
PAI-​1
52
70
7
PAI-​2
47, 60
0.0518
18
u-​PAR
50, 60
Fragment X
Fragment Y
Fragment D
Fragment E
Fragment D
Bβ 1-42, Aα fragments
Fibrin(ogen)
D
E
D
Fig. 22.7.1.16  Plasmin digestion of fibrin(ogen) results in fibrin
degradation products: X, Y, D, and E. The final protolytic fragments
resulting from plasmin degradation of fibrin are two molecules of
fragment D and one of E.

22.7.2 Evaluation of the patient with a bleeding t
22.7.2 Evaluation of the patient with a bleeding tendency 5509 Trevor Baglin
22.7.2  Evaluation of the patient with a bleeding tendency
5509
matrix. U-​PA associates with the urokinase plasminogen-​activator
receptor (u-​PAR).
Antiplasmin
Antiplasmin is the physiological inhibitor of plasmin. It has a mo-
lecular mass of about 58 kDa and is synthesized in the liver. As an in-
hibitor, it has three major functions: to inhibit plasminogen binding
to fibrin; to inhibit the proteolytic activity of plasmin; and to bind to
fibrin in a covalent manner by the action of factor XIIIa. By binding
to fibrin, antiplasmin competitively inhibits the binding of plas-
minogen to fibrin. However, when plasminogen within the fibrin
clot is converted to plasmin, the latter is protected from inhibition
by antiplasmin. On the other hand, free plasmin formed in the cir-
culation is rapidly inhibited.
TAFI
TAFI is also known as plasma procarboxypeptidase B, and it is ac-
tivated to carboxypeptidase B by large amounts of thrombin in a
reaction dependent upon thrombomodulin. TAFI down-​regulates
fibrinolysis after clot formation and serves as an important regula-
tory mechanism for the fibrinolytic system. TAFI acts primarily by
reducing the number of high-​affinity plasminogen binding sites on
fibrin, the end result of which is decreased fibrinolysis.
The fibrinolytic and coagulation systems are closely interrelated.
Under normal conditions, fibrin clot formation is always accom-
panied by fibrinolysis. The formation of the fibrin clot that contains
both tPA and plasminogen results in formation of plasmin within
the clot so that clot lysis eventually ensues. It also appears that ac-
tivated factor XI and factor XII enhance fibrinolytic activity. The
action of the protein C system to decrease thrombin formation
down-​regulates the TAFI which would favour increased fibrin-
olysis. Although much is still unknown, it is generally accepted that
both the coagulation and fibrinolytic systems are related to the gen-
eral process of inflammation involving several other host defence
mechanisms.
FURTHER READING
Cines DB, et  al. (1998). Endothelial cells in physiology and in the
pathophysiology of vascular diseases. Blood, 91, 3527–​61.
Collen D (1999). The plasminogen (fibrolytic) system. Thromb
Haemost, 82, 259–​70.
Coughlin SR (2005). Protease-​activated receptors in haemostasis,
thrombosis and vascular biology. J Thromb Haemost, 3, 1800–​14.
Crawley JT, et al. (2007). The central role of thrombin in haemostasis. J
Thromb Haemost, 5 Suppl 1, 95–​101.
Degen JL, Bugge TH, Goguen JD (2007). Fibrin and fibrinolysis in in-
fection and host defense. J Thromb Haemost, 5 Suppl 1, 24–​31.
Gailani D, Renne T (2007). The intrinsic pathway of coagulation: a
target for treating thromboembolic disease? J Thromb Haemost, 5,
1106–​12.
Griffin JH, et al. (2007). Activated protein C. J Thromb Haemost, 5
Suppl 1, 73–​80.
Heijnen H, van der Sluijs P (2015). Platelet secretory behaviour: as di-
verse as the granules ... or not? J Thromb Haemost 13, 2141–​51.
Lundblad RL, White GC 2nd (2005). The interaction of thrombin with
blood platelets. Platelets, 16, 373–​85.
Ma YQ, Qin J, Plow EF (2007). Platelet integrin alpha(IIb)beta(3): acti-
vation mechanisms. J Thromb Haemost, 5, 1345–​52.
Monroe DM, Key NS (2007). The tissue factor-​factor VIIa com-
plex: procoagulant activity, regulation, and multitasking. J Thromb
Haemost, 5, 1097–​105.
Mosesson MW (2005). Fibrinogen and fibrin structure and functions.
J Thromb Haemost, 3, 1894–​904.
Nemerson Y (2007). My life with tissue factor. J Thromb Haemost,
5, 221–​3.
Nurden AT (2005). Qualitative disorders of platelets and megakaryo­
cytes. J Thromb Haemost, 3, 1773–​82.
Peake I, Goodeve A (2007). Type 1 von Willebrand disease. J Thromb
Haemost, 5 Suppl 1, 7–​11.
Pober JS, Sessa WC (2007). Evolving functions of endothelial cells in
inflammation. Nat Rev Immunol, 7, 803–​15.
Roberts HR, Hoffman M, Monroe DM (2006). A cell-​based model of
thrombin generation. Semin Thromb Haemost, 32 Suppl 1, 32–​8.
Roberts HR, Monroe DM, Hoffman M (2006). Molecular biology and
biochemistry of the coagulation factors and pathways of haemo-
stasis. In: Lichtman MA et al. (eds) Williams hematology, 7th edi-
tion, pp. 1665–​93. McGraw-​Hill, New York.
White GC II, Rompietti R (2007). Platelet secretion: indiscriminately
spewed forth or highly orchestrated? J Thromb Haemost, 5, 2006–​8.
22.7.2  Evaluation of the patient
with a bleeding tendency
Trevor Baglin
ESSENTIALS
An apparent bleeding tendency is a common clinical problem, with
presentation varying from acute unexpected bleeding during or im-
mediately after surgery or dental extraction, to spontaneous unusual
or excessive bruising, purpura, epistaxis, or a chronic haemorrhagic
tendency. Long-​standing bleeding symptoms suggest a lifelong con-
dition, whereas recent-​onset bleeding suggests an acquired disorder.
If a bleeding disorder has been diagnosed and characterized in an-
other family member, then the cause of bleeding may be easily iden-
tified, but the absence of a family history does not exclude a heritable
disorder. The commonest cause of an acquired bleeding disorder is
antithrombotic therapy.
Investigations for bleeding disorder include full blood count and
film (severe bleeding rarely occurs in the absence of trauma with a
platelet count of more than 20 to 30 × 109/​litre), prothrombin time
(PT), activated partial thromboplastin time (APTT), fibrinogen level,
reptilase time (useful for determining if a prolonged APTT is due to
heparin), individual factor assays, mixing studies (can indicate if pro-
longation of PT or APTT is likely due to a factor deficiency or an in-
hibitor), platelet function analysis, and (rarely) bleeding time.
Aside from general supportive care, specific therapy can be
given when a defined haemostatic abnormality is identified. Drugs
that cause bleeding should be stopped. Overanticoagulation due
to a vitamin K antagonist can be reversed with vitamin K and/​or
prothrombin complex concentrate; dabigatran and be reversed
section 22  Haematological disorders
5510
with idarucizumab; and factor Xa-​inhibitors may be reverse with
andexanet alfa where this is approved for use. Vitamin K should also
be given to critically ill patients and those with liver disease. Early
and sufficient blood product support should be given to those with
massive blood loss and/​or dilutional coagulopathy. Judicious use of
fresh frozen plasma and platelets is required in patients with severe
coagulopathy such as disseminated intravascular coagulation while
the underlying condition is being treated. Patients with overt haem-
atological disorders will require specialist care.
Introduction
An apparent bleeding tendency is a common clinical problem.
A comprehensive history is needed to assess the nature and extent
of the bleeding, to guide clinical examination, and to logically pri-
oritize investigations. An acquired bleeding tendency is much more
common than a heritable genetic disorder and the most common
cause of an acquired bleeding disorder is antithrombotic therapy,
particularly oral anticoagulant therapy. Some patients who bleed ab-
normally during or after surgery have a mild underlying heritable
haemostatic defect and so an important aspect of assessment is to
determine if there is a heritable defect with late clinical presentation.
Effective treatment depends on identifying the underlying cause of
bleeding and anticipating when it is likely to be of clinical import-
ance so that preventive therapy can be given at times of risk to pre-
vent excess bleeding.
Clinical assessment
Presentation of a bleeding disorder varies from acute unexpected
bleeding during or immediately after surgery to spontaneous un-
usual or excessive bruising, purpura, epistaxis, or other bleeding
developing over several months. With increasing use of pharma-
cological thromboprophylaxis in both medical and surgical in-
patients and increasing indications for long-​term antithrombotic
therapy, it is imperative to consider drug-​induced bleeding
during the initial evaluation of abnormal bleeding. In all cases
a comprehensive history is needed. Most heritable disorders
of haemostasis are mild, for example, von Willebrand’s disease
(VWD), and abnormal bleeding may not become manifest until
there is a haemostatic challenge, such as surgery or menstruation.
Therefore, an important aspect of the assessment of a patient with
an apparent acquired bleeding disorder is to determine if it is
genuinely of recent onset. The major issues to be determined are
as follows:
•	Is haemostatic capacity reduced or is there a nonhaematological
cause for bleeding?
•	If haemostatic capacity is reduced, is it due to a heritable defect
with late clinical presentation or is it the result of a newly acquired
defect?
•	If newly acquired, is it due to an anticoagulant or antiplatelet
drug?
•	If not due to reduced haemostatic capacity then what are the
likely circumstances that resulted in abnormal bleeding?
Taking a history
Assessing haemostatic capacity
The main purpose of the history is to establish a clinical impres-
sion, or profile, which indicates whether haemostatic capacity is
normal or not and if there is a likely explanation for bleeding. Simple
bruising, epistaxis, or menorrhagia has low positive and negative
predictive value in isolation and unless a systematic history is taken
with consideration of the ‘overall picture’ one can easily be misled
into excluding or misdiagnosing a disorder. Rarely is it possible to
absolutely exclude a disorder or make a positive diagnosis exclu-
sively on the historical information. However, the history is critical,
for example, there are some patients who definitely bleed after sur-
gical challenges and yet the mechanism of their bleeding disorder
cannot be characterized by laboratory investigations. Nevertheless,
such patients should be considered to have a bleeding disorder and
empirical treatment to reduce bleeding before surgical procedures
should be planned. In contrast, patients with borderline abnormal
laboratory tests but with no clinical bleeding tendency at times of
haemostatic challenge should not be diagnosed as having a bleeding
disorder. This scenario is frequently encountered when interpreting
von Willebrand protein and factor XI results as there is a poor cor-
relation between these factor levels and bleeding tendency in indi-
viduals from families with a familial bleeding tendency. Recurrent
bleeding from a single site suggests a structural abnormality.
Bleeding at many different sites suggests a haemostatic defect.
Severity and pattern of bleeding
The circumstances of the bleeding episode should be carefully as-
sessed. Was bleeding spontaneous or provoked, for example, by
trauma or surgery? Was the degree of bleeding excessive or the
pattern of bleeding unusual? Did bleeding result in anaemia, pos-
sibly requiring transfusion? Was the site of bleeding unusual, for
example, a joint bleed in an adult with no previous history of ab-
normal bleeding? Was there bleeding with previous haemostatic
challenges such as surgery, trauma, dental extraction, and menstru-
ation? Bleeding symptoms over a long time period suggest a lifelong
bleeding condition while recent-​onset bleeding suggests an acquired
disorder, although a mild lifelong condition can be unmasked at
times of haemostatic stress such as surgery.
Purpura
Purpura describes bleeding into the skin. The extent of bleeding may
be small (petechiae) or larger (bruising, also called ecchymoses).
Bruising is common in patients with reduced haemostatic capacity
but is also very common in patients who have no apparent defect
of haemostasis. Very extensive bruising, particularly over soft areas
that are not likely to be traumatized, and very large bruises in the
absence of trauma (>5 cm diameter), are more likely in patients with
reduced haemostatic capacity. Bruising over bony areas does not
necessarily indicate an abnormality and many normal children fre-
quently have several bruises over their knees and shins.
When bruising is the result of a bleeding disorder, the pattern of
bleeding may be suggestive of the type of underlying disorder, for
example, ecchymoses suggesting coagulation factor deficiencies
such as classical haemophilia or liver disease and petechiae sug-
gesting thrombocytopenia or a vessel wall defect. However, these
22.7.2  Evaluation of the patient with a bleeding tendency
5511
distinctions are not absolute; for example, thrombocytopenia may
present with large bruises rather than petechiae. Thrombocytopenia
causes petechiae, typically when the platelet count is less than 20
× 109/​litre. Petechiae may occur when there is platelet dysfunction
or with vascular purpura, either of which can be congenital or ac-
quired. Petechiae are unusual with low von Willebrand protein levels
but may occur in some individuals with low levels when antiplatelet
drugs are prescribed or when there is an increase in hydrostatic pres-
sure, for example, in the arm after application of a blood pressure
cuff (Hess’ test or Rumpel–​Leede test).
Epistaxis
Nosebleeds are common in children in the absence of a bleeding dis-
order. They also occur in some adults with allergic rhinitis. Repeated
bleeding from the same nostril suggests a local cause. Lifelong re-
current epistaxis can occur in VWD, haemophilias, and hereditary
haemorrhagic telangiectasia (Osler–​Weber–​Rendu syndrome).
Recent-​onset epistaxis in adults may be due to an acquired disorder
such as thrombocytopenia but is surprisingly uncommon in adults
with lifelong bleeding disorders such as the haemophilias and VWD.
Gingival bleeding
Gum bleeding in the absence of any other abnormal bleeding is usu-
ally due to gingivitis requiring improved dental hygiene. Rarely is
isolated gingival bleeding due to an underlying disorder.
Menorrhagia
Menorrhagia is a common problem and not specific for a bleeding
disorder. Recent-​onset menorrhagia in older women is likely to be
due to a gynaecological cause. The main problem in assessing men-
orrhagia is subjectivity. For example, many women with VWD do
not complain of heavy periods because they consider their own ex-
perience as ‘normal’. It is important to determine the pattern, for ex-
ample, bleeding for several days with clots and particularly bleeding
that interrupts normal lifestyle such as absence from work, is more
likely to be abnormal and may be due to an underlying haemostatic
defect. Very prolonged menorrhagia, rather than heavy bleeding, is
more likely to be gynaecological.
Dental extraction
Bleeding after dental extraction is variable in the normal population.
Bleeding lasting more than an hour, rebleeding, or late bleeding re-
quiring suturing on more than one occasion suggests a bleeding dis-
order. Bleeding after extraction requiring blood transfusion even on
one occasion suggests a bleeding disorder. Prolonged bleeding after
dental extraction is typical with low von Willebrand factor (VWF)
levels and rebleeding is typical of haemophilia.
Surgery
It is important to ask specifically about all operations including cir-
cumcision and tonsillectomy. Abnormal bleeding during surgery or
in the postoperative period may be due to antithrombotic therapy,
such as warfarin or aspirin, that was not stopped. Abnormal sur-
gical bleeding may be the first presentation of a mild or moderate
heritable bleeding defect if there has been no previous haemostatic
challenge, for example, a male having their first operation. It is useful
to try and determine if the bleeding was just local, which might be
due to a local anatomical reason such as a failed suture, or if there
was evidence of more generalized bleeding such as oozing from the
wound or bruising at venepuncture or venflon sites. Increasingly,
patients are prescribed low-​dose heparin (nowadays usually low
molecular weight heparin) to prevent venous thrombosis and in a
minority of patients this can unmask a mild bleeding tendency, such
as that associated with low VWF levels. In most patients, low-​dose
heparin does not appreciably increase surgical bleeding. However,
it is important to review the drug charts from the time of the oper-
ation to ensure that the correct dose of heparin was given and that no
other drugs that might cause bleeding were administered.
Bleeding in unusual sites
Bleeding in an unusual site sometimes suggests a specific diag-
nosis. Joint bleeding rarely occurs when there is normal haemostatic
capacity. It usually indicates severe coagulation factor deficiency,
such as severe congenital factor VIII or IX deficiency or overdose
with an oral vitamin K antagonist (VKAs) with an international
normalized ratio (INR) in excess of 8.0. Umbilical stump bleeding
in the neonate is typical of severe congenital factor XIII deficiency,
although the condition is very rare (one per million of the popula-
tion). Intracerebral bleeding in an otherwise healthy neonate neces-
sitates exclusion of severe congenital factor VIII or IX deficiency,
factor XIII deficiency and severe thrombocytopenia such as occurs
in fetomaternal alloimmune thrombocytopenia (FNAIT).
Family history of bleeding
If a bleeding disorder has been diagnosed and characterized in an-
other family member then the cause of bleeding may be easily iden-
tified. The absence of a family history does not exclude a heritable
disorder. The penetrance of VWD is incomplete, and new mutations
account for one-​third of new patients with haemophilia A.
Drug history
The most common cause of an acquired bleeding disorder is
antithrombotic therapy. Increasingly, low-​dose heparin for inpatient
thromboprophylaxis and oral anticoagulant (warfarin, dabigatran,
rivaroxaban, apixaban, edoxaban) and antiplatelet drugs (aspirin,
clopidogrel, prasugrel, ticagrelor, cangrelor, nonsteroidal anti-​
inflammatory drugs) in both inpatients and outpatients are respon-
sible for bleeding.
Clinical examination
Skin
The skin should be inspected in its entirety for evidence of bleeding,
noting the distribution (bony or soft areas), pattern (random or sug-
gestive of nonaccidental injury), and size (petechiae or ecchymoses).
Senile purpura occurs predominantly on the extensor surfaces of
the hands and arms and the face. The lesions tend to persist for sev-
eral weeks becoming increasingly dark. Senile purpura is due to skin
atrophy and resultant blood vessel fragility. Senile purpura is not as-
sociated with an underlying systemic bleeding disorder.
Purpura occur with amyloid, which may cause bleeding due
to capillary fragility as a result of amyloid infiltration, or rarely
an acquired deficiency of factor X.  Vessels are extremely fra-
gile and bleed with very minor trauma. Amyloid may also cause
section 22  Haematological disorders
5512
proteinuria and excess bleeding may complicate renal biopsy in
these patients. Some of these patients also have myeloma, causing
thrombocytopenia.
Petechiae with a perifollicular distribution occur in scurvy,
which may also present with more widespread bleeding into
joints, gastrointestinal bleeding, and intracerebral bleeding.
Other features include xerostomia, keratoconjunctivitis sicca, and
hyperkeratosis. Dental decay is common in patients with scurvy.
Treatment with vitamin C results in improvement within hours.
Scurvy may be the cause of bleeding in elderly patients with a very
poor diet.
Purpura occurs with infections including meningococcal septi-
caemia and diphtheria, chickenpox, measles, and the haemor-
rhagic fevers of Ebola virus and Lassa fever. Purpura fulminans
describes necrotic skin lesions which occur with overwhelming
infection and the development of disseminated intravascular
coagulation (DIC).
Allergic purpura may follow exposure to chemicals and toxins.
Henoch–​Schönlein purpura is the most common allergic purpura
and involves principally skin, joints, gastrointestinal tract, and kid-
neys. It typically occurs in children after an upper respiratory tract
infection due to streptococcus. The rash consists of purpuric papules
over the shins, thighs, and buttocks, sometimes with small ulcers, and
the rash is associated with arthritis, nephritis, and abdominal pain.
IgA-​containing immune complexes are deposited in the vessel walls.
Mixed cryoglobulinaemia in patients with hepatitis C infection
can produce extensive purpura in association with arthropathy and
glomerulonephritis.
Psychogenic purpura refers to unexplained bruising with pre-
ceding pain in association with anxiety. It has also been referred to
as ‘auto-​erythrocyte sensitization’ following reports that subcuta-
neous injection of the patient’s own red cells can induce the lesions.
However, it is uncertain if this is a genuine clinical sign.
Telangiectasia may occur in the skin and the mucous membranes.
In patients with hereditary haemorrhagic telangiectasia, they occur
predominantly in the skin of the hands and fingertips and are evident
on the lips. The lesions blanch on pressure in contrast to purpura.
Telangiectasias also occur in pregnancy and liver disease, usually on
the face and chest. Rarely, large cavernous haemangiomas or aortic
aneurysms can cause local consumption of coagulation factors and
platelets resulting in a systemic bleeding disorder.
Skin hyperelasticity, scars, papules, and plaques may indicate a
collagen vascular disorder. Ehlers–​Danlos syndrome, Marfan syn-
drome, pseudoxanthoma elasticum, and osteogenesis imperfecta
are associated with a bleeding tendency due to abnormal platelet–​
vessel wall collagen interaction. Unusual scars may be due to a
dysfibrinogenaemia, Ehlers–​Danlos syndrome, or pseudoxanthoma
elasticum. Long-​term steroid therapy and Cushing disease cause
skin atrophy and bruising typically on the extensor surfaces of the
hands and arms and on the thighs.
Mucosa
Telangiectasias are dilated small vessels that may be found in the
skin and in the mucous membranes of the respiratory, gastrointes-
tinal, and urinary tracts, vagina, eye, liver, and brain in patients
with hereditary haemorrhagic telangiectasia (Osler–​Weber–​Rendu
syndrome). Recurrent epistaxis and gastrointestinal bleeding cause
iron deficiency.
Musculoskeletal
Severe haemophilia A (factor VIII deficiency) and B (factor IX defi-
ciency) are characterized by repeated spontaneous bleeds into joints,
muscles, and soft tissue. The most common joints that bleed are the an-
kles, knees, hips, and elbows. Acute haemarthrosis presents as an acutely
swollen painful joint resulting in joint immobilization. Repeated bleeds
into a joint (target joint) produces chronic haemophilic arthropathy
with features of both osteoarthritis (mechanical pain on movement)
and rheumatoid arthritis (inflammatory pain at rest). Muscle haema-
tomas occur in the iliopsoas, gluteal, calf, and forearm muscles and are
more insidious than joint bleeds. Compartment syndrome can com-
plicate large bleeds and fibrosis and contractures produce dysfunction
and deformity. Large haematomas can cause pseudotumours, par-
ticularly when there is chronic rebleeding. These large, expanding soft
tissue cysts produce mass effects including neuropathy and bone ero-
sion and may produce chronic fistulas.
Splenomegaly
Splenomegaly can cause hypersplenism with thrombocytopenia.
Patients with myeloproliferative disorders may have impaired platelet
function. Essential thrombocythaemia is a myeloproliferative dis-
order which is particularly associated with impaired platelet func-
tion but both bleeding and thrombosis occur. In patients with very
high platelet counts, there can be increased consumption of von
Willebrand protein causing an acquired von Willebrand syndrome.
Patients with polycythaemia are particularly prone to chronic
gastrointestinal bleeding. Splenomegaly may be due to portal hyper-
tension in patients with liver disease. In these patients, bleeding
may be due to thrombocytopenia, platelet dysfunction, coagulation
factor deficiency, and production of dysfunctional factors. In add-
ition, there may be local bleeding sites such as oesophageal varices.
General aspects of examination
In addition to identifying signs that may indicate the likelihood
and type of a bleeding disorder, it is important to consider broader
aspects. For example, is a patient anaemic due to iron deficiency
as a consequence of chronic gastrointestinal blood loss? Patients
with severe bleeding disorders who have been treated with human-​
derived blood products, in particular pooled products that have
not been virally inactivated, may have chronic infections including
hepatitis C and HIV. Chronic liver disease, opportunistic infections,
and other complications may be evident. Bleeding as a result of liver
failure, renal failure, or paraproteinaemia should be considered. In
children, the possibility of nonaccidental injury should be con-
sidered, particularly with multiple bruises around the head and
neck, or a pattern of bruising in keeping with gripping or shaking.
The retina should be examined for haemorrhages. In a drowsy pa-
tient or a patient with raised intracranial pressure or an acute focal
neurological deficit, there is the possibility of an intracranial bleed.
Investigations
Laboratory tests
Coagulation tests include:
•	prothrombin time (PT)
•	activated partial thromboplastin time (APTT)
22.7.2  Evaluation of the patient with a bleeding tendency
5513
•	fibrinogen level
•	thrombin time (TT)
•	reptilase time (RT)
•	factor assays
•	mixing studies
•	platelet function analysis
Coagulation tests are typically performed on plasma that has been
separated from a venous blood sample by centrifugation. Thrombin
generation takes place on phospholipid surfaces (provided by plate-
lets normally) and so an artificial lipid preparation is added as the
platelets are removed by the centrifugation. Most routine clotting
tests use the time taken for a clot to appear as the endpoint of the
assay. Blood is usually taken into tubes containing citrate, which che-
lates calcium and thereby prevents clotting. After centrifugation, the
plasma is isolated and recalcified during the clotting assay. Clotting
tests are indicated in patients with a personal or family history of
bleeding. They are not generally indicated as routine preoperative
screening tests as they have very low sensitivity and specificity for
surgical bleeding in unselected patients. Furthermore, a prolonged
APTT on a ‘screening sample’ is most likely to be due to contact
factor deficiency or an incidental lupus anticoagulant, neither of
which will cause bleeding in a patient with no personal bleeding his-
tory but may lead to an unnecessary delay in surgery or an inter-
ventional procedure. Preoperative assessment of bleeding risk is
better determined by identification of a personal or family bleeding
history, which should then be investigated accordingly with specific
tests including coagulation factor assays, von Willebrand protein
level and function, and platelet count and function.
Assessment of haemostasis
The history is of primary importance in determining if haemostatic
capacity is genuinely reduced. A ‘bleeding score’ derived from that
used to quantify bleeding in patients with VWD can be a useful tem-
plate for ensuring that a systematic history is taken from patients
with an apparent bleeding tendency (Table 22.7.2.1). There is no ab-
solute score that indicates an unequivocal diagnosis of a bleeding
disorder but the higher the score, the more likely there is an under-
lying tendency to bleeding.
Blood count and film examination
Bleeding tendency increases with thrombocytopenia with a platelet
count of less than 80 × 109/​litre but severe bleeding rarely occurs in
the absence of trauma with a platelet count greater than 20 to 30 ×
109/​litre. Film examination is mandatory when a bleeding disorder
is suspected. Pseudothrombocytopenia due to platelet clumping
can lead to an erroneous diagnosis of true thrombocytopenia if
the film is not examined for clumps. Conversely, a normal platelet
count may be occasionally reported in patients with true thrombo-
cytopenia, the normal count being an artefact, for example, due
to the presence of a cryoglobulin. In these patients, the thrombo-
cytopenia is readily apparent from examination of the blood film.
Abnormal platelet morphology may suggest an underlying heritable
platelet defect such as Bernard–​Soulier syndrome, May–​Hegglin
abnormality (MYH9-​related disease), or a grey platelet syndrome.
Alternatively, a leukaemia or an acquired myeloproliferative dis-
order or myelodysplastic syndrome may be apparent from the blood
count and film examination.
Prothrombin time
The PT is the time taken in seconds for a fibrin clot to form after
recalcification and addition of thromboplastin (a preparation of tissue
factor which is the protein that activates factor VII). The normal PT
is about 11 to 14 s, depending on the type of thromboplastin used.
The PT is prolonged by:
•	oral anticoagulant therapy (typically warfarin)
•	direct oral anticoagulants (DOACs; dabigatran, rivaroxaban,
apixaban, edoxaban)
•	vitamin K deficiency
•	liver disease
•	DIC
•	dilutional coagulopathy (massive blood transfusion)
Anticoagulant therapy with oral VKAs is monitored by the INR.
The INR is derived from the PT ratio and is a standardized method
of reporting which permits comparability between laboratories. It is
inappropriate to use the INR for any purpose other than measuring
the intensity of oral anticoagulant therapy with a VKA (such as war-
farin|). For all other purposes, the PT or PT ratio should be used.
The PT and APTT are variably prolonged by DOACs with the PT
being more sensitive to factor Xa inhibitors (rivaroxaban, apixaban,
edoxaban) and the APPT more sensitive to thrombin inhibitors
(dabigatran). However, these tests cannot be used to quantify the
anticoagulant effects of DOACs and should be considered qualita-
tive at best. The INR is not used to monitor these drugs as it is for
VKA. Laboratories should be aware of the sensitivity of their own
assays to each drug. It is likely that coagulation tests will be per-
formed on patients taking DOACs as part of clinical assessment (e.g.
admission to the accident and emergency department), and the re-
sults might wrongly be interpreted if it is not known that the patient
has taken one of these drugs.
Activated partial thromboplastin time
The APTT is the time taken in seconds for a fibrin clot to form after
recalcification and exposure to a contact factor activator, such as
kaolin. The normal APTT is about 32 to 38 s, depending on the type
of contact factor activator used.
The APTT is prolonged by:
•	unfractionated heparin (low molecular weight heparin has min-
imal effect at therapeutic levels)
•	oral anticoagulant therapy (variably)
•	vitamin K deficiency (mildly, the PT is more sensitive)
•	liver disease (mildly, the PT is more sensitive)
•	DIC
•	dilutional coagulopathy (massive blood transfusion)
•	severe and moderate deficiencies of clotting factors VIII, IX, or XI.
•	lupus anticoagulant activity due to antiphospholipid antibodies
•	contact factor deficiency (including factor XII and prekallikrein,
none of which cause bleeding)
A low factor XII level produces a marked prolongation of the APTT
but deficiency is not associated with any bleeding tendency. This re-
flects the fact that the APTT is a nonphysiological test which while
useful for screening for deficiency of some clotting factors (VIII, IX,
and XI) does not actually reflect physiological haemostasis.
section 22  Haematological disorders
5514
Table 22.7.2.1  International Society on Thrombosis and Hemostasis Bleeding Assessment Tool
Symptom
−1
0
1
2
3
4
Epistaxis
Number episodes/​year
Average duration
Medical attention required?
<5 and <10 min
5 or >10 min
Required medical
consultation
Packing or cauterization or
antifibrinolytic therapy
Blood transfusion/​replacement
therapy or DDAVP
Bruising
Exposed or unexposed sites
Size on average
Minimal or no trauma
Specify medical attention
No or trivial
(<1 cm)
1 cm and no trauma
Medical
consultation
required
Bleeding from
minor wounds
Number per year
Duration of average episode
Medical attention required
No or trivial
(<5/​year)
5/​year or episodes
average >5 min
Consultation only
Surgical haemostasis
Blood transfusion/​replacement
therapy or desmopressin
Oral cavity bleeding
Tooth eruption
Gums, spontaneous
Gums, after brushing
Bites to lip and tongue
Medical attention required
No
Bleed with at least one of:
Tooth eruption
Gums (spontaneous)
Gums (brushing)
Bites to lip/​tongue
Consultation only
Surgical haemostasis or
antifibrinolytic
Blood transfusion/​replacement
therapy or desmopressin
Post dental extraction
Number of extractions
Number complicated by
bleeding
Medical attention required
No bleeding in
≥2 extractions
None done or no
bleeding in
1 extraction
Reported, no consultation
Consultation only
Resuturing or packing
Blood transfusion/​replacement
therapy or desmopressin
GI bleeding
Ulcer, portal hypertension,
haemorrhoids
Spontaneous
Treatment required
No
Associated with ulcer,
portal HTN, haemorrhoids,
angiodysplasia
Spontaneous
Surgical haemostasis, blood
transfusion, replacement
therapy, DDAVP, antifibrinolytic
Surgery
Number of surgeries
Number complicated by
bleeding
Postoperative medical
attention
No bleeding in
≥2 surgeries
None done or no
bleeding in
1 surgery
Reported, no consultation
Consultation only
Surgical haemostasis or
antifibrinolytic
Blood transfusion/​replacement
therapy or desmopressin
Menorrhagia
Duration of menstruation
Duration of heavy
menstruation
How often change pads/​
tampons on heaviest/​
average days
What type of feminine
product used
Medical attention and
treatment
No
Consultation only
Antifibrinolytics, pill
D&C, iron therapy, ablation
Blood transfusion/​replacement
therapy or desmopressin or
hysterectomy
22.7.2  Evaluation of the patient with a bleeding tendency
5515
Postpartum haemorrhage
Number of deliveries
Number complicated by
bleeding
Medical attention required
No bleeding in
≥2 deliveries
None done or
no bleeding in 1
delivery
Consultation only
D&C, iron therapy,
antifibrinolytics
Blood transfusion/​replacement
therapy or desmopressin
Hysterectomy
Muscle haematomas
Post trauma and spontaneous
Treatment required
Never
Post trauma or therapy
Spontaneous, no
therapy
Spontaneous or traumatic,
requiring desmopressin or
replacement therapy
Spontaneous or traumatic,
requiring surgical intervention
or blood transfusion
Haemarthrosis
Post trauma and spontaneous
Treatment required
Never
Post trauma, no therapy
Spontaneous, no
therapy
Spontaneous or traumatic
requiring DDAVP or
replacement therapy
Spontaneous or traumatic
requiring surgical intervention
or blood
CNS bleeding
Subdural or intracerebral
Never
Subdural, any intervention
Intracerebral, any intervention
CNS, central nervous system; D&C, dilation and curettage; DDAVP, 1-​deamino-​8-​d-​arginine vasopressin; GI, gastrointestinal; HTN, hypertension.
Notes on the bleeding score:
Mark the highest scoring box for each symptom (mark one box only for each symptom).
Sum the score obtained from each of the 12 symptoms.
A bleeding score of ≥4 is considered positive. The higher the score, the more likely there is an underlying tendency to bleeding.
Standard laboratory tests such as the prothrombin time (PT) and activated partial thromboplastin time (APTT) are insensitive to mild and moderate reductions of levels of coagulation factors which may be clinically
significant and cause bleeding. These tests are not influenced at all by levels of von Willebrand factor. Therefore, it is on the basis of the history that a decision is made on the extent of laboratory testing. The blood count and
film, PT, APTT, and platelet count should be measured. If these are normal but the history is suggestive of an underlying bleeding disorder, a more comprehensive laboratory assessment of haemostatic capacity is indicated.
Measurement of the platelet count gives no indication of platelet function.
section 22  Haematological disorders
5516
Fibrinogen level
Fibrinogen levels are low in:
•	DIC
•	dilutional coagulopathy (massive blood transfusion)
•	advanced liver disease
•	following thrombolytic therapy
•	congenital hypofibrinogenaemia (very rare)
Thrombin time
The TT is the time taken in seconds for a fibrin clot to form after
addition of thrombin.
The TT is prolonged by:
•	unfractionated heparin
•	direct thrombin inhibitors (dabigatran, argatroban)
•	hypofibrinogenaemia (see earlier for low fibrinogen levels)
•	fibrin degradation products (high levels may occur in DIC and
after thrombolysis)
Reptilase time
This is a snake venom-​based test. It is prolonged by low fibrinogen
levels but not by heparin and so comparison of the TT and RT
is useful for determining if a prolonged APTT is due to heparin;
a long TT with normal RT indicates heparin. Heparin contam-
ination is common on samples from hospitalized patients, even
though the use of heparin flushes and sampling from catheters is
often denied by clinical staff. Weak heparin flushes significantly
prolong the APPT on samples taken from indwelling catheters. In
many cases it is necessary to obtain a venous sample from a fresh
venepuncture.
Factor assays
Individual factor assays are useful in patients with a bleeding his-
tory and are guided by PT and APTT results. The cascade model of
coagulation is no longer considered to represent the physiological
process involved in coagulation. The cascade model was derived
from observation of results using the PT and APTT assays but these
are not ‘physiological’ tests. While the cascade model may not be
‘physiologically true’, it is still a useful framework for interpreting
PT and APTT results. For example, in a patient with a bleeding his-
tory with a normal PT and a long APTT (not due to heparin or
a lupus anticoagulant), there may be deficiency of factor VIII, IX,
or XI (Fig. 22.7.2.1). Table 22.7.2.2 summarizes the interpretation
of laboratory investigations.
Mixing studies
If the PT or APTT are prolonged then a 50:50 mix of patient plasma
with normal plasma will indicate if the prolongation is likely due
to a factor deficiency (the mix corrects the abnormality) or an in-
hibitor, such as heparin or a specific factor inhibitor (the mix does
not correct the abnormality).
Platelet function analysis
Platelet function can be assessed at high and low shear. The platelet
function analyser (PFA-​100) is an automated technique that meas-
ures the ability of platelets to occlude an aperture under conditions
of high shear. The test is performed on a citrated blood sample
within 4 h of sample collection and is abnormal in the presence of
low von Willebrand protein activity or platelet function defects.
Thrombocytopenia causes prolonged closure and so the test re-
quires a normal platelet count in order to assess platelet function.
Platelet function at low shear rate is assessed by platelet aggregation.
Prothrombin
Thrombin
Fibrinogen
Fibrin
VIIa
IX
IXa
VIIIa
X
Xa
Va
XI
XIa
Tissue factor
XIIa
Contact factors
XII
VII
APTT
PT
Common
pathway
Extrinsic
pathway
Intrinsic
pathway
TT
Fig. 22.7.2.1  Cascade model of coagulation. APTT, activated partial thromboplastin time; PT,
prothrombin time; TT, thrombin time.
22.7.2  Evaluation of the patient with a bleeding tendency
5517
Aggregation studies are performed typically on platelet rich plasma
prepared by slow centrifugation of citrated blood within 4 h of
sample collection. There is a poor correlation with bleeding ten-
dency except in specific congenital disorders characterized by severe
platelet dysfunction, for example, Glanzmann thrombasthenia and
Bernard–​Soulier syndrome.
•	Agonists used for aggregation studies include ADP, collagen, ara-
chidonic acid and adrenaline (epinephrine).
•	Response to ristocetin is an agglutination response dependent on
induced conformational change of platelet membrane proteins
(e.g. glycoprotein Ib–​IX–​V, promoting interaction with VWF).
•	Ristocetin-​induced platelet agglutination (RIPA) is carried out at
high (1.2 mg/​ml) and low ristocetin concentrations (0.5 mg/​dl).
Positive RIPA at 0.5 mg/​dl is an abnormal result and is observed in
type 2B VWD and with high VWF levels (e.g. pregnancy).
Platelet storage pool disorders are characterized by absent platelet
α or δ granules. α-​Granule proteins include β-​thromboglobulin
and platelet factor 4 which can be measured by enzyme-​linked im-
munosorbent assay—​these are deficient in α-​granule storage pool
defects. Nucleotides (ADP/​ATP) can be measured by a variety of
techniques including high-​pressure liquid chromatography and are
deficient in δ-​storage pool disorders. These analyses are beyond the
scope of most haematology laboratories. Many patients with in-
herited thrombocytopenias are misdiagnosed as immune thrombo-
cytopenia (ITP) but complex laboratory techniques, including flow
cytometry and targeted gene sequencing, are required for diagnosis.
Bleeding time
The bleeding time is used much less frequently now that platelet
function analysis at high shear is readily available. The bleeding
time does not predict surgical bleeding and is largely no longer con-
sidered a test with clinical utility.
Other investigations
Global tests of haemostasis
Further tests of haemostasis which assess the dynamic interaction
of the individual components of haemostasis rather than the
Table 22.7.2.2  Interpretation of laboratory investigations
Coagulopathy
PT
APTT
Fibrinogen
TT
Platelets
(PFA)
Heritable
VWF—​mild
N
N
N
N
N
Abnormal
VWF—​moderate
N
N or ↑
N
N
N
Abnormal
VWD—​severe
N
↑
N
N
N
Abnormal
VWD—​type 2B (rare)
N
N
N
N
↓
Abnormal
VIII
N
↑
N
N
N
N
IX
N
↑
N
N
N
N
XI
N
↑
N
N
N
N
VII
X, V, or II
↑
↑
N
N
N
N
Fibrinogen
N
N
↓
↑
N
N or abnormal
XIII
N
N
N
N
N
N
Thrombocytopenia
N
N
N
N
↓
Abnormal
Glanzmann’s disease
N
N
N
N
N
Abnormal
Bernard–​Soulier syndrome
N
N
N
N
↓
Abnormal
Acquired
Heparin
N or ↑
↑
N
↑a
N
N
Warfarin
↑
N or ↑
N
N
N
N
Aspirin/​clopidogrel
N
N
N
N
N
Abnormal
Dilutional coagulopathy
↑
↑
N or ↓
N or ↑
↓
abnormal
Renal disease
N
N
N
N
N or ↓
Abnormal
Liver disease
↑
↑
N or ↓
N or ↑
N or ↓
N or abnormal
DIC
↑
↑
N or ↓b
N or ↑
↓
abnormal
Acquired von Willebrand syndrome
N
N or ↑
N
N
N
Abnormal
Hyperfibrinolysis
N or ↑
N or ↑
N or ↓
N or ↑
N
N or abnormal
Surgical bleeding
N
N
N
N
N
N
N, normal. PFA (platelet function analysis on the PFA-​100) only performed in selected cases. In cases of complex coagulopathy, PFA is abnormal primarily due to thrombocytopenia.
a With heparin, the thrombin time is prolonged but the reptilase time is normal.
b With DIC, the level of fibrin degradation products (such as D-​dimer) is very high and this interferes with fibrin polymerization which contributes with low fibrinogen levels to the
prolonged thrombin time.
section 22  Haematological disorders
5518
amount or function of specific components in isolation are avail-
able in some specialized laboratories. These techniques include
thromboelastography, thrombin generation, and tests of fibrinolysis.
Anatomical imaging
Specific bleeding points may be due to structural abnormalities
and imaging directly by endoscopy or by CT, magnetic reson-
ance imaging, or angiography will be indicated in some patients.
Interventional radiology with arterial embolization can be used to
stop localized bleeding, for example, from the gastrointestinal tract.
Specific issues
Drug-​induced bleeding
The most common cause of an acquired bleeding disorder is anti-
coagulant therapy. Approximately 1 in 100 of the United Kingdom
population are now receiving long-​term oral anticoagulant therapy.
Increasingly, patients are being treated with DOACs rather than
warfarin. Overanticoagulation with oral anticoagulants (VKA or
DOACs) is responsible for most life-​threatening bleeds attributable
to antithrombotic therapy.
Vitamin K antagonists
Overanticoagulation due to a VKA such as warfarin is indicated by
a high INR, often due to intercurrent illness and antibiotic use. The
INR is particularly influenced by the factor VII level which is not
the main determinant of bleeding risk. When over-​anticoagulaiton
due to an oral VKA is reversed by intravenous vitamin K the INR
may correct over several hours but a significant bleeding tendency
remains as the factor VII level rises more quickly than the other
vitamin K-​dependent factors. Therefore, in patients with signifi-
cant bleeding, reversal requires a combination of factor replacement
(a prothrombin complex concentrate (PCC) containing factors II,
VII, IX, and X for immediate and effective reversal of bleeding) and
vitamin K (for a sustained reversal when the response to factor re-
placement has decayed).
Direct oral anticoagulants
Overanticoagulation with a DOAC may not be associated with a par-
ticularly prolonged PT or APTT and a quantitative assay is required
to determine the intensity of anticoagulation with these drugs. A di-
lute plasma thrombin time such as the Hemoclot assay is useful
for thrombin inhibitors (dabigatran) and drug-​specific calibrated
anti-​Xa assays are required for factor Xa-​inhibitors (rivaroxaban,
apixaban, edoxaban).
There is no evidence that the anticoagulant effect of DOACs can
be reversed by administration of plasma-​derived factors, including
PCC, but their short half-​life (<12 h) makes management of
bleeding less problematic than that associated with VKA. A spe-
cific antidote to dabigatran (idarucizumab, a monoclonal antibody
fragment that binds to dabigatran) is licensed, and factor Xa-​
inhibitors (e.g. andexanet alfa, a truncated form of enzymatically
inactive factor Xa which acts as a decoy receptor, binding to and
reversing the anticoagulation action of factor Xa inhibitors) are
becoming available.
Antiplatelet drugs
Platelets are integral to thrombin generation and antiplatelet drugs
can be considered as anticoagulants, hence their ability to prevent
thrombosis. Bleeding risk is in part determined by the potency of
antiplatelet activity, for example, the bleeding risk associated with
a fibrinogen receptor antagonist (IIb–​IIIa inhibitor) is far greater
than with aspirin or an ADP receptor antagonist such as clopidogrel.
The individual response to antiplatelet therapy is extremely variable
and even aspirin or an ADP receptor antagonist may produce a sig-
nificant bleeding tendency in some patients. The pharmacological
half-​life of antiplatelet drugs is determined by the mechanism of in-
hibition of platelet function. Aspirin and ADP antagonists irrevers-
ibly inhibit platelet function and so recovery is dependent on new
platelet production. The lifespan of a platelet is typically 10 days and
so 10% of the platelet pool is replaced each day. Once there is 50%
replenishment (5 days after stopping therapy), platelet-​dependent
haemostatic capacity is usually adequate for normal blood coagu-
lation. NSAIDs reversibly inhibit platelet function and there is no
effect 24 to 48 h after stopping treatment.
Surgical bleeding
Postoperative bleeding is a common clinical problem. It is essential to
examine the drug and infusion charts and check that the dose of any
drug that may affect haemostasis is not excessive. It is also imperative
to determine if the site of surgery is the only site of bleeding. If this is
the case, for example, there is no bleeding from venepuncture sites or
an endotracheal tube, and there is no history of previous abnormal
bleeding, then depending on the results of coagulation tests it is im-
portant to keep the possibility of anatomical surgical bleeding as a
likely possibility. In some cases of severe bleeding the patient may
have to return to theatre to look for a bleeding point. Severe surgical
bleeding may result in a dilutional coagulopathy due to fluid volume
replacement or DIC due to hypotensive shock with a severe exacerba-
tion of bleeding and a complex secondary coagulopathy.
Critically ill patients
There are many potential acquired disorders of haemostasis in crit-
ically ill patients. A coagulopathy due to vitamin K deficiency oc-
curs within a few days in critically ill patients with no oral intake.
Parenteral vitamin K supplementation should be used routinely to
prevent bleeding in the critical care setting. Many critically ill pa-
tients develop DIC.
Massive transfusion and dilutional coagulopathy
A dilutional coagulopathy resulting in deficiency of clotting factors
and platelets will cause abnormal bleeding in patients receiving large
amounts of plasma expanders and red blood cells even in the ab-
sence of DIC. It is important to give replacement therapy with fresh
frozen plasma and platelet concentrates guided by repeated meas-
urement of the PT, APTT, and platelet count. Standardized protocols
for the early administration of blood products in patients receiving
massive blood transfusion are increasingly used to ensure adequate
replacement of clotting factors and platelets in order to prevent or at
least limit the development of coagulopathy.
Disseminated intravascular coagulation
The major manifestations of DIC are end-​organ damage due to
microvascular thrombosis but the most readily apparent clin-
ical manifestation is often bleeding due to the consumptive
coagulopathy. DIC is a clinical diagnosis supported by the results
of laboratory investigations with a prolonged PT and APTT, a low
22.7.2  Evaluation of the patient with a bleeding tendency
5519
fibrinogen and platelet count, and elevated fibrin degradation prod-
ucts (such as D-​dimer). Persistent oozing from venepuncture sites
in patients with sepsis or in obstetric patients suggests DIC. A fi-
brinogen level less than 1 g/​litre suggests DIC in a patient with an ac-
quired severe bleeding disorder not due to dilutional coagulopathy.
The most important aspect of treatment is that of the underlying
cause (e.g. sepsis), although fresh frozen plasma and platelet concen-
trates are used to treat bleeding or prevent haemorrhage associated
with planned invasive procedures. A chronic form of DIC occurs in
patients with malignancy.
ABO blood group-​associated low von Willebrand
protein levels and von Willebrand disease
The most common heritable bleeding tendency is due to a low VWF
level. This may be due to genetic mutation of the VWF gene (often
designated VWD), or more commonly the effect of epigenetic fac-
tors such as blood group O (designated blood group O-​associated
low VWF level). Regardless, the level of von Willebrand protein ap-
pears to be an important continuous variable influencing the coagu-
lation phenotype. The first apparent manifestation of this may be
excessive surgical bleeding and as a result the patient is considered to
have an acquired bleeding disorder. The history may be informative,
such as a detailed menstrual history in women. It can be difficult to
establish a diagnosis of a mild reduction in von Willebrand protein
in the immediate postoperative period as levels rise due to the stress
response. Consequently, it is prudent to re-​evaluate patients several
weeks after an episode of abnormal surgical bleeding. VWF levels
should be interpreted in relation to blood group and the clinical cir-
cumstances at the time a blood sample was taken. VWF levels rise
with age and so a mild bleeding disorder associated with low levels
may be attenuated as the patient ages.
Thrombocytopenia
Many drugs result in a reversible idiosyncratic thrombocytopenia.
In most cases, drug-​induced thrombocytopenia is mild and does not
cause bleeding. Notable exceptions are quinine and gold-​induced
thrombocytopenia which are severe. An evaluation of drug history
and cessation of possibly implicated drugs is essential in patients with
acquired bleeding who are found to be thrombocytopenic. Other com-
monly used drugs for which there is good evidence for drug-​induced
thrombocytopenia include amiodarone, atorvastatin, carbamaze-
pine, cimetidine, diclofenac, digoxin, ranitidine, co-​trimoxazole,
and vancomycin. Cytotoxic drugs produce a dose-​dependent sup-
pression of bone marrow platelet production and thrombocytopenic
bleeding is common in oncology practice. Bone marrow suppression
and bone marrow failure syndromes, such as aplastic anaemia and
myelodysplasia, often result in production of dysfunctional platelets
and the bleeding tendency is significantly greater than in patients
with thrombocytopenia and an uncompromised marrow, such as
occurs in ITP in which the bleeding risk is relatively low. Inherited
thrombocytopenias are often misdiagnosed as ITP in adults but
correct diagnosis of these disorders is difficult. A family history of
thrombocytopenia or a lifelong history of a relatively stable, though
low, platelet count is suggestive. The normal platelet count decreases
with age and platelet counts less than 150 × 109/​litre may be ‘normal’
in older patients. Therefore, abnormal bleeding should not neces-
sarily be attributed to a mild reduction in platelet count. Over the
age of 65 years the lower limit of normal is around 120 × 109/​litre and
by the age of 80 years around 100 × 109/​litre. In addition, abnormal
bleeding is not usually apparent until the platelet count is below 80 ×
109/​litre, when platelet function is normal.
Thrombocytopenia is common in HIV infection. Gestational
thrombocytopenia occurs in 5% of pregnancies but the platelet
count is rarely less than 80 × 109/​litre and it is not associated with an
increased bleeding tendency.
Renal disease
Bleeding risk increases with the degree of renal impairment and is
due to a defect of platelet–​vessel wall interaction as well as impaired
platelet function. In most patients, laboratory tests of haemostasis
are normal. Platelet function tests may give variable results that do
not correlate with bleeding risk. The cause of the bleeding is an ac-
cumulation of dialysable uraemic toxins including urea and phenols
which inhibit platelet function and possibly VWF activity. Anaemia
contributes to the bleeding tendency due to a reduction in platelet–​
vessel wall contact as the haematocrit decreases. Bleeding is most
commonly into the skin and mucous membranes and gastrointes-
tinal haemorrhage is common. Haemodialysis and peritoneal dia-
lysis reduce the bleeding tendency without any appreciable effect
on tests of platelet function. 1-​deamino-​8-​d-​arginine vasopressin
(DDAVP) improves haemostasis. Correction of anaemia by transfu-
sion or erythropoietin therapy to maintain the haemoglobin above
100 g/​litre is beneficial. Administration of conjugated oestrogens
has also been reported to reduce the bleeding tendency.
Liver disease
The PT and APTT are frequently abnormal in patients with ad-
vanced liver disease but correction of abnormalities with fresh
frozen plasma is only indicated if there is active bleeding or in an-
ticipation of an invasive procedure. The liver is the site of synthesis of
the majority of proteins involved in haemostasis. In cirrhosis, there
is deficient production of coagulation factors compounded by pro-
duction of dysfunctional factors (due to defective post-​translational
carboxylation) with thrombocytopenia due to portal hypertension
with hypersplenism and in some case defective marrow platelet
production. Platelet function is defective. In obstructive jaundice,
there is production of dysfunctional factors which respond initially
to intravenous vitamin K. In acute hepatitis, there is predominantly
a consumptive coagulopathy due to DIC. In advanced liver dis-
ease and hepatoma, there may be additional dysfibrinogenaemia.
Hypofibrinogenaemia with a fibrinogen level less than 1 g/​litre
occurs with fulminant hepatic failure. Reduced clearance of tissue
plasminogen activator may cause hyperfibrinolysis which contrib-
utes to bleeding. A variable degree of DIC may be present in patients
with chronic liver disease and acute DIC may be precipitated by
infection. Transfusion of platelets, fresh frozen plasma, PCC (con-
taining factors II, VII, IX, and X), and fibrinogen (or cryoprecipitate
as a source of fibrinogen) will depend on individual circumstances.
Parenteral vitamin K should always be considered and frequently a
trial of therapy is required, for example, 10 mg daily for 3 days.
Acquired haemophilia and acquired von
Willebrand syndrome
Acquired inhibitors are rare and most often autoantibodies. Platelet
autoantibodies result in shortened platelet survival and thrombo-
cytopenia (ITP). The bleeding manifestations of ITP are variable

22.7.3 Thrombocytopenia and disorders of platelet
22.7.3 Thrombocytopenia and disorders of platelet function 5520 Nicola Curry and Susie Shapiro

22.7.4 Genetic disorders of coagulation 5532 Elean
22.7.4 Genetic disorders of coagulation 5532 Eleanor S. Pollak and Katherine A. High
section 22  Haematological disorders
5532
Disorders of platelet procoagulant activity
Scott syndrome is an extremely rare autosomal recessive condition.
Reduced negatively charged phospholipids on the surface of the
platelet, normally an important surface for coagulation reactions,
results in reduced tenase and prothrombinase activity.
FURTHER READING
Arnold DM, et al. (2007). Systematic review: efficacy and safety of
rituximab for adults with idiopathic thrombocytopenic purpura.
Ann Intern Med, 146, 25–​33.
Bolton-​Maggs PH, et al. (2006). A review of inherited platelet disorders
with guidelines for their management on behalf of the UKHCDO.
Br J Haematol, 135, 603–​33.
Liang Y, et al. (2012). Rituximab for children with immune thrombo-
cytopenia: a systematic review. PLoS One, 7, e36698.
Neunert C, et al. (2011). The American Society of Hematology 2011
evidence-​based practice guideline for immune thrombocytopenia.
Blood, 117, 4190–​207.
Provan D, et al. (2010). International consensus report on the inves-
tigation and management of primary immune thrombocytopenia.
Blood, 115, 168–​86.
Scully M, et al. (2012). Guidelines in the diagnosis and management
of thrombotic thrombocytopenic purpura and other thrombotic
microangiopathies. Br J Haematol, 158, 323–​35.
Shih A, et al. (2014). Novel treatments for immune thrombocytopenia.
Presse Med, 43, e87–​95.
Stasi (2012). How to approach thrombocytopenia. Hematology Am Soc
Hematol Educ Program, 2012, 191–​7.
22.7.4  Genetic disorders
of coagulation
Eleanor S. Pollak and Katherine A. High
ESSENTIALS
Much of what is understood about specific coagulation proteins
has emerged from the careful study of hereditary disorders of blood
coagulation.
Haemophilia
Haemophilia is a familial X-​linked disorder due to deficiency of ei-
ther factor VIII (haemophilia A) or factor IX (haemophilia B), compo-
nents of the intrinsic enzymatic complex that activates factor X. The
severity of the disease correlates with predicted concentrations of
activated factor protein, and those with activity levels below 1% are
defined as having severe disease.
Clinical features and diagnosis—​the main manifestations are
bleeding into joints and soft tissues, with haemophilic arthropathy
and joint deformity being inevitable complications in untreated
patients. Other features include pseudotumours, bleeding into the
urinary system, and bleeding following clinical procedures (e.g.
dental extractions). Laboratory diagnosis is based on a modification
of the classic activated partial thromboplastin time (APTT) assay, with
inhibitor screening used to exclude other causes of prolonged APTT
(e.g. lupus anticoagulant).
Treatment—Previous methods of administering therapy to Factor
VIII and FIX deficient patients continue to advance. Now, rather than
providing ‘on demand’ therapy at specific strenuous times, such as
prophylactically before surgery, current treatment seeks to rebal-
ance the hemostatic deficiency. The use of recombinant factors is
preferable to preparations derived from pooled human plasma sam-
ples, which have led to numerous infectious complications (hepatitis
B and C, HIV, and parvovirus. The development of inhibitory anti-
bodies is a significant problem, particularly in patients with haemo-
philia A. Trials of gene therapy are being performed.
Von Willebrand disease
Von Willebrand disease is a common autosomal dominant dis-
order of platelet function caused by a functional deficiency of von
Willebrand factor (VWF). VWF, normally synthesized by mega­
karyocytes, prevents degradation of factor VIII; VWF, also made by
endothelial cells, enhances platelet activation and recruitment at
sites of tissue damage. It may be due to quantitative deficiency of
VWF (types 1 and 3), or to defect in platelet binding affinity (type 2).
Clinical features and diagnosis—​typical presentation includes
nosebleeds, menorrhagia, and easy bruising. Laboratory diagnosis
involves both an antigenic test and an activity test (ristocetin co-
factor), in which formalin-​fixed platelet aggregation is induced due
to ristocetin-​enhanced VWF binding to glycoprotein complex Ib–​IX.
Treatment—​mild von Willebrand disease is treated with
desmopressin 1-​deamino-​8-​d-​arginine vasopressin (DDAVP), which
releases factor VIII and VWF from endothelial cells. Other treatments
include ε-​aminocaproic acid (for patients who require dental sur-
gery, and women with menorrhagia), oestrogens, and factor VIII
concentrates.
Other hereditary disorders of coagulation
These include (1) hereditary deficiency of the plasma metalloproteinase
ADAMTS13, which predisposes to thrombotic thrombocytopenic
purpura; (2) combined deficiency of coagulation factors V and VIII,
caused by single-​gene defects in the coordinated machinery for pro-
tein trafficking and secretion; (3) factor XI deficiency—​an autosomal re-
cessive diathesis of variable severity frequently occurring in Ashkenazi
Jews; (4) inherited deficiencies of factors II, V, VII, and X—​these cause
bleeding tendencies of varying severity and are inherited as recessive
disorders; and (5) deficiency of the contact activating factors, factor
XIII, and fibrinogen.
Hypercoagulable diseases due to deficiencies of
anticoagulants or propensity to thrombosis
Typical presentations occur with deep venous thrombosis and/​or
pulmonary embolism, and hypercoagulable states should be con-
sidered particularly with presentation of ‘unusual’ thrombosis (e.g.
superficial thrombophlebitis, mesenteric vein thrombosis, and cere-
bral vein thrombosis).
Antithrombin III deficiency—​diagnosis poses difficulties in the
post-​thrombotic period when patients frequently have lower levels
of antithrombin III due either to consumption of antithrombin III
during clot formation or to the decreased function seen with heparin
administration. Treatment is typically with therapeutic or prophy-
lactic low molecular weight heparin or warfarin; antithrombin III
22.7.4  Genetic disorders of coagulation
5533
concentrate may be given during an acute event or as a prophylactic
treatment to prevent further disease.
Deficiencies of protein C and protein S—​in addition to thrombotic
manifestations, protein C deficiency may also manifest as warfarin-​
induced skin necrosis and dangerously life-​threatening purpura
fulminans in the homozygous or compound heterozygous protein
C-​deficient neonate.
Factor V Leiden—​a single mutation, in the gene encoding the
factor Va protein, leads to prolonged factor Va activity and resist-
ance to activated protein C. The factor V Leiden mutation occurs
in about 5% of people of European ancestry and may predispose to
thrombotic disease.
Prothrombin 20210 mutation—​the frequent allelic variant
(G20210A) in populations of European ancestry increases the con-
centration of prothrombin, thus biasing haemostatic balance to-
wards excess thrombin formation.
Introduction
Haemostasis, the physiological process of blood clot formation,
involves a coordinated interaction between the wall of the blood
vessel, platelets, and blood coagulation proteins. The haemostatic
mechanism maintains a state of readiness to respond to a multi-
tude of haemostatic stressors to prevent haemorrhage while also
preventing inappropriate clot formation. Although acquired dis-
eases of the coagulation system frequently occur with liver disease
and other pathological disease states, this chapter focuses specific-
ally on genetic disorders resulting from abnormalities and/​or de-
ficiencies of the blood coagulation proteins. More specifically, this
chapter covers haemophilia, von Willebrand disease, and deficien-
cies/​abnormalities of fibrinogen and factors II, V, VII, X, XI, XII, and
XIII. The role of an inherited increased risk for excess clotting will
also be addressed. These conditions may result from either the loss
of function of anticoagulant proteins (antithrombin III, protein C,
and protein S) or a gain of function of procoagulant proteins (factor
V Leiden and prothrombin 20210G to A).
Additionally, we briefly describe haemostasis-​related genes: 
LMAN1 (previously ERGIC-​53) and MCFD2 linked to com-
bined factor V/​factor VIII deficiency, ADAMTS13, associated
with thrombotic thrombocytopenic purpura, and the gene for
vitamin K epoxide reductase (VKORC1), the enzyme respon-
sible for recycling vitamin K 2,3-​epoxide to the enzymatically
activated form.
The coagulation cascade as a haemostatic
mechanism
The human blood coagulation system involves a coordinated array
of reactions that generates a stable fibrin clot when needed and pre-
vents unnecessary clot formation. The system involves numerous
proteins that interact, principally on phospholipid surfaces, to
create a meshwork of fibrin fragments entrapping haematopoietic
cells (Fig. 22.7.4.1). The majority of coagulation enzymatic com-
plexes involve protease enzymes. Many of these enzymes are serine
proteases, and a subset have the distinguishing feature that their
functional synthesis requires vitamin K to enable post-​translational
modification of glutamic acid residues in the N-​terminal region; this
property provides the basis of the therapeutic mechanism by which
the drug warfarin prevents proper synthesis of functional factors.
The principal enzyme balancing the pro-​ and anticoagulant forces
is prothrombin, thought to be the evolutionary forerunner of the
mammalian coagulation proteins. In addition to its procoagulant
functions, prothrombin, once activated, provides anticoagulant and
cellular mobility functions as well.
In 1905, Morawitz first described the importance of thrombin,
thromboplastin, and calcium in cleaving fibrinogen to create a fibrin
clot. In the early 1930s and 1940s, laboratory tests were developed
that relied on in vitro fibrin clot formation to analyse the adequacy
of a patient’s clotting system. The waterfall cascade of sequential ac-
tivation steps resulting in a fibrin clot was elegantly described in the
early 1960s, delineating separate pathways to account for the pro-
thrombin time (PT) and the partial thromboplastin time that the
earlier laboratory tests measure. However, the set of activation steps
is now better described as an interwoven, reinforcing set of reactions
(Fig. 22.7.4.2). The unique specificities of the coagulation enzymes
summarized in the classical coagulation cascade have been found to
be more versatile in activating diverse proteins under varied condi-
tions. However, the separate pathways, now termed the tissue factor
(extrinsic) and the intrinsic pathways, help define the steps involved
in the principal tests used in clinical medicine for evaluation of
haemostatic proteins. For the series of reactions and specific factors
involved, the time to clot formation defines the principal parameter
used in clinical evaluation of the health of a patient’s coagulation
system. The assays (the PT, the activated partial thromboplastin time
(APTT), and activity levels of specific individual clotting factors)
Fig. 22.7.4.1  Scanning electron micrograph of a whole blood clot.
There is a meshwork of fibrin fibres emanating from platelet aggregates in
which erythrocytes, lymphocytes, and other cells are trapped.
Courtesy of John W Weisel and Chandrasekaran Nagaswami, Department of
Cell and Developmental Biology, University of Pennsylvania School of Medicine,
Philadelphia, PA.
section 22  Haematological disorders
5534
compare the time needed for clot formation in a patient’s plasma
with that in a control pool of plasma from normal donors.
Endothelial injury and tissue damage first trigger clot formation.
The response of the platelets forms the primary phase of healing by
temporarily patching the site of vascular injury. Subsequent to this
initial platelet phospholipid patch, a fibrin clot provides a more solid
framework for the necessary but slower cellular repair. Secondary
haemostasis begins with injury-​induced exposure of the integral
membrane protein tissue factor to plasma proteins, enabling for-
mation of the active enzymatic complex tissue factor–​factor VIIa.
The generation of tissue factor–​factor VIIa then catalyses clotting by
activating both factor X to factor Xa and factor IX to factor IXa. This
activation primarily involves the cleavage of an arginine–​isoleucine
bond in a secreted plasma protein zymogen to form a two-​chain ac-
tive protein. Thus, once tissue injury has signalled the need for fibrin
clot formation and tissue factor–​factor VIIa has initiated coagula-
tion, the haemostatic process amplifies through the generation of
factor IXa from factor IX, which is 10 times more abundant than
factor VII and consequently leads precipitously to thrombin gener-
ation. Among thrombin’s numerous roles is the activation of the es-
sential procoagulant cofactors factors V and VIII. This process then
further amplifies clotting by generating more thrombin through the
active cofactors Va and VIIIa which then form the tenase (factor IXa/​
factor VIIIa) and prothrombinase (factor Xa/​factor Va) complexes
(Fig. 22.7.4.1). Thrombin also activates the cross-​linking enzyme
(factor XIIIa) and the fibrinolytic inhibitor (TAFIa), and triggers
platelet recruitment. Importantly, thrombin generation simultan-
eously counterbalances its procoagulation activities by inciting lysis
of the clot via the release by endothelial cells of tissue plasminogen
activator converting plasminogen to plasmin, the enzyme respon-
sible for fibrin clot lysis. Thrombin also dampens the clotting process
by activating protein C that actively breaks down the critical pro-
coagulant cofactors factors Va and VIIIa.
The basis for initiating clot formation in the PT and APTT tests is
titration of calcium into an anticoagulated plasma specimen along
with a source of phospholipid. In addition, in the PT test, the source
of phospholipid is a thromboplastin reagent that provides tissue
factor to enable the tissue factor–​factor VIIa complex to catalyse
clot formation. The variation in PTs, due to differences in the source
of reagent tissue factor, has led to development of the International
Sensitivity Index which creates an international normalized ratio
(INR) for clinical management and the increasing ability to syn-
thesize a thromboplastin with an International Sensitivity Index
approaching 1.0. In the APTT test the phospholipid reagent lacks
tissue factor and thus prevents formation of the tissue factor–​factor
VIIa complex. An activator, such as silica particles, also greatly de-
creases the time required for clot formation through activation of
factor XII via the contact activation system.
Deficiencies of specific clotting proteins
Haemophilia
Deficiency of either factor VIII (haemophilia A) or factor IX
(haemophilia B), which together make up the factor VIIIa/​factor IXa
intrinsic tenase enzymatic complex, results in the clinical pheno-
type commonly known as haemophilia. A sex-​linked bleeding di-
athesis, now thought to be haemophilia, was described in Talmudic
writings as a cause of fatal haemorrhage at circumcision. In the
modern era, the disease may cause bleeding at circumcision, but
haemophilia principally presents with haematoma formation, easy
bruising, and bleeding at the site of venepuncture during the tod-
dler period. The disease exists in severe, moderate, and mild forms
classified as such on the basis of a clinical laboratory blood coagu-
lation test performed to assess the level of functional coagulant
protein (percentage activity of factor VIII or factor IX). The patho-
logical problem in both haemophilia A (factor VIII deficiency) and
haemophilia B (factor IX deficiency; also called Christmas disease),
is the inability to form a functional tenase complex to activate factor
X to factor Xa. Although factor X can still be activated to factor Xa
by tissue factor–​factor VIIa, the available quantities of factor VII
(400 ng/​ml) do not allow sufficient activation of factor X to enable
clotting to occur in a physiologically timely fashion. Although pa-
tients with haemophilia may have some difficulties with immediate
haemorrhage subsequent to a cutaneous or superficial injury, they
characteristically have joint and deep tissue bleeding problems as
discussed later. The severity of disease is very well predicted by
an in vitro assay for evaluation of the deficient protein level such
that patients with severe disease have levels of factor activity of less
than 1%, patients with moderate disease have activity levels of 1 to
5%, and patients with mild disease have activity levels of 6 to 30%.
Normal factor VIII and IX activity levels are 50 to 200% and 75 to
125%, respectively.
Numerous genetic mutations have been described accounting for
the factor deficiencies causing haemophilia. In part because of the
considerable difference in size between the factor VIII gene (186 kb),
and the factor IX gene (34 kb), the ratio of the frequency of factor
VIII to factor IX deficiency is between 4 and 5 to 1 (c.186/​34 kb).
PT
APTT
FXIIa
FXI
Contact
Thr
FXIa
FIX
FVIII
FX
FIXa
FVIIIa
Tenase
complex
Prothrombin
Fibrinogen
FXIII
Fibrin dimers
FXIIIa
Fibrin monomers
Thrombin
(Thr)
FV
Thr
FXa
FVa
Prothrombinase
complex
TF-FVIIa
TISSUE INJURY
EXPOSING
TISSUE FACTOR (TF)
HMWK
PL-Ca2+
PL-Ca2+
PL-Ca2+
PL-Ca2+
Thr
Thr
Fig. 22.7.4.2  Schematic representation of the enzymatic reactions
involved in blood clot formation. APTT, activated partial thromboplastin
time; F, factor; PL-​Ca2+, phospholipids/​calcium; PT, prothrombin time; Thr,
thrombin.
22.7.4  Genetic disorders of coagulation
5535
Thus, the frequency of haemophilia A is approximately 1 in 5000 to
6000 and that of haemophilia B is approximately one-​fifth of that.
Among affected cases, approximately one in three to one in four pa-
tients presents spontaneously without a familial inheritance pattern.
One of the only differences between factor VIII and IX deficiencies
is the frequency of severe disease, which occurs more commonly
in factor VIII deficiency (60% of cases as compared with 45% in
haemophilia B). This difference is largely attributed to the frequency
of mutation due to a factor VIII gene inversion in intron 22 of the
26-​exon factor VIII gene. At this locus of the factor VIII gene, a re-
gion of homology to sequences telomeric to the factor VIII gene, a
recombination event results in the inability to synthesize any func-
tional factor VIII, thus leading to severe disease (<1% functional
protein activity). A less common inversion in intron 1 has also been
described in 2 to 3% of severe haemophilia A patients. In both factor
VIII and factor IX deficiency, milder disease is commonly due to
missense mutations.
The clinical features of haemophilia predominantly include
bleeding into joints and soft tissues. The incidence of central ner-
vous system bleeding has dramatically decreased with concentrate
therapy. The average life expectancy of people with severe haemo-
philia has increased from 11 years at the beginning of the 20th cen-
tury to approximately 70 years at the beginning of the 21st century.
However, there was a marked decrease to 60 years in the 1980s, when
the devastating effects of blood-​borne viral disease again shortened
average life expectancy.
In the untreated patient with severe disease, haemophilic arthro­
pathy and joint deformity are inevitable complications. In decreasing
order of involvement, the most commonly affected joints include
the knee, elbow, ankle, shoulder, wrist, and hip. Recurrent bleeding
episodes create a hypertrophic synovial lining with chronic inflam-
mation; however, the pathophysiology responsible for recurrent
joint bleeding remains unknown. Arthropathies commonly neces-
sitate replacement of affected joints for pain control and improve-
ment of mobility. Soft-​tissue haemorrhages frequently complicate
haemophilia; further complications due to these haemorrhages
include compartment syndrome, neurological damage, and exten-
sive blood loss from retroperitoneal bleeds. Haematoma formation,
a frequent complication of haemophilia, may arise spontaneously
or with trauma and require extensive factor replacement and
fasciotomy, the necessity for which can be assessed by mean arterial
pressure in a compartment. Intracranial haemorrhage, occurring
in approximately 5% of patients, warrants immediate evaluation
and treatment within the first 6 to 8 h of presentation; however, the
majority of children presenting to an emergency department with
central nervous system symptoms have not suffered from intracra-
nial haemorrhage.
A pseudotumour, an encapsulated collection of blood most com-
monly originating in bone or soft tissues, is a rare but extremely ser-
ious consequence of haemophilia occurring in approximately 2% of
patients. This complication is difficult to manage but may sometimes
be treated with surgery at specialized haemophilia centres.
Frequently, patients with haemophilia have haematuria, the se-
verity of which may range from self-​limited episodes to gross haema-
turia with significant blood loss. Protease inhibitors for HIV therapy
may lead to haematuria with flank pain or renal stones. Physicians
should be aware of the possibility of nephrotic syndrome in patients
with severe disease and high titres of antibody inhibitors receiving
high-​dose intravenous replacement therapy and other agents to in-
duce tolerance.
Dental procedures warrant involvement of a haemophilia specialist.
Factor replacement levels of 25 to 100% are suggested depending on
the complexity of the dental procedure. Antifibrinolytics such as ε-​
aminocaproic acid (6-​aminohexanoic acid) or tranexemic acid and
fibrin sealants may be a helpful adjuvant to replacement therapy.
On account of the sex-​linked inheritance pattern, haemophilia
is rarely found in women unless extensive lyonization takes place
in the normal factor gene, or the woman is born to a haemophilic
father and a carrier mother. Normal vaginal delivery is considered
to be relatively safe in the case of a haemophilic infant; however,
vacuum extraction, midcavity forceps deliveries, and invasive fetal
monitoring should be avoided because of the increased risk of for-
mation of subgaleal and cephalic haematomas.
The laboratory diagnosis of haemophilia is based on a modifica-
tion of the classic APTT assay used as a standard test for the haemo-
static system. Normally patients are evaluated due to bleeding
symptomatology or because of a prolonged APTT result. The APTT
is a very sensitive but poorly specific screening test for haemo-
philia. All patients, even those with mild disease, will normally
have a prolonged APTT unless there is a problem with specimen
acquisition or the insensitivity of the APTT reagent. Once sus-
pected, haemophilia can be evaluated by an inhibitor screen which
involves performing a 50:50 mix of patient and normal plasma to
evaluate whether the prolongation is due to a deficiency of a clot-
ting protein or alternatively to the presence of an inhibitor. There
are many causes of a prolonged APTT other than haemophilia (see
Table 22.7.4.1). Classically, a phospholipid inhibitory antibody,
called a lupus anticoagulant, will cause a prolongation of the APTT
of the 50:50 mix due to the effect of the phospholipid inhibitory
antibody on the normal pooled plasma. A  lupus anticoagulant
which causes a prolonged APTT may also result in a low factor
VIII or factor IX activity level. In such cases, further testing for a
lupus anticoagulant is necessary to rule out a low factor VIII due to
a lupus anticoagulant as opposed to a deficiency.
Previously, management of haemophilia involved administer­
ing the deficient protein (factor VIII or factor IX) to a patient in
a so-called ‘on-demand’ protocol. Subsequent studies established
the superiority of prophylactically administered factor concen-
trates in terms of clinical outcomes such as joint health, and most
patients are now managed of prophylactic regimens.
The treatment landscape continues to adjust as new therapy be-
comes commercially available. Novel therapy in the form of a
bispecific factor IXa- and factor X-directed antibody, administered
subcutaneous weekly or every other week, bypasses the need for
intravenous fVIII. This innovative therapy has been licensed for pa-
tients (adult and paediatric haaemophilia A) both with or without
factor VIII inhibitors.
Before the development of stringent purification and viricidal
procedures, the transmission of viral disease was almost inevitable
as each vial of plasma-​derived concentrate was pooled from approxi-
mately 60 000 to as many as 400 000 donors, although the number
has recently been reduced to 15 000. Tragically, most patients with
severe disease treated before 1985 developed HIV. Rates of develop-
ment of hepatitis B and C are also extremely high. Although dras-
tically reduced, the potential for transmission of infectious disease
has not been totally eliminated. Many recombinant preparations are
section 22  Haematological disorders
5536
prepared with human serum albumin, thus leaving a possible source
of transfusion of a blood-​borne disease.
Treatment
Acute bleeding episodes
Safe and effective treatment options continue to improve for the
management of acute bleeding episodes for patients with haemo-
philia A and B. Blood products available include fresh frozen plasma
which contains both factors VIII and IX, prothrombin complex
concentrates containing factors II, VII, IX, and X, activated pro-
thrombin complex concentrates (factors IIa, VIIa, IXa, Xa), mono-
clonal antibody-​purified factor VIII and factor IX, and recombinant
factor VIII and factor IX. Recombinant factor VIIa is now approved
for use in patients with inhibitors during acute bleeds. Currently,
trials using gene therapy approaches are underway and may pro-
vide a method for continuous prophylaxis against bleeding (see
‘Future directions: gene transfer as a method of treating haemo-
philia’). Recombinant or highly purified products are the optimal
therapy because of the great benefit:risk ratio. Availability, ease of
administration, cost, viral safety, and thrombotic risk, particularly
in patients undergoing high-​dose therapy or procedures with a high
risk of thrombotic complications, dictate the choice of product.
Cryoprecipitate, made from the precipitate of thawed frozen plasma,
contains factor VIII but does not contain factor IX. Cryoprecipitate
and fresh frozen plasma should only be used in the haemophilia
patient in an emergency setting where concentrates are not avail-
able. Inhibitor formation, the development of antibodies to the defi-
cient protein, arises subsequent to transfusion of a blood product or
factor replacement and is the major complication of treatment. An
inhibitor presents an extremely difficult situation for patient man-
agement (see ‘Complications of therapy’).
Several immunoaffinity-​purified plasma-​derived factor VIII and
factor IX products are available in the United States of America
and Europe and currently have excellent records of viral safety, effi-
cacy, and lack of thrombogenicity. When concentrate is unavailable,
fresh frozen plasma is readily available in most emergency settings.
Viricidal methods using solvent detergent treatment may now be
applied in production of fresh frozen plasma; furthermore, each
unit is from a single screened donor, thus the risk of transfusion-​
transmitted disease is low.
Recombinant factor VIII and factor IX have been licensed for over
a decade. These proteins are produced in cultured mammalian cells
and purified from conditioned medium. Recombinant factor IX is
devoid of human plasma whereas the recombinant factor VIII con-
centrates utilize human plasma-​derived albumin for stabilization.
As in vivo coagulant activity of recombinant factor IX is only 80% of
in vitro estimates used for labelling of product in international units
(IU)/​mg, it is recommended that the calculated factor IX dosage
be multiplied by a factor of 1.2 for dose calculation when using re-
combinant factor IX. A plausible explanation for this discrepancy
is a difference in post-​translational modifications compared with
plasma-​derived factor IX.
During severe and critical bleeds it is optimal to achieve 50 to
100% factor activity levels for 7 to 10  days (e.g. for pharyngeal,
retropharyngeal, retroperitoneal, and central nervous system
bleeds). More modest levels of 20 to 50% for 2 to 7 days are generally
adequate for dental extractions, haematuria, intramuscular, or soft-​
tissue bleeds with dissection, or bleeds into mucous membranes.
Levels of 20 to 30% for 1 to 2 days are recommended for uncompli-
cated haemarthroses, superficial muscle, or soft-​tissue bleeds. The
frequency of dosing is every 12 to 24 h for factor IX concentrates
and every 8 to 12 h for factor VIII concentrates. At 24 h for factor IX
and 12 h for factor VIII, the calculated amount to infuse would be
one-​half the initial amount of factor IX, as the half-​life of factor IX is
approximately 18 to 24 h.
The timing of factor level determination should be 15 to 30 min
after the loading dose and immediately prior to subsequent doses
Table 22.7.4.1  Coagulation laboratory testing, plasma concentrations, and chromosomal location of coagulation proteins
Deficient clotting
factor
PT
APTT
Half-​life of
protein
Plasma
concentration
Inheritance (chromosome)
Fibrinogen
Increased
Increased
2–​4 days
200–​400 mg/​dl
Autosomal (1)
Prothrombin (factor II)
Increased
Increased
3 days
100 µg/​ml
Autosomal (11)
Factor V
Increased
Increased
36 h
10 µg/​ml
Autosomal (1)
Factor VII
Increased
Normal
2–​6 h
0.5 µg/​ml
Autosomal (13)
Factor VIII
Normal
Increased
8–​12 h
0.1 µg/​ml
X-​linked (X)
von Willebrand’s factor
Normal
Mildly increased in approximately
50% of cases
Several hours
10 µg/​ml
Autosomal (12)
Factor IX
Normal
Increased
24 h
5 µg/​ml
X-​linked (X)
Factor X
Increased
Increased
40 h
10 µg/​ml
Autosomal (13)
Factor XI
Normal
Increased
60–​80 h
5 µg/​ml
Autosomal (4)
Factor XII
Normal
Increased
50 h
30 µg/​ml
Autosomal (5)
Factor XIII
Normal
Normal
9–​19 days
10 µg/​ml
Autosomal (6: α chain) and (1: β chain)
Protein C
Normal
Normal
6 h
5 µg/​ml
Autosomal (2)
Protein S
Normal
Normal
30 h
25 µg/​ml
Autosomal (3)
Antithrombin III
Normal
Normal
48 h
150 µg/​ml
Autosomal (1)
PT, prothrombin time; APTT, activated partial thromboplastin time.
22.7.4  Genetic disorders of coagulation
5537
for appropriate dose adjustments. When factor concentrates are
used for patients with inhibitors, higher doses will most likely be re-
quired. Additionally, some authors have also reported good success
and reduced cost with constant infusion regimens.
Calculation of the optimal factor concentration
for administration
The number of IUs of factor required is equal to:
(kg bodyweight) × (desired % factor level increase) ×
  C (kg bodyweight) × (desired % factor level increase) × C
where C is a constant depending on the product and the source of the
product. The value of C is 0.5 for administration of plasma-​purified
and recombinant factor VIII, 1 for administration of plasma-​purified
factor IX, and 1.2 for administration of recombinant factor IX.
Surgery in patients with haemophilia
When possible, treatment should be instituted by caregivers aware
of major and minor adverse reactions and complications occurring
in the haemophiliac population. Care should also be given in as-
sociation with an experienced reference laboratory able to provide
timely evaluation of a patient’s response to treatment. Therapeutic
factor levels should be obtained before surgery. Depending on the
type of surgery, the factor level should reach levels of 50 to 100%
of normal and should be maintained 2–​7 days post procedure. For
brain or prostate surgery, a factor level approaching 100% is re-
commended because of the higher risk of bleeding. In patients with
milder haemophilia A, administration of the synthetic octapeptide
desmopressin (1-​deamino-​d-​arginine vasopressin (DDAVP)) may
be helpful in increasing factor levels. However, this is not the case
for haemophilia B.
In addition to factor concentrates, fibrin glue has been recom-
mended with circumcision, antifibrinolytics with dental proced-
ures, and recombinant factor VIIa and/​or apheresis in patients with
high-​titre inhibitors. Aprotinin may be considered with caution in
cardiac procedures; however, the use of aprotinin has been proposed
to increase the risk of thrombogenicity.
Complications of therapy
The main adverse outcomes related to treatment with concentrates
include transmission of viruses when using plasma-​derived prod-
ucts and development of inhibitory antibodies seen with both re-
combinant and plasma-​derived products. Thrombosis has also been
a complication of early complex concentrates used for patients with
inhibitors.
The development of purification schemes that inactivate viruses,
and the development of recombinant products, has dramatically de-
creased the incidence of transmission of viral disease. Early prepar-
ations of prothrombin complex concentrates presented a significant
risk of thrombotic complications, but this risk has now been mark-
edly reduced.
Inhibitor formation, the development of antibodies that in-
hibit clotting activity, occurs subsequent to transfusion of a blood
product or factor replacement. Development of inhibitors presents
difficult challenges for the management of haemophilia. Therapeutic
strategies largely rely on the ability to bypass the factor VIIIa–​factor
IXa tenase complex. Inhibitor formation almost exclusively arises
in severely affected patients and occurs in approximately 7 to 52%
of patients with haemophilia A but in only about 1 to 3% of patients
with haemophilia B. This difference in inhibitor formation is not
completely understood, but possible explanations include the higher
incidence of severe disease in haemophilia A, prenatal exposure to
maternal factor IX but not factor VIII antigens due to the former’s
ability to pass through the placenta membrane, the structural simi-
larity between factor IX and other vitamin K-​dependent proteins,
the higher plasma levels of factor IX, and the greater inherent im-
munogenicity of factor VIII due to its larger size. One very rare but
severe complication that may occur with the development of a factor
IX inhibitor is the development of a potentially life-​threatening ana-
phylactic complication following first treatment.
Therapeutic strategies for treatment of patients with inhibitors
need to address the acute management of the bleeding episode as
well as a longer-​term treatment directed toward suppression of
antibody production. Quantification of the titre of factor inhibitor
involves mixing the patient’s plasma with test plasma containing a
known amount of factor, normally from a pool of healthy donors.
After incubation, factor activity levels present in the patient’s incu-
bation mixture are compared with that in a control mixture so that
the amount of inhibitory antibody can be calculated. The Bethesda
unit (BU), the standard unit used to report a titre of factor inhibitor,
represents the amount of inhibitor that inactivates 50% of factor ac-
tivity. Acute management of bleeding in a patient with an inhibitor
relies first on quantifying the BU of the inhibitor. With low-​titre in-
hibitors (<5 BU), it may be possible to overwhelm the inhibitor with
aggressive concentrate therapy. With high-​titre inhibitors, it is usu-
ally necessary to bypass the inhibitor using either prothrombin com-
plex concentrates, activated prothrombin complex concentrates,
porcine factor VIII, or recombinant factor VIIa. These bypassing
agents, with the exception of porcine factor VIII, largely work by dir-
ectly activating factor X to factor Xa, thus bypassing the need for the
intrinsic tenase complex. With porcine factor VIII, it is wise to first
perform testing to ensure that the patient’s inhibitor does not cross-​
react with the porcine factor VIII. Because of the life-​threatening
bleeding complications in patients with inhibitors, immune toler-
ance regimens have been designed with the aim of eradicating the
inhibitor in the long term. Therapeutic regimens involve infusion
of high-​dose factor concentrates enabling a tolerance to the defi-
cient factor. This protocol is highly effective in approximately 80%
of patients.
Viral diseases
Severely affected patients treated with plasma-​derived concentrates
before 1985 had an extremely high rate of viral disease. Over the
course of a 70-​year lifespan, a patient with severe haemophilia may
be exposed to donations from 70 million individuals as a result of
the pooling of thousands of donor units for concentrate production.
Specific laboratory tests to screen for HIV, hepatitis C, and hepa-
titis B, in addition to much improved donor screening procedures,
have dramatically limited the number of contaminations from indi-
viduals carrying viral diseases. Solvent detergent treatment proced-
ures, which inactivate enveloped viruses (HIV and hepatitis viruses
B, C, D, and G), and heat treatment procedures used to eliminate
nonenveloped viruses (hepatitis A  and E viruses and parvovirus
B19) have radically decreased the risk of viral infection from plasma-​
derived products. The viral inactivation procedures in current
section 22  Haematological disorders
5538
use include pasteurization, vapour heating, high-​dry heating, and
nanofiltration. Recently, β-​propiolactone ultraviolet inactivation has
been discontinued because of ineffective viricidal technique.
HIV  In the late 1970s and early 1980s, HIV appeared in the blood
supply before routine laboratory testing was developed to detect
its presence. The leading cause of death in American haemophilia
patients in 1982 was haemorrhage; however, contaminated blood
products during the period between 1979 and 1983 led to a sharp
rise in viral disease shortly thereafter. A large proportion of patients
with haemophilia became infected with HIV and have subsequently
died from AIDS. Risk factors for infection included the severity of
the disease (severely affected patients were much more commonly
affected than those with moderate or mild disease), the type of
concentrate used (factor VIII vs factor IX concentrate), the viral
inactivation procedures used in product preparation, and the geo-
graphical location of the patient with regard to percentage of blood
products contaminated.
The incidence of HIV infection in American patients who received
plasma-​derived concentrates between 1979 and 1984 was lower
in patients receiving factor IX complex concentrates (55%) than
in those receiving factor VIII concentrates (approximately 90%).
Despite the devastating consequences of HIV for affected individ-
uals and families, the projected impact on births of patients with
haemophilia over the next two centuries is small (1.79% reduction).
Hepatitides  Contaminated plasma-​derived products also led to
significant morbidity and mortality due to hepatitis viruses. Effective
viricidal techniques have greatly reduced the incidence of hepatitic
viral disease in this population. In the United States of America in
the late 1980s, 87% of the 345 HIV-​negative and more than 99% of
the HIV-​positive patients showed evidence of prior infection with
hepatitis B, hepatitis C, or hepatitis D viruses. Infection due to hepa-
titis A virus has rarely been reported in patients with haemophilia in
the United States of America. Solvent/​detergent inactivation of con-
centrates has been associated with a high prevalence of antibodies to
hepatitis A virus.
Hepatitis B was commonly seen in patients with haemophilia until
routine screening of liver enzymes and the subsequent availability
of hepatitis-​specific antibody and antigen tests in the 1980s. Most
patients are now vaccinated against hepatitis B so that it is difficult
to estimate hepatitis B infection from concentrate administration.
The hepatitis delta virus, dependent on coinfection with hepatitis
B virus, has also been a significant cause of illness in patients with
haemophilia; its prevalence is largely attributed to the administra-
tion of prothrombin complex concentrates.
Routine testing for hepatitis C, instituted in the early 1990s, has
reduced but not eliminated hepatitis C contamination in the donor
pool. A variable susceptibility and morbidity is seen in response to
hepatitis infection; cirrhosis was estimated at approximately 20%
and liver failure at 10 to 20% 20 years after infection. Concurrent
infection with HIV can accelerate complications of hepatitis C virus.
There is also an increased likelihood of hepatocellular carcinoma
with long-​term infection with viral hepatitis.
Other infectious agents  The great majority of patients with haemo-
philia have antibodies to parvovirus B19. This is a small, nonlipid-​
enveloped, highly heat-​resistant virus found to contaminate
plasma-​derived products. Methods that inactivate other nonenveloped
viruses in products have not proved to be routinely effective against
this virus. Although parvovirus B19 infection is often mild and self-​
limited, infection with parvovirus B19 has the potential to severely
compromise the health of an infected immunodeficient patient.
There is experimental evidence in animal models that cellular
blood components, plasma, and plasma components have a po-
tential, though minimal, risk of transmitting the prion disease
Creutzfeldt–​Jakob disease (CJD). To date, no definitive direct infec-
tion of a recipient of a blood product or blood product concentrate
has been documented, although transmission has been reported
from corneal and dura mater grafts and cadaveric pituitary hor-
mones. The American Red Cross currently administers a question-
naire to screen donors for risk of prion disease. There is now concern
that transmission of new variant Creutzfeldt–​Jakob disease (vCJD)
may differ from classical CJD and could potentially be transmitted
through plasma-​derived concentrates. This new variant has been
associated with outbreaks of bovine spongiform encephalopathy,
potentially from dietary exposure. Experimental evidence shows
that bovine spongiform encephalopathy and vCJD are caused by the
same infectious agent, and prion-​related protein has been found in
lymphoid tissue of patients with vCJD. In Europe, large amounts of
plasma products have been removed from the market as a result of
concerns about the risk of vCJD.
Treatment of patients infected with hepatitis virus
and/​or HIV
Vaccination against hepatitis A and B is highly recommended for
patients who receive concentrates and lack viral antibodies indi-
cative of past infection. Treatment of hepatitis C with direct acting
antivirals has excellent rates (>80%) of sustained virological re-
sponse. Liver transplantation has been successful in many cases
for patients with liver failure who are unresponsive to treatment.
The liver transplant fortuitously corrects the deficiency of clotting
protein due to synthesis of clotting factors by the orthotopic liver.
However, the possibility of reinfection with viral disease is signifi-
cant and must be included in management decisions.
Therapies continue to rapidly improve for HIV treatment.
Healthcare providers for patients with both haemophilia and HIV
must keep abreast of changing HIV therapies and work alongside
infectious disease specialists. Drug-​related hepatitis in haemophilia
patients has been reported subsequent to treatment of HIV, particu-
larly in response to indinavir. Additionally, complications in HIV-​
positive haemophilia patients taking protease inhibitors include
haematuria, intracranial bleeds, and excessive bleeding often re-
quiring hospitalization and administration of higher than expected
doses of factor concentrate to correct the bleeding. Protease inhibitor
therapy should not be withheld from HIV-​positive individuals with
haemophilia. A 6-​month prospective study of 20 haemophilia pa-
tients receiving protease inhibitors revealed only one unusual bleed,
which was corrected by factor infusion.
Future directions: gene transfer as a method
of treating haemophilia
The development of clotting factor concentrates resulted in a
dramatic improvement in life expectancy for individuals with
haemophilia. Nonetheless, this treatment strategy has a number of
disadvantages. The protein must be infused intravenously, and has
22.7.4  Genetic disorders of coagulation
5539
a relatively short half-​life in the circulation. This makes chronic
prophylaxis difficult, especially in small children where venous ac-
cess may present a problem. In addition, the product is so expensive
that only about one-​quarter of the world’s patients with haemophilia
(those in the developed world) have access to it. Although current
viral inactivation techniques have largely eliminated the risk of HIV
and hepatitis in plasma-​derived products, there are ongoing con-
cerns about the risk of other blood-​borne diseases (CJD, transfusion-​
transmitted viruses) that are not easily eradicated using current
techniques. These concerns have fuelled interest in the development
of a gene transfer approach to the treatment of haemophilia. Such
an approach, if successful, would result in continuous production
of a level of clotting factor adequate to prevent bleeds rather than
treating bleeds after they have occurred. The level of clotting factor
required for this goal can be predicted based on a generation of ex-
perience with clotting factor concentrates. Thus, in Swedish prophy-
laxis studies, it has been shown that maintenance of trough factor
levels in the range of 1 to 3% are adequate to prevent all the life-​
threatening bleeds and most of the joint bleeds in boys with severe
haemophilia. Data from a large natural history study show that pa-
tients with mild haemophilia born with circulating FVIII levels 12%
have annualized bleeding rates of zero. These data provide bench-
mark goals for levels of expression in gene therapy.
Successful gene transfer approaches require three elements:  a
therapeutic transgene, a means of delivering it (i.e. a vector), and an
appropriate target cell type in which gene transfer and expression will
exert a therapeutic effect. Of the inherited diseases for which gene
transfer approaches have been attempted, haemophilia has a number
of advantages. First, tissue-​specific expression is not required.
Although clotting factors are normally synthesized in the liver, bio-
logically active material can be synthesized in a variety of tissues,
including fibroblasts, muscle cells, and endothelial cells. This allows
latitude in the choice of target cell. Second, the therapeutic window
is wide, since even small increases in circulating levels of factor are
likely to result in some improvement in symptoms, and ­increases up
to 100% would provide the patient with physiological normal activity
levels. Excellent small and large animal models of the diseases exist
(murine and canine), and determination of therapeutic efficacy is in
the case of haemophilia relatively straightforward, since levels of
circulating factor correlate well with symptoms of the disease.
A number of different strategies for gene therapy for haemophilia
have been investigated in preclinical studies, and five of these were
evaluated in early phase clinical trials conducted in the United States
of America beginning in 1999. The plethora of approaches suggests
that there may eventually be more than one successful combination
of vector and target tissue that is safe and effective for haemophilia;
this is an advantage, since the haemophilia population is a hetero-
geneous one, and approaches that work well in some instances (e.g.
delivery of a viral vector to the liver), may be impossible for indi-
viduals with severe liver disease due to prior hepatitis C infection.
One of these original studies, in which an AAV vector expressing
human factor IX was infused into the hepatic artery in men with
severe haemophilia B, resulted in expression of therapeutic levels of
factor IX, in the range of 10 to 12%, for a period of several weeks, but
expression was gradually lost. The decline in factor IX levels was ac-
companied by an asymptomatic and self-​limited rise in serum trans-
aminases. Studies of the immune response to vector in a subsequent
subject documented an expansion of a population of capsid-​specific
CD8+ T cells following vector infusion; this population contracted
after the loss of factor IX expression. In vitro studies demonstrated
that peptides derived from the AAV vector are displayed on the sur-
face of the transduced cell in the context of MHC class I molecules;
the molecular basis of the loss of expression and rise in liver trans-
aminases then appeared to be presentation of AAV-​derived capsid
sequences on the surface of the transduced hepatocyte, which
flagged them for destruction by circulating lymphocytes. Because
the only source of capsid is the material that is infused, that is, there
is not ongoing synthesis of capsid proteins, the conclusion was that
a short course of immunomodulatory therapy, administered in re-
sponse to a rise in liver enzymes or a fall in factor IX levels, would
block the response and allow long-​term expression. In a subsequent
trial that used an AAV8 vector with a self-​complementary DNA
structure and a codon-​optimized sequence, provision was made
to institute a short course of high-​dose prednisolone if indicated.
This proved effective and participants infused at the highest dose,
2 × 1012 vg/kg, have shown long-term expression (at last report, up
to 8 years with observation ongoing) of circulating factor IX levels in
the range of 4 to 6%, enough to convert severe haemophilia B to mild
disease. A key advance in this trial was the realization that the strong
tropism of AAV vectors for the liver permitted intravenous infusion
of the vector rather than infusion through the hepatic artery in an
interventional radiology procedure. Subsequent trials have substi-
tuted a high specific activity Factor IX molecule (Factor IX Padua)
for native Factor IX; this has driven durable expression of Factor IX
in the range of 30%, at lower doses of AAV vector (5 × 1011 vg/kg).
The large size of the factor VIII cDNA slowed efforts to extend this
approach to haemophilia A, but multiple trials of AAV-mediated
gene transfer, some in late phase testing, are now underway. The re-
sults in the AAV studies underscore the iterative nature of advances
in gene transfer, with problems encountered in early phase testing in
humans requiring additional studies in the laboratory or in preclin-
ical animal models to inform the next generation of clinical studies.
Von Willebrand’s disease
In 1926, Erik von Willebrand first described what we now know
as von Willebrand’s disease upon finding an autosomally in-
herited bleeding diathesis in a large kindred on the Åland islands
in the Gulf of Bothnia between Sweden and Finland. Although
the bleeding disorder in this family resulted in haemorrhagic
death in multiple family members, the bleeding diathesis in pa-
tients with von Willebrand’s disease is usually much milder. Most
commonly, patients with von Willebrand’s disease manifest mu-
cosal platelet-​type bleeding tendencies of varying severity. Nose
bleeds, menorrhagia, and easy bruising are the most common
manifestations.
The pathophysiology of von Willebrand’s disease involves a
functional deficiency of von Willebrand factor (VWF), a 270-​kDa
monomer that forms a large multimeric plasma glycoprotein com-
prised of several up to 100 subunits. Synthesized in the megakaryocyte
and endothelial cell and stored in subcellular granules, VWF enables
proper two-​chain factor VIII formation and serves as a carrier, thus
preventing degradation of factor VIII and lengthening the half-​life
of the labile factor VIII protein to around 8 h. VWF secreted by
endothelial cells also binds to heparin glycosaminoglycan and to
the platelet glycoprotein complex Ib–​IX enhancing platelet activa-
tion and further platelet recruitment at sites of tissue damage. The
section 22  Haematological disorders
5540
interaction between platelets and VWF is thought to provide the
explanation for the mucosal bleeding phenotype occurring in pa-
tients with von Willebrand’s disease. These patients frequently have
reduced levels of factor VIII. However, the remaining factor VIII
is normally sufficient to prevent the haemophilia-​type symptom-
atology of arthropathy and deep tissue bleeding.
The 180-​kb gene for VWF is located on chromosome 12 and
consists of 52 exons. There are three types of von Willebrand’s dis-
ease: types 1, 2, and 3. Types 1 and 3 are quantitative deficiencies of
the VWF, but type 2 is a qualitative deficiency due to binding defects
of the VWF. The inheritance of types 1 and 3 are autosomal dom-
inant and autosomal recessive, respectively. However, rare reports
of an autosomal dominant inheritance pattern for type 3 have been
published. There are four principal subtypes of type 2 classified as
follows: 2A, absence of high molecular weight VWF species causing
decreased platelet-​dependent function; 2B, increased affinity of
VWF for platelet glycoprotein Ib–​IX; 2M, platelet functional defect
not caused by the absence of high molecular weight multimers; and
2N, a factor VIII binding abnormality (Table 22.7.4.2).
Laboratory diagnosis of von Willebrand’s disease involves as-
saying the plasma for VWF. The two principal tests are an antigenic
test (VWF antigen) and an activity test (VWF ristocetin cofactor)
in which formalin-​fixed platelet aggregation is induced due to the
ristocetin-​enhanced VWF binding to glycoprotein complex Ib–​IX.
Comparison of the tests helps identify the enhanced ristocetin-​
induced aggregation seen in type 2B von Willebrand’s disease
where VWF ristocetin cofactor is typically much lower than VWF
antigen. Other tests performed in the evaluation of von Willebrand’s
disease include the level of factor VIII, which is often decreased,
and the APTT, which is elevated in approximately half of cases
of von Willebrand’s disease due to the low activity of factor VIII.
Nonreducing gel immunoelectrophoresis is employed to assay the
distribution of multimeric subunits of VWF with a gel containing
antibody to VWF antigen. This assay is particularly relevant for visu-
alization of the presence of low, intermediate, and high molecular
weight VWF subunits. The intermediate and high molecular weight
species are markedly decreased in subtypes of type 2 disease. A de-
creased normal pattern is seen in type 1 disease, although the de-
creased visual intensity may be difficult to quantify. Type 3 disease
shows near absence of all subunit molecular weights. The lower
limit of the normal range of VWF varies with blood type (A, B, O,
AB). Thus, symptomatology must be evaluated based on normal
ranges for each blood type. The bleeding time in a patient with von
Willebrand’s disease is most often prolonged; however, the test is no
longer routinely necessary because of the nonspecific nature of a
positive result and the higher specificity of other testing.
The specific treatment for von Willebrand’s disease varies with a
patient’s symptoms, the circumstances of the need for treatment, the
subtype of von Willebrand’s disease, laboratory results indicating the
potential success of increased VWF with nonprotein-​based treat-
ment, and the clinical experience with a particular patient and the
biological family members. When possible, treatment is preferred
without blood products. The mainstay of treatment for mild disease
is treatment with the synthetic octapeptide desmopressin, DDAVP.
This causes release of factor VIII and VWF from endothelial cells
raising the plasma VWF by approximately two-​ to tenfold. Thus
treatment with DDAVP relies on a partial quantitative deficiency of
VWF. Intravenous and nasal preparations are available. The nasal
preparation allows a patient to self-​administer medication at ei-
ther regular intervals or on an as-​needed basis. The phenomenon of
tachyphylaxis, the decreased effectiveness of repeated doses of the
compound, does occur, and there is usually little response after three
consecutive doses. In the past, DDAVP was considered contraindi-
cated in type 2B von Willebrand’s disease because of the thrombo-
cytopenia sometimes observed with DDAVP infusion. However,
this recommendation is controversial and should be assessed on a
case-​by-​case basis.
Patients with type 3 von Willebrand’s disease may lack sufficient
intracellular reserves for effective therapy; thus alternative meas-
ures for such patients are usually necessary. A better understanding
of VWF function in vivo exposes limitations with current clinical
testing for VWD activity measured during static conditions. These
testing difficulties largely result from a lack of simulated physiologic
flow environments. Using functional flow-​based tests along with
collagen and platelets may help better analyse patient disease ac-
tivity. New practical tests may help safely classify patients as higher
or lower risk for clinical bleeding.
A trial of effectiveness of DDAVP is often indicated, particu-
larly prior to prophylactic surgical use of the compound. The trial
is normally performed after subtyping the VWF disease to ensure
that DDAVP is not contraindicated, as in type 2B. Optimally, the
test should not be given within 24 h of the last DDAVP infusion nor
at a time of environmental stress in order to minimize problems as-
sociated with tachyphylaxis or depletion of intracellular reserves.
A therapeutic trial entails measurement of VWF antigen before and
1 h after DDAVP infusion of 0.3 μg/​kg. The patient should be moni-
tored carefully during this period because of possible flushing, mild
anaphylactoid reactions, and possible hyponatraemia.
ε-​Aminocaproic acid is frequently administered in the setting of
dental surgery to inhibit fibrinolysis. However, care must be taken
in administration to patients with a predisposition to thrombosis
because of the potential deleterious effects of ε-​aminocaproic acid in
this setting. Other compounds which may be administered include
oestrogens in women because of the natural positive regulation of
synthesis of VWF with oestrogen compounds. This may ameliorate
menorrhagia in such patients. Components in cryoprecipitate in-
clude factor VIII, fibrinogen, and factor XIII, in addition to VWF.
Cryoprecipitate had been the mainstay of plasma-​based therapy
until the recent availability of factor VIII concentrates with pre-
served VWF protein such as Alphanate and Humate P. The use of
cryoprecipitate, which does not undergo viral inactivation, has thus
fallen out of favour.
Treatment of von Willebrand’s disease with DDAVP is the method
of choice in patients who respond to this therapy. DDAVP for intra-
venous or subcutaneous use is supplied as either a 4-​μg/​ml 10-​ml
vial or a 15-​μg/​ml 1-​ or 2-​ml vial preparation. The recommended
Table 22.7.4.2  Subtypes of von Willebrand disease type 2
Subtype
Change in binding affinity
Characteristics
2A
Decreased platelet binding
Loss of high molecular
weight multimers
2B
Increased platelet GP1α binding
Thrombocytopenia
2N
Decreased factor VIII binding
Low factor VIII levels
2M
Decreased factor VIII binding
Normal multimers
22.7.4  Genetic disorders of coagulation
5541
dose is 0.3 μg/​kg, mixed in 30 ml normal saline, infused slowly over
30 min or 0.4 μg/​kg subcutaneously. This dose may be repeated after
12 to 24 h. A DDAVP nasal spray is available in a metered dose pump
which delivers 0.1 ml (150 μg) per actuation. The bottle is at a con-
centration of 1.5 mg/​ml and contains 2.5 ml with a nasal spray pump
which can deliver twenty-five 150-​μg or twelve 300-​μg doses. For
administration, patients who weigh less than 50 kg should deliver
one 150-​μg spray in one nostril. For those weighing over 50 kg, one
spray should be delivered in each nostril for a total dose of 300 μg.
Administration may be repeated after 24 h. Precautions to take with
the medication include administration no more than every 24 h or
for three consecutive days unless under the supervision of personnel
from a haemophilia treatment centre. The medication should not
be used in pregnant women or in children under 2 years of age. The
medication should be used with caution in the elderly and in indi-
viduals with a history of cardiovascular disease.
ADAMTS13 deficiency—​thrombotic
thrombocytopenic purpura
The gene encoding the novel metalloproteinase ADAMTS13 (a
disintegrin-​like and metalloproteinase with thrombospondin type-​
1 motifs) was discovered in 2001. ADAMTS13, located on chromo-
some 9q34, cleaves the peptide bond at Tyr1605 to Met1606 in
the A2 domain of VWF. Normally, ADAMTS13 rapidly degrades
‘unusually large’ VWF multimers into smaller multimers. Lack of
ADAMTS13 due to familial absence or acquired inhibition may
result in thrombotic thrombocytopenic purpura with an increase
of the ultra-​large multimers and formation of platelet clumps and
microthrombi (see also Chapter 22.7.4).
Combined deficiency of coagulation cofactors factor V
and factor VIII
Combined deficiency of coagulation cofactors factor V and factor
VIII (F5F8D) is a rare autosomal recessive disorder due to genetic
mutations in the coordinated system of protein trafficking. The dis-
order results from mutations in the genes for the transmembrane
lectin LMAN-​1 (ERGIC-​53) on chromosome 18 (18q21.3–​q22) or
its protein complex partner MCFD2 (multiple coagulation factor
deficiency 2) located on chromosome 2 (2p21). MCFD2 recruits
glycoproteins factors V and VIII for endoplasmic reticulum–​Golgi
transport by the molecular chaperone ERGIC-​53. LMAN1 or
MCFD2 mutations that cause this rare disorder in patients result
in indistinguishable clinical manifestations with mild to moderate
bleeding symptomatology. Plasma antigen and activity levels of both
factors V and VIII measure between 50 and 300 U/​L.
Factor XI deficiency
Factor XI deficiency is an autosomal recessive bleeding diathesis
of variable severity. It was first described in 1953 as a third type of
haemophilia and is thus sometimes referred to as haemophilia C
or Rosenthal syndrome. The deficiency predominantly occurs in
eastern European Ashkenazi Jews, accounting for more than 50%
of cases. In Ashkenazi Jews, the disorder is reported to occur in 5 to
11% of individuals in the heterozygous state and 0.1 to 0.3% in the
homozygous state. Genetically, the mutations are grouped into three
types: type I, abnormalities in the intron–​exon splice boundaries;
type II, mutations that result in a premature stop in translation; and
type III, mutations resulting from a missense mutation.
The protein itself is an 80-​kDa protein that circulates in the plasma
as a zymogen in a noncovalent association with high molecular
weight kininogen. It contains four apple domains in its protein
structure, and although factor XIa is a cleaving protease, its structure
differs from the vitamin K-​dependent serine protease coagulation
proteins. Factor XI is principally activated by factor XIIa in the pres-
ence of a negatively charged surface (contact activation). The lack
of any bleeding diathesis related to a severe deficiency of factor XII
suggests the importance of thrombin as an alternative mechanism of
in vivo factor XI activation.
The in vitro factor XI activity level does not correlate well with
clinical phenotype. Family history of the bleeding complications and
the specific mutated sites are more predictive. Bleeding manifest-
ations are rare in heterozygotes and occur in approximately 50% of
homozygous patients.
Factor XI activity levels are assayed in an APTT-​based test.
Bleeding problems include easy bruising, epistaxis, haematuria, post-
partum haemorrhage, haematomas, and menorrhagia. Haemophilia
symptoms, including haemarthroses and intramuscular bleeding,
are rare. Bleeding most frequently occurs after trauma or surgery.
Damage to tissues rich in fibrinolytic activity, such as oral mucosa
and the prostate, are more commonly associated with bleeding
problems.
Therapy for patients with factor XI deficiency is indicated for
symptomatic bleeding and prophylactically for surgery in pa-
tients with markedly reduced levels (i.e. <20%), unless there is
no personal or family history of any bleeding complication. Fresh
frozen plasma should be readily available at surgery for infusion
in case of a bleeding emergency. Factor XI has a half-​life of 60 to
80 h; 10 ml plasma/​kg per day is usually adequate for maintaining
haemostasis. Prophylactic therapy for most surgery includes re-
placement of factor XI with plasma at a loading dose of 15 ml/​kg
followed by 3 to 6 ml/​kg every 24 h. The protective level for surgical
prophylaxis is suggested as 45% for major surgery and 30% for
minor surgery.
Antifibrinolytic therapy with ε-​aminocaproic may be a helpful
adjunct to plasma therapy; however, antifibrinolytics should be
avoided in patients with haematuria or bleeding in the bladder be-
cause of possible obstruction by clots.
Deficiencies of proteins in the tissue factor and
common pathways
The autosomally inherited deficiencies of factors II, V, VII, and X re-
sult in bleeding diatheses of varying severity. Such deficiencies of co-
agulation factor correlate poorly with tests of in vitro factor activity;
these are thus quite different disorders from haemophilia, in which
in vitro assessment predicts the clinical phenotype very well. These
factor deficiencies can best be assessed by an initial screen using
the PT as a measurement of the tissue factor pathway. Although the
APTT may be prolonged with deficiencies of factors II, V, and X, but
not VII, the PT is most often much more sensitive.
Factors II, VII, and X, are structurally homologous containing a
signal peptide, a propeptide region necessary for recognition by the
post-​translational modifying enzyme γ-​glutamyl carboxylase, an
intermolecular binding region (two epidermal growth factor (EGF)
domains in factors VII, IX, and X and two kringle domains in the
prothrombin molecule), and a catalytic domain in the C-​terminal
of the molecule.
section 22  Haematological disorders
5542
Deficiency of prothrombin (factor II) results from a lack of pro-
thrombin or a malfunctional prothrombin protein. Deficient patients
present with haemorrhagic manifestations. All reported patients
with a prothrombin deficiency retain some prothrombin, suggesting
that complete prothrombin deficiency is incompatible with life. This
is consistent with the knockout mouse model which results in em-
bryonic lethality at 9.5 to 11.5 days postcoitum in over 50% of fetuses;
however, for some unknown reason, some murine knockout pro-
thrombin fetuses are able to survive to birth but promptly die within
2 days due to haemorrhage. Patients with heterozygous prothrombin
deficiency most commonly are either asymptomatic or have min-
imal bleeding. Bleeding manifestations include easy bruising, soft-​
tissue haemorrhage, excessive postoperative bleeding, epistaxis, and
menorrhagia in women. Haemarthroses are uncommon.
Congenital disease is characterized by a lifelong and a family
bleeding history. Levels of 20 to 30% prothrombin normally prevent
symptomatic bleeding. When necessary, administration of plasma is
recommended at doses of 15 to 20 ml/​kg followed by 3 ml/​kg every
12 to 24 h. Prothrombin complex concentrates can be administered
for serious bleeds and as prophylaxis before surgery. Transmission
of viral disease and thromboembolic phenomena are risks of the ad-
ministration of prothrombin complex concentrates.
Factor V deficiency occurs in fewer than one in a million indi-
viduals. Approximately 20% of the body’s factor V reserve resides
in the platelets. Thus, it is not surprising that patients with factor V
deficiency tend to have mucosal bleeding manifestations including
epistaxis, gastrointestinal bleeds, and menorrhagia in women.
Haemarthroses, although a possible complaint, are much less
common than in haemophilia. Mild to moderate bleeding may be
treated by raising the factor V activity to about 20% of normal with
a plasma dose of approximately 15 to 20 ml/​kg followed by 3 to 6 ml/​
kg every 24 h. Due to the large amount of factor V stored in α gran-
ules, platelet transfusions may be an appropriate therapy. However,
patients should be monitored for the possibility of generation of
antiplatelet antibodies.
Factor VII deficiency presents as a variable bleeding disorder
ranging from mild to severe, with a possibility of fatal intracranial
haemorrhage. Patients with homozygous or compound hetero-
zygous mutations manifest symptoms similar to those of a patient
with haemophilia. However, unlike the correlation between activity
levels and severity of disease in haemophilia, the in vitro factor VII
activity clotting test provides only a relative indication of possible
disease manifestations. This, in part, is caused by different tissue
factor origins utilized within clinical laboratory thromboplastins.
Manifestations of factor VII deficiency include haemarthrosis,
arthropathies, haematoma formation, and retroperitoneal bleeding.
Fatal intracranial haemorrhage is estimated to occur in approxi-
mately 16% of patients with severe disease. Levels below 10% ac-
tivity most often result in bleeding manifestations. Therapy includes
replacement of factor VII activity levels to 10 to 25% for patients
undergoing most types of surgery. Treatment options include plasma
at 5 to 10 ml/​kg for 6 to 12 h for 1 to 2 days for minor episodes. For
surgery, the recommended dose is administration of 15 to 20 ml/​kg
followed by maintenance doses of 3 to 6 ml/​kg every 12 h.
Prothrombin complex concentrates may frequently be used to
supply the factor VII along with the other vitamin K-​dependent
proteins. Although thrombogenicity has only been reported on rare
occasions, this does remain a minor yet potential complication. In
July 2005, recombinant coagulation factor VIIa (NovoSeven) was
officially approved in the United States of America for treating pa-
tients with factor VII deficiency. Intravenous doses of 15 to 30 µg/​kg
are therapeutically effective in this setting for acute bleeds, a signifi-
cantly lower dose than that used for treatment of haemophilia pa-
tients with inhibitors. However, the possible development of a factor
VII inhibitor must be considered, as this has been reported. The
product is administered every 2 h for prophylaxis during surgery for
the first 24 h, then reduced to every 3 h 24 to 48 h postoperatively,
and then further reduced according to patient symptomatology and
necessity, depending on the risk of bleeding into the surgical site.
Factor X deficiency may present with symptomatology similar to
that of a patient with severe haemophilia. Haemarthroses, soft tissue
haemorrhages, retroperitoneal bleed, central nervous system haem-
orrhages, pseudotumours, and menorrhagia may occur.
Therapy with fresh frozen plasma includes a loading dose of 10 to
15 ml/​kg followed by approximately 50% of that at 24 h.
Deficiency of the contact activating factors, factor XIII,
and fibrinogen
Although the APTT is grossly prolonged (often >150 s) with de-
ficiencies of the contact activating factors—​factor XII, high mo-
lecular weight kininogen, and prekallikrein—​these deficiencies are
not associated with bleeding manifestations and will not be covered
further here.
Factor XIII deficiency often presents shortly after birth with
bleeding of the umbilical cord. Patients with clinical manifest-
ations typically have factor levels of less than 1%. Factor XIII is a
transglutaminase that cross-​links fibrin monomers, thus stabilizing
a forming fibrin clot. Patients with deficiency of factor XIII therefore
have delayed wound healing and often suffer from soft tissue haem-
orrhages, haemarthroses, haematomas, and excessive bleeding from
poorly healed wounds. Up to 25% of individuals deficient in factor
XIII may experience intracranial bleeding. For unknown reasons,
affected men may have oligospermia and affected women may suffer
from repeated spontaneous abortions. Since routine clotting tests
are normal in factor XIII deficiency, a physician must specifically
request a test for factor XIII deficiency which entails a clot solubility
test using 2% chloroacetic acid on a formed clot. Treatment of factor
XIII deficiency involves administration of small amounts of factor
XIII required to minimize bleeding complications. Prophylaxis in-
cludes using 2 to 3 ml/​kg of fresh frozen plasma every 4 to 6 weeks
or one bag of cryoprecipitate per 10 to 20 kg every 3 to 4 weeks. To
prevent spontaneous abortions, products containing factor XIII can
be administered every 14 to 21 days.
Afibrinogenaemia may cause dangerous haemorrhagic episodes.
However, it is somewhat surprising that the mutation does not lead
to embryonic death in light of the fact that the blood is incoagulable
in vitro. The lack of necessity for fibrinogen during fetal develop-
ment is supported by the viable fibrinogen knockout mouse model.
Prolonged bleeding from the umbilical cord often permits early
recognition of an affected child. The leading cause of death in
afibrinogenaemia is intracranial haemorrhage. Haemorrhages from
mucous membranes occur frequently, and haemarthroses occur
in approximately 20% of patients. Pregnancy-​related problems in-
clude first-​trimester abortion, placental abruption, and postpartum
bleeding complications and may be markedly reduced by adminis-
tration of fibrinogen. However, fibrinogen replacement may cause
22.7.4  Genetic disorders of coagulation
5543
thromboembolic phenomena. The target fibrinogen level for re-
placement therapy is approximately 50 to 100 mg/​dl. One bag of
cryoprecipitate contains approximately 250 mg of fibrinogen; thus
dosing of cryoprecipitate usually necessitates 5 to 10 bags per 70 kg
person. Therapeutic complications include allergic reactions and
the development of antifibrinogen antibodies. Thromboembolic
phenomena may occur in conjunction with fibrinolytic inhibitors
or oral contraceptives.
Dysfibrinogenaemia results from a functional deficiency of fi-
brinogen associated with a malfunctioning molecule, although
some degree of antigen remains present. Approximately 55% of pa-
tients with dysfibrinogenaemia remain asymptomatic, 25% have a
bleeding tendency, and 20% may experience thrombotic episodes
ranging from mild to fatal events.
Combined defects
Numerous combined deficiencies have been described; the under-
lying mutation for several of these combined deficiencies has been
determined. Combined deficiency of the two structurally similar
proteins factor V and factor VIII is an autosomal recessively inherited
disorder of variable bleeding severity (mentioned previously). Other
combined deficiencies for which a genetic mechanism has been de-
scribed include deficiency of factors II, VII, IX, and X caused by a
mutation in the γ-​glutamyl carboxylase gene, required for a critical
post-​translational modification in vitamin K-​dependent factors.
Vitamin K clotting deficiency subtypes 1 and 2 result from func-
tional deficiencies of enzymes γ-​glutamyl carboxylase and VKORC,
respectively. Symptoms occur not only from defective functions of
all vitamin K-​dependent coagulant factors (both pro and inhibitory)
but defective protein γ-carboxylation results in nonhaemostatic de-
velopmental and skeletal abnormalities.
Hypercoagulable disease due to deficiencies
of anticoagulant
Pathological diseases resulting from inappropriate clot formation
in either the arterial or venous circulation is a major cause of mor-
bidity and mortality worldwide. The genetic contribution to this
pathophysiology is not well understood, particularly concerning
thrombosis in the arterial circulation. Clearly cardiovascular disease
represents a complex multifactorial process.
The contribution of genetic factors to venous thrombotic disease
is better understood; it may be associated with either an isolated
deficiency of an anticoagulant protein, a malfunctioning procoagu-
lant protein, or a combination of these processes. The functional
deficiencies become particularly relevant during times of increased
environmental stress such as in the puerperium or in postsurgical,
traumatic, or immobilized states. In addition to deficiency states,
several common mutations involving a gain of function have also
been described which can disrupt the delicate balance of coagula-
tion by shifting the balance toward greater procoagulant function.
Procoagulant and anticoagulant plasma proteins interact with
platelets and cellular phospholipids to promote physiological co-
agulation. Regulation of thrombin formation is the key step in
the proper balance between pro-​ and anticoagulant functions.
Anticoagulant proteins are particularly important in areas where
there may be prolonged exposure of procoagulant factors and
platelet phospholipids to the vessel wall, predisposing an individual
to thrombotic disease. Deficiencies of anticoagulant proteins thus
place a patient at an increased risk for thrombosis in the slowly
flowing venous circulation. In the rapidly flowing arterial circula-
tion, laminar flow largely prevents prolonged interaction between
platelets and vessel walls.
The principal anticoagulant proteins that keep the procoagulant
proteins in check include thrombomodulin, tissue factor pathway in-
hibitor, antithrombin III, protein C, and protein S. Thrombomodulin,
an integral membrane protein expressed by endothelial cells, plays a key
role in tempering the action of thrombin. Despite attempts to discover
mutations in the thrombomodulin gene, only rare reports have impli-
cated thrombomodulin in the pathophysiology of disease, although
some recent studies suggest the existence of polymorphic regulation
variants in the promoter region. Recently, a mutation in the small but
critical protein known as tissue factor pathway inhibitor, which inhibits
procoagulant function by binding to factor Xa either alone or in asso-
ciation with tissue factor–​factor VIIa, has been suggested to be associ-
ated with a ninefold increased risk of venous thrombosis.
Deficiencies leading to a hypercoagulable state are most fre-
quently caused by deficiencies of antithrombin III, protein C, and
protein S.  These anticoagulant deficiencies result from either a
quantitative deficiency (type I) or a qualitative deficiency (type II).
Deficiencies of any of these factors may cause life-​threatening deep
venous thromboses and pulmonary emboli, or may be asymptom-
atic. Clinical presentation relates to physical sequelae in the affected
organ. In addition to deep venous thromboses and pulmonary em-
boli, symptomatology may include superficial thrombophlebitis,
mesenteric vein thrombosis, and cerebral vein thrombosis.
Antithrombin III deficiency
A deficiency of antithrombin III was the first anticoagulant pro-
tein deficiency described to be associated with an increased risk of
thrombosis. Antithrombin III is a 60-​kDa glycoprotein found at high
concentrations in the plasma—​150 μg/​ml: approximately 15-​ to 30-​
fold higher than that of many other pro-​ and anticoagulant proteins.
Antithrombin III primarily inhibits thrombin but also inhibits fac-
tors IXa, Xa, XIa, XIIa, kallikrein, and plasmin. The ability to inhibit
thrombin requires interaction with heparin, which increases the in-
hibitory activity several thousandfold. Historically, the risk of throm-
bosis in individuals deficient in antithrombin III has been thought to
be higher than that seen with deficiencies of protein S or protein C, or
than that seen with increased functionality of the procoagulant pro-
teins factor V and prothrombin. Clearly the influences of gene–​gene
and gene–​environment interactions contribute to this risk. A normal
activity range for most procoagulant/​anticoagulant proteins may be
as low as 50%. However, the critical requirement for antithrombin III
can be surmised from the 80% lower limit of a normal antithrombin
III level, significantly higher than that for other coagulation proteins.
This makes the diagnosis of antithrombin III deficiency particularly
difficult in the post-​thrombotic period when patients frequently
have lower levels of antithrombin III due either to consumption of
antithrombin III during clot formation or to the decreased func-
tion seen with heparin administration. Additionally, the presence of
homozygous disease of antithrombin III deficiency has only been re-
ported with rare type II deficiencies resulting from impaired heparin
binding mutations. No homozygous type I deficiencies have been re-
ported, probably because of their incompatibility with life.
section 22  Haematological disorders
5544
The frequency of antithrombin III deficiency in patients with
thrombophilia varies widely between studies. The cause of these
differing frequencies has recently been carefully addressed by van
Boven and colleagues. Their study clearly shows the strong influ-
ence of acquired and genetic factors which modulate the baseline
risk due to one specific genetic mutation, highlighting the role of
additional factors when combined with genetics. In thrombophilic
family studies, the risk of thrombosis is 20 times greater than in con-
trol populations. The most frequent presentation is deep venous
thrombosis with a pulmonary embolism, particularly after an
inciting environmental influence such as surgery or immobiliza-
tion, or the start of oral contraceptives or pregnancy/​postpartum in
women. The average age of first onset is 33 years. In patients deficient
in antithrombin III without a known acquired risk, the rate of inci-
dence of thrombosis is less than 1% per year.
Therapy for antithrombin III deficiency includes prophylactic
treatment with warfarin, low molecular weight heparin, and treat-
ment of an acute event with heparin or another anticoagulant therapy,
for example administration of a fibrinolytic agent in the patient pre-
senting early enough during an acute episode. Antithrombin III
concentrate may be administered for therapy of deficiency during an
acute event or as a prophylactic treatment to prevent further disease.
Deficiencies of proteins C and S
Deficiencies of proteins C and S present with thrombotic mani-
festations similar to those seen with antithrombin III deficiency.
However, in protein C deficiency an additional complication
­includes warfarin-​induced skin necrosis and life-​threatening pur-
pura fulminans in the homozygous or compound heterozygous
protein C deficient neonate. A diagnosis of protein C deficiency is
found in approximately 33% of individuals with warfarin-​induced
skin necrosis, a development that may lead to skin necrosis several
days after initiation of warfarin therapy. The proposed mechanism
for this condition is due to the earlier decrease in protein C com-
pared with decreases in procoagulant proteins following initiation
of warfarin therapy (due to the short half-​life of protein C, c.6 h). It is
thus standard clinical practice to begin warfarin only after a patient
has first been anticoagulated with heparin or another immediately-​
acting anticoagulant therapy.
Protein C acts in concert with its cofactor protein S to inacti-
vate the active forms of the procoagulant cofactors, factors Va and
VIIIa. Protein C is a vitamin K-​dependent serine protease structur-
ally similar to factors VII, IX, and X. Protein S is also vitamin K
dependent because of the conserved N-​terminus but lacks enzym-
atic function because of the existence of a sex-​hormone binding
globulin domain instead of a catalytic domain at the C-​terminus.
Thrombin activates protein C to activated protein C when bound to
thrombomodulin, a protein that acts like an endothelial cell receptor
for thrombin. Symptomatic manifestations of protein C or protein S
deficiencies are similar to those of antithrombin III deficiency. Deep
venous thrombosis with or without pulmonary embolism occurs in
50% of patients by the age of 30 to 45 years, depending on the study
population. Environmental and gene–​gene interactions are par-
ticularly important. As with antithrombin III deficiency, superficial
thrombophlebitis, cerebral vein thrombosis, and mesenteric vein
thrombosis are all possible complications. Postphlebitic syndrome
presents as a complication after deep venous thrombosis in up to
50% of patients.
Deficiencies of protein Z
Like coagulant protein C and protein S, protein Z serves as a vitamin
K-​dependent anticoagulant protein regulated by membrane surface-​
associated procoagulant proteins. The 62-​ kDa vitamin K-​dependent
single-​chain glycoprotein Z acts as a cofactor to the enzyme protein
Z-​dependent protease inhibitor (ZPI), a catalytically active serpin.
Clinical, animal, and meta-​analytic studies provide evidence that ZPI
and/​or PZ deficiency can be associated with thrombotic complications.
The risk of problems amplifies with compounding haemostatic stresses
including pregnancy, and concurrent arterial or venous abnormalities.
Hypercoagulable mutations
Factor V Leiden and prothrombin 20210 mutation
Since 1994, two additional common mutations have been described
leading to an increased risk of thrombosis. These mutations, unlike the
anticoagulant protein deficiencies, are due to gain-​of-​function muta-
tions causing either an increased resistance to inactivation in factor V
(factor V Leiden) or increased levels of a procoagulant protein (pro-
thrombin) which results in higher levels of thrombin formation.
Activated protein C (APC) resistance was first described by
Dahlback in a 42-​year-​old man with a history of recurrent throm-
boses. Dahlback noted an absence of prolongation of the APTT
found after addition of APC, which is normally prolonged due to
inactivation of factors Va and VIIIa. Soon thereafter, Poort and col-
leagues identified a single mutation as the principal cause of APC
resistance in the vast majority (over 90%) of patients. The mutation
leads to a decreased ability of APC to inactivate the cofactor Va due
to an amino acid substitution (arginine for glutamine) at a critical
hydrolysis point in the factor Va protein normally enabling inacti-
vation. Other causes of APC resistance not due to factor V Leiden
include a haplotype in the factor V molecule, the H2 haplotype.
Factor V Leiden leads to thrombotic disease as described for
hypercoagulable states due to deficiencies of anticoagulant protein.
Due to the extremely high incidence of factor V Leiden in the white
population (c.5%), gene–​gene interactions play a particularly im-
portant role in manifestation of disease. It should be noted that the
frequency of factor V Leiden in most nonwhite populations is low.
The prothrombin 20210 mutation reported in 1996 results in an
increased concentration of prothrombin, also tipping the balance
towards excess thrombin formation. The cause of this increase is as-
sociated with a guanine to adenine mutation (G20210A) at the last
base of the 3′ untranslated region in the factor V gene. The mech-
anism by which this influences prothrombin levels is thought to be
post-​transcriptional.
Factor IX Padua
The occurrence of a nonhaemophilia mutation at amino acid 338
had been predicted in 1993 due to its location at a cytosine–​guanine
(CpG) hotspot. In Padua in 2009, such an X-​linked factor IX muta-
tion was identified resulting in an arginine (R) to leucine (L) (factor
IX Padua) substitution. The (R338L) substitution does not cause
factor IX deficiency but conversely correlates with a gain-​of-​function
mutation resulting in a 5-​ to 10-​fold elevated factor IX coagulant ac-
tivity. The hyperfunctional (R338L) substitution in animal experi-
mentation reveals the potential to utilize this mutation in gene-​based
future designs to treat haemophilia B patients with inhibitors.
22.7.4  Genetic disorders of coagulation
5545
Vitamin K epoxide reductase complex, subunit 1
The gene encoding vitamin K epoxide reductase (VKOR), the en-
zyme that completes the vitamin K cycle, was identified in 2004.
VKOR converts vitamin K 2,3-​epoxide to the enzymatically activated
form. The gene was independently identified by two distinct tech-
niques, a traditional positional cloning approach and a novel small
interfering RNA-​aided functional screen of the predicted locus on
chromosome 16. VKOR is a multisubunit complex; VKOR complex,
subunit 1 (VKORC1) is a 163-​amino acid integral membrane protein
(18 kDa) highly expressed in the liver. The oral anticoagulant war-
farin, used in the prophylaxis of acute and chronic thromboembolic
conditions, inhibits the action of vitamin K-​dependent proteins.
Pharmacogenetics-​based dosing algorithms based on VKORC1
genotyping have been developed to account for interindividual
variation in patient response to warfarin. Additionally, differences
in warfarin metabolism, largely based on variant cytochrome P450
complex alleles, are used when initiating warfarin dosing.
In addition, a free website (http://​www.warfarindosing.org) has
been created to help pharmacogenetically guide and adjust warfarin
doses for patients and their prescribing physicians.
Direct-​acting oral anticoagulants
Better understanding of dangers and drawbacks of warfarin have
paralleled a timely appearance of new direct-​acting oral anticoagu-
lants (DOACs). In general, these new drugs have more predictable
pharmacogenetic responses, fewer interactions with food and other
medications, and display much wider therapeutic windows.
Compared to warfarin, DOACs display shorter half-lives and
achieve their full therapeutic effect in hours as opposed to days.
Varia­tions exist between available DOACs (dabigatran, apixaban,
rivaroxaban, and edoxaban) regarding creatinine clearance, me-
tabolism by cytochrome P450 isoforms, and protein binding.
Additionally, DOACs require little to no regular monitoring, under
known conditions.
Therapy using the new drugs has become a recommended
treatment prophylactic option for both short term and long-term
thrombophrophylaxis for prevention and management of atrial fib-
rillation, stroke as well as venous thrombosis. The prior major con-
cern of the lack of reversibility has been resolved with availability
of new agents that counteract several DOACs. Time of last dosage
remains vital in knowing if, when, and how to manage excess
drug. Treatment of treat life-threatening and uncontrolled bleeds
has been established with Idarucimab to counteract dabigatran
as well as andexanet alpha to overcome an excess of apixaban and
rivaroxaban.
Issues still remain regarding higher medication cost compared with
warfarin therapy, in addition to a small but increased risk of bleeding
with some agents. However, the decrease in time and expense due to
previous laboratory monitoring in addition to the availability of re-
versal agent protocols has helped outweigh most drawbacks.
FURTHER READING
General articles about coagulation
Colman RW, et al. (2006). Overview of hemostasis. In: Colman RW,
et al. (eds) Thrombosis and haemorrhage, 5th edition, pp. 3–​16. JB
Lippincott, Philadelphia.
Davie EW, Ratnoff OD (1964). Waterfall sequence for intrinsic blood
clotting. Science, 145, 1310–​12.
Gage BF, Lesko LJ (2008). Pharmacogenetics of warfarin: regulatory,
scientific, and clinical issues. J Thromb Thrombolysis, 25, 45–​51.
Li T, et al. (2004). Identification of the gene for vitamin K epoxide re-
ductase. Nature, 427, 493–​4.
Levy GG, et al. (2001). Mutations in a member of the ADAMTS gene
family cause thrombotic thrombocytopenic purpura. Nature, 413,
488–​94.
MacFarlane RG (1964). An enzyme cascade in the blood clotting mech-
anism and its function as a biochemical amplifier. Nature, 202, 498–​9.
Nichols WC, et al. (1998). Mutations in the ER-​Golgi intermediate
compartment protein ERGIC-​53 cause combined deficiency of co-
agulation factors V and VIII. Cell, 93, 61–​70.
Roberts HR, Lozier JN (1992). New perspectives on the coagulation
cascade. Hosp Pract, 27, 97–​105, 109–​12.
Zhang B, et al. (2008). Genotype-​phenotype correlation in combined
deficiency of factor V and factor VIII. Blood, 111, 5592–​600.
Haemophilia and von Willebrand’s disease
Berntorp E, et al. (2006). Inhibitor treatment in haemophilias A and
B: summary statement for the 2006 international consensus confer-
ence. Haemophilia, 12 Suppl 6, 1–​7.
Bolton-​Maggs PH (2006). Optimal haemophilia care versus the reality.
Br J Haematol, 132, 671–​82.
Bolton-​Maggs PH, et al. (2004). Evidence-​based treatment of haemo-
philia. Haemophilia, 10 Suppl 4, 20–​4.
Ljung R, et al. (2019). Inhibitors in haemophilia A and B: Management
of bleeds, inhibitor eradication and strategies for difficult-to-treat
patients. Eur J Haematol, 102(2), 111–22.
Ota S, et al. (2007). Definitions for haemophilia prophylaxis and its
outcomes: the Canadian consensus study. Haemophilia, 13, 12–​20.
Plug I, et al. (2006). Mortality and causes of death in patients with
hemophilia, 1992–​2001:  a prospective cohort study. J Thromb
Haemost, 4, 510–​6.
Ragni MV (2006). The hemophilias: factor VIII and factor IX deficien-
cies. In: Young N, Gerson S, High K (eds) Clinical hematology,
pp. 814–​29. Elsevier, Philadelphia.
Sadler JE, et al. (2006). Update on the pathophysiology and classifica-
tion of von Willebrand disease: a report of the Subcommittee on von
Willebrand Factor. J Thromb Haemost, 4, 2103–​14.
Administration of factor concentrates
Ewenstein BM, et al. (2004). Consensus recommendations for use of cen-
tral venous access devices in haemophilia. Haemophilia, 10, 629–​48.
Federici AB, Mannucci PM (2007). Management of inherited von
Willebrand disease in 2007. Ann Med, 39, 346–​58.
Manco-​Johnson MJ, et al. (2007). Prophylaxis versus episodic treat-
ment to prevent joint disease in boys with severe hemophilia. N Engl
J Med, 357, 535–​44.
Yasunaga, H (2007). Risk of authoritarianism: fibrinogen-​transmitted
hepatitis C in Japan. Lancet, 370, 2063–​7.
Infectious diseases associated with haemophilia therapy
Ponte ML (2006). Insights into the management of emerging infec-
tions: regulating variant Creutzfeldt-​Jakob disease transfusion risk
in the UK and the US. PLoS Med, 3, e34.
Posthouwer D, et al. (2006). Treatment of chronic hepatitis C in patients
with haemophilia: a review of the literature. Haemophilia, 12, 473–​8.
Rumi MG, et  al. (2004). Hepatitis C in haemophilia:  lights and
shadows. Haemophilia, 10 Suppl 4, 211–​15.

22.7.5 Acquired coagulation disorders 5546 T.E. Wa
22.7.5 Acquired coagulation disorders 5546 T.E. Warkentin
section 22  Haematological disorders
5546
Gene therapy in haemophilia
High KA (2005). Gene transfer for hemophilia: can therapeutic ef-
ficacy in large animals be safely translated to patients? J Thromb
Haemost, 3, 1682–​91.
Manno CS, et al. (2006). Successful transduction of liver in hemophilia
by AAV-​factor IX and limitations imposed by the host immune re-
sponse. Nat Med, 12, 342.
Murphy SL, High KA (2008). Gene therapy for haemophilia. Br J
Haematol, 140, 479–​87.
Nathwani AC, et  al. (2006). Self-​complementary adeno-​associated
virus vectors containing a novel liver-​specific human factor IX ex-
pression cassette enable highly efficient transduction of murine and
nonhuman primate liver. Blood, 107, 2653.
Nathwani AC, et  al. (2011). Adenovirus-​associated virus vector-​
mediated gene transfer in hemophilia B. N Engl J Med, 365, 2357–​65.
Thrombotic disease
Bucciarelli P, Rosendaal FR, Tripodi A (1999). Risk of venous thrombo-
embolism and clinical manifestations in carriers of antithrombin,
protein C, protein S deficiency, or activated protein C resistance: a
multicenter collaborative family study. Arterioscl Thromb Vasc Biol,
19, 1026–​33.
Gage B, et al. (2008). Use of pharmacogenetic and clinical factors to pre-
dict the therapeutic dose of warfarin. Clin Pharmacol Ther, 84, 326–​31.
Mannucci PM (2005). Laboratory detection of inherited thrombophilia: a
historical perspective. Semin Thromb Hemost, 31, 5–​10.
Moake JL (2004). Thrombotic thrombocytopenic purpura: survival by
‘giving a dam’. Trans Am Clin Climatol Assoc, 115, 201–​19.
Nicolaes GA, Dahlbäck B (2003). Congenital and acquired activated
protein C resistance. Semin Vasc Med, 3, 33–​46.
Shih AW, Crowther MA (2016). Reversal of direct oral anticoagulants:
a practical approach. Hematology Am Soc Hematol Educ Program,
2016(1), 612–19.
Genetic databases
Ensembl. http://​www.ensembl.org
National Center for Biotechnology Information (NCBI). http://​www.
ncbi.nlm.nih.gov/​
UCSC Genome Browser. http://​genome.ucsc.edu
22.7.5  Acquired coagulation
disorders
T.E. Warkentin
ESSENTIALS
Acquired disorders of coagulation may be the consequence of many
underlying conditions, and although they may share abnormality of
a coagulation test, for example, a prolonged prothrombin time (PT),
their clinical effects are diverse and often opposing.
General clinical approach
Diagnosis—​most acquired disorders of coagulation can be identi-
fied by screening haemostasis tests, including (1) PT; (2) activated
partial thromboplastin time (APTT); (3) thrombin clotting time; (4)
fibrin degradation products (FDPs), including (5) the cross-​linked
fibrin assay (D-​dimer); and (6) complete blood count with examin-
ation of a blood film. Few bleeding disorders give normal results in
all these tests, but disorders predisposed to thrombosis as a result
of deficiency of natural anticoagulants (e.g. antithrombin, protein
C, and protein S) or certain mutations (e.g. factor V Leiden) must be
specifically sought.
Treatment—​patients with coagulopathies who are bleeding or
who require surgery are usually treated with blood products such
as platelets and frozen plasma. Other treatments used in par-
ticular circumstances include (1) vitamin K—​required for the post-​
translational modification of factors II, VII, IX, and X as well as the
anticoagulant factors, protein C, and protein S; (2) cryoprecipitate—​
used principally for the treatment of hypofibrinogenaemia; (3) con-
centrates of specific factors—​used in isolated deficiencies (e.g. of
factors VIII, IX, XI, VII, or fibrinogen); (4)  antifibrinolytic agents
(e.g. ε-​aminocaproic acid and tranexamic acid); (5) desmopressin
(1-​deamino-​8-​d-​arginine vasopressin (DDAVP))—​increases factor
VIII and von Willebrand factor.
Prohaemorrhagic coagulation disorders
Vitamin K deficiency—​most haemostatic factors are produced exclu-
sively by the liver, including the vitamin K-​dependent factors II, VII,
IX, and X, deficiency of which can be caused by (1) malabsorption
of fat-​soluble vitamins, or (2) coumarin overanticoagulation—​minor
bleeding episodes occur in about 6 to 10% of patients per year and
major bleeding episodes in 1 to 3%.
Liver disease—​abnormalities include a disproportionately pro-
longed PT, reduced/​normal fibrinogen levels, and/​or pancytopenia
(indicating hypersplenism) in an appropriate clinical setting.
Disseminated intravascular coagulation (DIC)—​clinical manifest-
ations range from generalized haemorrhage to widespread micro-
vascular thrombosis, predisposing to multisystem organ dysfunction
and ischaemic limb necrosis. Initiated by numerous triggers, for ex-
ample, the extrinsic coagulation pathway (tissue factor) or interleukin-​
6 in the context of systemic inflammation. May be caused by a wide
variety of conditions, including trauma and cardiogenic shock, in-
fection/​septic shock, obstetric complications, acute haemolysis,
immunological disorders, and vascular anomalies. The presence of
DIC is often indicated by abnormal coagulation tests associated with
thrombocytopenia and red cell abnormalities on examination of the
blood film: FDPs and fibrin D-​dimers are usually greatly increased.
Immunoglobulin-​mediated factor deficiency—​(1) acquired factor VIII
deficiency—​this is suggested by the occurrence of bleeding, either
spontaneously or after minor trauma, in association with a prolonged
APTT and a normal PT, with mixing experiments with normal pooled
plasma indicating the presence of an inhibitory antibody. The condi-
tion is of unknown cause in 50% of cases, with the remainder associated
with other autoimmune disorders (e.g. systemic lupus erythematosus),
lymphoid and other malignancies, penicillin treatment, or the post-
partum state. Aside from treatment with DDAVP (mild bleeding) or
purified human factor VIII (or VIIa) concentrates (severe bleeding),
patients with high antibody titres may require immunosuppressive
therapy (e.g. prednisone ± cyclophosphamide, rituximab). (2) Other
acquired coagulation-​factor deficiencies caused by antibodies.
Other acquired coagulation-​factor deficiencies—​these include
(1)  haemodilution and massive transfusion; (2)  heparin and
22.7.5  Acquired coagulation disorders
5547
acquired heparin-​like anticoagulants; (3)  coagulopathies sec-
ondary to plasma cell dyscrasias; (4) hyperfibrinolysis—​which may
be a result of thrombolytic therapy, malignancy, cardiopulmonary
bypass procedures, or advanced liver disease; and (5) heterogeneous
coagulopathies induced by venoms (snake bites).
Prothrombotic coagulation disorders
Heparin-​induced thrombocytopenia—​caused by IgG antibodies
which recognize complexes of platelet factor 4 and heparin, typ-
ically leading to a fall in the platelet count beginning 5 to 10 days
after starting the drug (but more abruptly in patients who have re-
cently been exposed to it). Thrombosis is caused by several factors,
including activation of platelets and stimulation of tissue factor ex-
pression on monocytes. Clinical manifestations include (1) venous
thrombosis (deep vein thrombosis (including venous limb gangrene),
pulmonary embolism); (2)  arterial thrombosis (major limb artery
thrombosis, stroke, myocardial infarction). Protamine administered
after cardiac surgery (to reverse heparin anticoagulation) can trigger
acute thrombocytopenia and thromboembolic complications in a
patient with platelet-​activating antiprotamine/​heparin antibodies.
Adenocarcinoma-​associated chronic DIC—​metastatic adenocar-
cinoma and other tumours may be associated with a prothrombotic
state and large vessel thromboses. Tissue factor and prothrombotic
cysteine proteases have been found in tumour extracts. Heparin is
the preferred treatment; coumarins (e.g. warfarin) can cause venous
limb gangrene in a limb with deep vein thrombosis.
Antiphospholipid antibody syndrome—​caused by antibodies that
are usually directed against protein cofactors such as β2-​glycoprotein
I and prothrombin. Clinical manifestations include intermittent throm-
boses and (rarely, but most dramatically) sudden life-​threatening
arterial occlusions. Lupus anticoagulant activity is shown by dem-
onstrating inhibition of phospholipid-​dependent coagulation assays
(most commonly by prolongation of the APTT), with antiphospholipid
antibodies also detected by enzyme-​immunoassay using purified
phospholipids as the target antigen (e.g. anticardiolipin antibody
assay). Most patients require long-​term anticoagulation.
Other conditions associated with microvascular thrombosis—​these
include disorders predominantly affecting small venules, for ex-
ample, (1) coumarin-​induced skin necrosis; (2) coumarin-​induced
venous limb gangrene; (3)  symmetric peripheral gangrene; and
(4) purpura fulminans; in contrast, (5) thrombotic microangiopathy
(e.g. thrombotic thrombocytopenic purpura or haemolytic uraemic
syndrome) typically affects arterioles.
Introduction
A coagulopathy is a disorder associated with an abnormal coagula-
tion assay result, such as a prolonged prothrombin time (PT) (often
expressed as the international normalized ratio (INR)), activated par-
tial thromboplastin time (APTT), or thrombin clotting time (TCT).
Coagulopathies can be associated with either bleeding or throm-
bosis, and have many causes (Table 22.7.5.1). The importance of the
clinical context is illustrated by two contrasting patient scenarios that
have in common a prolonged international normalized ratio (INR)
(6.0; usual therapeutic range, 2.0–​3.0) during oral anticoagulant
therapy: one patient has a life-​threatening intracranial haemorrhage
complicating warfarin therapy given for a prosthetic heart valve; in
contrast, another patient, who was treated for deep vein thrombosis
(DVT) complicating heparin-​induced thrombocytopenia (HIT) has
the limb-​threatening complication of warfarin-​induced venous limb
gangrene, caused by microvascular thrombosis.
Table 22.7.5.2 lists common screening tests for coagulopathy.
Only a few coagulopathies give normal results in all these screening
assays (e.g. α2-​antiplasmin deficiency, factor XIII deficiency, mild
von Willebrand’s disease including type 2A von Willebrand’s syn-
drome associated with monoclonal gammopathy or aortic stenosis).
Agents for treating acquired disorders
of coagulation
Blood products are usually indicated for the treatment of patients
with coagulopathies who are bleeding or who require a major inva-
sive procedure.
Fresh frozen plasma or frozen plasma
Fresh frozen plasma is plasma that is frozen within 8 h of collec-
tion; it contains all the haemostatic factors at concentrations be-
tween 0.7 and 1.0 U/​ml. Frozen plasma is plasma frozen within 24 h
of collection, and is similar to fresh frozen plasma, except that it
contains significantly less factor VIII; however, isolated VIII defi-
ciency is treated with factor VIII concentrate (rather than plasma),
and so for virtually all clinical situations where fresh frozen plasma
use is appropriate, frozen plasma can be given instead (this is be-
cause factor VIII is an acute phase reactant, and is usually not sig-
nificantly reduced in most coagulopathic disorders). Accordingly,
either fresh frozen plasma or frozen plasma can be used to treat
coagulopathy of liver disease, haemodilution from massive trans-
fusion, and disseminated intravascular coagulation (DIC). In some
jurisdictions, frozen plasma has replaced fresh frozen plasma, and
the latter product is no longer available. For a 70 ​kg adult with a
3 ​litre plasma volume, 1 litre of frozen plasma (or fresh frozen
plasma) will increase the coagulation factors by about 0.25 U/​ml. In
most patients, this should lead to levels greater than the minimum
required for adequate haemostasis (>0.30 U/​ml for most factors).
Repeat frozen plasma transfusion (e.g. 500 ml every 6 h) may be ne-
cessary if the haemostasis defect is ongoing. Frozen plasma is being
supplanted by cryosupernatant as a replacement fluid for throm-
botic thrombocytopenic purpura (TTP). Solvent/​detergent-​treated
plasma, in which most blood-​borne pathogens are inactivated (but
not nonenveloped viruses such as hepatitis A, parvovirus B19, or the
agent that causes Creutzfeldt–​Jakob disease, a potential blood-​borne
pathogen), has become available, but is limited by its high cost.
Cryoprecipitate
This contains fibrinogen (0.10–​0.25 g/​unit), factors VIII and XIII,
von Willebrand factor (VWF), and fibronectin. Its principal indica-
tion is the treatment of hypofibrinogenaemia, where it increases fi-
brinogen levels using just one-​quarter of the volume of blood product
compared with fresh frozen plasma. Cryoprecipitate is appropriate
for patients with significant hypofibrinogenaemia, for example,
DIC, primary fibrinolysis, and congenital hypofibrinogenaemia. For
a bleeding patient whose fibrinogen level is about 0.5 g/​litre, 10 U
of cryoprecipitate would probably increase the fibrinogen to above
section 22  Haematological disorders
5548
Table 22.7.5.1  Acquired coagulopathies that cause bleeding or thrombosis
Acquired coagulopathies
Comment
Prohaemorrhagic disorders
Vitamin K deficiency or pharmacological antagonism
by coumarin
Reduced levels of vitamin K-​dependent procoagulant factors (II (prothrombin), VII, IX, X)
Liver disease
Multiple factor deficiencies, especially factors XI and XII (although VIII levels are usually normal/​elevated);
low fibrinogen levels can indicate hyperfibrinolysis
Severe haemodilution/​massive transfusion
Multiple factor deficiencies; concomitant DIC in some patients
Acute DIC: haemorrhagic
Certain forms of DIC, e.g. acute head trauma, placental abruption, can lead to bleeding secondary to
generalized coagulopathy, especially with fibrinogen depletion
Acquired coagulation factor inhibitor (autoimmune)
Anti-​VIII autoantibodies are most common
Direct oral anticoagulants (anti-​IIa (antithrombin),
anti-​Xa)
Dabigatran (direct thrombin inhibitor) tends to prolong the APTT; in contrast, the direct Xa inhibitors,
rivaroxaban and edoxaban, tend to prolong the INR, although apixaban has minimal effect on the INR
Heparin and related drugs
Marked APTT prolongation with heparin overdose (extreme overdose also prolongs INR); low molecular
weight heparin overdose only minimally prolongs APTT
Heparin-​like anticoagulants
Rare; associated with plasma cell disorders; minimal or no prolongation in APTT
Paraprotein-​induced coagulopathies
See text and Box 22.7.5.3
Hyperfibrinolysis
Associated with prostate adenocarcinoma, advanced liver disease, post-​thrombolytic therapy, after cardiac
surgery, or aortic aneurysm
Snake venom
See text and Table 22.7.5.6
Prothrombotic disorders
Heparin-​induced thrombocytopenia (HIT)
Strong association with venous and arterial thrombosis; about 10 to 20% of patients have decompensated
DIC (elevated INR, low fibrinogen, and/​or microangiopathic blood film)
Protamine-​induced thrombocytopenia (PIT)
Potential explanation for post-​cardiac surgery thrombocytopenia and thrombosis in susceptible patient
with platelet-​activating antiprotamine/​heparin antibodies (which could be present if preoperative heparin
is given to diabetic patient receiving protamine–​insulin)
Chronic DIC secondary to adenocarcinoma
Strong association with venous and arterial thrombosis; improves with (low molecular weight) heparin;
predisposes to coumarin-​induced microthrombosis (see later in table), especially venous limb gangrene
(acral limb necrosis with associated DVT)
Acute DIC associated with symmetric peripheral
gangrene or purpura fulminans
Certain forms of DIC, e.g. cardiogenic or septic shock, are associated with microthrombosis and acral
ischaemic limb necrosis, especially in setting of ‘shock liver’ (acute ischaemic hepatitis)
Antiphospholipid syndrome (APS)
Prolonged APTT due to ‘lupus anticoagulant’ (‘nonspecific inhibitor’); associated with venous and arterial
thrombosis, spontaneous abortions, thrombocytopenia
Coumarin-​induced necrosis
Central skin or acral limb necrosis resulting from microvascular thrombosis; pathogenesis includes
depletion of vitamin K-​dependent natural anticoagulant, protein C, in setting of hypercoagulability
Thrombotic microangiopathy (TMA)
Thrombocytopenia and microangiopathic haemolysis (red cell fragmentation), platelet–​VWF
microthrombi within arterioles; elevated INR and APTT are occasionally seen
Table 22.7.5.2  Screening haemostasis tests
Assay
Comment
Prothrombin time (PT), often expressed as
international normalized ratio (INR)
Screen for deficiency of factors VII, X, V, II, and/​or fibrinogen (e.g. vitamin K deficiency/​coumarin therapy,
liver disease)
Activated partial thromboplastin time (APTT)
Screen for deficiency of factors VIII, IX, X, V, II, contact factors, and/​or fibrinogen; monitor certain
anticoagulants, e.g. heparin, lepirudin, argatroban
Thrombin clotting time (TT or TCT)
Screen for hypofibrinogenaemia and/​or presence of heparin; some TCT assays are also sensitive to FDPs
Serum fibrin(ogen) degradation products (FDPs)
Requirement to clot blood sample can lead to false-​positive results due to incomplete blood clotting
(e.g. residual heparin)
Cross-​linked fibrin assay (D-​dimer)
Detects fibrin degradation products generated after thrombin, factor XIII, and plasmin have acted upon
fibrinogen (marker for DIC and/​or thrombosis)
Paracoagulation assay (e.g. protamine sulphate test)
Positive paracoagulation assay often means DIC is clinically significant and may require blood products or
anticoagulant therapy
Bleeding time
Assesses primary haemostasis, i.e. VWF-​ mediated platelet adhesion to endothelium with secondary aggregation
of platelets within haemostatic plug, now replaced in many centres by platelet function analyser (PFA100)
Complete blood count; blood film examination
Platelet enumeration, and assessment of causes for thrombocytopenia, e.g. red cell fragments indicating
microangiopathy
22.7.5  Acquired coagulation disorders
5549
1.0 g/​litre, although a lower than expected increment could occur if
the patient had a higher volume of distribution (e.g. a cirrhotic patient
with ascites). Where available, fibrinogen concentrates (see later) are
increasingly being used for treatment of hypofibrinogenaemia.
Specific factor concentrates
These are available for use in patients with an isolated deficiency
in certain factors, such as VIII or IX. Prothrombin complex con-
centrates (PCCs) contain the vitamin K-​dependent factors (ei-
ther three-​factor PCCs containing procoagulant factors II, IX,
and X, or four-​factor PCCs that additionally contain factor VII),
and four-​factor PCC is appropriate for the rapid reversal of severe
coagulopathy related to coumarin use. Activated PCC (e.g. factor
VIII inhibitor bypassing activity (FEIBA)) and factor VIIa are other
specialized concentrates with specific uses, for instance, to manage
a bleeding patient with an acquired factor VIII inhibitor. Certain
other isolated factor deficiencies can be managed by specific factor
concentrates, such as recombinant factor VIIa, factor XI, factor XIII,
and fibrinogen. Protein C concentrates are available in some juris-
dictions for treatment of congenital protein C deficiency.
Pharmacological therapies
These include the antifibrinolytic agents ε-​aminocaproic acid and
tranexamic acid. ε-​Aminocaproic acid and tranexamic acid bind to the
lysine-​binding sites of plasminogen; paradoxically, although increasing
the susceptibility of plasminogen to proteolysis by plasminogen acti-
vator, these lysine analogues also prevent plasminogen from binding to
fibrin, thus impeding fibrinolysis. Oral dosing for ε-​aminocaproic acid
is about 7 g (100 mg/​kg) initially, followed by 3.5 g (50 mg/​kg) every 4 h;
similar doses are used for intravenous administration. For tranexamic
acid, 1.0 to 1.5 g is given every 8 h by mouth; the dose is reduced to
between 0.5 and 1.0 g every 8 h if given intravenously (higher dosing
is appropriate if given prior to cardiac surgery). Both drugs are avail-
able in 500-​mg capsules. These drugs are appropriate for the treatment
of hyperfibrinolysis, for instance, bleeding following thrombolytic
therapy or associated with cardiac or hepatic surgery. These drugs are
generally contraindicated in patients with DIC, however, as blocking
secondary fibrinolysis could lead to microvascular thrombosis.
Desmopressin
Desmopressin or 1-​deamino-​8-​d-​arginine vasopressin (DDAVP),
a synthetic vasopressin analogue, leads to an increase in factor VIII
and VWF levels that peak between 45 and 90 min after intravenous
infusion (0.3 μg/​kg in 50 ml normal saline over 20–​30 min; maximum
dose, 20 μg). Although repeat DDAVP can be given at 12-​ to 24-​h
intervals, the drug becomes less effective over time (tachyphylaxis) as
endothelial stores of VWF are depleted, limiting the usual number of
injections to no more than three with any treatment course. Flushing,
tachycardia, mild hypotension, free-​water retention (leading to
dilutional hyponatraemia), and angina are occasional side effects.
Prohaemorrhagic acquired coagulation disorders
Vitamin K deficiency disorders
Vitamin K-​dependent coagulation factors
Vitamin K is required for the post-​translational modification of
six haemostatic factors, four with procoagulant activity (factors
II, VII, IX, and X), and two with anticoagulant activity (protein
C and protein S). The physiological relevance of a seventh factor,
factor Z, remains unclear. The enzyme vitamin K-​dependent γ-​
glutamylcarboxylase adds a carboxyl group to each member of a
cluster of glutamyl residues, thereby forming the γ-​carboxyglutamyl
residues crucial for enabling these six haemostatic factors to interact
with phospholipid membranes in a calcium-​dependent fashion.
During this γ-​carboxylation reaction, the reduced form of vitamin
K is oxidized to vitamin K epoxide; oral anticoagulants inhibit the
enzyme complex (vitamin K epoxide reductase complex) that regen-
erates the reduced form of vitamin K.
Diet and absorption of vitamin K
Vitamin K1 (phylloquinone) is exclusively derived from plants;
vitamin K2 (menaquinone) is synthesized by bacteria. Green, leafy
vegetables, such as broccoli, lettuce, cabbage, and spinach, are very
good dietary sources of vitamin K (100–​500 μg/​100 mg). Vitamin K
is fat-​soluble, and absorption occurs primarily in the small bowel.
Serum vitamin K concentrations are only between 150 and 800 pg/​
ml and, as hepatic storage is limited (half-​life is just a few days), a
regular daily intake of about 0.1 to 0.5 μg/​kg is required. Although
bacterial synthesis is not a major source of vitamin K in humans, anti-
biotic treatment nevertheless predisposes to vitamin K deficiency.
Vitamin K deficiency
Malabsorption of fat-​soluble vitamins caused by biliary tract disease,
or primary bowel disorders such as coeliac or inflammatory bowel
disease, can cause vitamin K deficiency. An inadequate diet, par-
ticularly when combined with antibiotic therapy, is another cause.
Indeed, coagulopathy can arise during a brief period of decreased
intake (e.g. 1 week postoperatively).
A disproportionately prolonged PT/​INR in the appropriate clin-
ical setting suggests vitamin K deficiency (Table 22.7.5.3). The diag-
nosis is usually confirmed by assessing the response to vitamin K
administration. Compared with the treatment of a coumarin over-
dose, small amounts of vitamin K are effective, for example, 1 mg
vitamin K given orally or by slow intravenous infusion (over at least
30 min to minimize risk of an anaphylactoid reaction). For serious
bleeding, frozen plasma, fresh frozen plasma, or especially PCCs
provide a more rapid (but transient) correction of the coagulopathy.
Coumarin overanticoagulation
Oral anticoagulants (e.g. coumarins such as warfarin and
phenprocoumon) are widely used to prevent and treat thrombosis
via their vitamin K antagonism. An INR target range between 2.0
and 3.0 is appropriate for most clinical indications, although a
higher therapeutic range (INR 2.5–​3.5) is appropriate for patients
at very high risk for thrombosis (e.g. with mechanical prosthetic
heart valves).
Bleeding is the major complication of coumarin, with minor and
major bleeding episodes occurring in about 6 to 10% and 1 to 3%
of patients per year respectively; the intracranial haemorrhage rate
is between 0.25 and 1% per year. Changes in diet or alcohol con-
sumption, poor patient compliance, and the introduction of new
drugs (Table 22.7.5.4) can cause bleeding by producing coumarin
overanticoagulation. In contrast, recurrent gastrointestinal or
urinary tract bleeding at therapeutic levels of anticoagulation often
indicates an occult gastrointestinal or renal lesion, respectively.
section 22  Haematological disorders
5550
The treatment of nontherapeutic (elevated) INRs depends on
the clinical setting. Oral vitamin K use is appropriate in many non-
urgent conditions as it avoids the risk of anaphylactoid reactions to
intravenous use, and has more predictable effects than subcutaneous
injection. Much larger and prolonged vitamin K dosing (100–​
150 mg/​day) is required to treat accidental or deliberate overdoses of
long-​acting second-​generation rodenticides (‘superwarfarins’), such
as brodifacoum.
Table 22.7.5.3  Results of screening haemostasis assays in various clinical settings
PT/​INR
APTT
Fibrinogen
TCT
Fibrin D-​dimers,
fibrin monomers
Platelets
Vitamin K deficiency or antagonism (coumarin)
↑↑
↑
N
N
N
N
Liver disease
N, ↑, ↑↑
N, ↑
↓, N
N, ↑
N, ↑, ↑↑
N, ↓, ↓↓
Heparin
N, ↑a
↑↑
N
↑↑
N
N, ↓b
LMWH, danaparoid
N
N, sl↑
N
N, sl↑
N
N
Thrombin inhibitors (argatroban, dabigatran)
N, ↑
↑, ↑↑
↓c, N
↑↑
N
N
Factor Xa inhibitors (rivaroxaban, edoxaban, apixaban)
N d, ↑
N, ↑
N
N
N
N
Thrombolytic therapy
sl↑
N
↓, ↓↓
↑, ↑↑
↑, ↑↑
N, ↓
Renal disease
N
N
N
N
N
N, sl↓
Acute DIC
↑, ↑↑
N, ↑, ↑↑
↑e, N, ↓, ↓↓
N, ↑, ↑↑
↑↑
↓, ↓↓
Chronic DIC
N, ↑
N, sl↑
N, ↓
N, ↑
↑, ↑↑
N, ↓, ↓↓
Primary fibrinolysis
N, sl↑
N, sl↑
↓, ↓↓
↑, ↑↑
↑, ↑↑
↓, N
Lupus anticoagulant
N, ↑f
Ng, ↑, ↑↑
N
N
N
N, ↓, ↓↓h
Factor VIII inhibitor
N
↑, ↑↑
N
N
N
N
Haemodilution
↑
↑, ↑↑
↓, ↓↓
↑, ↑↑
N
↓
a An elevated PT/​INR secondary to heparin indicates very high heparin levels (e.g. dosing used in cardiac surgery, overdose).
b Unfractionated heparin is associated with early, mild, transient thrombocytopenia secondary to weak platelet-​activating effects (nonimmune heparin-​associated thrombocytopenia);
thrombocytopenia that begins 5 or more days after starting UFH or LMWH can indicate HIT.
c Thrombin inhibitors can spuriously indicate low fibrinogen levels due to interference with certain fibrinogen assays.
d Apixaban has the least effect on the INR among the commercially available direct factor Xa inhibitors.
e Elevated plasma fibrinogen levels despite DIC-​associated fibrinogen declines can occur when DIC complicates inflammatory disorders with hyperfibrinogenaemia.
f An elevated INR can indicate hypoprothrombinaemia.
g A normal APTT can be found if the laboratory chooses to use an APTT reagent that is insensitive to lupus anticoagulant activity.
h Many patients with antiphospholipid syndrome have concomitant thrombocytopenia; severe thrombocytopenia can indicate catastrophic antiphospholipid syndrome (CAPS).
Table 22.7.5.4  Drugs, food, and dietary supplement interactions with warfarin by level of supporting evidence and direction of interaction
Potentiation of warfarin’s anticoagulant effect
Inhibition of warfarin’s anticoagulant effect
Anti-​infectives: amoxicillin/​clavulanateb, azithromycinb, ciprofloxacina, clarithromycinb,
cotrimoxazolea, erythromycina, fluconazolea, isoniazida, itraconazoleb, levofloxacinb, metronidazolea,
miconazole oral gela, miconazole vaginal suppositorya, ritonavirb, tetracyclineb, voriconazolea
Anti-​infectives: dicloxacillinb, griseofulvina, nafcillina,
ribavirina, rifampicina, ritonavirb
Cardiovascular: amiodaronea, aspirinb, clofibratea, diltiazema, fenofibratea, fluvastatinb, propafenonea,
propronolola, quinidineb, ropiniroleb, simvastatinb, sulfinpyrazone (biphasic with later inhibition)a
Cardiovascular: bosentanb, cholestyraminea
Analgesics/​anti-​inflammatories and immunologics: acetaminophenb, aspirinb, celecoxibb,
dextropropoxypheneb, interferonb, phenylbutazonea, piroxicama, tramadolb
Analgesics/​anti-​inflammatories and immunologics:
azathioprineb, mesalaminea
CNS drugs: alcohol (if concomitant liver disease)a, citaloprama, chloral hydrateb, disulfiramb,
entacaponea, fluvoxamineb, phenytoin (biphasic with later inhibition)b, sertralinea
CNS drugs: barbituratesa, carbamazepinea,
chlordiazepoxideb
GI drugs and food: cimetidinea, fish oila, grapefruitb, mangoa, omeprazolea
GI drugs and food: high vitamin K content foodsa,
avocadoa, soy milkb, sucralfateb
Herbal supplements: boldo-​fenugreeka, danshenb, don quaib, lycium barbarum Lb, PC-​SPESb, quilinggaoa,
Herbal supplements: ginsengb
Other drugs: anabolic steroidsa, fluorouracil,b gemcitabineb, levamisole/​fluorouracilb, paclitaxelb,
tamoxifenb, tolterodineb, zileutona
Other drugs: chelation therapyb, influenza vaccineb,
mercaptopurinea, multivitamin supplementb, raloxifeneb
CNS, central nervous system; GI, gastrointestinal.
Level of causation: a highly probable, b probable. See Ageno et al. (2012) for other drugs for which level of causation is listed as ‘possible’ and ‘highly improbable’.
Modified from Ageno et al. (2012). Oral anticoagulant therapy. Antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians Evidence-​based
Clinical Practice Guidelines. Chest, 141 (Suppl), e44S−e88S.
22.7.5  Acquired coagulation disorders
5551
Urgent reversal of coumarin anticoagulation
When urgent reversal of coumarin anticoagulation is required (e.g.
life-​threatening bleeding or need for surgery within 6 h), blood prod-
ucts should be given in addition to intravenous vitamin K. Although
either frozen plasma or fresh frozen plasma are options, four-​factor
PCCs (containing factors II, VII, IX, and X), where available, are
strongly preferred, as reversal can be achieved more reliably, and
with much lower volumes of blood product, compared with plasma
(Box 22.7.5.1).
Direct oral anticoagulant overanticoagulation
Direct oral anticoagulants (DOACs), which either inhibit thrombin
(dabigatran) or factor Xa (rivaroxaban, apixaban, edoxaban), are now
available; these variably prolong the PT (or INR) and PTT, have half-​
lives of approximately 12h (assuming normal renal function), and
(except for dabigatran) have no specific antidotes. Bleeding risk is
increased in the elderly and in patients with renal dysfunction. Life-​
threatening bleeding should be managed by general measures (dis-
continuing the anticoagulant, red cell transfusions) and, possibly, by
specialized prohaemostatic therapies such as four-​factor PCC or acti-
vated PCC (e.g. FEIBA) or recombinant factor VIIa (e.g. in our centre,
we use as off-​label therapy 2000 units of four-​factor PCC for an adult
with life-​threatening bleeding associated with a DOAC that inhibits
factor Xa) (Box 22.7.5.1). Haemodialysis has been reported to re-
duce dabigatran levels. Idarucizumab, a dabigatran-​binding antibody
fragment, is approved by the Food and Drug Administration in the
United States of America when reversal of the anticoagulant effects of
dabigatran is needed for emergency surgery/​urgent procedures or in
life-​threatening or uncontrolled bleeding.
Liver disease
Most haemostatic factors are produced exclusively by the liver.
Exceptions include factor VIII (hepatic and extrahepatic synthesis),
VWF (endothelium, megakaryocytes), and several factors produced
by endothelium (e.g. plasminogen activator and plasminogen acti-
vator inhibitor type I (PAI-​1)). Box 22.7.5.2 lists the multiple effects
on haemostasis caused by liver disease. Often, bleeding is primarily
related to anatomical factors, such as oesophageal varices or gastric/​
duodenal ulcers, though reduced hepatic synthesis of coagulation
factors can be a contributing factor. Increased susceptibility to DIC
via superadded illness (e.g. bacterial peritonitis), impaired clearance
of activated coagulation factors, and hyperfibrinolysis are other fac-
tors. The key role of acute ischaemic hepatitis (‘shock liver’) in ex-
plaining microthrombosis is discussed later in this chapter.
A prolonged PT/​INR is the most frequent laboratory abnormality
(Table 22.7.5.3). The fibrinogen level is usually normal or increased;
when hypofibrinogenaemia occurs, it generally indicates severe
liver disease or hyperfibrinolysis. Fibrin(ogen)-​degradation product
(FDP) (also known as fibrin split product) and fibrin D-​dimer levels
Box 22.7.5.1  Management of bleeding or need for urgent
surgery in patients receiving vitamin K antagonists
Indications
4-factor prothrombin complex concentrate (PCC) is indicated for treat-
ment of severe or life-threatening acute bleeding associated with war-
farin (or other vitamin K antagonists [VKAs]) OR for rapid reversal of
warfarin [or other VKAs] for urgent surgical procedures (generally, when
surgery is required in less than 6 hours).
PCC has also been used for treatment of severe or life-threatening
acute bleeding associated with factor Xa inhibitors (apixaban,
rivaroxaban, or edoxaban).
Dosing (VKA reversal)
Dosing for target INR ~1.5
Patient INR
Dose of 4-factor PCC*
INR 2–3
20 IU/kg
INR 3–​6
30 IUkg
INR >6
40 IU/kg
Add 10 IU/kg for target INR < 1.2.
Also give vitamin K 5–10 mg IV over 30 minutes; may repeat in 12–24
hours.
Under normal circumstances the PCC dose should not exceed 3000 IU
(maximum infusion speed, 8 mL/min).
For severe acute bleed with unknown INR, give 2000 IU.
For vials containing 500 IU, round dose to the nearest 500 IU (e.g., 2780 IU
rounds up to 3000 U administered).
4-factor PCC refers to product containing (procoagulant) factors II, VII,
IX, and X; in contrast, 3-factor PCC contains factors II, IX, and X, but not
factor VII.
Dosing (Factor Xa inhibitor reversal)
Give 4-factor PCC, 2000 IU.
Relative contraindication
Patients with heparin-induced thrombocytopenia (HIT) or suspected HIT
(product contains heparin).
Source data from McMaster University, Department of Medicine
(Hematology and Thromboembolism), Clinical Protocols (and Reversals):
Prothrombin Complex Concentrate, at: http://fhs.mcmaster.ca/medicine/
hematology/anticoag_octoplex.htm.
Box 22.7.5.2  Causes of bleeding and thrombosis in liver disease
Predispose to bleeding
•	 Effects of portal hypertension:
—​	 Oesophageal varices (bleeding site)
—​	 Splenomegaly (thrombocytopenia)
•	 Decreased thrombopoietin production (thrombocytopenia)
•	 Decreased procoagulant factor synthesis
•	 Abnormal coagulation factor synthesis:
—​	 Dysfibrinogenaemia (increased sialic acid content)
—​	 Decarboxylated vitamin K-​dependent factors
•	 Decreased clearance of plasmin, plasminogen activators, and fibrin
(ogen) degradation products
•	 Vitamin K malabsorption
•	 Platelet dysfunction
•	 Increased susceptibility to adverse hepatic effects of alcohol or other
drugs
•	 Decreased α2-​antiplasmin synthesis (predisposes to hyperfibrinolysis)
Predispose to thrombosis
•	 Decreased natural anticoagulant synthesis (e.g. protein C, antithrombin)
•	 Decreased clearance of activated coagulation factors
•	 Physician reluctance to prescribe antithrombotic therapy
section 22  Haematological disorders
5552
are often increased; thus, the laboratory picture can resemble that of
DIC even in a patient who is otherwise clinically stable.
Management of hepatic coagulopathy should include a trial of
vitamin K (e.g. 10 mg once daily for 3 days), although this will not
benefit most patients. Frozen plasma (or fresh frozen plasma) may
be given to bleeding patients with a prolonged INR, or who require
major invasive procedures. Retrospective studies suggest that minor
invasive procedures (e.g. paracentesis and pleurocentesis) can usu-
ally be performed safely with an INR as high as 1.8. For patients
suspected to have significant fibrinolysis, antifibrinolytic therapy
can be tried. PCCs should only be used in emergencies, given
their prothrombotic potential in this group of patients. Platelet
transfusions usually provide minimal increase in the platelet count
in patients with platelet sequestration caused by hypersplenism.
DDAVP improves haemostasis in patients with prolonged bleeding
time secondary to hepatic platelet dysfunction.
Haemodilution and massive transfusion
Coagulopathies occur in most patients who receive crystalloids, col-
loids, or red cell concentrates following trauma, surgery, or fluid re-
suscitation for other major illnesses. In many patients, no bleeding
results despite moderate abnormalities in the INR, APTT, TCT, and
platelet count. The reason is that all the individual coagulation fac-
tors remain at haemostatically effective levels, even though the la-
boratory assays are abnormal when all the factor levels are uniformly
reduced.
Massive transfusion is defined as the transfusion of blood prod-
ucts equivalent to the patient’s total blood volume within 24 h. Red
cell concentrates do not provide significant amounts of platelets or
coagulation factors. Thus infusions of platelets, frozen plasma (or
fresh frozen plasma), and, sometimes, cryoprecipitate are often
needed as well. Massive transfusion protocols that timely admin-
ister blood using a predetermined ratio, for example, plasma, plate-
lets, and red blood cells in a 1:1:2 ratio (i.e. 5 units of frozen plasma,
1 pack platelets (5 units), and 10 units of red blood cells) may be
life-​saving.
Disseminated intravascular coagulation
DIC is a group of clinicopathological syndromes characterized by
widespread activation of coagulation; there results intravascular gen-
eration of thrombin, formation of fibrin, and reactive fibrinolysis.
Clinical consequences range from coagulation factor and platelet
depletion, resulting in generalized haemorrhage, to widespread
microvascular thrombosis, predisposing to multisystem organ dys-
function or limb necrosis. ‘Acute’ DIC, caused by septicaemia, trauma,
and obstetrical complications, is most frequent; ‘chronic’ DIC, typ-
ically caused by malignancy, is often associated with a dramatic
hypercoagulable state (Table 22.7.5.5). Although DIC is usually a sys-
temic process, sometimes a localized abnormality (such as a vascular
malformation or aortic aneurysm) leads to the regional activation of
coagulation and results in the depletion of haemostatic factors.
DIC is usually triggered by the extrinsic coagulation pathway: tissue
factor and factor VIIa (Fig. 22.7.5.1). The proinflammatory cytokine
interleukin-​6 (IL-​6) is a principal mediator of DIC in septicaemia
and other systemic inflammatory responses, and impairs natural
anticoagulant and fibrinolytic pathways. For example, a sustained
increase in PAI-​1 impairs plasmin formation despite intravascular
fibrin generation.
Diagnostic and treatment approach to DIC
One or more prolonged clotting times and thrombocytopenia in a pa-
tient with one of the disorders listed in Table 22.7.5.5 suggests DIC.
However, similar test results are seen in patients following major sur-
gery, emphasizing the need to interpret the laboratory data in the ap-
propriate clinical context. Typically, cross-​linked fibrin degradation
products, such as D-​dimers, are greatly increased in DIC. Sometimes,
specialized haemostasis assays are useful, for example, protein C and
antithrombin activity levels in DIC complicated by shock liver and
symmetrical peripheral gangrene/​purpura fulminans.
The cornerstone of management is treating its underlying cause
and providing supportive measures. For bleeding patients, replace-
ment of depleted haemostatic factors with frozen plasma (or fresh
frozen plasma), cryoprecipitate (or fibrinogen concentrate), and
platelet transfusions may be needed. Heparin may benefit patients
with large-​vessel thrombosis or acral ischaemia. The routine use of
vitamin K and folate will avoid coagulation and platelet count dis-
turbances in some patients.
Trauma and shock
Tissue injury due to trauma, burns, or hypoperfusion (e.g. cardiogenic
shock) can cause DIC. Head injury in particular can result in DIC with
hypofibrinogenaemia, probably because of intravascular release of
tissue thromboplastin from injured brain.
Acute ischaemic hepatitis (shock liver)
The combination of acute ischaemic hepatitis (‘shock liver’) and
DIC (e.g. secondary to cardiogenic or septic shock) can explain is-
chaemic limb gangrene despite palpable or Doppler-​identifiable
Table 22.7.5.5  Main causes of disseminated intravascular
coagulation
Acute DIC
Trauma, burns
Cardiogenic shock
Infection, especially septic shock
Obstetrical complications:
• Placental abruption
• Amniotic fluid embolism
• Pre-​eclampsia/​eclampsia
• Puerperal sepsis
Malignancy, promyelocytic leukaemia
Allergic reactions
Severe heparin-​induced thrombocytopenia
Severe haemolysis
Envenomation (e.g., snake bite)
Chronic DIC
Malignancy, especially metastatic adenocarcinoma
Obstetrical complications:
• Dead fetus syndrome
Chronic liver disease
Vascular anomalies:
• Giant haemangioma (Kasabach–​Merritt syndrome)
• Aortic aneurysm
22.7.5  Acquired coagulation disorders
5553
peripheral arterial pulses. The term ‘symmetrical peripheral gan-
grene’ is used when tissue necrosis primarily affects the distal ex-
tremities (invariably, the feet; in one-​third of patients, fingers/​hands
as well). When there is additional nonacral skin necrosis, the term
‘purpura fulminans’ is applicable. Acral (distal extremity) necrosis
results from microthrombosis (capillaries, venules, arterioles).
The pathophysiology includes (a) acute DIC; (b) depletion of nat-
ural anticoagulants, protein C and antithrombin, secondary to in-
creased consumption (DIC) and decreased synthesis (shock liver);
and (c) poor acral blood flow secondary to hypotension. A major
role for vasopressor therapy in causation of limb necrosis is probably
overstated. Ischaemic limb necrosis typically begins approximately
2 to 5 days after onset of shock liver; thus, preceding ‘shock liver’
can be regarded as a ‘warfarin equivalent’ (as coumarin-​induced
microthrombosis too usually begins approximately 2 to 5 days after
starting warfarin or another coumarin anticoagulant).
Infection
Gram-​negative and Gram-​positive bacteria can cause DIC, either
from procoagulant bacterial components (e.g. endotoxin and
Staphylococcus aureus toxin) or via the host response to infection
(e.g. interleukin-​6). The clinical spectrum ranges from prom-
inent thrombocytopenia with minimal activation of coagulation,
to marked coagulation factor and natural anticoagulant depletion.
Certain infections, such as meningococcaemia and Capnocytophaga
canimorsus (from dog bites), sometimes produce severe acquired
consumptive protein C and/​or antithrombin deficiency (usually
with concomitant shock liver), which leads to widespread ischaemic
necrosis of the extremities (symmetric peripheral gangrene) and
elsewhere (purpura fulminans). Postvaricella purpura fulminans
can be caused by acquired antiphospholipid antibodies that inter-
fere with protein S.
Obstetrical complications
Acute DIC can be caused by thromboplastin-​like materials re-
leased during placental abruption or amniotic fluid embolism. Pre-​
eclampsia too can be accompanied by DIC, although there can be
clinical and laboratory overlap with other life-​threatening compli-
cations of pregnancy (e.g. fatty liver of pregnancy and HELLP syn-
drome (haemolysis, elevated liver enzymes, low platelets)). Bleeding
due to hypofibrinogenaemia is often prominent in pregnancy-​
associated DIC. Chronic DIC can be caused by fetal death.
Acute haemolysis
Haemolysis caused by incompatible blood transfusions, certain infec-
tions (e.g. Clostridium perfringens septicaemia), or microangiopathic
disorders such as TTP and HELLP, can sometimes be associated
with DIC.
Immunological disorders (including HIT)
Severe allergic reactions (e.g. anaphylaxis), transplant rejection, glom-
erulonephritis, and other vasculitic disorders are sometimes associated
with DIC. Severe HIT can also be associated with overt DIC; in such
patients, APTT-​monitored therapies (e.g. argatroban, bivalirudin)
can fail because the concomitant HIT-​associated coagulopathy re-
sults in supratherapeutic APTT levels upon beginning anticoagulant
therapy, which leads to inappropriate dose interruptions/​reductions
and associated treatment failure (‘APTT confounding’).
Vascular anomalies
Giant haemangiomas cause overt DIC in about 25% of those affected
(Kasabach–​Merritt syndrome). Although activation of coagulation
and fibrinolysis is localized to the vascular anomaly, depletion of
haemostatic factors produces a clinical and laboratory profile indis-
tinguishable from DIC. Eradication of haemangioma by radiation,
embolization, or surgery is curative. Medical therapies have included
heparin, antifibrinolytic drugs (combined with cryoprecipitate to
thrombose the vascular tumour), glucocorticoids, and interferon.
DIC also occurs in about 0.5 to 1% of patients with abdominal
aortic aneurysms, which usually contain adherent thrombi.
Immunoglobulin-​mediated factor deficiency
Coagulation factor inhibitors are usually IgG antibodies that
bind to specific coagulation factors, and either neutralize their
Tissue factor
+
factor VIIa
Factor XIII
Platelet
activation
THROMBIN
Fibrinogen
Increased
thrombin generation
via tissue factor
Impaired
natural anticoagulant
mechanisms
Factor IXa
(factor VIII)
Factor Xa
(factor V)
PC system
X-FDPs
(D-dimer)
PLASMIN
Impaired
fibrinolysis via
increased PAI-1
Plasminogen
Plasminogen
activator
PAI-1
X–FDPs
X–FDPs
ATIII
B
A
TFPI
+
+
+
+
C
Soluble
fibrin
Cross-linked (X)
fibrin
α2AP
Fig. 22.7.5.1  Pathogenesis of thrombosis in DIC. (a) DIC is usually
triggered by tissue factor, which activates coagulation by complexing
with factor VIIa, ultimately resulting in the generation of thrombin.
(b) Impaired natural anticoagulant mechanisms (e.g. excessive
consumption of natural anticoagulants, or cytokine-​mediated
downregulation of natural anticoagulant pathways) predispose to
microvascular thrombosis. (c) Impaired fibrinolysis via increased PAI-​1
leads to greater microvascular thrombosis. Sometimes, hyperfibrinolysis
is caused by increased plasminogen activator release, or low levels of
α 2-​antiplasmin. α 2AP, α 2-​antiplasmin; ATIII, antithrombin III; fDPs,
fibrinogen degradation products; FDPs, fibrin degradation products;
PAI-​1, plasminogen activator inhibitor type 1; PC, protein C; TFPI,
tissue-​factor pathway inhibitor.
section 22  Haematological disorders
5554
activity (most coagulation factor inhibitors) or result in accel-
erated clearance (e.g. antiprothrombin antibodies associated
with the antiphospholipid antibody syndrome). Acquired in-
hibitors against coagulation factors are rare in otherwise normal
(nonhaemophiliac) individuals. Even the most common auto-
immune coagulation factor deficiency (factor VIII) has an esti-
mated incidence of only 1 per 1 000 000 per year.
Acquired factor VIII inhibitor
Acquired factor VIII deficiency should be suspected in a patient with
spontaneous bleeding, or bleeding following minor trauma, that oc-
curs in association with a prolonged APTT and a normal PT/​INR
(Table 22.7.5.3). Most commonly, muscle or cutaneous haematomas
occur, but life-​threatening retroperitoneal or intracranial haemor-
rhages are described; haemarthrosis is uncommon (cf. congenital
haemophilia). The disorder occurs most commonly in older people
(median age 60 years), affects men and women equally, and is idio-
pathic in 50% of cases. Other autoimmune disorders (e.g. systemic
lupus erythematosus), lymphoid and other malignancies, penicillin
treatment, or the postpartum state, have been observed in some pa-
tients. About 20% of patients die of bleeding, often from their initial
bleeding episode.
A rapid screening test for a coagulation factor inhibitor is per-
formed by repeating the APTT after mixing patient plasma 50:50
with normal pooled plasma. An inhibitor is suggested by a prolonga-
tion time more than 4 s over the control, although some inhibitors
require a 2-​h incubation at 37°C to show inhibition. Confirmation
is obtained by a specific factor assay showing reduced levels of factor
VIII; inhibitor quantitation is most often performed by the Bethesda
assay, in which various dilutions of patient plasma are mixed with
normal plasma and incubated for 2 h at 37°C: a Bethesda unit is de-
fined as the reciprocal of the plasma dilution that yields a 50% reduc-
tion in residual factor VIII activity in the test system. Unfortunately,
the Bethesda assay tends to underestimate the amount of inhibitor
in nonhaemophiliac patients with acquired factor VIII inhibitors.
Therapy of bleeding depends upon its severity and the amount
of inhibitor present, if known. For patients with minor bleeding,
detectable factor VIII levels, and low inhibitor levels (<5 Bethesda
units), desmopressin (DDAVP) can be tried. Peak factor VIII levels
occur between 45 and 90 min post DDAVP, and repeat levels should
be measured to assess efficacy. In other patients with low inhibitor
levels but with more severe bleeding, purified human factor VIII
concentrates are usually effective. One approach is to give an ini-
tial intravenous bolus of 100 U/​kg, followed by a continuous infu-
sion of factor VIII at 10 U/​kg per h, with factor VIII levels measured
again 4 to 6 h later. Careful clinical and laboratory assessment for
response is needed, since inhibitor levels may have been under-
estimated, or higher inhibitor levels stimulated by factor VIII use.
Either PCCs or recombinant factor VIIa can be given for patients
refractory to human factor VIII. Activated PCC (e.g. FEIBA or
Autoplex) are more effective than nonactivated PCCs, but con-
comitant antifibrinolytic therapy should be avoided to reduce risk
of thromboembolic complications. Recombinant factor VIIa may be
preferable for perioperative management, since the risk for inducing
postoperative thrombosis is probably lower. In desperate situations,
extracorporeal immunoadsorption using staphylococcal protein
A may be helpful in removing the antibodies.
Spontaneous disappearance of the inhibitor occurs in about 10
to 30% of patients, most commonly in the patient who developed
her inhibitor postpartum. Nevertheless, the unpredictable clin-
ical course, and the potential for life-​threatening bleeding, means
that immunosuppressive therapy should be given to most patients.
The most widely adopted treatment is with corticosteroids (pred-
nisolone, 1 mg/​kg daily) in combination with cyclophosphamide
(1–​2 mg/​kg daily), which eradicates the inhibitor in about 70%
of cases. A more recent alternative is the anti-​CD20 monoclonal
antibody rituximab, which has been used successfully in many
autoimmune conditions; the regimen consists of four separate
intravenous infusions (375 mg/​m2 each), given at weekly intervals.
Other options include combination chemotherapy (prednisone,
cyclophosphamide, vincristine); ciclosporin; or high-​dose intra-
venous IgG (1 g/​kg for 2 days, or 0.4 g/​kg for 5 days). Even partial
remission can help reduce bleeding. Women with postpartum
factor VIII inhibitors usually develop remission within 30 months,
and only rarely develop recurrent factor VIII inhibitors with later
pregnancies. They also may be less likely to respond to corticoster-
oids or other immunosuppressive therapy.
Other acquired coagulation factor deficiencies
Hypoprothrombinaemia
This should be suspected in patients with the antiphospholipid anti-
body syndrome, particularly if bleeding occurs or the PT/​INR is
prolonged. Typically, these pathogenic antifactor II antibodies are
non-​neutralizing, and therefore mixing patient plasma 50:50 with
normal pooled plasma can produce correction of the APTT, in con-
trast to other coagulation factor inhibitors.
Thrombin inhibitors
These are rare, but may cause severe bleeding. More often, pa-
tients have antibodies that react preferentially against bovine
thrombin: these are formed following the use of ‘fibrin glue’, which
contains various bovine clotting factors. Patients have prolonged
PT/​INR, APTT, and TCT (especially using bovine thrombin).
However, it is more likely that any bleeding is the result of clinically
significant antibovine factor V antibodies.
Factor V inhibitors
Rarely, IgG antibodies against factor V arise spontaneously or
following treatment with topical bovine thrombin used at sur-
gery. Fresh frozen plasma usually does not provide enough factor
V to treat bleeding; however, platelet transfusions can be ef-
fective, as platelet activation causes factor V to be released into
haemostatic plugs.
Factor XIII inhibitors
These inhibitors, which sometimes occur in association with iso-
niazid therapy, cause bleeding via impaired factor XIII-​mediated
cross-​linking of fibrin. Factor XIII should be measured in a patient
with unexplained bleeding and normal results of screening coagu-
lation assays.
Factor X inhibitors
Factor X inhibitors are a rare cause of bleeding in patients with pro-
longed PT/​INR and APTT. The differential diagnosis also includes
22.7.5  Acquired coagulation disorders
5555
amyloidosis of the AL (amyloid light chain) variety, caused by ad-
sorption of factor X to amyloid fibrils.
Factor IX inhibitors
In nonhaemophiliac patients, factor IX inhibitors are rare and usu-
ally associated with autoimmune disease. Treatment includes PCCs
or purified factor IX, and immunosuppression. The differential diag-
nosis of acquired, isolated, factor IX deficiency includes the neph-
rotic syndrome (urinary loss of factor IX).
Factor XI inhibitors
These rare inhibitors are most often observed in association with
systemic lupus erythematosus, and usually do not cause bleeding or
require specific treatment.
Factor VII inhibitors
Factor VII inhibitors are extremely rare, and usually do not cause
bleeding or require treatment. The diagnosis is suggested by an iso-
lated prolonged PT/​INR in the absence of coumarin or vitamin K
deficiency.
Acquired von Willebrand syndrome
Rarely, bleeding is caused by a severe acquired deficiency of VWF,
most often in the setting of a monoclonal gammopathy, benign
or malignant. Typically, there is disproportional deficiency of
the largest VWF multimers due to antibody-​mediated clearance
(acquired type 2A von Willebrand syndrome). Aortic stenosis
and obstructive cardiomyopathies are other causes of type 2A
von Willebrand syndrome:  this explains why aortic valve re-
placement can cure Heyde’s syndrome (aortic stenosis associ-
ated with recurrent gastrointestinal haemorrhage secondary to
angiodysplasia).
Heparin and acquired heparin-​like anticoagulants
Bleeding is a complication of heparin treatment, particularly when
the APTT is above the therapeutic range. In patients with massive
accidental or deliberate heparin overdose, intravenous protamine
can be given to treat bleeding complications.
Rarely, patients with spontaneous bleeding and prolonged APTT
and TCT measurements have circulating heparin-​like anticoagu-
lants. Usually associated with plasma cell myeloma and other plasma
cell dyscrasias, the coagulopathy does not necessarily respond even
to large-​dose protamine infusion, and fatal haemorrhage can ensue.
Circulating dermatan sulphate glycosaminoglycan appeared to ex-
plain the bleeding in a patient with renal failure.
Coagulopathies secondary to plasma cell dyscrasias
Plasma cell myeloma, macroglobulinaemia, and other plasma
cell dyscrasias such as primary amyloidosis can cause various
coagulopathies (Box 22.7.5.3). Usually, the TCT is prolonged,
most often because of paraprotein-​induced interference with fibrin
polymerization. A distinct syndrome is monoclonal paraprotein-​
associated acquired von Willebrand syndrome type 2A, in which
high-​dose intravenous IgG corrects VWF levels for several days
or a few weeks (helpful for managing acute bleeding or before
surgery). In some patients, apheresis can improve haemostasis
by quickly reducing paraprotein levels, as antineoplastic chemo-
therapy is initiated.
Hyperfibrinolysis
Activation of fibrinolysis occurs normally when fibrin clots are
formed during physiological or pathological haemostasis. However,
primary fibrinolysis (Table 22.7.5.3) is sometimes the major cause
for bleeding, and requires specific treatment.
Thrombolytic therapy
About 0.5 to 0.7% of patients with myocardial infarction who re-
ceived thrombolysis with either streptokinase or tissue plasminogen
activator develop an intracranial haemorrhage. The thrombolytic
agent should be stopped immediately in any such patient, and
they should receive cryoprecipitate and an antifibrinolytic drug
(e.g. tranexamic acid); platelets and frozen plasma (or fresh frozen
plasma) can help to increase factor V and VIII levels that may have
been reduced by plasmin generated by thrombolysis. It can take
between 24 and 36 h for fibrinogen levels to recover after stopping
thrombolytic therapy.
Malignancy
Cancer-​associated DIC usually causes a hypercoagulable state.
However, promyelocytic leukaemia and prostatic adenocarcinoma
are two malignancies commonly associated with prominent
hyperfibrinolysis. Laboratory abnormalities include prolonged PT/​
INR, APTT, TCT, and hypofibrinogenaemia. The use of all-​trans-​
retinoic acid during induction chemotherapy of promyelocytic
leukaemia has reduced the frequency of life-​threatening bleeding.
Antifibrinolytic therapy can control bleeding in cancer-​associated
hyperfibrinolysis.
Cardiopulmonary bypass surgery
Excess bleeding, defined as more than 1 litre per procedure, is a
common problem following heart surgery utilizing cardiopul-
monary bypass (extracorporeal circulation). About 20% of all red
cell concentrates in the United States of America are given for cardiac
surgical bleeding. About 5% of patients require urgent resternotomy
for critical rates of blood loss (defined as >500 ml in the first 1 h;
400 ml/​h in the first 2 h; >300 ml/​h in the first 3 h; or >1 litre in 4 h).
Re-​exploration reveals bleeding vessels in two-​thirds of patients; the
remainder have diffuse oozing.
Thrombocytopenia, transient platelet dysfunction, and hyper­
fibrinolysis are the principal haemostatic defects. Typically, the platelet
count falls by between 30 and 60% mainly from haemodilution, al-
though platelet losses from bleeding and within the extracorporeal
perfusion device also occur. The thrombocytopenia persists for 3 to
Box 22.7.5.3  Haemostatic abnormalities associated
with dysproteinaemias
•	 Interference with fibrinogen polymerization
•	 Isolated factor deficiency:
—​	 Factor X, fibrinogen, or α2-​antiplasmin deficiency (amyloidosis)
—​	 Acquired von Willebrand syndrome (monoclonal gammopathy)
•	 Hyperviscosity (compromising vascular integrity)
•	 Circulating glycosaminoglycan (heparin-​like inhibitor)
•	 Thrombocytopenia secondary to:
—​	 marrow failure (disease or treatment related)
—​	 autoimmune thrombocytopenia
•	 Platelet dysfunction
section 22  Haematological disorders
5556
4 days, followed by recovery of the platelet count to values exceeding
the preoperative baseline. Marked prolongation of the bleeding time
(>30 min) quickly improves to under 15 min shortly after surgery,
and to normal several hours later. Some platelet function defects
are ‘extrinsic’ and reversible (e.g. hypothermia, heparin), whereas
others indicate longer-​lasting ‘intrinsic’ changes (surface glycopro-
tein deficiency, acquired granule depletion). Preoperative treatment
with aspirin, clopidogrel, or ticagrelor also increases bleeding; un-
like with aspirin or clopidogrel, platelet transfusions are not usually
effective for bleeding associated with ticagrelor.
The importance of hyperfibrinolysis in postcardiac surgical
bleeding is highlighted by meta-​analysis of studies of high-​dose
aprotinin, a plasmin inhibitor derived from bovine lung: a two-​thirds
reduction in blood transfusion, and 50% reduction in resternotomy.
However, aprotinin is now infrequently used because of concerns
regarding its adverse effect profile. Other antifibrinolytic drugs that
reduce bleeding include the lysine analogues, tranexamic acid (e.g.
10 mg/​kg bolus before cardiopulmonary bypass; then 1 mg/​kg per
h, although dosing regimens vary widely) and ε-​aminocaproic acid
(total dose up to 20 g). Although these therapies are usually given
before cardiopulmonary bypass, they may also provide benefit when
used postoperatively for bleeding patients.
Management of postcardiac surgical bleeding also includes blood
transfusions, especially platelets and frozen plasma, although their
benefit is unproven. Residual heparin, including heparin ‘rebound’,
can respond to additional protamine. Desmopressin probably is in-
effective. No universally accepted algorithm for management exists.
Liver disease
Hyperfibrinolysis complicating liver disease is discussed elsewhere.
Venom-​induced coagulopathies (snake bites)
Envenomations can harm or kill humans generally through sys-
temic effects (e.g. profound hypotension) (see also Chapter 10.4.2).
Sometimes, however, life-​threatening coagulopathies result.
Snake bites
In the United States of America, about 8000 bites from venomous
snakes occur each year, resulting in 10 to 20 deaths. This relatively
low mortality reflects the less lethal character of New World snakes,
as well as the victim’s usual close proximity to medical facilities and
antivenin therapy. Pit vipers (rattlesnakes, copperheads, cotton-
mouths, massasaugas) account for 99% of snakebite poisonings in
the United States of America. Worldwide, annually over 100 000
people are estimated to die from snakebite, many in India. Although
death usually results from multiple mechanisms (such as circulatory
shock, rhabdomyolysis, renal failure, pulmonary failure, and neuro-
toxicity), bleeding is sometimes the major factor.
Venoms contain multiple digestive enzymes with a broad spec-
trum of activity that can include effects on human haemostasis
(Table 22.7.5.6). Within a species, haemostatic effects of envenom-
ation vary with snake age, diet, and other factors. North American
rattlesnakes typically cause the ‘defibrination syndrome’; despite
even profound hypofibrinogenaemia, bleeding is uncommon. In
contrast, venom from Old World vipers frequently cause generalized
activation of the coagulation system (DIC), with a greater chance of
bleeding or microvascular thrombosis. Bleeding can also result from
platelet inhibitors present within venom; for example, the platelet
fibrinogen receptor antagonist echistatin (from Echis carinatus), or
‘haemorrhagins’ such as jararhagin (from Bothrops jararacussu) that
damage endothelium.
Immediate treatment of a snake bite includes efforts to limit the
venom spread (immobilizing and placing a constriction band prox-
imal to the bite site). Rapid transport to medical facilities is crucial
since antivenin therapy is the mainstay of treatment. Antivenin
treatment is indicated for patients with significant pain or swelling,
as well as suspected or proven haemostasis abnormalities, as these
indicate envenomation rather than a ‘dry bite’. Hypersensitivity
testing to the antivenin should be performed to rule out pre-​existing
hypersensitivity to horse serum. The treatment of snake bite is dis-
cussed in Chapter 10.4.2.
Coagulation studies should include complete blood count
(including platelets), PT/​INR, APTT, TCT, fibrinogen, and FDPs.
Abnormal results indicate envenomation, and are an indication for
antivenin therapy. The bedside assessment of defibrination involves
placing a few millilitres of blood in a clean, dry test tube at room
temperature for 20 min; incoagulable blood indicates defibrination.
Usually, blood products should only be given to patients with
bleeding. A small clinical trial found that heparin was ineffective in
patients with DIC caused by a Russell’s viper bite.
Laboratory and therapeutic uses of snake venoms
Snake-​venom fractions are useful for certain laboratory assays. For
example, the thrombin-​like enzyme batroxobin (reptilase, Bothrops
atrox and moojeni), cleaves fibrinopeptide A from fibrinogen even in
the presence of heparin. Thus, a prolonged reptilase time indicates
hypofibrinogenaemia even in heparin-​containing plasma.
Ecarin activates prothrombin irrespective of its γ-​carboxylation
status; thus, it can be used to detect proteins induced by vitamin K ant-
agonists to document vitamin K deficiency or dysprothrombinaemia.
An ecarin clotting time is superior to the APTT for monitoring
therapy with hirudin (no longer marketed) or other direct thrombin
inhibitors (e.g., argatroban). Differences in phospholipid dependency
of venom prothrombin activators have led to the use of a Textarin/​
ecarin ratio to detect lupus anticoagulants; a ratio over 1.3 is a sensitive
and relatively specific test for lupus anticoagulants.
Russell’s viper venom contains a potent activator of factor X (RVV-​
X); the dilute Russell’s viper venom time (dRVVT), performed by
adding RVV-​X and diluted rabbit brain phospholipid to test plasma
prior to recalcification, measures the rate of formation and activity
of the phospholipid-​dependent prothrombinase complex in produ-
cing thrombin. The dRVVT is thereby prolonged in the presence of
a lupus anticoagulant.
A commercially available protein C activator (Protac) from
Agkistrodon contortrix contortrix (the southern copperhead) has
greatly simplified assays for protein C activity, as well as in screening
for defects in the protein C anticoagulant pathway.
The defibrinogenating snake venom ancrod (Arvin, derived from
the Malayan pit viper Calloselasma [Agkistrodon] rhodostoma),
which proteolyses fibrinopeptide A, was formerly used for manage-
ment of HIT, acute stroke, thrombotic nephropathy, and priapism.
The inability to control thrombin generation is a potential drawback
of this therapy. Batroxobin (Defibrase) is another defibrinogenating
venom that has seen limited clinical applications.
22.7.5  Acquired coagulation disorders
5557
Prothrombotic-​acquired coagulation disorders
Some acquired coagulation disorders are characterized by an in-
creased risk for thrombosis, rather than bleeding. Accordingly, the
appropriate treatment usually involves anticoagulant therapy, even if
there are abnormal coagulation or platelet count values.
Macrovascular thrombosis
Some acquired coagulation disorders typically cause thrombosis
in large veins and arteries, although small-​vessel thrombi can also
result.
Heparin-​induced thrombocytopenia
HIT is caused by IgG antibodies that recognize multimolecular com-
plexes of platelet factor 4 (PF4) and heparin. Thrombosis results from
IgG-​induced platelet activation (via platelet Fc receptors), resulting
in the generation of procoagulant, platelet-​derived microparticles,
tissue factor expression by monocytes, and inactivation of heparin
by PF4 released from platelets. Increased thrombin–​antithrombin
complex levels indicate DIC in almost all patients with this condi-
tion, although a prolonged INR and/​or APTT and/​or low fibrinogen
level occurs in only approximately 20% of cases.
Typically, the fall in platelet count begins 5 to 10 days after starting
heparin (‘typical-​onset’ HIT); however, in patients who received
heparin within the past 5 to 100 days, the platelet count can fall
abruptly upon resuming heparin therapy (‘rapid-​onset’ HIT), be-
cause of residual circulating antibodies. HIT occurs in approxi-
mately 0.2% of heparin-​treated patients (but up to 5% of certain
high-​risk populations: e.g. postoperative orthopaedic patients re-
ceiving unfractionated heparin for over 1 week). HIT is less frequent
in patients initially treated with low molecular weight heparin or
fondaparinux. HIT antibodies are remarkably transient; moreover,
HIT does not usually recur with future heparin exposure, although
Table 22.7.5.6  Venom-​induced coagulopathies (selected examples)
Animal source of venom
Main biological effects (trivial name of
venom component in bold)
Comments
Main distribution
Venomous snakes
Family Viperidae
Subfamily crotalinae (pit vipersa)
Crotalus adamanteus (Eastern
diamondback rattlesnake)
Crotalase: cleaves FPA, but not FPB,
from fibrinogen (decreased fibrinogen,
plasminogen; increased FDPs)
‘Thrombin-​like’ based upon fibrinopeptide
A cleavage, but does not activate platelets or
factor XIII; despite ‘defibrination syndrome’,
bleeding is uncommon
USA (coastal plain from
Florida to Mississippi)
Crotalus atrox (Western
diamondback rattlesnake)
Catroxobin: cleaves FPA from fibrinogen;
other fibrinogenase activities
Also causes defibrination syndrome, usually
without bleeding; venom also contains
catrocollastatin-​C (platelet inhibitor)
USA (California to
Arkansas); Mexico
Calloselasma [Agkistrodon]
rhodostoma (Malayan pit viper)
Ancrod: cleaves FPA from fibrinogen
Purified ancrod previously used as an
antithrombotic agent
Southeast Asia
Subfamily viperinae (true vipersa)
Echis carinatus (saw-​scaled viper)
Ecarin: activates prothrombin and platelets
Causes DIC, often with bleeding; most
common cause of snake-​bite mortality in the
African savannah
India, Africa, Asia
Daboia russelli (Russell’s viper,
formerly, Vipera russelli)
Russell’s viper venom: activates factor X
Causes DIC, often with bleeding; venom also
causes direct nephrotoxicity
Far East
Bothrops jararacussu (jararacucu,
lance-​headed pit viper)
•	 Botrocetin: platelet agglutination via VWF;
•	 Jararhagin: haemorrhagin
Venom also contains thrombin-​like and
factor Xa-​activating enzymes, and can cause
severe bleeding
Brazil
Family Elapidaeb
Notechis scutatus (tiger snake)
Notecarin: activates prothrombin
Fatal haemorrhage has been reported
Australia
Family Colubridaec
Nonsnake envenomations that cause coagulopathy
Lonomia achelous (caterpillar)
Proteolysis of factor XIII; reduced fibrinogen,
factor V, plasminogen, and increased FDPs
also observed
Severe bleeding in humans (wound
site, mucous membranes, and internal
haemorrhage)
Venezuela, Brazil
Loxosceles reclusa (brown recluse
spider)
Activation of endothelium, with resulting
dysfunction of interactions with PMNs
Potential for severe skin lesions; systemic
effects (DIC, haemolytic anaemia) occur in
small minority of patients
Midwest USA
Two other families of venomous snakes (Hydrophiidae and Atractaspididae) do not cause coagulopathies.
a Pit vipers are New World snakes named for the heat-​sensitive pit located between the eye and the nostril that enables the snake to detect warm-​blooded prey even in darkness: the
three genera of the Crotalidae family that inhabit the USA are Crotalus (rattlesnakes), Agkistrodon (moccasins, including the copperheads and cottonmouths), and Sistrurus
(massasaugas and pigmy rattlesnakes).
b With the exception of several Australian species, such as taipan, tiger snakes, brown snakes, and black snakes, elapid snake bites usually cause neurotoxicity, and only occasionally
result in haemostatic abnormalities.
c The colubrid family includes boomslang, vine snake, keel backs, and the South American ‘green snake,’ which can also cause bleeding.
section 22  Haematological disorders
5558
deliberate re-​exposure is usually restricted to special situations
(e.g. cardiac or vascular surgery), and only if platelet-​activating anti-
bodies are no longer detectable.
Most patients with HIT develop venous or arterial thrombosis
(Fig. 22.7.5.2), most commonly a DVT, pulmonary embolism, major
limb artery thrombosis, stroke, or myocardial infarction. Acute or
chronic adrenal failure from bilateral adrenal haemorrhagic necrosis
(manifestation of adrenal vein thrombosis) has been described. The
thrombocytopenia is typically moderate in severity (median platelet
count nadir 60 × 109/​litre), and in only 10% of patients does the platelet
count fall to less than 20 × 10 /​litre. In at least 10% of patients, the platelet
count never drops below 150 × 109/​litre. This degree of thrombocyto-
penia in HIT is much less marked than observed in classic immune-​
mediated drug-​induced thrombocytopenia (Fig. 22.7.5.2).
Laboratory testing for HIT antibodies includes activation and
antigen assays. The former assays detect antibodies via their platelet-​
activating properties; the best platelet activation assays utilize
washed platelets, for example, the serotonin-​release assay (SRA) and
the heparin-​induced platelet activation (HIPA) test. Commercially
available antigen assays detect antibodies that bind to surface-​immo-
bilized PF4 complexed to heparin or polyvinylsulphonate. Antigen
assays are more likely to detect clinically insignificant antibodies,
with the potential for a false-​positive diagnosis of HIT. Recently,
automated HIT assays that give results within 30 minutes of plasma
preparation have become available.
Treatment includes stopping heparin and instituting alternative
nonheparin anticoagulation. Coumarin should not be given to pa-
tients during the acute (thrombocytopenic) phase of HIT; particu-
larly in those with associated DVT, there is substantial risk of limb loss
due to microvascular thrombosis (coumarin-​induced venous limb
gangrene). Thus, coumarin therapy should be postponed until the
platelet count has recovered to at least 150 × 109/​litre, and only then
cautiously overlapped (over at least 5 days) with an agent that inhibits
thrombin (or its generation). Suitable rapidly-​acting anticoagulants
include danaparoid (a low molecular weight mixture of glycosa-
minoglycans with predominant antifactor Xa activity), fondaparinux
(a synthetic antithrombin-​dependent factor Xa inhibitor modelled
after the crucial pentasaccharide sequence within active heparin),
argatroban (a synthetic small-​molecule direct thrombin inhibitor),
and bivalirudin (a synthetic 20-​amino acid analogue of hirudin).
Argatroban and bivalirudin dose adjustments are generally per-
formed using APTT monitoring. In contrast, fondaparinux and
danaparoid do not require APTT monitoring, an advantage that
avoids potential for ‘APTT confounding’, which refers to the situation
where APTT-​monitored therapies (e.g. argatroban and bivalirudin)
fail in patients with HIT-​associated DIC, as supratherapeutic APTT
levels (reflecting HIT-​associated coagulopathy rather than indicating
overanticoagulation) lead to inappropriate dose interruptions/​reduc-
tions, with subsequent progression of thrombosis (including micro-
vascular thrombosis/​limb necrosis). DOACS (e.g., rivaroxaban,
3
10
20
50
100
200
500
1000
5
No. of Patients (arbitrary units, increasing from bottom to top)
Bleeding
Thrombosis
~10 x 109/L
(median)
Heparin-induced
thrombocytopenia
Nadir Platelet Counts (x10−9/L) Shown on a Log10 Scale
Drug-induced immune
thrombocytopenia
~60 x 109/L
(median)
Fig. 22.7.5.2  Nadir platelet counts shown on a log10 scale: comparison of heparin-​induced
thrombocytopenia versus ‘classic’ drug-​induced immune-​mediated thrombocytopenic
purpura (e.g. caused by quinine or vancomycin). Whereas the latter typically produces
severe thrombocytopenia (median platelet count nadir c.10 × 109/​litre), heparin-​induced
thrombocytopenia usually results in mild-​to-​moderate thrombocytopenia (20–​150 × 109/​litre in
c.80% of patients; median platelet count nadir c.60 × 109/​litre). Thrombosis occurs in 50% or more
of patients with heparin-​induced thrombocytopenia, whereas drug-​induced thrombocytopenia
manifests as purpura and other mucocutaneous haemorrhage.
From Warkentin TE (2007). Drug-​induced immune-​mediated thrombocytopenia—​from purpura to thrombosis.
N Engl J Med, 356, 891–​3, with permission.
22.7.5  Acquired coagulation disorders
5559
apixaban) can be used to treat HIT; their fixed dosing regimens avoid
this problem of PTT confounding. Among patients with HIT, low
molecular weight heparin (LMWH) treatment has a high risk for
clinical cross-​reactivity, and should be considered a contraindicated
treatment for acute HIT. Some patients benefit from selected ad-
junctive treatments, such as high-dose intravenous immunoglobulin
(which interrupts HIT antibody-induced platelet activation). The
dramatic natural history of HIT, with a risk for subsequent throm-
bosis of about 50% even after stopping heparin, means that an al-
ternative anticoagulant, together with DVT surveillance, should be
considered for all patients strongly suspected to have HIT. Future use
of heparin and LMWH is usually avoided in patients with a history
of HIT (as suitable nonheparin anticoagulant options usually exist);
however, the risk of HIT recurrence appears to be low, and for some
situations (particularly cardiac and vascular surgery), intraoperative
anticoagulation with UFH is recommended, provided that platelet-​
activating antibodies are no longer present.
Protamine-​induced thrombocytopenia
Recently, a prothrombotic disorder associated with platelet-​activating
IgG antibodies that recognize protamine/​heparin complexes has
been reported. Cases have included diabetic patients who developed
marked thrombocytopenia and other complications after receiving
protamine sulphate to reverse heparin anticoagulation after cardiac
surgery; apparent triggers of immunization may have included pre-
operative LMWH thromboprophylaxis in the setting of protamine-​
insulin therapy.
Adenocarcinoma-​associated chronic DIC
Metastatic adenocarcinoma sometimes presents as venous or ar-
terial thrombosis accompanied by DIC. The diagnosis is suggested
by an unexpected rise in the platelet count during heparin treatment,
followed by an abrupt platelet count fall, together with new or pro-
gressive thrombosis, when heparin is stopped, despite therapeutic
anticoagulation with warfarin. The clinical situation can mimic HIT
(‘pseudo-​HIT’), but HIT antibodies are absent (or weakly detect-
able), and the platelet count recovers during resumption of heparin
(Fig. 22.7.5.3). Oral anticoagulants are ineffective, and may even
cause venous limb gangrene (discussed subsequently). Heparin, es-
pecially LMWH, is the preferred treatment. Tissue factor-​containing
tumour vesicles, and factor Xa-​activating enzymes found in tumour
extracts, are two possible explanations for these procoagulant effects
of adenocarcinoma.
Antiphospholipid antibody syndrome (‘lupus anticoagulant’)
This clinicopathological syndrome is characterized by large-​vessel
venous and/​or arterial thrombosis, recurrent miscarriages, and
thrombocytopenia. An associated ‘lupus anticoagulant’ (or ‘non-
specific inhibitor’) is a prolonged APTT that results from the inter-
ference by antibodies against phospholipid-​dependent coagulation
reactions; these antiphospholipid antibodies are usually directed
against protein cofactors such as β2-​glycoprotein I and prothrombin.
Sometimes a prolonged PT/​INR is caused by non-​neutralizing
antiprothrombin antibodies that cause hypoprothrombinaemia by
increased prothrombin clearance.
Despite these laboratory abnormalities, bleeding is unusual,
since severe thrombocytopenia or hypoprothrombinaemia is un-
common. More often, antiphospholipid antibodies are associated
with intermittent thrombosis, and patients often require long-​term
anticoagulation. The explanation for the paradoxical association
with thrombosis remains elusive, but it could be caused by antibody
interactions with other protein cofactors described (e.g. activated
protein C, protein S, and thrombomodulin). Many patients have a
thrombocytopenia that is typically mild and intermittent. Other less
common complications include cardiac valvulitis and microvascular
thrombosis, which can manifest as acrocyanosis, digital ulceration/​
gangrene, and livedo reticularis. Rarely, the abrupt onset of life-
threatening multiple large- and (especially) small-vessel vascular oc-
clusions occurs (‘catastrophic antiphospholipid syndrome’ [CAPS]);
potential treatments include heparin, high-dose corticosteroids,
high-dose intravenous immunoglobulin, and/or plasma exchange.
Antiphospholipid antibodies are detected by enzyme-​immunoassay
using purified phospholipids as the target antigen (e.g. the anti­
cardiolipin antibody assay). Lupus anticoagulant activity is shown
by demonstrating inhibition of phospholipid-​dependent coagulation
assays. Several assays should be performed, as anti-​β2 glycoprotein
I antibodies especially interfere with the conversion of prothrombin
to thrombin (i.e. best detectable by the dilute Rusell’s viper venom
time), whereas antiprothrombin antibodies interfere most with global
coagulation assays (e.g. kaolin clotting time). The coagulation times
remain prolonged following mixing with normal plasma; confirm-
ation involves adding excess phospholipid to neutralize the effects of
the antiphospholipid antibodies. Not all APTT reagents are sensitive
to antiphospholipid antibodies, and so these phospholipid-​dependent
coagulation assays should be performed in the appropriate clinical
situation, even if the APTT is normal.
The term ‘lupus anticoagulant’ refers to the frequent occurrence
of these antibodies in patients with systemic lupus erythematosus;
nevertheless, most patients with the antiphospholipid antibody
400
1000
800
600
Days after starting heparin
Platelet fall and
new PE when
UFH held for
liver biopsy
Clinical and platelet
count improvement
upon restarting UFH
200
0
0
4
8
12
16
20
24
28
32
36
40
44
48
52
56
60
UFH
Warfarin
Ancrod
Negative HIT tests (SRA)
R leg phlegmasia
and new PE
when INR = 6.5
Abrupt platelet count
fall off heparin
Rising platelet count on UFH
R leg DVT, PE
Pseudo-HIT cycle repeated
Recurrent R leg phlegmasia
and new PE with an INR = 4.0
UFH
UFH
Warfarin
UFH
Fig. 22.7.5.3  Pseudo-​HIT. Adenocarcinoma with thrombocytopenia
and phlegmasia cerulea dolens after stopping unfractionated heparin.
The timing of thrombocytopenia onset suggested heparin-​induced
thrombocytopenia prompting the use of an alternative anticoagulant
(ancrod). Heparin was restarted when PF4/​heparin antibodies were
not detected by activation assay (serotonin-​release assay (SRA)).
Subsequently, heparin discontinuation led to the recurrence of
thrombocytopenia and warfarin-​associated phlegmasia cerulea dolens
(repeat of pseudo-​HIT cycle). DVT, deep venous thrombosis; INR,
international normalized ratio; PE, pulmonary embolism.
section 22  Haematological disorders
5560
syndrome do not have systemic lupus erythematosus. Some pa-
tients have other autoimmune disorders, malignancy, infections,
or procainamide treatment, but usually no associated condition is
identified (primary antiphospholipid antibody syndrome). Many
patients require long-​term anticoagulation. Corticosteroids can
benefit patients with bleeding caused by hypoprothrombinaemia.
Microvascular thrombosis
Some disorders of haemostasis are characterized by small-​vessel
thrombi, affecting either small venules (e.g. coumarin-​induced ne-
crosis) or capillaries/​post-​capillary venules (DIC with severe de-
pletion of natural anticoagulants, protein C, and antithrombin) or
arterioles (e.g. TTP).
Coumarin-​induced skin necrosis
Coumarin-​induced skin necrosis (CISN) is characterized by ne-
crosis of the skin and underlying subcutaneous tissues that typ-
ically begins 2 to 5  days after commencing warfarin or coumarin
anticoagulants. CISN results from failure of the protein C natural
anticoagulant system to downregulate thrombin generation in the
microvasculature. The relatively short half-​life of protein C, com-
pared with prothrombin, explains the temporal profile of CISN—​that
is to say, a transient period of disproportionately reduced protein C
activity soon after starting coumarin (Table 22.7.5.7). Furthermore,
a relatively high proportion of affected patients have a hereditary ab-
normality of the protein C anticoagulant pathway, especially protein
C deficiency. Other disorders associated with CISN include con-
genital deficiency in protein S or antithrombin, factor V Leiden, and
HIT. The pathology is a predominantly noninflammatory, small-​
vessel thrombosis affecting the subcutaneous postcapillary venules
and small veins.
CISN characteristically affects central (nonacral) sites with sub-
stantial underlying fatty tissues, such as the breast, buttocks, hips,
and thighs (Fig. 22.7.5.4). Less common areas include the anterior
abdomen, flank, back, penis, legs, arms, and face. About 75% of pa-
tients are women; one-​third have multiple lesions that can be sym-
metrical. The earliest features are localized pain, induration, and
erythema; over the next few hours, the skin lesions progress to cen-
tral purplish or black discoloration, with blistering, subsequently
demarcating to full-​thickness skin necrosis. CISN is rare (1/​10 000
patients treated with warfarin).
Prompt reversal of anticoagulation with vitamin K may prevent
incipient CISN if recognized early. However, the diagnosis is usually
not made until necrosis is established; at this point, it is unknown
whether vitamin K, fresh frozen plasma, or protein C concentrates
alter its natural history. In patients without HIT, warfarin is usually
replaced by heparin. Many patients require surgical treatment, such
as skin grafting or tissue amputation. Following recovery, it is usu-
ally safe to reintroduce warfarin provided certain precautions are
taken, for example, the gradual initiation of oral anticoagulation.
Coumarin-​induced venous limb gangrene
Venous limb gangrene involves the acral (peripheral) regions of the
body—​most often the toes, feet, and legs, but sometimes also the
fingers, hands, and arms—​usually in association with DVT. The
severity ranges from an initial stage of phlegmasia caerulea dolens
(‘swollen, blue, painful’ limb) to extensive venous limb gangrene re-
quiring limb amputation. Two disorders predispose to coumarin-​
induced venous limb gangrene:  HIT and cancer-​associated DIC.
Recent data suggest that the supratherapeutic INR (typically >3.5)
that characterizes venous limb gangrene is caused by a severe re-
duction in factor VII, which parallels a severe reduction in protein
C activity that explains the microvascular thrombosis underlying
this syndrome. Essentially, coumarin interferes with the protein C
anticoagulant pathway, while at the same time it is unable to con-
trol the increased thrombin generation characteristic of HIT or
cancer-​associated DIC.
Purpura fulminans and symmetrical peripheral gangrene
Purpura fulminans is a rare syndrome of DIC and microvascular
thrombosis that results in multicentric ischaemic necrosis of the
skin and subcutaneous tissues, predominantly affecting the extrem-
ities (Fig. 22.7.5.5). The most common cause is overwhelming septi-
caemia, especially with meningococcus. A severe, acquired reduction
Table 22.7.5.7  Half-​lives of vitamin K-​dependent procoagulant and
anticoagulant factors
Procoagulant factors
Half-​life (h)
Anticoagulant
factors
Half-​life (h)
Factor II (prothrombin)
60
Protein C
9
Factor X
40
Protein S
40−60
Factor IX
24
Factor VII
4−6
The longer half-​life of the major procoagulant vitamin K-​dependent factor (factor II, or
prothrombin), compared with the major vitamin K-​dependent natural anticoagulant
factor (protein C), is relevant to the pathogenesis of CISN (see text).
‘Classic’ CISN
(central skin necrosis)
DVT
DVT Venous
limb
gangrene
Fig. 22.7.5.4  Coumarin-​induced skin necrosis: ‘classic’ syndrome
(usually affecting central tissue sites) and coumarin-​induced venous limb
gangrene. Typically, an active DVT subtends the distal extremity affected
by venous limb gangrene.
From Warkentin TE (1996). Heparin-​induced thrombocytopenia IgG-​mediated
platelet activation, platelet microparticle generation, and altered procoagulant/​
anticoagulant balance in the pathogenesis of thrombosis and venous limb gangrene
complicating heparin-​induced thrombocytopenia. Transfus Med Rev, 10, 249–​58,
with permission.
22.7.5  Acquired coagulation disorders
5561
in protein C activity complicating DIC is the most likely cause for
the microvascular thrombosis, and some experts recommend treat-
ment with protein C concentrates, if available. Autoantibodies
against protein S have been implicated in patients with postvaricella
purpura fulminans. In other patients with apparent ‘idiopathic’
purpura fulminans, autoantibodies that interfere with the protein
C anticoagulant system have been described. Peripheral symmetric
gangrene is a term sometimes used when acral regions of two or
more limbs are affected (Fig 22.7.5.5). More recently, a role for acute
ischaemic hepatitis (‘shock liver’) in predisposing to microvascular
thrombosis and ischaemic limb necrosis/​central skin necrosis sec-
ondary to natural anticoagulant (protein C, antithrombin) depletion
in DIC of critical illness with circulatory shock has been reported.
Septicaemia and other systemic inflammatory
response syndromes
Multiple organ failure often complicates septicaemia and other
systemic inflammatory disease syndromes, including adult re-
spiratory distress syndrome, fat embolism, and acute pancreatitis.
Thrombocytopenia and coagulopathy are common, and some pa-
tients have DIC that could contribute to organ dysfunction via
microvascular thrombosis. However, a prothrombotic basis for
organ failure is usually speculative, as microthrombosis is rarely
documented pathologically, and nonthrombotic microvascular dis-
turbances that impair tissue oxygen delivery also occur.
Thrombotic microangiopathy
Thrombotic microangiopathy is a clinicopathological syndrome
of microangiopathic haemolysis and thrombocytopenia carrying
a risk for arteriolar occlusion by microaggregates of platelets and
VWF, particularly affecting the kidneys and central nervous system.
Microangiopathic red cell changes are characteristic, for example,
‘helmet cells’ (schistocytes) and small, triangular red cell fragments.
The prototypic illness is TTP, which typically affects adults and is
idiopathic. However, familial and secondary forms of TTP also exist.
The pathogenesis of TTP involves the formation of platelet–​VWF
microaggregates in high-​shear situations (arterioles). Platelet-​
bound VWF levels are increased during TTP. Patients with familial
TTP have ultralarge multimers of VWF during remission; these very
large multimers disappear during active disease. A constitutional
deficiency of a VWF-​cleaving metalloproteinase (ADAMTS13) has
been identified in patients with familial TTP. In many patients with
nonfamilial TTP, an IgG autoantibody, which inhibits the VWF-​
cleaving metalloproteinase, has been identified.
The mainstays of treatment for acute TTP are corticosteroids and
frozen plasma (or fresh frozen plasma) given by infusion or apheresis.
Corticosteroids, often given as prednisone 200 mg/​day, may treat the
autoimmune component of TTP. Provision of either frozen plasma,
or the cryoprecipitate-​depleted fraction of plasma (cryosupernatant),
has greatly reduced mortality in TTP, likely through several mechan-
isms, for example, apheresis helps clear the pathogenic autoantibody
and large VWF multimers. The monoclonal antibody rituximab,
which recognizes CD20 (surface antigen on B-​cell precursors), ap-
pears to be effective in many patients with refractory or relapsing TTP.
The haemolytic uraemic syndrome (HUS) is a nephrotropic
thrombotic microangiopathy with a distinct pathogenesis, including
its association with verocytotoxin-​producing Escherichia coli usu-
ally acquired from eating undercooked meat (hamburger disease).
Although the majority of cases of ‘typical’ HUS is post-​diarrhoeal
(or D+ HUS), the remaining 25% of ‘atypical’ (or D−) HUS lacks a
diarrhoeal prodrome, with many patients having a hereditary de-
fect in an alternate complement pathway protein, such as gain-​of-​
function mutations in C3 or factor B, or loss-​of-​function mutations
in factor H or factor I.
Haemostasis in the newborn
Neonatal vitamin K deficiency
Haemorrhagic disease of the newborn caused by vitamin K defi-
ciency was once a relatively common cause of bleeding during the
first week of life, particularly in breastfed infants. Low vitamin K
levels in mother’s milk, and insufficient colonization of the newborn
bowel by bacteria producing vitamin K, predispose to the inability
to meet the infant’s vitamin K requirements (1 μg/​kg per day). The
routine administration of vitamin K, either 1 mg given intramuscu-
larly immediately after birth, or three oral doses of vitamin K, has led
to the near disappearance of this problem in developed countries.
Bleeding within 24 h of birth can occur in certain high-​risk settings,
for example, mothers receiving anticonvulsants or warfarin; in these
cases, the mother should receive vitamin K, 10 mg by mouth, each
day for 2 weeks prior to delivery. Vitamin K deficiency occurring
later in infancy despite appropriate neonatal vitamin K prophylaxis
can indicate hepatobiliary or bowel disease.
Neonatal DIC
In neonates, DIC commonly complicates infection, asphyxia, re-
spiratory distress syndrome, aspiration of meconium or amniotic
fluid, maternal hypertensive syndrome, hypothermia, and brain
Protein C and antithrombin
deficiency caused by increased
consumption (disseminated
intravascular coagulation)
Detectable pulses
on palpation or
Doppler signal
Nonacral
skin necrosis
(purpura
fulminans)
Nonacral skin necrosis
(purpura fulminans)
Protein C and antithrombin
deficiency caused by acute
ischaemic hepatitis (“shock liver”)
Symmetric peripheral
gangrene (acral skin
necrosis) with reduced
circulation to extremities
associated with
hypotension or use
of vasopressors
Fig. 22.7.5.5  Clinical profile of symmetric peripheral gangrene.
DIC, disseminated intravascular coagulation; FDP(s), fibrin(ogen) degradation
product(s); FPA, fibrinopeptide A; PMNs, polymorphonuclear leucocytes; VWF, von
Willebrand factor.
Adapted from Warkentin TE (2015). Ischaemic limb gangrene with pulses. N Engl J
Med, 373, 642–​55. Copyright © (2015) Massachusetts Medical Society. Reprinted
with permission.
section 22  Haematological disorders
5562
injury. This condition poses a significant risk of bleeding or throm-
bosis, as the immature liver has an impaired capacity to synthesize
coagulation factors, and the mononuclear phagocyte system has a
limited ability to clear activated coagulation factors. Treatment is
aimed at the underlying cause of the DIC, with blood product given
for the bleeding.
Neonatal purpura fulminans
Purpura fulminans can begin within hours or days following birth,
often first affecting the heels or venepuncture sites. The underlying
cause is usually a congenital abnormality affecting the protein C
anticoagulant system (homozygous deficiency of protein C or pro-
tein S). Frozen plasma or protein C concentrates given every few
days prevents a recurrence in some patients.
FURTHER READING
Bakchoul T, Jouni R, Warkentin TE (2016). Protamine (heparin)-​
induced thrombocytopenia: a review of the serological and clin-
ical features associated with anti-​protamine/​heparin antibodies.
J Thromb Haemost, 14, 1685–​95.
Bevan DH (1999). Cardiac bypass haemostasis: putting blood through
the mill. Br J Haematol, 104, 208–​19.
Cole MS, Minifee PK, Wolma FJ (1988). Coumarin necrosis—​a review
of the literature. Surgery, 103, 271–​7.
Galli M (2013). Treatment of the antiphospholipid syndrome. Auto
Immun Highlights, 5, 1–​7.
George JN, Nester CM (2014). Syndromes of thrombotic micro­
angiopathy. N Engl J Med, 371, 654–​66.
Holbrook A, et al. (2012). Evidence-​based management of anticoagu-
lant therapy: antithrombotic therapy and prevention of thrombosis,
9th ed:  American College of Chest Physicians Evidence-​Based
Clinical Practice Guidelines. Chest, 141 Suppl, e152S–​84S.
Holcomb JB, et al. (2015). Transfusion of plasma, platelets, and red blood
cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe
trauma: the PROPPR randomized clinical trial. JAMA, 313, 471–​82.
Hunt BJ (2014). Bleeding and coagulopathies in critical care. N Engl J
Med, 370, 847–​59.
Janbain M (2015). Acquired hemophilia A: emerging treatment op-
tions. J Blood Med, 6, 143–​50.
Kitchens CS (1992). Hemostatic aspects of envenomation by North
American snakes. Hematol Oncol Clin North Am, 6, 1189–​95.
Levi M, van der Poll T (2014). A short contemporary history of dis-
seminated intravascular coagulation. Semin Thromb Hemost, 40,
874–​80.
Ortel TL, et al. (1994). Topical thrombin and acquired coagulation
factor inhibitors: clinical spectrum and laboratory diagnosis. Am J
Hematol, 45, 128–​35.
Sane DC, et al. (1989). Bleeding during thrombolytic therapy for acute
myocardial infarction: mechanisms and management. Ann Intern
Med, 111, 1010–​22.
Warkentin TE (2001). Venous limb gangrene during warfarin treat-
ment of cancer-​associated deep venous thrombosis. Ann Intern Med,
135, 589–​93.
Warkentin
TE
(2007).
Drug-​induced
immune-​mediated
thrombocytopenia—​from purpura to thrombosis. N Engl J Med,
356, 891–​3.
Warkentin TE (2011). Fondaparinux treatment of acute heparin-​
induced thrombocytopenia confirmed by the serotonin-​release assay:
a 30-​month, 16-​patient case series. J Thromb Haemost, 9, 2389–​96.
Warkentin TE (2011). How I diagnose and manage HIT. Hematol Am
Soc Hematol Educ Program, 2011, 143–​9.
Warkentin TE (2015). Ischemic limb gangrene with pulses. N Engl J
Med, 373, 642–​55.
Warkentin TE (2019). High-dose intravenous immunoglobulin for the
treatment and prevention of heparin-induced thrombocytopenia: a
review. Exp Rev Hematol.
Warkentin TE, Greinacher A (2013). Heparin-​induced thrombocyto-
penia, 5th edition. CRC Press, Boca Raton, FL.
Warkentin TE, Pai M, Linkins LA (2017). Direct oral anticoagulants
for treatment of HIT: update of Hamilton experience and literature
review. Blood, 130, 1104–13.
Warkentin TE, et al. (1995). Heparin-​induced thrombocytopenia in pa-
tients treated with low-​molecular-​weight heparin or unfractionated
heparin. N Engl J Med, 332, 1330–​5.
Warkentin TE, et al. (1997). The pathogenesis of venous limb gangrene
associated with heparin-​induced thrombocytopenia. Ann Intern
Med, 127, 804–​12.
Warkentin TE, et al. (2015). Warfarin-​induced venous limb ischemia/​
gangrene complicating cancer: a novel and clinically distinct syn-
drome. Blood, 126, 486–​93.

22.8.2 Haemopoietic stem cell transplantation 5579
22.8.2 Haemopoietic stem cell transplantation 5579 E.C. Gordon- Smith and Emma C. Morris
22.8.2  Haemopoietic stem cell transplantation
5579
22.8.2  Haemopoietic stem
cell transplantation
E.C. Gordon-​Smith and Emma C. Morris
ESSENTIALS
Haemopoietic stem cells (HSCs) give rise to the blood cell lineages
and the cells of the immune system, and their transplantation may
be an appropriate part of the management of conditions including
(1) malignant haematological disorders (e.g. leukaemia, lymphoma,
myeloma); (2) bone marrow failure syndromes (e.g. aplastic anaemia);
and (3) congenital disorders—​(a) haematological (e.g. Fanconi an-
aemia); (b) immunological—​inherited immunodeficiency syndromes;
and (c) metabolic (e.g. lysosomal storage diseases).
Transplantation of HSCs uses either autologous HSCs (patient’s
own stem cells) or allogeneic HSCs (harvested from an appropriately
matched sibling or unrelated healthy donor).
Successful engraftment of allogeneic HSCs depends upon
(1) overcoming immune rejection by the recipient; (2) preventing or
suppressing graft-​versus-​host disease (GVHD), in which donor cells
mount an immune attack against recipient tissues; and (3) supporting
the patient through periods of profound cytopenias and immune
deficiency with susceptibility to infection.
Identification and sources of haemopoietic stem cells
HSCs are principally identified by expression of the surface antigen
CD34. Sources include (1) bone marrow; (2) peripheral blood—​fol-
lowing stimulation by cytokines (e.g. granulocyte colony-​stimulating
factor); and (3) umbilical cord blood.
Autologous haemopoietic stem cell transplantation
The rationale behind autologous haemopoietic stem cell transplant-
ation is to facilitate the delivery of higher doses of chemotherapy than
would otherwise be possible. As the patient’s haemopoietic stem
cells are removed prior to transplant, cryopreserved, and stored, they
can be reinfused following myeloablative chemotherapy. As there is
no immunological disparity when the patient’s own stem cells are
re-​infused there is no requirement for immune suppression or risk
of GVHD.
However, as the harvested stem cells may be contaminated
with a low level of malignant cells, there is a risk of relapse. The
relapse risk depends on the underlying disease and response to
pretransplantation chemo/​radiotherapy. Studies attempting to purge
the collected stem cells of contaminating malignant cells have been
performed, but have failed to show any improvements in overall
survival, relapse rate, or disease-​free survival compared to using
unmanipulated stem cells.
Allogeneic haemopoietic stem cell transplantation
Selection of donors
Fully HLA-​matched sibling donors are only available for about
one in three recipients. Minor disparity between the HLA types
of donor and recipient are allowable, but the greater the disparity,
the higher the frequency of complications. For those without such
a donor, possible sources are (1) volunteer donor banks, which
can provide acceptably HLA-​matched donors for about 80% of
recipients with the same genetic disequilibrium as the donor
pool; (2) umbilical cord blood banks; and (3) haploidentical family
donors.
Conditioning regimen
Conditioning regimens include measures to induce immunosup-
pression (required for all allogeneic transplants, excepting those
from identical twin donors) and, when appropriate, eradicate dis-
eased bone marrow. They may be (1)  myeloablative (no autolo-
gous recovery possible)—​cyclophosphamide with (a)  total body
irradiation (TBI), or (b) busulphan, or (c) antilymphocyte globulin;
or (2) nonmyeloablative—​fludarabine with varying combinations of
(a) antilymphocyte globulin or alemtuzumab (anti-​CD52); (b) low-​
dose cyclophosphamide, melphalan, or busulphan; and (c) low-​dose
TBI. After transplantation, donor lymphocyte infusions may be given
to patients with malignant disorders to enhance the graft-​versus-​leu-
kaemia or graft-​versus-​tumour effect.
Graft-​versus-​host disease
Acute GVHD—​this major cause of transplant-​related morbidity and
mortality may develop at any time within the first 6 weeks after trans-
plantation. The clinicopathological manifestations are secondary to
the recognition of alloantigens in recipient tissue, by donor-​derived
T lymphocytes, hence the importance of HLA typing. Manifestations
involve (1)  skin—​maculopapular rash, generalized erythroderma,
desquamation, and bullae; (2)  liver—​severity graded according to
level of serum bilirubin; and (3) gut—​diarrhoea, persistent nausea,
and pain/​ileus.
Chronic GVHD—​may follow acute GVHD or emerge de novo sev-
eral months after transplantation. Mainly affects the skin, but al-
most any organ may be affected, for example, lung (bronchiolitis
obliterans), gut, liver, eyes (sicca syndrome), buccal mucosa, skin
(sclerodermatous changes), and musculoskeletal system. Chronic
GVHD parallels graft-​versus-​leukaemia effect and total suppression
is associated with increased relapse in some leukaemia or lymphoma
transplants.
Prevention/​treatment—​these present a major challenge. Standard
prevention is with ciclosporin, with or without methotrexate and/​or
antilymphocyte globulin. Initial treatment of both acute and chronic
GVHD is with high-​dose steroids.
Other complications and prognosis
Conditioning regimens have considerable toxicity. (1) Acute—​the pa-
tient is particularly vulnerable to infectious complications in a period
of intense neutropenia, usually lasting 2 to 4 weeks after transplant-
ation and lymphopenia, which may last some months. Patients are
managed in isolation facilities, with (as appropriate) prophylactic
measures against fungal infection, herpes simplex, Gram-​negative
bacterial infection, and Pneumocystis jirovecii. Prophylactic regimens
vary according to local experience with resistant organisms and
availability of newer antimicrobials. Broad-​spectrum intravenous
antimicrobial therapy is used to treat fevers empirically. Ganciclovir
is given if routine monitoring shows evidence of cytomegalovirus
reactivation. (2)  Chronic—​common manifestations include retard-
ation of growth (particularly if transplanted in childhood), endocrine
impairment, infertility, intellectual impairment (following cranial
irradiation).
section 22  Haematological disorders
5580
Prognosis—​once the graft is fully established and tolerance is re-
constituted, immunosuppression may be stopped. In the absence
of GVHD, immune suppression can be weaned from as early as
3 months after transplantation. The procedure offers hope to many
patients with life-​threatening marrow failure or malignant disease for
which no other treatment is available, but in the long term recipients
have a reduced life expectancy due to relapse of the underlying dis-
ease, infection, and chronic GVHD. There is a small increased risk of
second malignancies.
Introduction
The idea that haemopoietic stem cells (HSCs) from the bone
marrow could be transferred from a normal individual to a pa-
tient to replace defective bone marrow has a long history. With
the exception of rare instances where marrow was obtained from
an identical twin, such attempts in humans universally failed until
an understanding of the immune processes involved in tolerance
and rejection became available. Much of the pioneering work in
making possible human bone marrow transplantation was carried
out by E. Donnall Thomas and colleagues in the United States of
America, work for which Thomas received the Nobel Prize jointly
in 1990. Experiments on inbred mice had shown that lethally ir-
radiated animals could be rescued by intravenous transfusion of
bone marrow from unirradiated mice and that this protection
was the result of engraftment of the normal marrow in the re-
cipient. Successful engraftment depended upon the donor marrow
being genetically acceptable by the recipient mouse or the re-
cipient mouse being sufficiently immunosuppressed. Engraftment
when there was immunological disparity between the donor and
recipient was followed after a period of 2 weeks or so by a ‘sec-
ondary’ disease in which the recipient mouse failed to thrive and
developed gastrointestinal disorders, liver failure, and skin disease,
leading to poor further development and eventual death from in-
fection. This so-​called runt disease is the murine equivalent of
graft-​versus-​host disease (GVHD) in humans, in which immuno-
competent cells from the immunologically disparate donor mount
an attack against recipient tissues. From these and other experi-
ments in outbred animals it was recognized that transplantation
of bone marrow would carry the special risk of GVHD and that
histocompatibility would be a critical requirement for successful
transplantation. Furthermore, considerable immunosuppression
would be required to achieve engraftment in all transplants except
those from syngeneic (identical twin) donors. However, it was also
recognized that not only the haemopoietic but also the immune
system would be replaced (reconstituted from donor stem cells)
following myeloablation and that the need for immunosuppres-
sion would cease, except for the management of GVHD, once
donor immune reconstitution was complete.
Transplants in dogs demonstrated that total-​body irradiation
and cyclophosphamide were sufficiently immunosuppressive to
permit engraftment, and that GVHD could be controlled to some
extent with methotrexate, if there was not great disparity between
the histocompatibility antigens of donor and recipient. The elu-
cidation of the major histocompatibility complex (MHC) on
chromosome 6 in humans, with the identification of the histocom-
patibility antigens at the A, B, or C (class I) and DR (class II) loci
of the HLA system, finally allowed the identification of appropriate
donors for human transplantation. The paramount importance
of histocompatibility in haemopoietic stem cell transplantation
(HSCT) has been confirmed subsequently by extensive clinical
practice. The first successful transplant from a nonidentical, but
HLA-​compatible, sibling was carried out in 1968 for a patient with
severe combined immune deficiency where the underlying dis-
ease prevented rejection. Further successful allogeneic transplants
from sibling donors using conditioning with total-​body irradiation
and cyclophosphamide carried out in 1969 in Seattle (Washington,
United States of America) by the group led by Thomas. Many
thousands of such transplants have been carried out subsequently,
though the precise indications and timing for transplant (during
the disease course) particularly in malignant disease, are not al-
ways as clear as they might be (Table 22.8.2.1) and the problems
of GVHD, graft failure, and infection remain hazards which con-
tribute to transplant-​related mortality. Allogeneic HSCT must al-
ways be compared with best available alternative therapies. On the
other hand, better support with blood products and antibiotics,
improved tissue typing techniques, and the introduction of less
toxic methods of delivering conditioning to control rejection and
Table 22.8.2.1  Main disorders for which haemopoietic stem cell
transplantation may be appropriatea
Acquired disorders
Haematological
malignancies
Acute leukaemias
Chronic myeloid leukaemia
Non-​Hodgkin lymphoma (including CLL)
Hodgkin lymphoma
Myeloma and other plasma cell dyscrasias
Solid tumours
Although HSCT has been used as adjunct to
chemotherapy in solid tumours, results to date
are universally disappointing
Bone marrow failure
syndromes
Myelodysplastic syndromes
Myeloproliferative neoplasms
Aplastic anaemia
Paroxysmal nocturnal haemoglobinuria
Congenital disorders
Haematological
Fanconi anaemia
β-​thalassaemias (an increasingly important
disorder for considering HSCT)
Diamond–​Blackfan anaemia
Kostmann syndrome
Immunological
Severe combined immune deficiency and other
primary immune deficiencies
Chronic granulomatous disease
Metabolic
Malignant osteopetrosis
Lysosomal storage diseases
CLL, chronic lymphocytic leukaemia.
a Stem cell transplantation may be considered an option according to availability of a
suitable donor, the stage or severity of the disease, and the availability and effectiveness
of other forms of management.
22.8.2  Haemopoietic stem cell transplantation
5581
GVHD, as well as better selection of recipients, have improved out-
comes steadily over the last 40 years.
Histocompatibility complex and haemopoietic
stem cell transplantation
The organization of the MHC on chromosome 6, and its importance
in transplantation, is described in detail in Chapter 4.7. The close-
ness of the relevant genes in the complex means that within families
there is little crossing-​over in germ-​line cells and inheritance more
or less follows the autosomal pattern, so that the chances of a sibling
having the same HLA type as a patient is about one in four. This is
genotypic identity in which not only the HLA types but also many
unidentified sequences in the MHC are identical. At each HLA locus
there are large numbers of possible alleles in humans leading to a
potential of many millions of different histocompatibility profiles.
However, within populations, certain HLA alleles tend to be asso-
ciated and segregate together, ‘genetic disequilibrium’, so that it is
theoretically and practically possible to find phenotypically identical
pairs within an unrelated population.
The identification of phenotypes was originally based upon sero-
logical testing for A, B, and DR antigens. The introduction of mo-
lecular techniques for identifying DNA sequences directly has shown
that there may be a large number of HLA gene products whose cog-
nate protein molecules are assigned to the same phenotype by sero-
logical methods. HLA typing of individuals within populations has
made possible the creation of large volunteer donor panels of individ-
uals prepared to supply HSCs. However, matching the MHC between
unrelated pairs even by molecular techniques gives at best pheno-
typic identity and there are likely to be many fine genetic differences.
Selection of donors by improved typing techniques has reduced the
risks associated with unrelated transplants but restricted the range of
appropriate donors. It has also become clear that there are very wide
variations in the linkage disequilibria at MHC loci between different
populations of the world so that a donor panel of one ethnic type may
have much reduced chances of providing stem cells for another.
Where there is an HLA disparity between donor and recipient,
HSCT is possible, but the incidence of complications rises steadily
as the degree of disparity increases. It is also apparent that the major
antigens of the MHC (HLA class I and II antigens) are not the only
ones important in determining the incidence and severity of GVHD.
GVHD is mediated by donor-​derived CD4+ helper and CD8+ cyto-
toxic T lymphocytes. These T cells recognize allogeneic MHC mol-
ecules and minor histocompatibility antigens (self proteins where
a polymorphism exists between donor and recipient) expressed on
normal recipient tissues. The role of specific major and minor re-
cipient antigens in the pathogenesis of GVHD is only now being
worked out in any detail. As discussed later, the cell-​mediated im-
mune attack on normal tissues causing GVHD is also linked to an
ability to attack abnormal, particularly malignant cells, producing
a graft-​versus-​leukaemia/​lymphoma (GVL) effect. Much effort has
been directed at identifying the donor cells, which mediate the GVL
effect and the antigen-​presenting cells of the recipient, which facili-
tate the GVL effect in order to separate GVL from GVHD, so far with
inconclusive results. It appears that many target antigens of GVL and
GVHD are shared. The problems and benefits of immunological
disparity obviously only apply in the allogeneic transplantation
setting and are absent when autologous stem cells are used to restore
haemopoiesis after intensive chemotherapy.
Haemopoietic stem cells
Stem cells are defined by their ability to proliferate and differentiate
into one or more specific cell lineages and also to maintain the stem
cell pool by self-​renewal. HSCs give rise to the blood cell lineages—​
red cells, granulocytes, and platelets as well as the cells of the im-
mune system. It seems probable that a single stem cell can repopulate
the blood and immune systems of an entire animal. HSCs can be
identified by immunophenotyping and their ability to repopulate
marrow. The best in vitro techniques have suggested that the human
HSC is closely related to precursors that carry an antigen designated
CD34, lack other haemopoietic markers including CD33, and have
no lineage-​specific markers. Whether such cells are truly the most
primitive cells that are capable of giving rise to both haemopoietic
and immunological precursors is not of practical importance since
successful haemopoietic reconstitution, both in allogeneic and au-
tologous transplants, is closely related to the number of such cells
present in the donation. The CD34+CD33–​ cells represent some
1 × 10–​3 to 10–​4 of the cells of normal human haemopoietic marrow.
Sources of haemopoietic stem cells
HSCs develop in specific sites within the bone marrow, designated
niches, which include specialized cells of the bone marrow stroma.
The stem cell is not fixed in its environment, and may leave the
marrow, enter the circulation, and home once again to the marrow
or to other sites depending on the chemokine and cytokine signals
it receives (see Chapter 22.2.1). Small numbers of CD34+ HSCs cir-
culate in normal blood, and this number is greatly increased during
the marrow recovery after cytotoxic chemotherapy. Administration of
certain cytokines, particularly granulocyte colony-​stimulating factor
(G-​CSF), increases the number of circulating CD34+ cells enor-
mously, such that for a period of a few days following treatment there
are adequate numbers in the circulation to use as a source of cells for
transplantation. Homing and mobilization, and recirculation of HSCs
is a continuous, dynamic process—​even under normal conditions.
In the early development of the fetus, haemopoiesis takes place in
the liver. Fetal liver cells have been used as a source of HSCs, mainly
for the treatment of inherited severe combined immune deficiency.
The logistics of such transplants, which require 11-​week-​old fetal
livers, make this an impractical approach. However, research on
embryonic stem cells suggests that there may be other important
sources of stem cells, not only for haemopoiesis but for other types
of tissue replacement. Of more immediate practical importance was
the finding that umbilical cord blood (UCB) contained large num-
bers of cells with high proliferative potential and characteristics of
stem cells. UCB has become a third practical source of donor cells.
Each of these sources—​bone marrow, peripheral blood, and cord
blood—​has advantages and disadvantages that impinge on clinical
management. A critical requirement for successful transplantation
is that there should be a sufficient number of stem cells. The ability to
expand stem cells ex vivo would solve this and other requirements,
but so far this has not proved to be useful clinically.
section 22  Haematological disorders
5582
HSCs from bone marrow
Until about 1993, most transplants were conducted using bone
marrow stem cells, but peripheral blood mobilized HSCs are now the
preferred option. Much of the data concerning the success and prob-
lems of stem cell transplantation are derived from the use of bone
marrow. Bone marrow is harvested with the patient/​donor under
general anaesthetic by aspiration from the posterior and superior
iliac crests, and if necessary the sternum. Experience has shown that
some 3 × 108 nucleated cells/​kg recipient body weight are required
for successful engraftment and this usually involves collecting 1 to
1.5 litres of bone marrow (mixed, of course, with blood). Donors
may have a unit of blood collected before harvesting, which is re-
turned at the end of the procedure to ameliorate the anaemia. The
procedure takes approximately 1 h and the donor usually requires
brief admission to hospital to recover. Serious complications are ex-
tremely rare and are those associated with the general anaesthetic or
local complications such as osteomyelitis or abscess formation. The
advantage of this source of stem cells from the donors’ viewpoint is
that collection is rapid, with a maximum of 48 h involvement. The
disadvantage is the need to be admitted to hospital for a general an-
aesthetic and the pain or discomfort and anaemia that follow the
procedure.
HSCs from peripheral blood
HSCs may be mobilized into the peripheral blood following ex-
posure to G-​CSF. For allogeneic transplantation, donors receive
G-​CSF (filgrastim or lenograstim) at a dose of 10 µg/​kg subcuta-
neously daily for 5 days. The peripheral granulocyte count rises to
30 × 109/​litre or higher and CD34+ cells appear in the peripheral
blood reaching a maximum 5 to 6 days after the start of treatment.
Leucocytes are collected by leucapheresis with the objective of
reaching more than 2 × 106 CD34+ cells/​kg body weight of the re-
cipient. Sufficient cells can usually be collected from normal donors
in one procedure, although poor mobilizers are found in the normal
population and are not infrequent when autologous stem cell collec-
tion from patients is required following chemotherapy. Plerixafor,
an inhibitor of the CXCL12/​CXCR4 axis, which retains stem cells
in the haemopoietic niche, is an important adjunct to G-​CSF in mo-
bilization and has the advantage of a more rapid effect. The main
disadvantage for donors of this type of stem cell collection is that
of bone pain or ache following the injections of G-​CSF and the
procedure of leucapheresis. Rare instances of splenic rupture have
been recorded. There has been some concern about the theoretical
potential of G-​CSF to cause cytogenetic abnormalities that could
lead to leukaemia; although no such effect has yet been found in
normal donors given the cytokine, long-​term follow-​up is no longer
mandated by donor registries. The main advantages are the avoid-
ance of admission to hospital and a general anaesthetic. When au-
tologous collection of stem cells is required, the concentration of
CD34+ cells may be increased further by giving cyclophosphamide
(or some other chemotherapeutic agents, such as etoposide) before
starting the G-​CSF. Autologous HSCT is used mainly for patients
with malignant disease, for marrow rescue following further inten-
sive chemotherapy.
The use of peripheral blood for harvesting stem cells for allogeneic
transplants provides high numbers of CD34+ cells and more rapid
engraftment than with bone marrow-​derived stem cells. On the
other hand, peripheral blood contains more T cells and, although
original concerns that acute GVHD (aGVHD) would be unaccept-
ably severe unless T cells were removed have proved unfounded,
chronic GVHD (cGVHD), especially extensive disease, is, more
prevalent than in bone marrow and UCB transplants and is the main
nonrelapse cause of mortality of peripheral blood stem cell trans-
plants. Nevertheless, the ease of collection and advantages of rapid
engraftment have meant that most autologous transplants (where
GVHD is not a problem), and the majority of allogeneic transplants,
are currently sourced from the peripheral blood.
HSCs from umbilical cord blood
Sourcing HSCs from UCB has several theoretical and practical
advantages. UCB is widely available with no risk to mother or in-
fant donor, there is low viral contamination, the immaturity of the
immune cells may theoretically reduce the risk of GVHD, and the
cells may readily be stored frozen and made available rapidly for
urgent transplantation. Furthermore, a balance of UCB stem cells
from different ethnic groups to take advantage of genetic disequi-
librium can be achieved and specific HLA types can be targeted.
A disadvantage is the relatively small volume of UCB and hence low
total numbers of HSCs. UCB donations have mostly been used to
transplant children. Recently the use of double or multiple UCB do-
nations to increase the infused HSC dose has been introduced for
adult transplants; further follow-​up is needed to assess the results,
though there is early evidence that the risk of cGVHD is less than
with bone marrow or peripheral blood donations. The storage and
administrative costs of UCB are high compared with sourcing from
unrelated donors at the time of transplant. A further difficulty is the
lack of any back-​up source of cells should the graft fail or disease
relapse occur. For adults with impaired thymic function the lack of
memory T lymphocytes specific for common latent viruses such as
cytomegalovirus (CMV) in the donation may result in repeated viral
reactivation. Nevertheless, UCB transplants are increasing as a prac-
tical source of HSCs in the future.
Plasticity of stem cells
The bone marrow and other tissues contain totipotent stem cells.
These bone marrow-​derived cells may differentiate to cardiac
muscle cells, nerve cells, striated muscle fibres, and many other tis-
sues, whether they are ectodermal, mesodermal, or endodermal in
origin. This potential is also present in embryonic stem cells. HSCs
have been used in phase I/​II studies examining the effect on cardiac
function after myocardial function and other insults. Although it
is possible to identify HSCs in the myocardium after such proced-
ures, practical benefits have not yet been clearly revealed. A second
class of bone marrow cells which have stem cell properties in vitro
and may be stem cells in vivo are the mesenchymal stromal (stem)
cells. In vitro they can differentiate into a variety of mesodermal tis-
sues. Their potential interest for HSC transplantations lies in their
immunomodulatory effects. Early clinical trials suggest they may be
able to modify aGVHD.
Donors for allogeneic stem cell transplantation
Problems of transplant-​related morbidity and mortality, graft rejec-
tion, GVHD, and infection increase with increasing donor disparity.
22.8.2  Haemopoietic stem cell transplantation
5583
HLA-​matched sibling donors are not only phenotypically matched
for the MHC, but have genotypic identity throughout most of the
MHC. This does not eliminate transplant-​related morbidity and
mortality, but reduces the incidence and severity of the problems
compared with unrelated volunteer donors matched only pheno-
typically for the MHC and mismatched at minor histocompatibility
antigens. HLA-​matched sibling donors are only available for about
one in three recipients in populations with an average of two or three
children per family. To overcome this shortfall, volunteer donor
banks have been established, now including more than 10 million
typed donors worldwide. These panels can provide HLA-​suitable
matches for about 80% of recipients with the same genetic disequi-
librium as the donor pool, though finding the right match may take
several weeks. Extensive immunosuppression of the recipient is re-
quired prior to transplantation to prevent graft rejection and after
transplantation to control GVHD. New methods of immunosup-
pression which allow the stepwise engraftment of donor marrow
may produce a greater degree of tolerance and permit successful
transplantation of HSCs with some degree of HLA disparity. UCB
banks have been established in many countries. Their use is likely
to increase if the efficacy of double or multiple donations remain ef-
fective for adults after longer-​term follow-​up. An advantage of using
haploidentical donors (typically parents or a child) is their ready
availability.
Management of transplant recipients
Conditioning regimen
The treatment of recipients prior to transplantation includes meas-
ures to induce immunosuppression and eradicate diseased bone
marrow. In HSC transplantation for malignant disease, most proto-
cols to date have contained cyclophosphamide combined either with
total-​body irradiation, single dose or fractionated, or with alkylating
agents such as busulphan. For nonmalignant conditions irradiation
should be avoided. For acquired aplastic anaemia, cyclophospha-
mide, either alone or combined with antithymocyte globulin, has
been the major conditioning regimen. Some of the more widely
used regimens are indicated in Table 22.8.2.2. The incidence and
severity of GVHD was reduced by giving methotrexate in a short
protocol after transplantation; the introduction of ciclosporin fur-
ther improved results. Such conditioning regimens, particularly
for malignant and genetic disorders, carry delayed as well as acute
toxicity, particularly for children. Where radiation is used (and to a
lesser extent busulphan), infertility is usual, growth is retarded, and
other endocrine functions may be impaired. Where transplantation
is used for patients who have already received irradiation or chemo-
therapy to the central nervous system, for example, patients with a
relapsed acute lymphoblastic leukaemia, intellectual impairment as
well as the earlier-​mentioned problems are common.
Success of stem cell transplantation in certain malignant condi-
tions is related to the cellular immune response to recipient cells and
tissue (GVL), provided by donor lymphocytes, rather than the direct
cytotoxic effect of conditioning. Repopulation of marrow by donor
HSCs does not require the immediate abolition of recipient marrow.
Full donor chimerism may be achieved over time rather than in one
step, and in some instances mixed stable chimerism of both donor
and recipient cells may occur. Conditioning regimens have been
introduced which do not rely on cytotoxic measures to obliterate
recipient marrow and immune system, but which have increased
immunosuppressive action. Such regimens include fludarabine,
a highly immunosuppressive drug that is not very cytotoxic, often
combined with antithymocyte globulin, monoclonal antibodies
(e.g. alemtuzumab, anti-​CD52), and/​or low-​dose total-​body irradi-
ation. Removal of T lymphocytes from the donor preparation (T-​
cell depletion), to reduce aGVHD with subsequent later add-​back
of donor lymphocytes to reduce the chances of relapse, is also used
(Table 22.8.2.2). Results using this approach have been encouraging
and reduced-​intensity transplantation regimens are widely used
for lymphomas and leukaemias, though long-​term follow-​up is re-
quired. The reduced toxicity has allowed the use of HSC transplant-
ations for patients up to 60 years of age or even older.
Recent modifications to conditioning regimens for recipients
of haploidentical donor HSC transplantations have demonstrated
that crossing the HLA barrier without unacceptably high inci-
dences of graft rejection, severe GVHD, and transplant-​related
mortality is achievable. This has included the use of high-​dose post-​
transplantation cyclophosphamide to selectively deplete in vivo
donor-​derived alloreactive T cells during the phase of maximal pro-
liferation in order to induce immune tolerance.
Graft-​versus-​host disease
GVHD is the consequence of immune attack on recipient tissues
by donor lymphocytes. Early experiments on mice showed that the
Table 22.8.2.2  Outline of examples of conditioning regimens
for allogeneic haemopoietic stem cell transplantation
Conditioning regimen
Indications
Myeloablative
Cyclophosphamide at 120 mg/​kg + TBI of
750–​1400 cGy
Acute leukaemia
Chronic myeloid leukaemia
Relapsed lymphoma
Cyclophosphamide at 120 mg/​kg +
busulphan at 16 mg/​kg
As above
β-​thalassaemia major
Other congenital bone marrow
disorders
Cyclophosphamide at 200 mg/​kg ± ALG
Acquired aplastic anaemia
Cyclophosphamide at 25–​100 mg/​kg + TBI
of 200 cGy
Fanconi anaemia
Reduced-​intensity conditioning
Fludarabine at 30 mg/​m2
Fanconi anaemia
± ALG or alemtuzumab
Congenital disorders of
haemopoiesis or immune
system
low-​dose cyclophosphamide or
melphalan or busulphan
± low-​dose TBI (200 cGy)
Acquired aplastic anaemia
With DLIa or virus-​specific T cellsb
Malignant disorders
ALG, antilymphocyte globulin; DLI, donor lymphocyte infusions; TBI, total body
irradiation.
Lower doses given in single fraction, higher doses fractionated.
a Given 3 months or longer post infusion to provide GVL or graft-​versus-​tumour effect.
b Current phase II and III clinical studies are exploring the role of virus-​specific T-​cell
infusions in the management of viral reactivation.
section 22  Haematological disorders
5584
murine equivalent (runt disease) developed when immune com-
petent donor lymphocytes were given to immunosuppressed, im-
munologically disparate recipients. Skin, gut, and liver were the
main organs affected. In human HSCT, mismatch in the MHC be-
tween donor and recipient correlates with the severity of GVHD.
However, the pathogenesis of GVHD involves more than the im-
munological disparity; inflammation, infection, tissue damage,
and the gut microbiota play a part in determining the severity of
the disease and which organs are affected. The complexity of the
pathogenesis has hindered a clear understanding of all the path-
ways involved and hence fully effective prevention and treatment.
Historically, GVHD was divided into aGVHD which appeared
within the first 100 days and cGVHD with a later onset. It is now
accepted that both may coexist and that there is overlap in the
pathogenesis. From a practical point of view, the distinction re-
mains a useful concept.
Acute GVHD
Originally defined aGVHD, manifest by various degrees of skin,
gut, and liver damage (Fig. 22.8.2.1), occurs within the first
3 months or so of the HSCT. For each organ, a score of 0 to 4+
was given, depending on the degree of damage and an overall
aGVHD score of grade 0 to IV devised to record overall severity
(Table 22.8.2.3). Some two-​thirds of patients transplanted from
HLA-​matched sibling or volunteer donors will experience some de-
gree of GVHD—​grade III to IV carrying a high risk of transplant-​
related mortality. Increasing donor–​recipient HLA mismatch is
associated with higher grades of GVHD but patient age, previous
chemotherapy, irradiation, and infections also increase the risk.
The basic classification of aGVHD has proven useful in comparing
the outcome of various studies of HSCT conditioning in different
diseases, but has limitations for understanding the pathogenesis.
The National Institutes of Health consensus criteria include ‘late-​
onset acute GVHD’ an overlap syndrome with cGVHD. A number
of steps are involved in the pathogenesis of GVHD. Animal models
suggest initiation involves activation of antigen-​presenting cells re-
lated to underlying disease, chemotherapy, conditioning regimen,
and radiation which may release proinflammatory cytokines such
as tumour necrosis factor (TNF)-​α. This may lead to an interaction
with donor T lymphocytes within lymphoid tissues such as Peyer’s
(a)
(c)
(e)
(b)
(d)
Fig. 22.8.2.1  Skin manifestations of acute and chronic GVHD. Acute GVHD: (a) grade I, skin +,
showing typical palmar maculopapular rash (recovered); (b) grade IV, skin 4+, generalized
erythroderma with early exfoliation; liver 3+, bilirubin greater than 250 µmol/​litre (fatal);
(c) grade III, skin 4+, bullous desquamation (recovered). Chronic GVHD: (d) sclerotic scarring
on back; (e) severe ulceration and contracting scleroderma-​like skin involvement.
22.8.2  Haemopoietic stem cell transplantation
5585
patches in the gastrointestinal tract. The reduction of tissue damage
in reduced-​intensity conditioning regimens may in part explain
the reduction in GVHD severity. Tissue damage may also explain
the activation of GVHD by virus infections such as CMV. The next
and fundamental stage is the activation, differentiation, and pro-
liferation of the donor T cells, principally CD4+ and CD8+ cells.
A variety of cell types including regulatory T cells, natural killer
cells, and mesenchymal stromal cells seem to have modifying roles
in aGVHD in ways that are not well understood but which offer
tantalizing clues to improved treatment. The final stage of aGVHD
is due to cytotoxic T-​cell-​mediated attack involving perforin and
granzyme pathways as well as other cytokine-​mediated inflamma-
tory pathways.
Chronic GVHD
cGVHD is usually defined as GVHD developing 100 days or more
after transplant. Following HSCT, some 40 to 70% of patients de-
velop some cGVHD. The pathogenesis is even more complex than
that of aGVHD. While skin, liver, and gastrointestinal tract remain
principal targets (Fig. 22.8.2.1), other targets such as lung, eyes
(Sicca syndrome), mucous membranes, and heart (pericarditis)
may be involved. Auto-​ and alloantibodies are common in cGVHD
though their pathogenetic role is not clear. However, it is likely that
B cells play a critical role in cGVHD. There is usually a degree of
induced immunodeficiency and susceptibility to infection. cGVHD
may be transient, persistent, or progressive. Diagnostic criteria for
cGVHD have been developed by the National Institutes of Health
consensus forum, based on extent of the process and severity of
damage. As with aGVHD a viral infection may trigger a relapse. In
HSCT for malignant disease, complete absence of cGVHD is associ-
ated with an increased risk of relapse, indicating the close relation-
ship between GVHD and GVL effect.
Prevention of GVHD
Prophylaxis has mainly relied on the use of ciclosporin with or
without methotrexate. Complete T-​cell depletion of the donor graft
may prevent GVHD but the increased incidence of graft failure or
malignant relapse means there is no overall survival advantage. The
polyclonal antithymocyte globulins have been used in addition to
ciclosporin after transplantation without clear-​cut benefit. Clearly
selecting the most appropriate donor possible to ameliorate GVHD
is an essential part of the prevention of GVHD.
Treatment of GVHD
The first approach to management of grade II to IV GVHD is high-​
dose corticosteroids, usually 1 to 2 mg prednisolone/​kg body weight,
together with ciclosporin or another calcineurin inhibitor. Topical
steroids may be effective in grade I to II skin GVHD. The difficul-
ties arise in steroid-​unresponsive or relapsing GVHD. About half
the patients so treated will achieve complete or partial responses
and failure to respond correlates with a poor prognosis. If there is
no response within 5 to 15 days or if there is early deterioration
on steroids, second-​line treatment needs to be started. The large
number of agents considered indicates the difficulty of managing
steroid-​resistant aGVHD. Antilymphocyte globulin may reduce
the severity but has not improved overall survival. Alemtuzumab
(Campath; anti-​CD52+ monoclonal antibody) may be effective in
some cases but carries a high risk of infection as it depletes both
T and B cells. Anti-​interleukin-​2 receptor antibodies (daclizumab,
basiliximab) may reduce aGVHD but most patients go on to develop
severe cGVHD. Anti-​TNFα agents (etanercept, infliximab) may also
produce responses in some cases but infections and failure in grade
III to IV aGVHD limit usefulness. For treatment of cGVHD, pred-
nisolone is usually started at a lower dose, 1 mg/​kg body weight,
but may need to be continued for months or years. Some steroid-​
sparing effect may be achieved by adding ciclosporin but this has
no significant overall effect on morbidity or mortality. Rituximab
(monoclonal chimeric anti-​CD20 antibody) produces mainly par-
tial responses in most patients with skin cGVHD or musculoskel-
etal problems and imatinib may reduce fibrosis. Extracorporeal
photopheresis has been used with success in patients with lung, gut,
and sclerodermatous cGVHD, although its cost and availability is
Table 22.8.2.3  Clinical staging of acute GVHD
Clinical stage
Organ involvementa
Skin
Liver
Gut
+
Maculopapular rash <25% of
body surface
Bilirubin 30–​50 µmol/​litre (2–​3 mg/​dl)
Diarrhoea 0.5–​1 litre/​day and/​or
persistent nausea
++
Maculopapular rash 25–​50% of
body surface
Bilirubin 50–​100 µmol/​litre (3–​6 mg/​dl)
Diarrhoea 1–​1.5 litre/​day
+++
Generalized erythroderma
Bilirubin 100–​250 µmol/​litre (6–​15 mg/​dl)
Diarrhoea >1.5 litre/​day
++++
Desquamation and bullae
Bilirubin > 250 µmol/​litre (>15 mg/​dl)
Pain ± ileus
Clinical grade
Stage
Skin
Liver
Gut
Functional
impairment
0 (none)
0
0
0
0
I (mild)
to 2+
0
0
0
II (moderate)
to 3+
III (severe)
2+ to 3+
2+ to 3+
2+ to 3+
2+
IV (life-​threatening)
2+ to 4+
2+ to 4+
2+ to 4+
3+
a Confirmation may require biopsy.
section 22  Haematological disorders
5586
limiting. There is a need for randomized controlled trials to properly
assess these second-​line treatments to identify the most effective.
Immune reconstitution and infection in HSCT
The immunosuppression of the conditioning, the cytotoxicity of
the agents used, and above all the severe depression of the immune
response caused by GVHD, both acute and chronic, all contribute
to the prolonged deficiency of cellular and humoral immunity and
hence the high risk of infection in the post-​transplantation period
(Fig. 22.8.2.2). The duration may be exacerbated by prior chemo-
therapy, T-​cell depletion, and subsequent immunosuppressive
treatment of GVHD. Agranulocytosis develops as soon as the condi-
tioning regimen is started and continues until the graft is established
and neutrophils return, usually after 2 to 4 weeks. During this phase,
bacterial and fungal infections are common. Patients are managed
in isolation facilities, preferably with laminar airflow to remove
environmental pathogens, particularly aspergillus. Prophylactic
antimicrobials including antifungals (e.g. fluconazole 100–​400 mg/​
day), and antivirals such as aciclovir 200 to 400 mg four times a day
to prevent herpes simplex reactivation are used to cover the neutro-
penic and lymphopenic phase. Antibacterials are occasionally used
prophylactically. The use of ciprofloxacin and other broad-​spectrum
antibiotics increases the risk of Clostridium difficile infection and
pseudomembranous colitis. Gram-​positive infection, particularly
by Staphylococcus epidermidis, is common because of the use of
indwelling catheters. Pneumocystis jirovecii (previously carinii) is a
major hazard, especially when steroid treatment is given for GVHD
and/​or CD4+ T-​cell count is low. Co-​trimoxazole 480 mg twice
daily three times a week should be given until the risk of cGVHD
has passed. CMV pneumonitis following CMV reactivation was a
major cause of transplant-​related mortality before the introduc-
tion of effective antiviral agents such as ganciclovir. Ganciclovir is
myelosuppressive so its use following early evidence of reactivation
rather than automatic prophylaxis is preferred when detection of
CMV by PCR or CMV antigenaemia are available. Newer antiviral
agents continue to reduce the risk of CMV disease after reactiva-
tion, together with novel cellular therapies including the adoptive
transfer of CMV-​specific T cells. Patients with cGVHD have im-
mune deficiency similar to splenectomized patients and should re-
ceive penicillin V 250 mg twice a day (or erythromycin if allergic to
penicillin) for lifelong prophylaxis against encapsulated organisms,
Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria
meningitides. A revaccination programme should be instituted once
cellular and humoral immunity are reconstituted, usually 1 to 2 years
after the transplant.
Blood transfusions
The intense immunosuppression of conditioning produces a risk of
engraftment by stem cells present in transfusion products. Where
transfusion is necessary, the products must be irradiated to prevent
proliferation of cells.
Long-​term follow-​up
Life expectancy of recipients alive 2 years after transplant is reduced
compared to matched controls, main causes of late death being
relapse, cGVHD, infection, and increased risk of solid tumours.
GVHD also increases the risk of later cardiovascular-​related events
(hypertension, diabetes, and dyslipidaemia). Immune suppression
may be stopped once the graft is fully established and tolerance
complete (usually 6 months to 2 years after transplant) but lifelong
follow-​up is still required.
Management of relapse
Patients transplanted for leukaemia, particularly chronic myeloid
leukaemia, who develop aGVHD and/​or cGVHD have less relapse,
though not better survival, than patients without GVHD. Relapse
may be effectively managed by giving donor lymphocyte infusions
though this increases the risk of cGVHD. There is a hierarchy of
the GVL effect: chronic myeloid leukaemia being the most respon-
sive to GVL, some effect in acute myeloid leukaemia, less in acute
lymphoblastic leukaemia, and variable in lymphomas and myeloma.
Indolent lymphoma and Hodgkin lymphoma are more responsive
Viral
Fungal
Bacterial
Gram-negative
Gram-positive
HSV
Candida   Aspergillus
Encapsulated
Days
0
50
100
150
Infections
Risk factors
Cellular and humoral immune deficiency
VZV/Late CMV
Adeno/RSV/CMV
Central lines
Acute GVHD
Neutropenia
Chronic GVHD
Fig. 22.8.2.2  Risk factors and timing of high-​risk infections following HSCT.
22.8.2  Haemopoietic stem cell transplantation
5587
to GVL. Donor lymphocyte infusions now form part of the man-
agement plan after transplantation for relapse in reduced-​intensity
transplantations when T-​cell depletion is used.
Indications for haemopoietic stem
cell transplantation
The main indications are shown in Table 22.8.2.1. They fall broadly
into two groups. In the first, donor stem cells are used for replace-
ment therapy—​a rather crude form of gene therapy—​for inherited
disorders. In the second group, donor stem cells are used in ma-
lignant disease as an adjunct to chemotherapy, both by providing
rescue from intensive cytotoxic chemo/​radiotherapy and through
the GVL effect. It is in this group that uncertainties remain as to the
most appropriate timing as well as effectiveness of allogeneic trans-
plantation. Randomized controlled trials have proved difficult
to complete and much of the evidence is based on registry data,
single-​centre studies, or historical controls. At the same time that
the results of HSCT have improved, the results of chemotherapy
and newer agents have also become better. Nevertheless, particu-
larly in children and younger adults, allogeneic transplantation is
widely used with significant success, particularly for relapsed con-
ditions where conservative management offers no potential for
cure. There is a very marked inverse relationship between success
of transplantation and age, children having much less transplant-​
related morbidity and mortality due to reduction in infectious
complications and GVHD. Children also tolerate a higher degree
of HLA mismatching than adults. The upper age limit for allo-
geneic transplant has continued to rise as results improve and the
use of reduced-​intensity conditioning regimens has become wide-
spread, with some centres routinely transplanting patients up to
the age of 65 to 70 years. However, the transplant-​related mor-
bidity and mortality at this age may be very marked. As would be
expected, results of allogeneic transplantation are best in low-​risk
groups, in first complete remission or with chemosensitive disease,
and are worst in relapsed and resistant disease. However, it was in
this last group that the potential benefits of allogeneic transplant-
ation were first clearly demonstrated by Thomas and his group in
Seattle. In most protocols for the management of leukaemias, the
inclusion of allogeneic transplantation, where a suitable sibling
donor is available, is considered either up-​front or as a form of
rescue in younger patients. The results of fully matched unrelated
donor transplantations continue to improve and are now closer to
those achieved with matched sibling donors. Recent advances in
developing tailored conditioning regimens and GVHD prophy-
laxis for haploidentical transplants has seen an increase in the use
of HLA antigen-​mismatched stem cells for patients where no other
options are available.
Indications for autologous transplantation
The use of autologous HSCs for treatment of malignant disease can
only be considered a form of rescue from increased chemotherapy
since there is no GVL effect. Where there may be tumour antigens
that are amenable to immune therapy, attempts have been made
to induce specific immunotoxicity, so far without clear-​cut benefit.
On the other hand, autologous stem cell rescue does allow greatly in-
creased intensity chemotherapy regimens for lymphoma, myeloma,
and a variety of solid tumours with shortening of hospital stay—​
indeed, in some cases treatment can be managed in an outpatient
setting—​and a prolonged course of therapy with repeated rescue
from stored cells. Autologous stem cells will also provide the vehicle
for gene therapy once techniques for gene insertion and long-​term
expression become practical.
Cellular therapies
Over the last 10 years, the development of antigen-​specific T-​cell
therapies has moved from research laboratories into the clinic and
there is now increasing evidence from early-​phase clinical trials for
their efficacy in treating chemotherapy-​resistant haematological
malignancies. Autologous tumour antigen-​specific T cells can be
isolated from tumour biopsies and expanded ex vivo (tumour infil-
trating lymphocytes) or stimulated ex vivo in the presence of the
relevant tumour antigen (tumour reactive autologous T cells) prior
to reinfusion into the patient following lymphodepleting condi-
tioning (to enhance homeostatic expansion of the infused T cells).
As the reinfused T cells are autologous there is no risk of rejection,
but their persistence and in vivo antitumour effects are determined
by their fitness and ability to survive in the nutrient-​poor and im-
munosuppressive tumour microenvironment. They may also be
subject to immunological tolerance mechanisms. More promising
are gene-​modified immune cells, typically CD8+ and/​or CD4+
T cells which have been genetically engineered to express antigen
receptors—​either a physiological T cell receptor (TCR) or a syn-
thetic chimeric antigen receptor (CAR), which is based on an anti-
body fragment and contains an engineered intracellular signalling
domain. The introduced receptors confer antigen specificity and as
such redirect patient’s own T cells to MHC-​presented peptide (rec-
ognized by TCRs) or cell surface antigens (recognized by CARs) on
the target malignant cells. Clinical-​grade retroviral and/​or lentiviral
vectors are used to introduce the antigen receptors and redirect spe-
cificity. Similar technology is now being used to enhance immune
cell function further by altering homing characteristics, differenti-
ation status, and susceptibility to immune suppressive agents.
Future directions for haemopoietic stem
cell transplantation
•	Better understanding and treatment of GVHD, particularly
cGVHD.
•	Wider use of alternative donors, such as UCB donations and
haploidentical donors.
•	Further clinical trials to establish best practice, conditioning
regimen, and indications.
•	Harnessing the GVL effect and preventing exhaustion of donor-​
derived T cells.
•	Development of cellular therapies after transplantation including
gene-​modified immune cells.
•	Continued exploration of genetic modification of HSCs for treat-
ment of inherited disorders.
section 22  Haematological disorders
5588
FURTHER READING
Attar EC (2012). Get out—​and stay out (Editorial on plerixafor).
Blood, 119, 3869–​70.
Bensinger WI (2012). Allogeneic transplantation: peripheral blood vs.
bone marrow. Curr Opin Oncol, 24, 191–​6.
Blazar BR, Murphy WJ, Abedi M (2012). Advances in graft-​vs.-​host
disease biology and therapy. Nature Rev Immunol, 12, 191–​6.
Bonini C, Mondino A (2015). Adoptive T-​cell therapy for cancer: the
era of engineered T cells. Eur J Immunol, 45, 2457–​69.
Bosch M, Khan FM, Storek J (2012). Immune reconstitution after
hematopoietic cell transplantation. Curr Opin Hematol, 19, 324–​35.
Brunstein CG, Setubal DC, Wagner JE (2007). Expanding the role of
umbilical cord blood transplantation. Br J Haematol, 137, 20–​35.
Chien JW, et  al. (2012). Evaluation of published single nucleotide
polymorphisms associated with acute GVHD. Blood, 119, 5311–​19.
Craddock C, Chakreverty R (2005). Stem cell transplantation.
In: Hoffbrand AV, Catovsky D, Tuddenham EGD (eds) Postgraduate
haematology, pp. 419–​35. Blackwell Publishing, Oxford.
Deeg HJ (2007). How I treat refractory GVHD. Blood, 109, 4119–​26.
Hough R, Cooper N, Veys P (2009). Allogeneic haemopoietic stem cell
transplantation in children: what alternative donor should we choose
when no matched sibling is available? Br J Haematol, 147, 593–​613.
June CH, Riddell SR, Schumacher TN (2015). Adoptive cellular
therapy: a race to the finish line. Sci Transl Med, 7, 280ps7.
Kanakry CG, Fuchs EJ, Luznik L (2016). Modern approaches to
HLA-​haploidentical blood or marrow transplantation. Nat Rev
Clin Oncol, 13, 10–​24.
Kennedy-​Nasser AA, Bollard CM. (2007). T cell therapies following
haematopoietic stem cell transplantation:  surely there must be a
better way than DLI? Bone Marrow Transplant, 40, 93–​104.
Kolb H, et al. (1995). Graft-​versus-​leukemia effect of donor lympho-
cyte transfusions in marrow grafted patients. European Group for
Blood and Marrow Transplantation Working Party for Chronic
Myeloid Leukemia. Blood, 86, 2041–​50.
Körbling M, Anderlini P (2001). Peripheral blood stem cell versus
bone marrow allotransplantation: does the source of hematopoietic
stem cells matter? Blood, 98, 2900–​8.
Lee SJ (2007). New approaches for preventing and treating chronic
graft-​versus-​host disease. Blood, 105, 4200–​6.
Nauta AJ, Fibbe WE (2007). Immunomodulatory properties of mesen-
chymal stem cells. Blood, 110, 3499–​506.
Peggs KS, et al. (2005). Clinical evidence of a graft-​versus-​Hodgkin’s-​
lymphoma effect after reduced-​intensity allogeneic transplantation.
Lancet, 365, 1934–​41.
Rizzo JD, et al. (2006). Recommended screening and preventive prac-
tices for long-​term survivors after hematopoietic cell transplant-
ation:  joint recommendations of the European Group for Blood
and Marrow Transplantation, Center for International Blood and
Marrow Transplant Research, and the American Society for Blood
and Marrow Transplantation (EBMT/​CIBMTR/​ASBMT). Bone
Marrow Transplant, 37, 249–​61.
Rocha V, et al. (2004). Transplants of umbilical-​cord blood or bone
marrow from unrelated donors in adults with acute leukaemia. N
Engl J Med, 351, 2276–​85.
Schmitt TM, et al. (2015). New strategies in engineering T-​cell receptor
gene-​modified t cells to more effectively target malignancies. Clin
Cancer Res, 21, 5191–​7.
Slavin S, et al. (1998). Nonmyeloablative stem cell transplantation and
cell therapy as an alternative to conventional bone marrow trans-
plantation with lethal cytoreduction for the treatment of malignant
and nonmalignant hematologic diseases. Blood, 97, 56–​63.
Socié G, et al. (1999). Long term survival and late deaths after allo-
geneic marrow transplantation. Late effects working committee of
the International Bone Marrow Transplant Registry. N Engl J Med,
341, 14–​21.
Socié G, et al. (2003). Non-​malignant late effects after allogeneic stem
cell transplantation. Blood, 101, 3373–​85.
Stenger EO, et  al. (2012). Dendritic cells and regulation of
graft-​versus-​host disease and graft-​versus-​leukemia activity. Blood,
119, 5088–​103.
Theurich S, et al. (2012). Polyclonal anti-​thymocyte globulins for the
prophylaxis of graft-​versus-​host disease after allogeneic stem cell or
bone marrow transplantation in adults. Cochrane Database Syst Rev,
9, CD009159.
Thomas ED (2005). Bone marrow transplantation from the personal
viewpoint. Int J Hematol, 81, 89–​93.
Thomson KJ, Potter M, Mackinnon S (2005). Non-​myeloablative
transplantation. In: Hoffbrand AV, Catovsky D, Tuddenham EGD
(eds) Postgraduate haematology, pp. 436–​48. Blackwell Publishing,
Oxford.
Thomson KJ, et al. (2009). Favorable long-​term survival after reduced-​
intensity allogeneic transplantation for multiple-​relapse aggressive
non-​Hodgkin’s lymphoma. J Clin Oncol, 27, 426–​32.
SECTION 23
Disorders of the skin
Section editor: Roderick J. Hay
23.1	 Structure and function of skin   5591
John A. McGrath
23.2	 Clinical approach to the diagnosis of skin
disease   5596
Vanessa Venning
23.3	 Inherited skin disease   5602
Thiviyani Maruthappu and David P. Kelsell
23.4	 Autoimmune bullous diseases   5612
Kathy Taghipour and Fenella Wojnarowska
23.5	 Papulosquamous disease   5621
Christopher E.M. Griffiths
23.6	 Dermatitis/​eczema   5630
Peter S. Friedmann, Michael J. Arden-​Jones, and
Roderick J. Hay
23.7	 Cutaneous vasculitis, connective tissue diseases,
and urticaria  5639
Volha Shpadaruk and Karen E. Harman
23.8	 Disorders of pigmentation   5677
Eugene Healy
23.9	 Photosensitivity   5688
Hiva Fassihi and Jane McGregor
23.10	 Infections of the skin   5695
Roderick J. Hay
23.11	 Sebaceous and sweat gland disorders   5699
Alison M. Layton
23.12	 Blood and lymphatic vessel disorders   5709
Peter S. Mortimer and Roderick J. Hay
23.13	 Hair and nail disorders   5724
David de Berker
23.14	 Tumours of the skin   5732
Edel O’Toole
23.15	 Skin and systemic diseases   5743
Clive B. Archer and Charles M.G. Archer
23.16	 Cutaneous reactions to drugs   5752
Sarah Walsh, Daniel Creamer, and Haur Yueh Lee
23.17	 Management of skin disease   5761
Rod Sinclair

ESSENTIALS
ESSENTIALS
section 22  Haematological disorders
5520
and often mild, as compared with thrombocytopenia associated
with inadequate platelet production (mentioned previously). Rarely,
an autoantibody to a specific clotting factor, such as factor VIII,
produces a severe acquired bleeding disorder. An acquired von
Willebrand syndrome may be due to either an autoantibody (often
an IgG paraprotein and occasionally IgM) or consumption of VWF
in patients with uncontrolled essential thrombocythaemia.
Hyperfibrinolysis
A bleeding tendency due to heritable defects of fibrinolysis is excep-
tionally rare. Pharmacological doses of fibrinolytic activators, such
as recombinant tissue factor activator, have an immediate lytic effect,
and hence clinical utility as a ‘clot-​busting drugs’, but their adminis-
tration is associated with a very high incidence of bleeding.
Bleeding in the neonate
Bleeding in the neonate may be due to a rare heritable defect or
an acquired abnormality occurring in utero or soon after delivery.
Thrombocytopenia is present in 30 to 40% of neonates in special
care baby units. Thrombocytopenia presenting in the first 72 h is
most often due to chronic fetal hypoxia, often with intrauterine
growth retardation, but is rarely due to transplacental passage of an
alloantibody to a platelet antigen, FNAIT, previously known as NAIT
(neonatal alloimmune thrombocytopenia, now fetal and neonatal).
Thrombocytopenia presenting after 72 h is most often in association
with sepsis or necrotizing enterocolitis and is often associated with
DIC. Bleeding from the umbilical stump or intracranial haemor-
rhage not explained by DIC or severe thrombocytopenia may be due
to severe deficiency of factors VIII, IX, XIII or afibrinogenaemia.
Management of bleeding
Acute bleeding
Effective treatment depends on a critical assessment of the extent
and nature of bleeding and the likely cause. When a defined haemo-
static abnormality is identified, specific therapy can be given. Drugs
that cause bleeding should be stopped and if appropriate (based on
an assessment of relative benefit and risk), specific reversal agents
administered, such as PCC for patients over-​anticoagulated with
VKA. Nonhaematological causes of bleeding should be managed ap-
propriately, for example, dialysis and red cell transfusion in patients
with renal failure. Vitamin K should be given to critically ill patients
and patients with liver disease. Early and sufficient blood product
support should be given to patients with massive blood loss and to
those with dilutional coagulopathy. Patients with overt haemato-
logical disorders such as myelodysplasia, ITP, or factor VIII inhibi-
tors will require specialist care. Pharmacological agents can be used
to increase haemostatic capacity but should be used by clinicians
with appropriate experience. Such drugs include DDAVP, tranex-
amic acid, and off-​licence use of drugs such as recombinant factor
VIIa. Aprotinin was used extensively in the past but is now used with
caution because of thrombotic complications, including death, and
renal impairment.
Nonacute bleeding
It is important to identify the circumstances that contribute to ab-
normal bleeding and to determine the likelihood of an underlying
persistent bleeding tendency as this will influence future man-
agement, for example, at times of surgery or decisions regarding
antithrombotic therapy. A comprehensive drug history should iden-
tify drugs that may have to be stopped. Some individuals are par-
ticularly sensitive to the usually mild anticoagulant effect of aspirin
or ADP antagonists or even NSAIDs. A single episode of abnormal
surgical bleeding may not be readily explained but this should be
taken into consideration at times of future surgery so that mechan-
ical rather than pharmacological thromboprophylaxis is used and
any antiplatelet or anticoagulant therapy stopped with certainty.
FURTHER READING
Balduini CL, Savoia A, Seri M (2013). Inherited thrombocytopenias
frequently diagnosed in adults. J Thromb Haemost, 11, 1006–​19.
Federici AB (2014). Clinical and laboratory diagnosis of VWD.
Hematology Am Soc Hematol Educ Program, 2014, 524–​30.
Harrison P, et al. (2011). Guidelines for the laboratory investigation of
heritable disorders of platelet function. Br J Haematol, 155, 30–​44.
Hunt B, et al. (2015). A practical guideline for the haematological man-
agement of major haemorrhage. Br J Haematol, 2015, 170, 788–​803.
Keeling D, et al. (2011). Guidelines on oral anticoagulation with war-
farin –​ fourth edition. Br J Haematol, 154, 311–​24.
Makris M, et al. (2012). Guideline on the management of bleeding in
patients on antithrombotic agents. Br J Haematol, 160, 35–​46.
Monroe DM, Hoffman M (2006). What does it take to make the perfect
clot? Arterioscler Thromb Vasc Biol, 26, 41–​8.
Toh CH, Alhamdi Y (2013). Current consideration and management
of disseminated intravascular coagulation. Hematology Am Soc
Hematol Educ Program, 2013, 286–​91.
22.7.3  Thrombocytopenia and
disorders of platelet function
Nicola Curry and Susie Shapiro
ESSENTIALS
The platelet is the smallest circulating blood cell. In health, it plays
a vital role in haemostasis, and in disease contributes to prob-
lems of bleeding and/​or thrombosis. The number of platelets pro-
duced is under tight homeostatic control, regulated by the cytokine
thrombopoietin. A normal platelet count lies within the range 150 to
450 × 109/​litre.
Vessel injury initiates a set of highly regulated and specialized
responses from the platelet, which include adhesion, activation,
granule release reactions, and platelet aggregation. A stable haemo-
static plug is thereby formed at the site of injury, where platelets are
bound tightly to fibrinogen via the glycoprotein IIb/​IIIa receptor.
Thrombocytopenia
Thrombocytopenia is defined as a reduction in the number of cir-
culating platelets to fewer than the normal reference range (typically
22.7.3  Thrombocytopenia and disorders of platelet function
5521
<150 × 109/​litre). Spontaneous bleeding is uncommon unless the
platelet count falls below 10 to 20 × 109/​litre or unless there is ab-
normal platelet function. Thrombocytopenia can be classified
according to three main pathologies: (1) increased platelet destruc-
tion, (2)  reduced platelet production, and (3)  increased platelet
sequestration.
Disorders of increased platelet destruction may be immune medi-
ated or nonimmune. Primary immune thrombocytopenia (ITP) is an
acquired disorder affecting both adults and children, characterized
by an isolated thrombocytopenia (platelet count <100 × 109 /​litre)
for which no precipitant can be found. Primary ITP is a diagnosis of
exclusion. Corticosteroids are the main first-​line therapy for adult ITP,
commonly prednisolone. Nonimmune causes of platelet destruction
include microangiopathic haemolytic disorders such as thrombotic
thrombocytopenic purpura, haemolytic uraemic syndrome, and dis-
seminated intravascular coagulation.
Decreased platelet production—​most cases are acquired, with
common or important causes being toxins (drugs, alcohol), nutri-
tional deficiencies (folate or vitamin B12), bone marrow infiltration,
and myelodysplastic syndrome.
Disorders of platelet distribution and platelet sequestration—​these
include splenomegaly and hypersplenism, haemodilution (in pa-
tients who have received large volumes of crystalloid solutions or
blood products), and extracorporeal circulation.
Disorders of platelet function
Disorders of platelet function are usually acquired. The most common
causes are medications and toxins (aspirin, nonsteroidal anti-​
inflammatory agents, ticlopidine, clopidogrel, glycoprotein IIb/​IIIa
inhibitors), systemic disorders (chronic kidney disease), and haemato-
logical diseases (chronic myeloproliferative disorders, myelodysplastic
syndromes, dysproteinaemias). Congenital disorders, which can affect
platelet adhesion and aggregation, secretion, or procoagulant activity,
are a rare cause of symptomatic bleeding.
Introduction
The platelet is the smallest circulating blood cell. It is a 2-​ to 4-​µm
diameter, discoid, anucleate cell that circulates in the blood fol-
lowing release from the megakaryocyte. In health, it plays a vital
role in haemostasis (see Chapter 22.7.1), and in disease contributes
to problems of bleeding as well as thrombosis. It plays important
roles in other processes such as promotion of vessel constriction as
well as vessel wall repair. Arguably, the most important additional
function for the platelet is the central role it takes in the cross-​talk
between coagulation and inflammation. Platelets are involved in
the promotion of leucocyte recruitment, the formation of platelet
microparticles, and the formation of neutrophil extracellular traps
which are increasingly being recognized as an important host re-
sponse to infection as well as playing a central role in thrombotic
and inflammatory diseases.
The number of platelets produced in the body is under tight
homeostatic control, regulated by the cytokine thrombopoietin
(TPO) which is produced by the liver. TPO binds to its receptor, c-​
MPL, found on both megakaryocyte and platelet surfaces and is then
internalized. TPO acts to stimulate megakaryopoiesis. A  normal
platelet count lies within the range 150 × 109/​litre to 450 × 109/​litre.
Platelets have a lifespan of approximately 7 to 10 days and they usu-
ally circulate in the bloodstream in a quiescent state. Their small
shape and size, in comparison to red blood cells and plasma proteins,
means that they circulate close to the endothelial surface facilitating
their role in clot formation following endothelial injury. Intact, un-
damaged vessel walls help to maintain the platelets in an inactive state
by releasing nitric oxide which acts both to dilate the vessel wall and
inhibit adhesion, activation, and aggregation of the platelet.
Vessel injury initiates a set of highly regulated and specialized
responses from the platelet which includes adhesion, activation,
granule release reactions, and platelet aggregation. In simple terms,
a platelet can be viewed as acting principally to form ‘haemostatic
bricks’ at the sites of vessel injury which are then ‘glued’ in place by
fibrin.
The term ‘platelet disorder’ covers a very large and heterogeneous
group of diseases with myriad causes. Platelet disorders can be in-
herited or acquired and may be classified in the following way:
1.	 An abnormality of platelet number (quantitative disorder), for
example, thrombocytopenia (platelet count <150 × 109/​litre) or
thrombocytosis (platelet count >450 × 109/​litre) (thrombocytosis
is covered in Chapter 22.3.6).
2.	 An abnormality of platelet function (qualitative disorder).
3.	 A combination of quantitative and qualitative abnormalities.
Acquired disorders are more common and therefore are more fre-
quently encountered in everyday clinical practice. This chapter will
focus on thrombocytopenia and disorders of platelet function.
Thrombocytopenia and platelet dysfunction
Thrombocytopenia is defined as a reduction in the number of circu-
lating platelets to fewer than a laboratory’s normal reference range
(typically <150 × 109/​litre). Spontaneous bleeding is uncommon un-
less the platelet count falls below 10 to 20 × 109/​litre or unless there
is abnormal platelet function. Thrombocytopenia can be classified
according to three main pathologies: (1) reduced platelet produc-
tion, (2) increased platelet destruction, and (3) increased platelet se-
questration (Table 22.7.3.1).
History and physical examination of the
thrombocytopenic patient/​patient
with platelet dysfunction
A single platelet count should not be evaluated in isolation and a
broader assessment of the clinical and laboratory picture must be
sought when investigating thrombocytopenia. The history will pro-
vide many of the important details and should focus on: bleeding
history, family history (if an inherited disorder is considered), drug
history and history of other symptoms related to a potentially causa-
tive acquired disorder.
Bleeding history
(See also Chapter 22.7.2.) Typical symptoms of thrombocytopenia or
platelet dysfunction relate to increased mucosal surface bleeding: skin
bruising/​bleeding (petechiae, ecchymoses), gum bleeding, epistaxis,
heavy menstrual bleeding, and gastrointestinal bleeding. However, in
some patients there may be no symptoms at all until a patient bleeds
section 22  Haematological disorders
5522
excessively after a haemostatic challenge such as surgery, dental ex-
traction, or childbirth. Less commonly, a patient may present with
symptoms of anaemia, due to continued occult blood loss and the
development of iron deficiency. A bleeding history is a subjective
method of assessment and mild platelet disorders can be difficult
to distinguish from normality. There are standardized guides or
‘bleeding questionnaires’ that can improve the accuracy of taking a
bleeding history and the International Society for Thrombosis and
Hemostasis (ISTH) recommend the ISTH Bleeding Assessment Tool
(2010). In women with heavy menstrual bleeding, pictorial men-
strual blood loss charts are a useful method of quantifying bleeding
and can be helpful as a means of monitoring response to treatment.
Drug history
It is important to also take a full drug history, and specifically to ask
about over-​the-​counter medicines that may contain aspirin or other
antiplatelet agents. The use of other drugs that may cause an im-
mune thrombocytopenia should be specifically sought (i.e. heparin,
sulfonamides, and quinine).
Family history
A patient is more likely to have an inherited disease if they give a
positive family history and present at a young age. Some of the in-
herited platelet disorders are autosomal recessive—​consider asking
if the patient’s parents are in a consanguineous relationship.
Previous platelet counts can prove invaluable as they may pro-
vide evidence of a longstanding thrombocytopenia suggesting
an inherited disease, or may show a rapid recent fall in a platelet
count more in keeping with an acquired problem. Von Willebrand
disease presents with symptoms that are very similar to a platelet dis-
order and should be excluded during the investigation, particularly
platelet-​type and type 2B von Willebrand disease where thrombo-
cytopenia is a common feature. Connective tissue disorders such
as Ehlers–​Danlos syndrome can present with problematic bruising
and should also be considered in the differential. Examination of pa-
tients with thrombocytopenia or platelet dysfunction should focus
on documenting the sites of bruising and bleeding—​particularly
noting sites of active bleeding. The examination should then focus
on evaluation of the lymph nodes, liver, spleen, and joints (to look
for signs of disease that can be associated with thrombocytopenia—​
see Box 22.7.3.1).
Table 22.7.3.1  Classification of thrombocytopenia by aetiology
Acquired
Platelet destruction
Immune:
Immune thrombocytopenic purpura (ITP)
Immune thrombocytopenia associated with other autoimmune disorders: Evan’s syndrome, systemic lupus erythematosus,
lymphoproliferative disorders
Drug induced: penicillins, bendroflumethiazide, digoxin, quinine, gold, heparin
Infection: HIV, hepatitis, malaria
Post-​transfusion purpura
Fetal and neonatal alloimmune thrombocytopenia
Nonimmune:
Thrombotic thrombocytopenic purpura
Haemolytic uraemic syndrome
Disseminated intravascular coagulation
Pregnancy related: haemolysis with elevated liver enzymes and low platelets (HELLP), pre-​eclampsia
Cardiopulmonary bypass
Reduced production
Marrow failure: aplastic anaemia
Marrow infiltration: metastatic cancer, haematological malignancies (leukaemia, lymphoma, myeloma), myelofibrosis,
storage disorders (Gaucher’s disease), granulomatous disorders (sarcoidosis)
Nutritional deficiency: vitamin B12 and folate
Toxins: alcohol, drugs (co-​trimoxazole, penicillamine), viral infections (HIV, hepatitis)
Altered distribution/​sequestration
Splenomegaly, massive transfusion, cardiopulmonary bypass
Congenital
Platelet destruction
Thrombotic thrombocytopenic purpura (TTP)
Reduced production ± altered
platelet function
Small platelets: Wiskott–​Aldrich syndrome
Normal-​sized platelets: congenital amegakaryocytic thrombocytopenia, thrombocytopenia with absent radius
Large platelets: MYH-​9-​related disorders including May–​Hegglin anomaly
Box 22.7.3.1  Secondary associations of immune
thrombocytopenia
•	 Infections:
—	 HIV
—	 Hepatitis C
—	 Varicella
—	 Epstein–​Barr virus
•	 Collagen vascular disease:
—	 Systemic lupus erythematosus
—	 Rheumatoid arthritis
•	 Lymphoproliferative disorders
—	 Hodgkin disease
—	 Non-​Hodgkin lymphoma, including CLL/SLL
•	 Other:
—	 Antiphospholipid antibody syndrome
—	 Autoimmune thyroid dysfunction
—	 Sarcoidosis
—	 Post bone marrow transplantation
22.7.3  Thrombocytopenia and disorders of platelet function
5523
Laboratory evaluation of the thrombocytopenic
patient/​patient with platelet dysfunction
Routine laboratory tests
First-​line laboratory tests and the reasons for performing these
tests are shown in Table 22.7.3.2. The blood film is essential.
Pseudothrombocytopenia is a common reason for a low platelet
count. This is an artefact that occurs when platelets clump to-
gether after blood is collected into ethylenediaminetetraacetic acid
(EDTA). This effect occurs in 0.1% of blood samples and is caused
by a clinically insignificant autoantibody which agglutinates platelets.
Often the artefact can be avoided by using an anticoagulant other
than EDTA (e.g. citrate). Fragmented red cells (schistocytes) may be
seen in thrombotic thrombocytopenic purpura (TTP), haemolytic
uraemic syndrome (HUS), and disseminated intravascular coagu-
lation (DIC). Leukoerythroblastic changes, such as teardrop-​shaped
red blood cells, nucleated red blood cells, and immature white cells
suggest infiltration of the bone marrow. The presence of abnormal
circulating cells such as lymphoblasts or myeloblasts suggests a
haematological malignancy. Typical changes on the peripheral smear
such as megaloblastic red blood cells and hypersegmented neutro-
phils suggest vitamin B12 or folate deficiency. Atypical lymphocytes
suggest a viral infection. Large platelets suggest the diagnosis of im-
mune thrombocytopenia but if there are many giant platelets a diag-
nosis of an inherited thrombocytopenia should be entertained.
Other laboratory investigations that may be indicated include
antinuclear antibody, rheumatoid factor, thyroid-​stimulating hor-
mone, antiphospholipid antibodies, and testing for HIV, hepatitis C,
and other infectious causes (Epstein–​Barr virus, varicella), vitamin
B12 and folate, and lactate dehydrogenase.
More specialist tests
Examination of the bone marrow should be considered if the aeti-
ology of the thrombocytopenia is uncertain. A bone marrow exam-
ination is required when abnormalities are seen on the blood film
suggestive of haematological malignancy or infiltration of the
marrow.
Assessing platelet function
Many platelet function tests are specialized and can only be per-
formed at certain laboratories. Tests should be performed within 2 h
of venepuncture, because platelets can undergo in vitro activation or
desensitization if left in blood bottles.
Methods such as the ‘Ivy’ or the ‘Template’ bleeding time were
used historically to assess primary haemostasis. This test is no longer
recommended as it is subject to many confounding factors. Another
test, the PFA-​100 or PFA-​200 test, goes some way to replacing the
bleeding time and provides an automated means of assessing pri-
mary haemostasis. However, it cannot be used to exclude a platelet
disorder when used alone as it is not sufficiently sensitive to reli-
ably detect a mild platelet disorder. Two cartridges are used for each
test; an ADP/​collagen and an adrenaline/​collagen cartridge. It is im-
portant to note that results are unreliable in the face of a platelet
count less than 80–100 × 109 /​litre.
Platelet aggregation is generally considered the gold standard test
of platelet function. It involves analysing the response of a patient’s
platelets to a variety of agonists. The aggregometer monitors changes
in light transmission through a sample of platelet-​rich plasma,
over time. Light transmission increases as platelet aggregation in-
creases. The following are commonly used agonists: ADP, collagen,
ristocetin, arachidonic acid, and epinephrine. Each agonist is tested
at multiple concentrations. In concert with aggregometry, platelet
secretion tests and/or platelet nucleotides are commonly evaluated.
Assessment of platelet ADP and ATP levels are useful for the diag-
nosis of storage pool disorders.
Many of these platelet tests are difficult to perform and patients
may need to be tested more than once. Other specialist tests in-
clude platelet flow cytometry and electron microscopy. Finally,
molecular genetic testing of families with inherited disorders of
platelets may be helpful. A United Kingdom-​wide clinical study
(BRIDGE-​BPD study; http://​www.bridgestudy.org) is recruiting
patients with platelet defects (and other rare bleeding disorders)
where the genetic cause is not yet elucidated. This study has used
exome sequencing to investigate the associations between geno-
type and phenotype.
General approach to management
Once a patient has been diagnosed with thrombocytopenia, gen-
eral measures to reduce their bleeding risk should be implemented.
These include avoidance of antiplatelet agents, in particular non­
steroidal medications; avoidance of trauma, and this should in-
clude consideration of contact sports such as rugby or martial arts;
consideration of hormonal inhibition of the menses; and control
of blood pressure. Tranexamic acid, an antifibrinolytic, is a useful
drug that patients can self-​administer for minor bleeding such as
gum bleeding or epistaxis and is also a helpful adjunctive therapy
for menorrhagia. Specific treatment for certain conditions are de-
scribed in the relevant sections.
If a patient requires a surgical intervention and has normal platelet
function, the United Kingdom British Committee for Standards in
Haematology (BCSH) guidelines recommend minimum platelet
thresholds for invasive procedures (Table 22.7.3.3).
Table 22.7.3.2  First-​line laboratory tests for thrombocytopenia
Historical full blood count (FBC)
Trends in platelet counts provide clues to underlying diagnosis
FBC and reticulocyte count
A reticulocyte count is helpful in the presence of anaemia, as if high it suggests haemolysis
and if low it points towards reduced red cell production. Cytopenias involving other cell
lineages are suggestive of disorders involving the bone marrow
Blood film
Essential (see text).
Prothrombin time (PT), activated partial thromboplastin
time (APTT), Clauss fibrinogen ± D-​dimers
Exclude disseminated intravascular coagulation or an additional coagulation abnormality
Renal and liver function
Exclude haemolytic uraemic syndrome, liver dysfunction
section 22  Haematological disorders
5524
Disorders of increased platelet destruction
Disorders of increased platelet destruction can be subdivided into
two principal categories: immune and nonimmune. Nonimmune
causes include the microangiopathic haemolytic disorders such as
TTP, HUS, and DIC.
Immune-​mediated platelet disorders
Immune-​mediated platelet disorders can be classified according
to the type of antibody involved. These include autoantibodies
(immune thrombocytopenic purpura), alloantibodies (NAIT or
post-​transfusion purpura (PTP)), and immune complex formation
(heparin-​induced thrombocytopenia). Heparin-​induced thrombo-
cytopenia is described in detail in Chapter 22.7.5 and will not be
covered here. Autoimmune thrombocytopenia is classified as pri-
mary (idiopathic) if there is no underlying cause and secondary if it
is associated with a systemic disease.
Primary immune thrombocytopenia
Primary ITP is an acquired immune-​mediated disorder af-
fecting both adults and children. It is characterized by an isolated
thrombocytopenia, defined as a platelet count of less than 100 ×
109 /​litre, for which no precipitant can be found. Primary ITP is a
diagnosis of exclusion and there are no reliable clinical or labora-
tory parameters that allow specific diagnosis. Historically, ITP was
thought to arise solely due to the effects of an IgG autoantibody that
coated the platelet and led to the increased clearance of platelets by
Fcɤ receptors in the liver and spleen. Recent research has recognized
that more complex mechanisms including direct T-​cell-​mediated
cytotoxic effects and impaired platelet production are also important.
ITP affects men and women equally, except between the ages of
30 to 60 years where it is more common in women. Adults who pre-
sent with ITP often describe an insidious onset of symptoms and are
more likely to have chronic disease. Childhood ITP often follows a
discrete viral illness (60% of cases), and presents with acute onset of
symptoms. Signs and symptoms of ITP vary widely. Many patients
have no symptoms and their thrombocytopenia may be picked up
by a routine full blood count. Other patients experience significant
bleeding. The severity of the thrombocytopenia, age (>60 years),
and a previous history of haemorrhage are predictors of increased
bleeding risk. Overall, adults with ITP have a four-​ to fivefold in-
creased risk of death from bleeding and infection although the ab-
solute rate of fatal haemorrhage from ITP is low, 0.016 to 0.039 cases
per adult patient-​year at risk.
Diagnostic approach
A thorough history and examination, as outlined previously, should
be completed. Examination should be normal, aside from signs of
mucosal bleeding. Mild splenomegaly may be evident in younger
patients, but significant splenomegaly, lymphadenopathy, or hepato-
megaly would point to another cause. Additional investigations for
ITP, over and above those set out previously, should be considered
(Table 22.7.3.4). A bone marrow examination is not of value rou-
tinely, but is recommended within the international ITP consensus
guidelines in the following groups: patients older than 60 years, pa-
tients presenting with systemic symptoms or abnormal signs, or in
patients where splenectomy is being considered.
Treatment of adult ITP
An individualized approach should be taken for the treatment of
patients. Patients with a platelet count greater than 50 × 109/​litre
rarely require therapy unless there is evidence of active bleeding or
Table 22.7.3.3  Recommended minimum platelet counts
for invasive procedures
Clinical indication
Treatment value
(×109/​litre)
Therapeutic
Massive transfusion
50
Massive transfusion and multiple trauma or TBI
100
DIC and bleeding
50
Intracerebral bleeding
100
Prophylactic
Pre-​invasive procedure, i.e. LP, CVC, epidural
50
Pre-​surgery
50-​75
Pre-​surgery at high-​risk sites: i.e. brain/​eye
100
CVC, central venous catheter; DIC, disseminated intravascular coagulation; LP, lumbar
puncture; TBI, traumatic brain injury.
Adapted from UK British Committee for Standards in Haematology guidelines and
European trauma guidelines.
Table 22.7.3.4  Additional tests for investigation of ITP
Test
Reason for test
Reticulocyte count and
Direct antiglobulin test
Determines presence of haemolysis
Important if considering anti-​D therapy
Immunoglobulin quantification
Looking for: common variable immunodeficiency and IgA deficiency
Blood group and Rh status
Informative if considering anti-​D therapy
Helicobacter pylori serology
Urea breath test
Eradication of H. pylori has been shown to lead to resolution of thrombocytopenia in some instances
Antiphospholipid antibodies and lupus anticoagulant
These are positive in 40% of individuals with ITP
Pregnancy test in women of childbearing age
Management may differ in pregnancy
Antinuclear antibodies
Can help diagnose systemic lupus erythematosus and is also an indicator of chronicity in childhood ITP
Viral PCR for CMV and parvovirus
Chronic infection can cause thrombocytopenia
CMV, cytomegalovirus; PCR, polymerase chain reaction.
22.7.3  Thrombocytopenia and disorders of platelet function
5525
the patient needs to undergo a surgery. Platelet transfusions should
not be used in patients with ITP unless bleeding exists which re-
quires immediate therapy.
First-​line therapy
Corticosteroids  Corticosteroids are the mainstay first-​line therapy
for adult ITP. Prednisone is the most commonly used medication,
and is effective and easy to administer. An alternative to prednisone
is the steroid dexamethasone (Table 22.7.3.5). The duration and
dose of steroids should be kept to a minimum.
Intravenous anti-​D  Reticuloendothelial blockade results in a more
rapid rise in the platelet count than corticosteroids. Anti-​D can be
used for Rh D-​positive, nonsplenectomized patients. Blood group,
direct antiglobulin test, and reticulocyte count must be known prior
to the use of this therapy. Anti-​D is a pooled plasma product and
patients should be made aware of this prior to treatment.
In 2010, the Food and Drug Administration issued a black box
warning against anti-​D highlighting the risk of haemolysis, DIC,
and renal failure. The incidence of severe haemolytic reactions is es-
timated at 1 in 1115 patients and it tends to occur within 4 h of treat-
ment. The risk appears highest in adults over the age of 65 years or in
patients with baseline evidence of haemolysis or renal impairment.
It is recommended that anti-​D is avoided in these groups of patients.
Intravenous immunoglobulin  Intravenous immunoglobulin (IVIg)
is often administered when a rapid rise in platelet count is needed, for
example, prior to surgery. Side effects are common during IVIg in-
fusions and concomitant corticosteroid therapy can both reduce side
effects and may enhance platelet response.
Second-​line therapies
Approximately one-​third of patients with ITP fail to respond to first-​
line therapy or relapse. Optimal second-​line therapy is uncertain
and patients should be actively involved in treatment decisions. The
aim of second-​line therapy is to achieve a sustained haemostatic
platelet response.
Rituximab  Rituximab is a chimeric monoclonal antibody directed
against CD20, a pan B-​cell antigen. Studies comparing standard-​
dose (375 mg/​m2) versus low-​dose rituximab (100 mg) found no
difference in response rate, although the duration of response was
shorter with the lower dose. A disadvantage of rituximab is that
the average time to a response is long, leaving patients vulnerable
to bleeding, as well as a lack of sustained response. If a response is
durable (i.e. lasts for ≥12 months), a patient is likely to respond to
a second cycle of treatment at relapse. Rituximab is contraindicated
in patients with active hepatitis B infection. The optimum time to
administer rituximab is not yet clear and informative randomized
controlled trials are lacking.
Newer-​generation anti-​CD20 treatments are being developed.
Only one has been tested so far in ITP—​veltuzumab, a humanized
monoclonal antibody. It led to an observable response in 55% of pa-
tients with some showing a durable response to 4.3 years. The poten-
tial benefit is that it can be administered subcutaneously.
Thrombopoietin receptor agonists  TPO receptor agonists bind
and activate the TPO receptor and initiate megakaryocyte differen-
tiation and platelet production. There are two main types of TPO
receptor agonists: TPO peptide mimetics (romiplostim) and TPO
nonpeptide mimetics (eltrombopag). These drugs are used as main-
tenance therapy and as soon as they are discontinued, platelet counts
return to baseline levels.
Splenectomy  Historically, splenectomy was the second-​line treat-
ment of choice. Now, with greater choice of drug therapies there
is an increasing reluctance to offer splenectomy. If used, it should
be deferred until at least 6 months, since spontaneous remission of
Table 22.7.3.5  Common treatments for immune thrombocytopenia
Therapy
Dose
Initial response rate and
speed of response
Side effects
Long-​term response rate
First-​line therapies
Corticosteroids:
   Prednisone
   Dexamethasone
1 mg/​kg (range 0.5–​2 mg/​kg)
given until platelets rise above
30–​50 × 109/​litre
40 mg/​kg for 4 days every
2–​4 weeks for 1–​4 cycles
70–​80% respond within
a few weeks
90% respond within a
few weeks
Variable and include weight
gain, lability of mood, diabetes,
hypertension, cataracts,
avascular necrosis, increased
infection risk
15% at 10 years
50–​60% at 2–​5 years
Intravenous anti-​D
50–​75 mcg/​kg
80% respond within
4–​5 days
Haemolysis, fever, rigors
Fatal intravascular haemolysis,
DIC, and renal failure are rare
Typical response lasts 3–​4 weeks
IVIg
0.4 g/​kg per day for 5 days
or
1 g/​kg per day for 1–​2 days
80% respond
within 2–​4 days
(occasionally by 24 h)
Headaches, fever, rigors,
fatigue, transient neutropenia,
thrombosis, and renal failure
Response duration short-​lived.
Baseline platelet counts seen at
4 weeks
Second-​line therapies
Rituximab
375 mg/​m2 intravenously every
week for 4 weeks
60% respond with median
time to response 5.5 weeks
Few side effects, most
commonly fevers, rigors, or
itching with first infusion
20% at 3–​5 years
TPO receptor agonists:
   Romiplostim
   Eltrombopag
1–​10 μg/​kg
subcutaneously weekly
25–​75 mg/​day orally
80–​90% within 1–​4 weeks
70–​80% within 2–​3 weeks
Headache, fatigue, arthralgia,
increased bone marrow reticulin
Headache, increased bone
marrow reticulin, thrombosis
Response seen at 4 years with
continued administration
Response seen at 1.5 years with
continued administration
Splenectomy
–​
80% respond within
3 weeks
Lifelong risk of infection
60–​70% response at 5 years
section 22  Haematological disorders
5526
ITP can occur up to 6 to 12 months from diagnosis. Splenectomy
can be performed as an open or laparoscopic procedure, with mor-
tality rates of 1.0 and 0.2%, respectively. Indium-​labelled platelet
scans can be used to look for platelet destruction in the spleen. If
splenic destruction is confirmed, approximately 90% of patients
respond to splenectomy. Splenectomized patients are at lifelong
increased risk of infection, particularly from encapsulated organ-
isms, and should be vaccinated at least 4 weeks prior to surgery.
Patients should be given a polyvalent pneumococcal, meningo-
coccal C conjugate, and Haemophilus influenzae B vaccine. Patients
in receipt of rituximab within 6 months of splenectomy may not
respond to vaccination and revaccination should be given after B-​
cell recovery. No consensus has been reached about whether long-​
term prophylactic antibiotics are useful, but asplenic patients are
often given phenoxymethylpenicillin 250 to 500 mg twice daily or
erythromycin 500 mg twice daily. An alternative approach is to offer
splenectomized patients a home supply of antibiotics to use in case
of need for a febrile illness. Patients need to be well educated about
their risk of infection.
Third-​line therapies
Approximately 20% of patients will not achieve an acceptable platelet
count after first-​ and second-​line treatments and splenectomy. Some
of these non-​ or poor-responders will maintain a good quality of life
with a low platelet count (sometimes as low as 10 × 109/​litre) and
will only require therapy for surgical intervention. Other patients
have increased bleeding rates and a higher risk of death. Options
for them are limited but include combination chemotherapy and
Campath-​1H.
Emergency treatment of ITP
Patients with ITP may require an urgent increase in platelet count,
most commonly for surgical procedures, and sometimes for signifi-
cant active bleeding. Combination therapy is of particular benefit in
these situations and prednisone plus IVIg are recommended for pa-
tients with uncontrolled bleeding. High-​dose methylprednisolone
may also be useful. Platelet transfusion is appropriate as treatment
for a patient with significant active bleeding (and responses may be
better if transfusion is given with IVIg), along with other standard
measures to treat major haemorrhage. Vinca alkaloid drugs (i.e.
vincristine) in combination with other therapies have been re-
ported to lead to a better and more rapid rise in platelet count and
may be considered for emergency treatment.
ITP during pregnancy
ITP is estimated to occur in 1 in 1000 to 1 in 10 000 pregnancies.
Women who have previously had a diagnosis of ITP may relapse
during pregnancy. During the first two trimesters, treatment is initi-
ated only if the patient is symptomatic, the platelet count falls below
20–​30 × 109/​litre, or the patient needs to undergo a procedure.
High-​quality data are lacking to guide safe platelet thresholds for
procedures, but as a general rule caesarean section or spontaneous
vaginal delivery are thought safe with a platelet count above 50 ×
109/​litre and neuraxial anaesthesia requires a count of at least 80 ×
109/​litre. There is no evidence to recommend a caesarean section
over vaginal delivery in maternal ITP.
Primary treatments for pregnant patients with ITP are similar to
nonpregnancy and corticosteroids and IVIg remain the first-​line
treatments. Prednisone is often started at a lower dose (10–​20 mg/​
day) and adjusted according to response. IVIg doses and response
rates are similar to nonpregnant patients and IVIg is commonly used
in situations where the platelet count needs to rise rapidly. First-​
line therapies can fail and combination therapy, such as high-​dose
methylprednisolone (1 g) in combination with IVIg or azathioprine
has been used especially in the weeks just prior to delivery to estab-
lish a safe platelet count. Splenectomy has been performed in preg-
nancy, and if required is recommended in the early second trimester
using a laparoscopic technique.
Neonatal ITP occurs in approximately 10% of cases. Procedures
during delivery that may increase intracranial haemorrhage,
such as Ventouse extraction, fetal scalp electrodes, or forceps de-
livery should be avoided. A cord blood sample should be drawn
immediately after delivery. The neonate should be monitored for
between 2 and 5 days after delivery. If a platelet count falls below
20 × 109/​litre or the neonate is bleeding, a single dose of IVIg is
recommended and if there is significant haemorrhage, a platelet
transfusion should be given. Transcranial ultrasound scanning
should be routine for all neonates with a platelet count less than
50 × 109/​litre.
ITP in childhood
The diagnosis of ITP in a child is one of exclusion. Patients com-
monly present with bruises and petechiae and only very few (3%)
have clinically significant bleeding. The incidence of intracerebral
haemorrhage in children is in the order of 0.1 to 3%. Most chil-
dren can be managed expectantly. This approach requires the
parents to understand the attendant risks of thrombocytopenia
for their child. Two-​thirds of children who are managed using a
watch and wait policy will spontaneously remit within 6 months
of diagnosis. Time to remission is highly variable. Admission to
hospital is only required for children with significant bleeding.
Treatment for children mirrors adult management but many
treatments including corticosteroids, immunosuppression, and
splenectomy, are not thought to be curative and may cause more
problems than thrombocytopenia. Prednisone is often used first
line at a dose of 1 to 2 mg/​kg per day for a maximum of 14 days or
4 mg/​kg per day for 4 days; 75% of patients respond and platelet
recovery is seen rapidly by 2 to 7 days. IVIg leads to a response in
80% of patients and recovery is rapid within 1 to 2 days. Children
are more commonly treated with a single dose of IVIg at 0.8 to
1.0 g/​kg. Anti-​D is also used in childhood ITP. Treatment of life-​
threatening bleeding mirrors adult therapy with the use of platelet
transfusions in combination with intravenous high-​dose cortico-
steroids and IVIg.
Some children develop chronic ITP. A child should be re-​evaluated
3–​6 months after primary diagnosis. Investigations should include
a bone marrow examination to look for haematological malig-
nancy, antinuclear antibodies, tests for antiphospholipid syndrome,
immunoglobulin quantification, and a review of medications. The
aim of therapy for chronic ITP is to maintain a haemostatic platelet
count while minimizing corticosteroid exposure. Splenectomy
may also be performed in children with ITP but risks of infection
are high (3% overwhelming sepsis) and this risk must be weighed
against the low overall mortality for childhood ITP (0.5%). If first-​
line therapies prove ineffective, treatments such as rituximab can
be used. A  systematic review of uncontrolled studies evaluating
22.7.3  Thrombocytopenia and disorders of platelet function
5527
rituximab therapy in 323 children reported a pooled complete re-
sponse of 39% (platelets >100 × 109/​litre) and an overall response
rate of 68% (platelets >30 × 109/​litre) with a median response dur-
ation of 12.8 months. TPO receptor agonist therapies have also been
used in the paediatric setting and a recently published placebo-​
controlled randomized controlled trial, PETIT2 (NCT01520909),
evaluated the safety and efficacy of eltrombopag in chronic ITP.
Eltrombopag led to a sustained response in 40% of patients and no
safety concerns were raised.
Secondary immune thrombocytopenia
A variety of medical disorders cause secondary immune thrombo-
cytopenia (Box 22.7.3.1). The treatment for secondary immune
thrombocytopenia is similar to that of ITP.
Alloimmune thrombocytopenia
Alloimmune thrombocytopenia is caused by alloantibodies directed
against platelet glycoproteins (GPs). There are two alloimmune
thrombocytopenic disorders: NAIT and PTP.
Fetal and neonatal alloimmune thrombocytopenia
NAIT is a condition in which maternal alloantibodies cross the
placenta and destroy fetal platelets. The mother’s immune system
recognizes paternal platelet antigens, expressed on fetal platelets,
as foreign. This disorder can cause severe and life-​threatening fetal
thrombocytopenia that may lead to intracranial haemorrhage and
death in utero. About 80% of NAIT cases are caused by anti-​HPA-​1a
and 15% anti-​HPA-​5b; other HPA antibodies are detected occasion-
ally. The diagnosis of NAIT requires the demonstration of maternal
platelet  alloantibodies that react against platelet-​specific antigens
present in the father and infant but not in the mother. There is no
laboratory parameter which reliably predicts the severity of fetal/​
neonatal thrombocytopenia.
Management of the bleeding neonate is with platelets which are
negative for the antigen (usually HPA-​1a negative); use random
donor platelets in an emergency. NAIT recurs in 85 to 90% of sub-
sequent pregnancies. For subsequent pregnancies, cordocentesis
is performed at about 20 weeks for platelet count and phenotype.
If the neonate is affected, the mainstay of treatment is IVIg given
to the mother every week (1 g/​kg). A beneficial effect of IVIg on
the fetal/​neonatal platelet count occurs in about 67% of cases.
Weekly intrauterine platelet transfusions may be used if other
treatment fails.
Post-​transfusion purpura
PTP is rare (<1 in 700 000 transfusions) and is manifest by a severe
thrombocytopenia (platelet count <10 × 109/​litre) developing 5 to
12 days after a blood transfusion. Bleeding is common and may be
fatal. PTP is seen after transfusion of packed red cells, whole blood
and platelets. It occurs in patients who have been previously sensi-
tized, either by pregnancy or a previous transfusion, to foreign platelet
antigens. These patients develop an alloantibody against a platelet
antigen that they lack, usually HPA-​1a (or less commonly HPA-​5b).
PTP occurs most often in multiparous women. The alloantibodies
cause platelet destruction. The diagnosis of PTP is confirmed by the
demonstration of IgG alloantibodies in the patient’s serum against
one of the HPA antigens. It is unclear why alloantibodies attack the
patient’s own, as well as the transfused platelets. IVIg therapy is the
primary treatment (1 g/​kg intravenously for 2 days) and 85% of pa-
tients respond. Platelet transfusion should be avoided, unless there is
significant bleeding. Plasmapheresis is an option for IVIg-​refractory
patients. The incidence of PTP has fallen since the introduction of
leucodepletion.
Drug-​induced thrombocytopenia
Many drugs can cause thrombocytopenia. The medications most
commonly implicated include heparin, penicillins, sulfonamides,
valproic acid, and quinine. However, virtually every medication
has been associated with thrombocytopenia. The diagnosis of
drug induced thrombocytopenia is mostly empiric. Laboratory
confirmation can be sought and involves the demonstration of
drug-​dependent antiplatelet antibodies using methods such as
enzyme-​linked immunosorbent assay (ELISA). Patients with
drug-​induced thrombocytopenia typically have moderate to se-
vere thrombocytopenia. Thrombocytopenia is usually seen after 2
to 3 days if the medication has previously been used, or after 1 to 3
weeks for a new medication. The thrombocytopenia usually resolves
within 5–​10 days of stopping the causative drug. In cases of severe
thrombocytopenia, the drug should be discontinued and the patient
may be treated using either IVIg or anti-​D. Treatment with cortico-
steroids is less effective. In cases of life-​threatening haemorrhage,
platelet transfusions may be required. Patients should not take the
drug causing the thrombocytopenia again as it will cause thrombo-
cytopenia with subsequent exposure.
Nonimmune platelet destruction
Microangiopathic haemolytic anaemia and related disorders
A microangiopathic haemolytic anaemia (MAHA) with thrombo-
cytopenia may be caused by TTP, HUS, DIC, HELLP (haemolysis,
elevated liver enzymes, and low platelets), and pre-​eclampsia/​
eclampsia.
Thrombotic thrombocytopenic purpura
This is a rare disorder characterized by thrombocytopenia and
microangiopathic haemolytic anaemia. Its incidence is approxi-
mately 6 per million per year. Recognition and early diagnosis is im-
portant as untreated the mortality is 90%. Even with optimal current
therapy mortality is about 20%. It may be acquired (95% of cases) or
congenital.
TTP is caused by a deficiency of the von Willebrand factor (VWF)
cleaving protease ADAMTS13 (a disintegrin and metalloprotease
with thrombospondin type 1 motif). Deficiency of ADAMTS13
results in circulation of ultra-​large VWF multimers. These are
haemostatically active and spontaneously bind to platelets, par-
ticularly under conditions of high shear. The VWF/​platelet aggre-
gates block small blood vessels. These microthromboses cause tissue
damage, consume platelets, and result in fragmentation of passing
red blood cells. Acquired TTP results from autoantibodies against
ADAMTS13. Most cases are idiopathic but TTP can be associated
with other diseases and conditions (Box 22.7.3.2).
Acquired TTP
Acquired TTP has a peak incidence of 40 years and slightly more
woman are affected then men. The only clinical criteria required for
section 22  Haematological disorders
5528
its diagnosis are thrombocytopenia and microangiopathic haemo-
lytic anaemia. Patients may also present with:
•	bleeding: bruising, petechiae, haematuria, retinal haemorrhage
•	neurological signs (in about 70% of patients) which may be tran-
sient: confusion, headache, visual problems, aphasia, paresis, coma
•	renal signs (about 30%):  acute kidney injury, proteinuria,
microhaematuria
•	cardiac signs (about 40%): chest pain, hypotension
•	gastrointestinal tract signs (about 30%)
•	nonspecific symptoms: fever, arthralgia, myalgia, pallor, jaundice,
abdominal pain
Diagnosis of TTP is challenging as there is significant overlap
with other disorders such as HUS, pregnancy-​related disorders,
DIC, and autoimmune disorders (Box 22.7.3.3). The initial diag-
nosis is made on clinical history and examination combined with
blood parameters consistent with microangiopathic haemolytic
anaemia: anaemia, thrombocytopenia, red cell fragments on blood
film, elevated reticulocytes/​bilirubin/​lactate dehydrogenase, and
low haptoglobin. A  negative Coombs’ test helps to exclude an
autoimmune haemolytic anaemia. A  coagulation screen should
be normal, rather than deranged as it is in DIC. The diagnosis is
confirmed by an ADAMTS13 activity assay less than 5 to 10% with
or without anti-​ADAMTS antibody, but treatment should not be
delayed while these results are awaited. Investigations should also
be sent to look for an underlying cause such as HIV, pregnancy,
pancreatitis, and malignancy (Table 22.7.3.6).
Acute acquired TTP should be treated as a medical emergency.
The mainstay of treatment is plasma exchange (PEX) with fresh
frozen plasma, although the optimum regimen has not yet been
determined. PEX significantly improves time to remission and
chance of survival compared to fresh frozen plasma infusion. PEX
removes the ultra-​large VWF and is a source of ADAMTS13. PEX
should be started within hours of the patient presenting. If there
is any delay than an infusion of fresh frozen plasma can be given
as a temporary holding measure. Patients require daily plasma
exchange of 1 to 1.5 blood volumes. This may be intensified in
patients who present with cardiac or neurological involvement,
or who are refractory to initial therapy. Plasma exchange should
be continued daily until the platelet count has normalized for
2 days and then it can be stopped (median time to remission is
about 2 weeks).
Given that autoantibodies are the most frequent pathogenic
mechanism, adjunctive therapies include corticosteroids and
rituximab (monoclonal anti-​CD20). High-​dose prednisone at 1
to 2 mg/​kg is commonly given although its efficacy has not been
unequivocally demonstrated. In prospective studies, with limited
patient numbers, rituximab has been shown to reduce time to re-
mission if given within the first 3 days of presentation, and may
prolong time between relapses. Other immunosuppressants have
also been tried, generally in the pre-​rituximab era and with less
evidence of effectiveness than is currently available for rituximab,
but they may be considered for refractory cases. These include
ciclosporin, cyclophosphamide and vincristine. Splenectomy is
associated with high mortality in the acute setting (40%) and has
limited proven benefit.
Major bleeding is rare and platelet transfusions should be
avoided as they can precipitate widespread thrombosis. Patients
can be supported with transfusion of packed red blood cells as
required. While the clinical efficacy of antiplatelets has not been
proven they are relatively safe—​and given the risk of micro-
vascular thrombosis, and the risk of venous thrombosis in an un-
well medical patient, it is generally recommended that patients
should be started on low-​dose aspirin (75 mg) and low molecular
weight heparin thromboprophylaxis as platelet count recovers
(>50 × 109/​litre) (expert opinion).
Relapse is defined as an episode of acute TTP more than 30 days
after remission. Relapse occurs in 20 to 50% of patients. It is recom-
mended that patients are counselled with regard to the risk of relapse
and advised to seek medical help early if they experience symptoms
of potential relapse. Patients should be monitored long term using
Box 22.7.3.3  Differential diagnosis of microangiopathic
haemolytic anaemia and thrombocytopenia
•	 Thrombotic thrombocytopenic purpura (TTP)
•	 Haemolytic uraemic syndrome (HUS)
•	 Disseminated intravascular coagulation (DIC)
•	 Pregnancy related: pre-​eclampsia, HELLP, HUS
•	 Vasculitis
•	 Catastrophic antiphospholipid syndrome
•	 Evans syndrome (autoimmune haemolytic anaemia and thrombo­­
cytopenia)
•	 Malignant hypertension
•	 Infections, typically viral (adenovirus, cytomegalovirus) or severe bac-
terial infections
•	 Disseminated malignancy
Table 22.7.3.6  Investigations in a patient with suspected
thrombotic thrombocytopenic purpura
To establish MAHA
Full blood count, blood film, reticulocytes, lactate
dehydrogenase, bilirubin, coagulation screen
(PT, APTT, fibrinogen), Coombs’ test
To look for organ
involvement
Renal function, troponin, ECG and consider
echocardiogram and CT/​MRI
To look for underlying
cause
Pregnancy test, HIV, hepatitis A/​B/​C serology,
autoantibody screen including antinuclear
antibodies, dsDNA, anticardiolipin antibody,
β2-​glycoprotein-​1, lupus anticoagulant screen,
vitamin B12, folate, thyroid function, amylase,
stool culture, CT chest/​abdomen/​pelvis
To confirm diagnosis
ADAMTS13 activity assay, anti-​ADAMTS13
antibody assay
ECG, electrocardiogram; MAHA, microangiopathic haemolytic anaemia; MRI, magnetic
resonance imaging.
Box 22.7.3.2  Conditions and diseases associated
with thrombotic thrombocytopenic purpura
•	 Pregnancy and postpartum
•	 Infections: HIV
•	 Drugs:  ciclosporin, quinine, ticlopidine, clopidogrel, interferon-​α,
simvastatin
•	 Connective tissue disorder: lupus erythematosus and scleroderma
•	 Allogenic bone marrow transplantation
22.7.3  Thrombocytopenia and disorders of platelet function
5529
ADAMTS13 activity assays and anti-​ADAMTS13 antibody assays.
It has been shown that patients with an ADAMTS13 activity less
than 10% or a detectable anti-​ADAMTS13 antibody have a threefold
increase risk of relapse at 1 year. In such cases, elective rituximab
has been successfully used to normalize ADAMTS13 activity in the
majority of patients.
Congenital TTP
Only about 100 cases of congenital TTP have been reported world-
wide, but this is likely to increase given the better understanding
of the disease and availability of ADAMTS13 assays. It is caused
by a mutation in the ADAMTS13 gene (chromosome 9) which re-
sults in a quantitative or qualitative deficiency. There is a wide spec-
trum of disease with the severest cases manifesting in the neonatal
period; milder cases may manifest in middle age or in pregnancy.
Treatment is with infusion of an intermediate-​purity FVIII concen-
trate (which contains ADAMTS13) or with infusion of fresh frozen
plasma.
Haemolytic uraemic syndrome
This syndrome includes microangiopathic haemolytic anaemia
and thrombocytopenia but with more marked renal failure
then is generally seen in TTP. It is discussed in more detail in
Chapter 22.7.3.
Classical HUS typically affects children and is associated with
verotoxin-​positive bloody diarrhoea, Escherichia coli serotype
O157:H7, or Shigella dysenteriae serotype I. Treatment is supportive
and may include renal dialysis.
Atypical HUS (aHUS) is generally not associated with diar-
rhoea. It may have multisystem symptoms and can be extremely
difficult to differentiate from TTP. The primary feature of HUS
is renal impairment/​failure, in association with thrombocyto-
penia and MAHA. As opposed to TTP, ADAMTS13 activity levels
are greater than 10%. Patients should be treated urgently with
plasma exchange, especially if TTP has not been excluded. There
is increasing evidence for complement dysfunction in aHUS and
many patients benefit from administration of the monoclonal anti-
body C5 inhibitor, eculizumab.
Pregnancy-​associated thrombocytopenia and
microangiopathic haemolytic anaemia
Several disorders may present for the first time in pregnancy or post-
partum period including congenital and acquired TTP, aHUS, and
DIC. Pregnancy-​specific disorders of pre-​eclampsia/​eclampsia and
HELLP are also associated with low platelets and MAHA. A multi-​
disciplinary approach with close liaison between obstetricians and
haematologists is vital.
Disseminated intravascular coagulation
DIC is an acquired syndrome characterized by widespread intra-
vascular activation of the coagulation cascade. The most frequent
clinical presentation of DIC is bleeding, although organ dysfunc-
tion can also result from microthrombi. Excessive thrombin over-
whelms the physiological inhibitors of coagulation and results in
excess fibrin, platelet activation, and fibrin/​platelet thrombosis,
and bleeding secondary to thrombocytopenia and coagulation
factor consumption. First-​line tests for its diagnosis include
platelet count, elevated fibrin degradation products, prolonged
prothrombin time, and low fibrinogen. DIC is discussed further
in Chapter 22.7.5.
Disorders of platelet distribution and
platelet sequestration
Splenomegaly and hypersplenism
Approximately 30% of circulating platelets are normally pooled in
the spleen. Splenomegaly results in an increase in the size of the pool
of platelets sequestered in the spleen and may result in moderate
thrombocytopenia (platelets >40 × 10/​9litre).
Haemodilutional disorders
Dilutional thrombocytopenia is seen after major surgery or large
volume blood transfusion. Incidental or gestational thrombocyto-
penia occurs in up to 75% of pregnancies and is usually mild—​
increased blood volume is likely to be a significant component.
Extracorporeal circulation
Thrombocytopenia associated with cardiopulmonary bypass is
multifactorial: haemodilution, blood loss, as well as activation of
platelets by the synthetic surface may contribute. The thrombocyto-
penia is usually mild but it is often accompanied by platelet dysfunc-
tion, secondary to activation on the synthetic surface as well as the
use of antiplatelet agents.
Disorders of decreased platelet production
Decreased platelet production results from abnormalities affecting
the megakaryocyte progenitor cells, megakaryocytes, or the bone
marrow stroma. An isolated reduction in platelet production is
rare—​it is generally associated with abnormal production of other
cell lines. Diagnosis is usually made by bone marrow to examine
megakaryocyte numbers and morphology. Causes may be acquired
or congenital.
Acquired disorders of decreased platelet production
Toxins
Many drugs and toxins can cause thrombocytopenia as a result of
bone marrow suppression. Common drugs include chemotherapy
agents, ionizing radiation, chloramphenicol, and nonsteroidal anti-​
inflammatory drugs (NSAIDs).
Alcohol directly suppresses platelet production, but additionally
thrombocytopenia may result from hypersplenism and nutritional
deficiency. The thrombocytopenia may be associated with a megalo-
blastic anaemia and ringed sideroblasts.
Nutritional deficiencies
Folate or vitamin B12 deficiency can result in thrombocytopenia
which may be severe. The thrombocytopenia may be associated
with a megaloblastic anaemia and hypersegmented neutrophils.
The platelet count recovers with replacement of the deficient
vitamin.
section 22  Haematological disorders
5530
Infection
Systemic infections (viral, bacterial, and fungal) may result
in thrombocytopenia of multifactorial aetiology. Viral infec-
tions may suppress platelet production directly by infection
of the megakaryocyte, toxic effects of viral proteins of cyto-
kines, haemophagocytosis, or immune destruction of platelets.
Thrombocytopenia has been associated with HIV, Epstein–​Barr
virus, adenovirus, measles, mumps, varicella, hepatitis, and parvo-
virus (erythrovirus) B19. Bacterial and fungal infections may also
cause thrombocytopenia by direct toxicity or haemophagocytosis.
Thrombocytopenia associated with malaria is thought to be due
to direct infection of the platelets, dysregulated cytokines and im-
mune function, and increased splenic removal. Treatment would
generally be supportive, with platelet transfusions if required,
although the thrombocytopenia is usually relatively mild and
platelet count increases as the infection resolves.
Infiltration of the bone marrow
Infiltration of the bone marrow generally results in a degree of pan-
cytopenia. Infiltration may be with nonhaematopoietic cells such as
metastatic cancer, granulomatous or storage disorders, or by haem-
atological malignancies (leukaemia, lymphoma, myeloma) and
myelofibrosis. Myelodysplasia may present initially with an isolated
thrombocytopenia.
Congenital disorders of decreased platelet production
Thrombocytopenia in infancy is usually secondary to platelet de-
struction. Inherited causes of reduced platelet production are rare
but should be considered when there is a family history of bleeding
or when thrombocytopenia in infants and children persists and is
otherwise unexplained. A few of the better characterized disorders
are outlined in the following sections, but these are only a fraction of
the inherited thrombocytopenias.
Treatment depends on the severity of the bleeding disorder and
associated platelet dysfunction. Options include local measures
including hormone treatment for menorrhagia, antifibrinolytics
such as tranexamic acid, desmopressin (DDAVP) for platelet dys-
function disorders, and platelet transfusions.
Inherited thrombocytopenia with reduced platelet size
Wiskott–​Aldrich syndrome:  an X-​linked disorder characterized
by micro-​thrombocytopenia, combined immunodeficiency, ec-
zema, and increased risk of developing autoimmune disorders and
malignancy.
Inherited thrombocytopenia with normal platelet size
Congenital amegakaryocytic thrombocytopenia:  an autosomal
recessive disorder due to mutations in the MPL gene, character-
ized by severe thrombocytopenia, and almost complete absence of
megakaryocytes in the bone marrow. Individuals develop progressive
bone marrow failure over 5 to 10 years although it can be more rapid.
Thrombocytopenia with absent radius:  an autosomal recessive
disorder characterized by a severe thrombocytopenia which clas-
sically improves throughout childhood. Individuals have bilaterally
absent radii and often other associated features including skeletal
defects of the lower limb, cow’s milk intolerance, and renal and car-
diac abnormalities.
Inherited thrombocytopenias with increased platelet size
MYH-​9 related disorders including May–​Hegglin anomaly: this
is an autosomal dominant macrothrombocytopenia caused by
deletion within the MYH9 gene, which encodes nonmuscle my-
osin II-​A heavy chain. Individuals usually have a mild bleeding
phenotype, although it may be more severe than expected from the
platelet count due to associated platelet dysfunction. Associated
features include sensorineural hearing loss, glomerulonephritis,
and cataracts. Neutrophils may have inclusions on blood film,
called Döhle-​like bodies.
Bernard–​Soulier syndrome and grey platelet syndrome are dis-
cussed under platelet function disorders, as the platelet dysfunction
is generally more marked than the thrombocytopenia.
Disorders of platelet function
Disorders of platelet function are usually acquired.
Acquired disorders of platelet function
Drugs
Numerous drugs have been shown to affect platelet function. Some
drugs have been designed for this purpose, while for others it is a
side effect.
Antiplatelet agents such as aspirin, thienopyridine derivatives,
and GPIIb/​IIIa inhibitors are used in cardiovascular disorders.
Clinical trial evidence is discussed further in Chapter  22.7.3.
Aspirin irreversibly inhibits cyclooxygenase within platelets
preventing formation of thromboxane, resulting in reduced
platelet aggregation. Thienopyridine derivatives (e.g. clopidogrel,
ticlopidine, and prasugrel) inhibit platelet function by inhibiting
the P2Y12 ADP receptor and therefore the ADP-​induced pathway
of platelet activation. Three commercially available GPIIb/​IIIa
inhibitors are abciximab (a monoclonal antibody), eptifibatide
(a synthetic cyclic heptapeptide), and tirofiban (a nonpeptide an-
tagonist). They block platelet aggregation by directly inhibiting
the platelet receptor for fibrinogen. Of note, there is a 5% risk of
thrombocytopenia and 1 to 2% risk of severe thrombocytopenia
(platelets <50 × 109/​litre).
NSAIDs inhibit cyclooxygenase (COX):  COX-​1 is found in
most cells, including platelets and the gastrointestinal epithe-
lium, and COX-​2 is induced by inflammation. COX-​2 selective
inhibitors, such as etoricoxib, have less activity on COX-​1 com-
pared to traditional NSAIDs. This reduces the effect on platelet
function and gastrointestinal side effects such as ulceration. If
an anti-​inflammatory is required in a patient with an inherited
bleeding disorder, COX-​2 selective inhibitors are therefore pre-
ferred. Other drugs described as potentially reducing platelet
aggregation are nitrates, calcium channel blockers, β-​blockers,
β-​lactam antibiotics, antiepileptics, tricyclic antidepressants, and
phenothiazines.
Chronic renal failure
Uraemia can result in defects in adhesion and aggregation of plate-
lets. The exact pathogenesis is unknown. DDAVP is occasionally
used to improve platelet function in a bleeding patient. Platelet func-
tion may also improve after dialysis.
22.7.3  Thrombocytopenia and disorders of platelet function
5531
Chronic myeloproliferative disorders and
myelodysplastic syndromes
Chronic myeloproliferative disorders and myelodysplasia (see
Chapters  22.3.5, 22.3.6, and 22.3.2) may be associated with ab-
normalities in platelet function as well as platelet count. Platelet
function tests may show impaired aggregation to a range of agon-
ists and storage pool defects. If the patient is bleeding then treat-
ment is supportive—​antifibrinolytics (tranexamic acid) and platelet
transfusions if necessary. The platelet function defect may respond
to treatment of the underlying disease.
Dysproteinaemias
Paraproteins, associated with Waldenström macroglobulinaemia,
monoclonal gammopathy of uncertain significance, and multiple
myeloma, may be associated with an acquired coagulation disorder
(acquired haemophilia and acquired VWD) but also can result in
abnormalities of platelet function. Nonspecific binding of the para-
protein may disrupt the platelet membrane receptors. Treatment
options include plasmapheresis to transiently remove the parapro-
tein, treatment of the underlying disorder, antifibrinolytics, and po-
tentially platelet transfusions.
Congenital disorders of platelet function
Inherited platelet function disorders are an uncommon cause of
symptomatic bleeding. They are heterogeneous in severity, difficult
to diagnose, and therefore mild platelet function disorders in par-
ticular are likely to be under-​diagnosed. Patients may present with
a history of easy bruising, epistaxis, menorrhagia, and prolonged
bleeding after surgery or dental procedures and other family mem-
bers may be affected.
These disorders may be classified functionally into abnormalities
of platelet adhesion, aggregation, signalling and secretion, and pro-
coagulant activity. A few of the most well-​characterized disorders
are outlined in the following sections. As for inherited thrombocyto-
penia disorders, treatment depends on the severity of the bleeding
disorder. Options include local measures including hormone treat-
ment for menorrhagia, antifibrinolytics such as tranexamic acid,
desmopressin (DDAVP) (see Chapter  22.7.4 for further detail of
practical administration), and platelet transfusions.
Disorders of platelet adhesion
Bernard–​Soulier syndrome is caused by a deficiency or abnor-
mality of platelet GPIb/​IX. This results in defective binding to
VWF, and a markedly reduced ability to adhere to sites of vascular
injury, where subendothelial VWF is exposed. It is rare, with an
approximate frequency of one in a million. It is autosomal re-
cessive and consanguinity is common in reported kindreds. It is
characterized by a mild–​moderate thrombocytopenia and giant
platelets. On platelet function testing, platelets do not agglutinate
in response to ristocetin, but show normal aggregation with other
agonists including ADP, collagen, and thromboxane. Diagnosis can
be confirmed by flow cytometry using analysis of GPIb-​α density.
It is a moderate to severe bleeding disorder. If antifibrinolytics and
local measures fail to control bleeding then platelet transfusions
are usually effective but there is a risk of alloimmunization by
HLA antigens or GPIb-​α. Normal infants of women who have
antiplatelet antibodies are at risk of alloimmune thrombocyto-
penia. Of note, one copy of the gene that encodes GP1b-​α is lost
in individuals with a chromosome 22q11 deletion (diGeorge and
velocardiofacial syndromes). Some of these individuals develop
macrothrombocytopenia although their platelet function is normal.
Disorders of platelet aggregation
Platelet aggregation occurs after the adhesion of platelets to the
damaged vessel wall, and occurs when activated platelets interact
with one another.
Glanzmann thrombasthenia is caused by quantitative or qualita-
tive abnormalities of GPIIb–​IIIa (platelet integrin αIIbβ3) resulting
in the absence of platelet aggregation. It is rare, autosomal recessive,
and consanguinity is common within reported kindreds. Platelet
function tests show absent aggregation with agonists such as ADP,
adrenaline, collagen, and arachidonic acid but agglutination with
ristocetin is present. Flow cytometry using antibodies to GPIIb
(CD41) and GPIIIa (CD61) is used for definitive diagnosis. It is
generally a severe bleeding disorder—​the clinical features are those
usually expected with platelet dysfunction: easy bruising, epistaxis,
and menorrhagia. Haemarthroses are very rarely reported. Similar
to Bernard–​Soulier syndrome, platelet transfusions are effective in
controlling bleeding, but there is a risk of alloimmunization espe-
cially to the missing GPs, resulting in refractoriness to subsequent
transfusions. Recombinant factor VIIa is licensed for use in those
patients who are refractory to platelet transfusions.
Disorders of signalling
Defects of platelet ADP receptors have been reported and associ-
ated with a bleeding tendency: G protein-​coupled receptors P2Y1
and P2Y12 and a ligand-​gated ion channel P2X1.
Storage pool disorders
Storage pool deficiency (SPD) syndromes result in secondary dis-
orders of aggregation. This is a heterogeneous group of disorders
characterized by a reduction in secretable substances stored in
platelet granules. It may result from reduced dense granules (δ-​SPD)
or reduced α-​granules (α-​SPD) or from both (αδ-​SPD).
Patients with δ-​SPD or αδ-​SPD usually have absent secondary
aggregation waves to ADP and adrenaline, although primary
waves are present. Collagen-​induced aggregation is usually ab-
sent or markedly reduced, but ristocetin-​induced agglutination is
present. In the majority of cases, SPDs are isolated disorders and
inheritance may be autosomal dominant, but this has not been de-
termined for many. These SPDs can also be associated with other
syndromes including Wiskott–​Aldrich syndrome, thrombocyto-
penia with absent radius, Hermansky–​Pudlak syndrome, and
Chediak–​Higashi syndrome.
Grey platelet syndrome is an α-​SPD. It is extremely rare, with
fewer than 100 cases reported worldwide. Both autosomal dom-
inant and autosomal recessive inheritance has been described. On
the blood film there is a thrombocytopenia, and platelets appear
agranular and misshapen. Electron microscopy demonstrates re-
duced or absent numbers of α-​granules

Transfusion and transplantation 5563 22.8.1 Blood
Transfusion and transplantation 5563 22.8.1 Blood transfusion 5563 D.S. Giovanniello and E.L. Snyder
CONTENTS
22.8.1	 Blood transfusion  5563
D.S. Giovanniello and E.L. Snyder
22.8.2	 Haemopoietic stem cell transplantation  5579
E.C. Gordon-​Smith and Emma C. Morris
22.8.1  Blood transfusion
D.S. Giovanniello and E.L. Snyder
ESSENTIALS
Transfusion of blood components is a life-​saving treatment for pa-
tients with severe haemorrhage and can also be used to replace co-
agulation factors and to ameliorate the effects of severe anaemia,
thrombocytopenia, and impaired platelet function. However, blood
transfusion has many hazards, hence its use should always be con-
sidered carefully and restricted to those who will gain benefit that
outweighs the risks.
General considerations
Safe administration of blood components requires secure processes
from vein to vein to prevent the incorrect blood product from being
given to the wrong patient. Reduction in transfusion risks is also
achieved by (1) robust arrangements for the collection, storage, and
delivery of appropriate supplies of blood products to their point of
need; (2) better understanding of the antigenic structures on blood
cells and the widespread introduction of advanced blood-​group
typing methods, screening for antibodies, and testing for compati-
bility before transfusion; (3) identification and screening for agents
present in donors, as well as the use of sterile disposable materials;
and (4)  comprehending the benefit of treating anaemia with the
need to avoid unnecessary transfusion, with its associated costs and
potential harm.
Blood group systems—​these include (1)  the ABO system—​the
codominantly expressed A and B genes code for glycosyltransferases
that add either N-​acetyl-​d-​galactosamine (A gene) or -​d-​galactose
(B gene) to the common precursor H antigen. Anti-​A and Anti-​B anti-
bodies are ‘naturally occurring’ and responsible for most haemolytic
transfusion reactions. (2)  The Rh system—​the most clinically im-
portant Rh antigen is D because it is strongly immunogenic; anti-​D
is responsible for immune reactions including haemolytic disease of
the newborn and immune-​mediated transfusion reactions. (3) Other
clinically significant blood group antigens—​these include Kell (K),
Duffy (Fy), Kidd (Jk), and the MNS systems; multiple antibodies can
develop when the range of red cell antigens in the donor population
differ from that of patients who require repeated transfusion.
Clinical use of blood components
Cellular components, namely those containing red blood cells,
platelets, and white blood cells, have specific clinical use and in-
dications. These products and their respective indications include
(1)  red blood cells—​symptomatic anaemia; (2)  leucocyte-​reduced
components (red blood cells and platelets)—​symptomatic an-
aemia, reduce febrile reactions from leucocyte antibodies, alterna-
tive to cytomegalovirus (CMV)-​negative components, prevent HLA
alloimmunization; (3)  washed components (red blood cells and
platelets)—​remove harmful plasma substances that contribute to al-
lergic transfusion reactions and remove antibodies that may lead to
adverse reactions; (4) platelet components—​thrombocytopenia with
bleeding, prophylactic transfusion, platelet function abnormalities;
and (5) granulocytes (obtained by apheresis)—​for neutropenic pa-
tients with an infection unresponsive to antibiotics but who have a
reasonable expectation for recovery (rarely used, given increased use
of haemopoietic growth factors in haematological practice); donor
lymphocyte infusions can induce remission of disease and improve
survival by exerting a graft-​versus-​leukaemia effect in some bone
marrow transplant recipients.
Noncellular products include plasma, cryoprecipitate, and
plasma derivatives. Respective indications include (1) fresh frozen
plasma—​replacement of plasma coagulation factors for which spe-
cific factor concentrates are not available, liver disease, dissemin-
ated intravascular coagulation, hypofibrinogenaemia, thrombotic
thrombocytopenic purpura, dilutional coagulopathy, and reversal
of vitamin K antagonists or vitamin K deficiency in the setting of
major bleeding; (2) cryoprecipitate—​fibrinogen and factor XIII re-
placement, factor VIII and von Willebrand factor replacement when
recombinant and virus-​inactivated concentrates are not available;
(3) albumin—​used principally in specialized surgical practice, replace-
ment fluid in therapeutic plasma exchange, and in the treatment of
22.8
Transfusion and transplantation
section 22  Haematological disorders
5564
liver disease; and (4) intravenous immunoglobulin—​used principally
for immunodeficiency syndromes, autoimmune rheumatic/​vascu-
litic diseases, Guillain–​Barré syndrome, and autoimmune haemo-
lytic anaemias; specific Rh (D) immunoglobulin is used to prevent
alloimmunization in D-​negative mothers; specific immunoglobulin
preparations are used as antivenoms and to treat viral infections (e.g.
hepatitis A and B).
Complications of transfusion therapy
Complications of transfusion therapy can be divided into two broad
categories—​immune and nonimmune complications. Immune
complications include (1) acute intravascular haemolytic reactions—​
usually caused by transfusions of ABO-​incompatible blood resulting
from patient identification or clerical errors; manifest with sudden
onset of back pain, hypotension, tachycardia, fever, chills, diaphor-
esis, and dyspnoea; treatment consists of immediately stopping the
transfusion and providing supportive care, but can be fatal despite
best management; (2) delayed haemolytic reactions—​usually caused
by an antibody that is initially of a titre below the limits of detection
on routine screening; (3) febrile nonhaemolytic reactions—​usually
attributed to the development of antibodies in the recipient dir-
ected against HLA and/​or leucocyte-​specific antigens on donor
white blood cells and platelets, or alternatively, may be attributed
to the passive transfusion of cytokines, such as interleukin (IL)-​1,
IL-​6, IL-​8, and tumour necrosis factor-​α, which are generated and
accumulate during the storage of blood components; (4)  allergic
reactions—​IgE mediated; IgA-​deficient patients are particularly
prone to anaphylactic reactions; (5) transfusion-​related acute lung
injury, most commonly attributed to the passive transfusion of blood
donor antibodies directed against HLA antigens to the recipient,
or less commonly, attributed to the passive transfusion of blood
donor antibodies directed against human neutrophil antibodies; and
(6) transfusion-​associated graft-​versus-​host disease, resulting from
an attack by viable immunocompetent donor lymphocytes on the
recipient’s antigen presenting tissues. This immunological assault is
manifested clinically by damage to skin, liver, gastrointestinal tract,
and bone marrow.
Nonimmune complications include (1) infection and septic trans-
fusion reactions, with organisms commonly implicated including
Gram-​positive (staphylococci) and Gram-​negative (enterobacter,
yersinia, pseudomonas) bacteria. Other infections that may be trans-
mitted from the donor include malaria, babesiosis, syphilis, leish-
mania, toxoplasmosis, and viral infections such as hepatitis B and
C, HIV-​1, HIV-​2, and West Nile virus. Immunocompromised recipi-
ents are also at risk from human cytomegalovirus and parvovirus
(erythrovirus) B19. A few patients have been shown to have acquired
variant Creutzfeldt–​Jacob disease, probably as a result of transfu-
sion from latently affected donors. (2) Other complications—​acute
problems can include circulatory overload, dilutional coagulopathy,
hypocalcaemia, and hypothermia; complications of multiple blood
transfusions include iron overload.
The focus of many haemovigilence programmes has been in
the prevention of these aforementioned complications. Risks of
alloimmunization from donor erythrocytes and leucocytes and trans-
mission of viruses can be avoided or reduced by (1) autologous blood
salvage during surgery or acute normovolaemic haemodilution im-
mediately before surgery; and (2) improved methods for leucocyte
reduction and inactivation of infectious agents (pathogen reduction/​
pathogen inactivation). Despite much research, the introduction of
blood substitutes has yet to be realized in clinical practice. Use of
purified recombinant haematopoietic growth factors (e.g. erythro-
poietin, granulocyte colony-​stimulating factor) has reduced the need
for transfusion of blood products in many patients.
Introduction
Transfusion medicine and blood banking is a speciality that has
evolved over the years. With greater understanding of red cell,
platelet, and leucocyte antigen structure and function, transfu-
sion therapy has improved. In addition, understanding current and
emerging infectious agents ensured patient safety. Blood banking
involves donor eligibility and testing, collection, processing, and
storage of blood components. The transfusion medicine service
in a hospital maintains necessary and vital activities contrib-
uting to successful blood transfusion. These functions include
pretransfusion testing, compatibility testing, additional processing
(e.g. washing, irradiation), and the evaluation of transfusion reac-
tions. Transfusion medicine has expanded over recent decades to
include multiple disciplines, such as therapeutic apheresis, cellular
therapy, and tissue banking. One of the most important techno-
logical improvements in transfusion therapy was the development
of sterile, disposable, and flexible plastic containers that allow sep-
aration of whole blood into cellular (e.g. red cells, platelets) and
noncellular (e.g. plasma, cryoprecipitate) components, known as
apheresis, derived from Greek word aphaeresis, meaning ‘to take
away.’ This technology allows the blood of a donor or patient to pass
through an apparatus that separates out one particular constituent
and returns the remainder to the circulation. Anticoagulants and
additives currently used to collect blood allow storage of liquid sus-
pensions of concentrated red cells for 35 to 42 days. These advances
have essentially eliminated the use of whole blood. Blood transfu-
sion is used to treat patients with severe anaemia, haemorrhage,
thrombocytopenia, and coagulation disorders. Although the haz-
ards of blood replacement are relatively small, the expected benefit
of a transfusion must outweigh the risk to the patient. Therefore, a
thorough understanding of the indications of blood transfusion is
required to minimize unnecessary blood replacement and to pre-
vent wastage of limited blood resources. Clinicians who prescribe
blood transfusion must also be familiar with the risks and be able to
recognize and treat transfusion reactions.
Blood collection and processing
Blood donation is either autologous or allogeneic, collected and
processed from whole blood or by apheresis procedure. Most of
the red blood cells (RBCs) and plasma manufactured in the United
States of America are from whole blood collections. Conversely,
the majority of the platelets manufactured in the United States of
America are collected by an apheresis procedure. The donor se-
lection process includes a health history, a directed mini phys-
ical exam, and donor questionnaire, which contains questions to
protect the recipient from acquiring a transfusion-​transmitted
22.8.1  Blood transfusion
5565
infection or immune-​mediated transfusion reaction, and to pro-
tect the donor from suffering an adverse reaction after donation.
This process also helps to ensure the donor’s ability to tolerate the
collection procedure. Whole blood is collected and subsequently
undergoes processing and separation into its components by cen-
trifugation. There are two schemata in processing whole blood: the
United States of America uses the platelet-​rich plasma method,
while Canada and Europe uses the Buffy coat method. The differ-
ence between them relates to use of different g forces during the
primary and secondary centrifugation steps. This is illustrated in
Fig. 22.8.1.1.
Pretransfusion testing
Pretransfusion testing is performed to prevent the transfusion of
incompatible blood that could result in a haemolytic transfusion
reaction. It includes ABO and Rh D typing, antibody detection,
identification, and compatibility testing. The steps of pretransfusion
testing prior to issuing a unit of RBCs are illustrated in Fig. 22.8.1.2.
When a patient blood sample is delivered to the blood bank and
appropriate patient identification is performed, the specimen is
centrifuged and tests are performed on either serum or plasma.
Processing schemas
WBD
USA
Canada and Europe
High centrifugation g force
High centrifugation g force
Low centrifugation g force
Low centrifugation g force
White blood cells
RBC
PC
Plasma
PC
PC
PC
PC
PC
PC
Pooled platelets
Cryoprecipitate
Cryo-reduced plasma
Plasma
Plasma
RBC
PRP
Buffy coat
Fig. 22.8.1.1  Processing schemas. WBD, whole blood donation; PRP,
platelet rich plasma; PC, platelet concentrate; RBC, red blood cell.
Transfusion order
Blood sample collection
ABO and Rh D typing
Antibody screen
Positive screen
Autocontrol
Positive
Autoantibody
Coombs’ test
lgG
Warm
Cold
Eluate
Phenotype
Identify the AB
Panel of cells
Alloantibody
Negative
C3
Crossmatch compatible
Negative screen
Crossmatch compatible/
Least incompatible/
Phenotypically matched
Antigen negative/
Crossmatch compatible
Fig. 22.8.1.2  The steps of pretransfusion testing prior to issuing a unit of RBC.
section 22  Haematological disorders
5566
The tests in the blood bank are serological, based on the agglutin-
ation reactions that result from antigens on the RBCs interacting
with antibodies. The first test is identifying patient ABO and Rh D
type, using commercially available reagents. During ‘front typing’,
red cells are reacted with antibodies directed against the A, B, and D
antigens. Blood grouping is confirmed during ‘back typing’ in which
serum/​plasma is tested for the presence of expected anti-​A and anti-​
B antibodies. This is followed by an antibody screen (known as the
indirect antiglobulin test or indirect Coombs’ test) performed by
incubating patient serum/​plasma with two to four reagent red cells
with a predetermined antigen phenotype that in sum will cover all
common and clinically significant alloantibodies. If an antibody is
present in the serum/​plasma, it will react with the screening cell(s)
and cause red cell agglutination. Antibody screening is commonly
performed at room temperature (immediate phase), after incu-
bating serum/​plasma and test red cells at 37°C, and after incuba-
tion with antihuman globulin serum (Coombs’ phase). The screen
will detect the presence of antibodies formed against foreign RBC
antigens (i.e. alloantibody) or against self (i.e. autoantibody). When
an alloantibody is suspected by a positive screen and a negative
autocontrol (patient plasma/​serum mixed with patient RBCs),
an antibody identification panel is performed that is an indirect
antiglobulin test using serum/​plasma tested against a larger com-
mercial ‘panel’ of group O reagent red cells from donors with pre-
determined phenotypes. If a pattern is detected, the specificity of
the alloantibody can usually be identified. This is followed by per-
forming a patient phenotype to detect the presence or absence of the
antigen on the patient RBCs to which the antibody is directed. The
patient phenotype in the case of an alloantibody should be nega-
tive. In some cases, a patient’s serum may react with all panel cells.
These ‘panagglutinins’ can be caused by (1) a single antibody dir-
ected against a high incidence antigen present on all panel test red
cells; (2) multiple antibodies that in total react with all test cells; or
(3) an autoantibody, in which case the patient’s serum will also react
with his or her own red cells. Autoantibodies will have a positive
screen, a positive autocontrol, and a positive direct antiglobulin test
(Coombs’ test). If complement is coating the RBCs (C3), the auto-
antibody is a cold reacting IgM. In the case of an IgG coating the
RBCs, it is a warm autoantibody and an eluate should be performed
to identify the specificity of the antibody that is coating the RBCs.
This is performed by dissociating antibodies from the sensitized
RBC (elution) and performing an antibody identification panel.
Once the antibody is identified, compatibility testing will follow
prior to issuing RBCs. Blood bank tests are performed using the
tube testing system in which red cell agglutination is identified in
standard test tubes. A number of newer systems are being used to
detect antigen–​antibody reactions. These include gel systems based
on the differential mobility of red cell agglutinates through gel col-
umns and capture systems in which test red cells are immobilized on
microtitre plates. Newer automated and semiautomated systems are
rapidly replacing tube testing for the majority of ABO grouping, Rh
typing, antibody screening, and crossmatching.
Blood group systems
Blood group antigens are proteins and carbohydrates attached to
lipids or proteins. These antigens are found on the surface of the
RBCs. Some of these antigens are found on other blood cells, tissues,
and secretions in addition to the RBCs. There are 33 blood group
systems known to date and over 290 antigens identified.
ABO system
The most clinically important blood group antigens belong to the
ABO system. The codominantly expressed A and B genes, located
on chromosome 9, code for glycosyltransferases that add either N-​
acetyl-​d-​galactosamine (A gene) or -​d-​galactose (B gene) to the
common precursor H antigen (Table 22.8.1.1). Group O individ-
uals lack functional transferases due to a single base deletion in
the A gene that eliminates its production. The AB antigens are of
critical importance because individuals who lack the A and/​or B
antigens form IgM and IgG antibodies directed against the missing
antigen(s). Circulating A and B antibodies can fix complement and
cause intravascular haemolysis. Anti-​A and Anti-​B antibodies are
‘naturally occurring’, that is, they are formed without prior clinical
antigenic stimulation. Presumably, individuals become immunized
following exposure to carbohydrate ABO antigenic determinants
commonly found in the bacterial environment. Accordingly, group
A individuals produce anti-​B, group B produce anti-​A, and group
O produce both anti-​A and anti-​B, as well as anti-​A,B, an IgG anti-
body. It is worth noting that this anti-​A,B antibody produced by
group O individuals can cross the placenta and potentially lead
to haemolytic disease of the newborn in non-​group O neonates.
Circulating A and B antibodies are of critical importance in blood
therapy because they are responsible for most major haemolytic
transfusion reactions.
Rh system
The Rh blood group system is composed of at least 50 distinct
antigens. The five major antigens in the Rh system (D, C, c, E, and
e) are responsible for most Rh-​related transfusion incompatibility. It
is now known that the D polypeptide is encoded at the RHD locus,
whereas the CcEe polypeptide is coded by alleles at the RHCE locus;
both are found on chromosome 1. Based on the D gene frequency in
Table 22.8.1.1  ABO blood group system
ABO type
Gene(s)
Enzyme coded by gene
Resulting
antigen
Antibody present in
plasma
Frequency (white
population)
O
H
L-​fucosyl transferase
H
Anti-​A, anti-​B, anti-​A,B
0.43
A
H,A
L-​fucosyl transferase and N-​acetyl-​d-​galactosamine transferase
H,A
Anti-​B
0.45
B
H,B
L-​fucosyl transferase and d-​galactosyl transferase
H,B
Anti-​A
0.09
AB
H,A and B
L-​fucosyl transferase and N-​acetyl-​d-​galactosamine and
d-​galactosyl transferase
H,AB
None
0.04
22.8.1  Blood transfusion
5567
North America and Europe, approximately 15% of individuals will
not produce D antigen and are ‘Rh negative’. Very rare individuals
who lack all Rh antigens are termed ‘Rh-​null’. Rh-​null red cells are
morphologically abnormal and typically have shortened survival,
resulting in a mild haemolytic anaemia.
The most clinically important Rh antigen is D because it is strongly
immunogenic; the likelihood of a D-​negative person developing
anti-​D following exposure to as little as 0.1 ml of D-​positive red
cells is extremely high. Anti-​D is responsible for immune reac-
tions including haemolytic disease of the newborn and immune-​
mediated transfusion reactions. Despite widespread use of Rh
immune globulin, anti-​D remains a most common cause of serious
haemolytic disease of the newborn. Rh-​negative women most com-
monly produce anti-​D after exposure to D-​positive red cells during
pregnancy, a miscarriage, or abortion. The anti-​D formed is of the
IgG class and therefore can cross the placenta where it may cause po-
tentially fatal intrauterine haemolysis in an Rh-​positive fetus.
Other blood groups
Other well-​characterized, clinically significant blood groups in-
clude Kell (K), Duffy (Fy), Kidd (Jk), and MNS systems. Antibodies
to these blood group antigens may form following exposure to the
corresponding antigens during transfusion or pregnancy and are
associated with immune-​mediated red cell destruction of trans-
fused cells and haemolytic disease of the newborn (Table 22.8.1.2).
In most cases, compatible blood can be found for patients with red
cell alloantibodies. Based on the high incidence of some red cell
antigens among specific donor populations, however, some patients
may be difficult to transfuse if they have developed multiple anti-
bodies. This is particularly true in patients with sickle cell disease
and other red cell disorders who require frequent transfusions.
Table 22.8.1.2 lists the antigen frequencies of the most clinically
relevant blood groups.
Certain blood groups are known to have particular disease asso-
ciations. The Kell system is linked to chronic granulomatous dis-
ease, a congenital disease in which a decreased oxidative capacity
of neutrophils leads to recurrent, severe bacterial infections. The
genetic defect seen in chronic granulomatous disease is located on
the X chromosome near the Kx Kell locus, mutation at that locus re-
sults in depressed expression of the Kell antigens. Abnormalities de-
scribed in this disease include acanthocytic red cells that are prone
to mild haemolysis, cardiomyopathy, areflexia, and skeletal myop-
athies: this is known as the McLeod phenotype. The Duffy antigens,
which occur at a much lower incidence among African populations,
have an interesting association with malaria. Specifically, Fy(a–​b–​)-​
negative red cells are resistant to Plasmodium vivax and P. knowlesi
infection. Red cells from most West African black people are Fy(a–​
b–​) and therefore resistant to these forms of malaria.
The antibodies to the Kidd antigens are unusual in that once
formed, they often fall below the level of detection and may not be
detected in an already immunized patient. In this situation, trans-
fusion of additional Kidd-​positive red cells may cause a rapid, sec-
ondary immunological response, leading to formation of high-​titre
anti-​Kidd with subsequent haemolysis and delayed transfusion re-
actions; this is illustrated in Fig. 22.8.1.3.
Antibodies
Alloantibodies
Alloantibodies are antibodies formed against foreign antigens
present on donor RBCs but absent on recipient RBCs. Some
alloantibodies are naturally occurring and some are formed fol-
lowing exposure to transfusion or by pregnancy. The clinical sig-
nificance of different alloantibodies varies. Antibody significance is
determined by their class and subclass, quantity, ability to activate
Table 22.8.1.2  Clinically significant blood groups
Red cell antigen
Antigen frequencya
Risk of haemolysis (immediate or delayed)
Risk of haemolytic disease of the newborn
A, B
Variable
High (immediate)
Moderate (anti-​A)
Low (anti-​B)
Rh
Variable
High (immediate and delayed)
Variable
High to low
K
0.09
High (immediate and delayed)
High
k
0.998
High (immediate and delayed)
Low
Fya
White 0.66
High (delayed)
Low
Black 0.10
Fyb
White 0.83
Low (delayed)
None
Black 0.23
Jka
0.77
Moderate (immediate and delayed)
Rare
Jkb
0.73
Moderate (immediate and delayed)
None
M
0.78
Low
Rare
N
0.72
None
None
S
0.55
Moderate (immediate and delayed)
Rare
s
0.89
Low (delayed)
Rare
a Antigen frequency in white population unless otherwise specified.
section 22  Haematological disorders
5568
complement, its thermal amplitude, the characteristics of the RBC
antigen, and the patient’s clinical status. A  clinically significant
alloantibody can cause a haemolytic transfusion reaction or haemo-
lytic disease of the newborn. If a clinically significant alloantibody
is identified, crossmatched compatible RBCs units that lack the
antigen should be provided.
Autoantibodies
Autoantibodies consist of immunoglobulins that react with a wide
range of self-​antigens including membrane and intracellular com-
ponents, adsorbed plasma proteins, and nuclear antigens. The
presence of an autoantibody does not necessarily cause increased
RBC destruction; up to 15% of hospitalized patients have a positive
Coombs’ test with no signs or symptoms of haemolysis. However,
patients with warm autoimmune haemolytic anaemia often require
transfusion. In this case, the blood bank may have difficulty finding
a ‘compatible’ unit of red cells because the patient’s serum not only
reacts with their own red cells, but also those of all donor red cells.
Additional time may be required by the blood bank to exclude the
presence of a significant underlying alloantibody that is obscured by
the autoantibody. Upwards of 25% of previously transfused patients
with warm autoimmune haemolytic anaemia may have an under-
lying alloantibody which can cause red cell haemolysis. Autoimmune
antibodies often appear to have specificity for Rh antigens (e.g. anti-​
e), but the transfusion of antigen negative red cells (e.g. e-​negative)
is not indicated, as in vivo red cell survival of antigen-​negative cells
is usually no better than antigen-​positive cells.
Compatibility testing
Routine compatibility testing is performed only on red cell units,
whole blood, and granulocytes prior to transfusion. Specifically,
donor red cells are reacted with patient serum, and if no reac-
tion is observed, the unit is considered ‘crossmatch compatible’. In
emergency situations, there may be insufficient time to perform
compatibility testing. Many hospitals will supply group O Rh-​
negative red cells until a patient sample is obtained and tested. If a
patient’s ABO Rh status is known with certainty, then type-​specific
uncrossmatched blood can be provided. In either case, compati-
bility testing is performed on these transfused units as soon as pos-
sible. It is critically important to realize that supplies of O-​negative
blood are limited and some centres adopted the strategy of trans-
fusing males with O Rh D-​positive units in an emergency situation
and preserve the O Rh D-​negative units for children and females.
‘Computer crossmatches’ have been instituted at many hospitals in
North America using a validated computer system. Patients with
known ABO and Rh type, and who have a negative antibody screen,
are provided with ABO-​compatible blood while omitting the cross-
match step described earlier. Although a true serological crossmatch
is not performed, the computer crossmatch is safe in the vast ma-
jority of transfusions.
Clinical use of blood components
Blood component therapy refers to transfusing only the specific
component that is required by a patient. Individual components are
stored under optimal conditions, which maintain the integrity and
potency of each respective blood component. RBCs are stored re-
frigerated at 1 to 6°C, platelets at room temperature, 22 to 24°C, due
to loss of function with refrigeration, and plasma is frozen at tem-
peratures equal to or less than −18°C to maintain functional clotting
factors, especially labile factors, which deteriorate with time. The
difference in storage requirements among blood components is one
of the reasons why whole blood usage has dropped. Other reasons
include the requirement to provide the best blood component for
the patient, cost, manufacturing logistics at the blood centre, and
Secondary
response
Primary
response
Alloantibody titers
First exposure to
antigen
Second exposure to
antigen
Time
Threshold of
detection
Fig. 22.8.1.3  The antibodies to the Kidd antigens are unusual in that once formed, they often
fall below the level of detection and may not be detected in an already immunized patient. In
this situation, transfusion of additional Kidd-​positive red cells may cause a rapid, secondary
immunological response, leading to formation of high-​titre anti-​Kidd with subsequent haemolysis
and delayed transfusion reactions.
22.8.1  Blood transfusion
5569
a limited blood supply. Plasma separated from whole blood can be
further fractionated into coagulation factor concentrates, albumin,
or gamma globulin. Cell separators (apheresis devices) capable of
collecting platelets, plasma, granulocytes, peripheral blood stem
cells, and RBCs, are also in widespread use across the United States
of America and Europe.
Red blood cells
RBCs account for approximately 75% of the annual cost of transfu-
sion therapy in the United Kingdom. Red cell units, prepared from
whole blood by removing most of the plasma, are indicated for pa-
tients with acute haemorrhage or chronic anaemias (Table 22.8.1.3).
Earlier preservative solutions composed of citrate, dextrose, and
phosphate buffers allowed storage of red cells from 21 to 35 days. It
was later observed that adenine improved cell viability by increasing
intracellular ATP levels. The haematocrit of RBC units varies from
55 to 70% depending on the specific anticoagulant/​preservative so-
lution used. Citrate contained in blood preservatives binds calcium
to inhibit clotting and may cause hypocalcaemia and alkalosis in
neonates and massively transfused patients.
Units of red cells stored refrigerated at 1 to 6°C have a shelf-​
life of 35 to 42 days depending on the ingredients of the preser-
vative. During storage, the following changes are observed in red
cell units: (1) a fall in pH, (2) decreases in red cell ATP and 2,3-​
diphosphoglycerate, (3)  increased supernatant potassium, and
(4) decreased supernatant glucose. RBCs with uncommon antigen
profiles can be frozen within 6 days of collection and stored for up to
10 years. They are frozen with glycerol to avoid cell dehydration and
damage during the freezing process.
The patient’s overall clinical status and laboratory parameters
should be considered as a whole when deciding to transfuse a pa-
tient. A  decision should not be based on the haematocrit alone.
Younger patients will usually tolerate a given degree of hypoxaemia
and hypotension better than older patients who may have underlying
coronary or myocardial disease. Evidence of symptomatic anaemia
includes excessive fatigue, malaise, headache, tachycardia, hypoten-
sion, and end-​organ damage. Hypovolaemic shock typically ensues
with acute loss (<24 h) of more than 30% of total blood volume.
Initially, the haematocrit will be falsely elevated in acute haemor-
rhage, but will then fall with fluid resuscitation. Slowly developing,
chronic anaemias are usually better tolerated than rapid-​onset
anaemias due to the ability of the body’s fluid compensatory mech-
anisms. Transfusion is rarely indicated when the haemoglobin (Hb)
level is greater than 100 g/​litre, and is often not considered until the
Hb level is less than 70 g/​litre. A patient’s cardiac and pulmonary
status must be considered when determining transfusion thresh-
olds. Patients with unstable angina or acute myocardial infarction
may require transfusion when the Hb level is less than 100 g/​litre.
In the absence of active red cell destruction, transfusing a single
unit will typically increase the Hb level by 10 g/​litre (haematocrit by
3%). RBCs are administered through a transfusion administration-​
infusion set containing a standard screen filter designed to remove
particles that are over 150 µm in size.
Platelets
Platelets are in most cases a by-​product of red cell and plasma separ-
ation. When manufactured from whole blood they are called platelet
concentrates. In the United States of America, platelets are prepared
by the platelet-​rich plasma method, whereas the buffy coat method
is used in Europe (Fig. 22.8.1.1). Each unit of ‘random donor’ plate-
lets prepared by differential centrifugation of a single whole-​blood
collection typically contains at least 5.5 × 1010 platelets suspended
in 50 ml of plasma or additive solution, the latter used mainly in
Europe. Platelets stored under agitation at 20 to 24°C in plastic con-
tainers that allow oxygen diffusion have a shelf life of 5 days. The risk
of bacterial growth and development of platelet function abnormal-
ities (platelet storage defect) has precluded longer storage. However,
in the United States, all platelets are now tested for bacterial contam-
ination using culture or surrogate methods (see ‘Septic reactions’).
‘Random donor’ whole blood-​derived platelets are usually admin-
istered in pools of 4 to 6 units. In the absence of conditions associ-
ated with decreased platelet survival, each unit can be expected to
raise the recipient’s platelet count by 5000 to 10 000/​µl. Pooled and
stored, leucoreduced, whole blood-​derived platelets are available
Table 22.8.1.3  Uses of blood transfusion components
Component
Indication for use
Red blood cells
Symptomatic acute and chronic anaemias, and as the replacement product in
erythrocytopheresis (red cell exchange)
Red blood cells frozen and deglycerolized
Symptomatic anaemia, storage of red cells of rare antigen composition for up to 10 years
Leucocyte-​reduced components (red blood cells and platelets)
Symptomatic anaemia and thrombocytopenia, reduce febrile reactions from leucocyte
antibodies, alternative to CMV-​seronegative components, prevent HLA alloimmunization
Washed components (red blood cells and platelets)
Prevent graft-​versus-​host disease in immunocompromised patients
Irradiated blood products (red blood cells and platelets)
Remove harmful plasma antibodies
Platelet components (pooled platelets and pheresis platelets)
Thrombocytopenia with bleeding, prophylactic transfusion, platelet function
abnormalities
HLA matched/​selected platelets and crossmatch-​compatible platelets
HLA-​alloimmunized thrombocytopenic patients with decreased platelet survival
Fresh frozen plasma
Replacement of plasma coagulation factors for which specific factor concentrates are not
available, liver disease, DIC, hypofibrinogenaemia, TTP
Cryoprecipitate
Fibrinogen and factor XIII replacement, factor VIII and VWF replacement when
recombinant and virus-​inactivated concentrates are not available
Granulocytes by apheresis
Neutropenic patient with infection unresponsive to antibiotics
DIC, disseminated intravascular coagulation; TTP, thrombocytopenic purpura, VWF, von Willebrand factor.
section 22  Haematological disorders
5570
in the United States of America. Licensed by the Food and Drug
Administration (FDA) in 2006, these products are usually manufac-
tured in pools of 5 units and offer the benefit of allowing the use of
culture techniques to detect bacterial contamination. Additionally,
platelets can be collected and manufactured by apheresis resulting
in single donor platelets. These products contain more than 3 ×
1011 platelets suspended in about 200 ml plasma or additive solu-
tion and they are equivalent to 4 to 6 average random donor pooled
platelet units.
Platelets are not normally crossmatched with the recipient’s
serum. ABO type-​specific platelets should be provided whenever
possible because transfusing out-​of-​type platelets may result in poor
platelet survival in the patient’s circulation. Rh antigens present
on the small number of contaminating red cells found in platelet
concentrates are capable of immunizing a Rh-​negative recipient. If
Rh-​negative platelet concentrates are not available for a Rh-​negative
patient, Rh-​positive platelets can be transfused followed by adminis-
tration of Rh immune globulin within 72 h of transfusion.
Platelets are provided to thrombocytopenic patients who are
bleeding or to severely thrombocytopenic patients as a prophylactic
measure. Spontaneous bleeding is rare when a patient’s platelet
count is above 20 000/​µl, and studies suggest that patients who re-
ceive chemotherapy can tolerate platelet counts as low as 5000 to
10 000/​µl. Postsurgical patients may require platelet transfusions to
control or prevent postoperative bleeding when the platelet count is
over 50 000/​µl. Overall coagulation status should also be considered
because patients with plasma coagulation factor disorders are more
likely to bleed at marginal platelet counts. Actively bleeding patients
receiving antiplatelet agents such as aspirin or clopidogrel, irre-
versible inhibitors of platelet function, may require transfusions at
higher platelet counts; however, transfused platelets will also be af-
fected if the drug is not discontinued.
Platelet refractoriness is a major issue for patients who are de-
pendent on platelet transfusions. Immune and nonimmune fac-
tors may be responsible for platelet refractoriness. Common causes
of diminished platelet survival post transfusion include spleno-
megaly, disseminated intravascular coagulation, bleeding, medi-
cation, and sepsis. Patients may become refractory to platelet
transfusions through either HLA or platelet  alloimmunization.
Once platelet  alloimmunization is documented, crossmatch-​
compatible platelets or HLA-​matched platelets should be con-
sidered. However, these special products are not readily available
in most blood banks. Increasing the dose of standard platelet con-
centrates can be considered until compatible platelets are identified.
Leucocyte reduced blood products should be provided to patients
who will require many platelet transfusions to decrease the risk of
HLA alloimmunization.
Plasma, cryoprecipitate, and plasma derivatives
Plasma therapy started in the late 1940s when fractionation tech-
niques were developed to separate plasma proteins from large pools
of human plasma. Fresh frozen plasma is prepared by separating
plasma from whole blood by centrifugation and then freezing the
plasma within 8 h of collection. This process maintains the activity
of labile coagulation factors, particularly factors V and VIII. Many
blood suppliers also prepare plasma that is frozen within 24 h of
collection; this product is considered an effective alternative to
fresh frozen plasma in most instances when plasma transfusion is
necessary. Plasma should not be transfused for volume expansion
because of the risk associated with plasma including transfusion-​
transmitted disease, transfusion-​related lung injury, and allergic
transfusion reactions, and because of the availability of other,
safer nonplasma substitutes. The indications for plasma transfu-
sion include inherited deficiencies of coagulation factors when
no factor concentrate is available. However, the primary indica-
tion is for acquired coagulopathies, such as deficiency of multiple
coagulation factors seen in liver disease, dilutional coagulopathy,
and disseminated intravascular coagulation. It has been used to re-
verse warfarin anticoagulation urgently though prothrombin com-
plex concentrate is preferred. Plasma is not particularly effective
in replacing individual clotting factors because of the large vol-
umes that would be required to obtain adequate factor levels. The
patient’s fluid and cardiovascular status may preclude the use of
large amounts of plasma.
Fresh frozen plasma is no longer the treatment of choice for
coagulopathies where virally inactivated or recombinant blood
products exist, such as for deficiencies of factor VIII (haemophilia
A) or factor IX (haemophilia B). Fears of transmitting infectious
disease with plasma transfusion remain of concern, particularly for
pooled products. In addition to donor screening and testing, other
strategies to decrease infectious risk that have been studied include
photoinactivation and solvent detergent treatment. Furthermore, in
order to decrease the risk of transfusion-​related acute lung injury,
in the United Kingdom, only male donor plasma has been used as a
source of fresh frozen plasma since 2003. A similar trend has started
in the United States of America.
Cryoprecipitate is prepared by thawing fresh frozen plasma be-
tween 1 and 6°C. The precipitate forms and the supernatant is
removed and labelled as cryoprecipitate-​reduced plasma; both
products are subsequently refrozen. Each 10-​ to 20 ​ml unit of cryo-
precipitate contains 100 to 350 mg fibrinogen/​unit, at least 80 IU/​
unit factor VIII, FXIII, fibronectin, and von Willebrand factor.
Use of cryoprecipitate is generally reserved for patients with se-
vere hypofibrinogenaemia (<1 g/​l). Cryoprecipitate is not used to
treat haemophilia or von Willebrand disease in developed countries
because safer alternatives are available that avoid the risk of viral
transmission. Cryoprecipitate and thrombin have been combined to
make ‘fibrin glue’. This biological sealant works well but exposes the
recipient to the risks of transfusion-​transmitted disease because of
the use of cryoprecipitate. Safer sealants have been developed that do
not expose patients to cryoprecipitate.
Albumin is available as a 5 or 25% solution and is used to treat
hypovolaemia and hypoalbuminaemia, primarily in surgical set-
tings and as a replacement fluid in plasmapheresis. Albumin is
virally inactivated by heat treatment plus other viral inactivation
steps, and is tested for hepatitis C virus RNA. Properly processed
albumin is not considered to transmit viral disease. Readily avail-
able nonplasma colloidal solutions have replaced albumin in many
situations requiring volume expansion. Intravenous immuno-
globulin is used to treat patients with immune thrombocytopenia,
Guillain–​Barré syndrome, and autoimmune haemolytic anaemias.
Prompt and adequate doses of Rh (D) immunoglobulin available
in intramuscular and intravenous preparations, are used to pre-
vent alloimmunization in D-​negative patients who are exposed to
D-​positive red cells through transfusion or pregnancy. Rapid ad-
vances in molecular techniques led to the cloning and purification
22.8.1  Blood transfusion
5571
of recombinant clotting factors. Recombinant factors VIII, IX, and
VIIa are available.
Granulocytes
Granulocytes are transfused primarily to neutropenic oncology
patients with an absolute neutrophil count less than 500/​µl and a
reasonable chance of marrow recovery, who develop bacterial or
fungal sepsis unresponsive to antimicrobial therapy or in patients
with functional neutrophil disorder. Granulocytes collected from
nonstimulated healthy donors by apheresis contain at least 1 ×
1010 neutrophils/​unit and can be stored for only 24 h at 20 to 24°C.
Higher numbers of granulocytes can be collected when donors are
stimulated by steroids and/​or growth factors. The product contains
a large number of red cells (20–​50 ml) and must be crossmatched
with the recipient’s serum. Granulocytes should be irradiated be-
cause of the large number of lymphocytes present in the product.
Due to their short half-​life, granulocytes are usually provided daily
until the patient can maintain an absolute neutrophil count above
500/​µl without transfusion or until the infection resolves. Infusion
of larger numbers of granulocytes allows measurable increases in
recipient neutrophil counts, but the optimal dose and frequency re-
main undefined. Febrile reactions to granulocytes are common, in
addition to the pulmonary symptoms that are found to be more se-
vere when amphotericin is administered near the time of granulo-
cyte infusion. Other complications include HLA alloimmunization
and transfusion-​associated graft-​versus-​host disease, eliminated by
irradiating the product. Overall, the additional benefit of granulo-
cyte transfusion for neutropenic patients compared to antibiotic
treatment alone remains unclear. A  randomized trial, Safety and
Effectiveness of Granulocyte Transfusions in Resolving Infection
in People with Neutropenia (The RING Study), was initiated in
2008 and concluded in 2014. This trial attempted to assess high-​
dose granulocyte transfusions for the treatment of infection in
neutropenia. This trial was unable to accrue sufficient numbers of
patients to determine whether outcomes were improved with gran-
ulocyte transfusions. To date, support for utilization of granulocyte
transfusions remains anecdotal.
Complications and management
of transfusion therapy
See Tables 22.8.1.4 and 22.8.1.5.
Acute intravascular haemolytic reactions
Acute intravascular haemolytic transfusion reactions are one of the
most serious transfusion complications. ABO incompatibility re-
mains the most common cause of immediate intravascular haemo-
lytic reactions. Donor erythrocytes carrying either A and/​or B red
cell antigens bind to the recipient’s anti-​A and/​or anti-​B antibodies,
resulting in complement fixation, formation of the C5b-​9 membrane
attack complex, and subsequent haemolysis. Biological response
modifiers, such as proinflammatory cytokines (interleukin (IL)-​1,
tumour necrosis factor α (TNFα)), chemokines (IL-​8), and com-
plement fragments (C3a, C5a), also play a role in the pathophysi-
ology of acute transfusion reactions. The sudden onset of back pain,
hypotension, tachycardia, fever, chills, diaphoresis, and dyspnoea
are clinical characteristics of acute intravascular transfusion re-
actions. The symptoms usually begin soon after the transfusion is
started. Laboratory studies reveal an increase in unconjugated bili-
rubin and a marked elevation of lactate dehydrogenase. Other evi-
dence of intravascular haemolysis includes haemoglobinuria and
haemoglobinaemia. The direct antiglobulin test (direct Coombs’
test) becomes reactive due to the coating of donor red cells with the
recipient’s antibodies.
Acute intravascular haemolytic transfusion reactions are usu-
ally caused by transfusions of ABO-​incompatible blood resulting
from patient identification or clerical errors, but they can also be
caused by incompatibility within other blood group (e.g. Duffy,
Kidd) systems. Proper labelling of samples used by the blood bank
for compatibility testing and careful identification of patients are
the best ways to prevent these potentially fatal reactions. Acute
haemolytic immune transfusion reactions are medical emergen-
cies and treatment consists of immediately stopping the transfu-
sion, close monitoring of vital signs, cardiac and airway support,
and maintenance of urine output with saline diuresis with or
without a loop diuretic. Dialysis should be considered in patients
with renal failure.
Delayed extravascular haemolytic reactions
Delayed haemolytic transfusion reactions occur in patients who
have a negative antibody screen on pretransfusion testing, but
who then experience accelerated destruction of transfused red
cells 3 to 14 days after transfusion. In most cases, red cell destruc-
tion is caused by an antibody that is initially of a titre below the
limits of detection on routine screening. The antibody then rap-
idly forms on re-​exposure to the offending antigen (Fig. 22.8.1.3).
The antibodies typically fix complement to C3 and stop, thus re-
sulting in extravascular haemolysis. Antibodies most commonly
implicated in delayed transfusion reactions are directed against
Rh (E, c), Kell, Duffy, and Kidd blood group antigens. Delayed
extravascular haemolytic transfusion reactions can be diagnosed
by an unexpected post-​transfusion fall in haematocrit, develop-
ment of unconjugated hyperbilirubinaemia, and appearance of a
positive direct antiglobulin test. A delay of 3 days to 2 weeks is
usually seen between transfusion and the onset of extravascular
Table 22.8.1.4  Major risks of blood transfusion therapy
Immune complications
Nonimmune complications
Acute haemolytic transfusion
reactions
Transfusion-​associated bacterial sepsis
Delayed extravascular
haemolytic reaction
Circulatory overload, cardiac failure
Febrile transfusion reaction
Viral transmission (hepatitis A, B, C, CMV,
parvovirus)
Allergic transfusion reaction
(urticaria and anaphylaxis)
Iron overload
Transfusion-​associated sepsis
Hypocalcaemia
Alloimmunization
Hypothermia
Transfusion-​associated graft-​
versus-​host disease
Dilutional coagulopathy due to factor
depletion, thrombocytopenia
Transfusion-​associated acute
lung injury
section 22  Haematological disorders
5572
haemolysis. Only rarely do delayed reactions result in intravas-
cular haemolysis.
Febrile nonhaemolytic reactions
Febrile nonhaemolytic transfusion reactions to RBC and platelet
transfusion are very common. They are classically attributed to the
development of antibodies in the recipient directed against HLA and/​
or leucocyte-​specific antigens on donor white blood cells and plate-
lets. Reactions between leucoagglutinins present in the transfused
product and recipient leucocyte antigens can also occur. Subsequent
formation of leucocyte antigen–​antibody complexes results in com-
plement binding and release of endogenous pyrogens such as IL-​1,
IL-​6, and TNFα. Cytokines generated by leucocytes during platelet
and red cell storage may also contribute to these febrile reactions to
transfusion. Symptoms may occur during or several hours after the
transfusion and typically include fevers (>1°C rise) accompanied
by shaking chills. Rarely, vomiting, dyspnoea, hypotension, and de-
creased oxygen saturation may develop. The severity of symptoms
is often directly related to the number of leucocytes in the product
or the rate or volume of transfusion. Leucoreduction of blood com-
ponents decreases the frequency of febrile transfusion reactions.
Premedication with an antipyretic can ameliorate mild febrile trans-
fusion reactions. Corticosteroids can also minimize febrile trans-
fusion reactions if they are administered several hours before the
transfusion. Intramuscular or subcutaneous meperidine will usually
resolve severe rigors within minutes. If symptoms do not resolve in
less than 4 h or are especially severe, other complications such as
sepsis due to contaminated blood products or a haemolytic reaction
should be considered.
Allergic reactions
Allergic reactions to plasma, platelets, and RBCs are relatively
common. They present as pruritus and/​or urticaria in the absence
of fever. Allergic reactions are IgE mediated and most symptoms are
attributed to histamine release. It may be difficult to distinguish al-
lergic and febrile transfusion reactions when urticarial symptoms
are accompanied by low-​grade fever. Common symptoms and signs
include erythema, papular rashes, weals, and pruritus. As in other
allergic responses, symptoms are not dose related and severe mani-
festations can occur following small exposures. Treatment of mild
allergic reactions consists of stopping the transfusion and adminis-
tering antihistamines. In a mild allergic reaction with only pruritus
and hives, it is acceptable to continue transfusing the same unit, pro-
vided the symptoms promptly resolve and no accompanying fever or
vasomotor instability is noted.
Severe allergic reactions with bronchospasm and cardiovas-
cular collapse are rare and should be treated like any other ana-
phylactic reaction with steroids, vasopressors, and airway support.
Anaphylactic transfusion reactions occur in IgA-​deficient patients
who have already developed anti-​IgA antibodies, and then re-
ceive plasma-​containing blood products. While IgA deficiency is
common in the general population (1 in 700 individuals), only a
subset of IgA-​deficient individuals are at risk since not all of them
develop the antibody. Patients with IgA deficiency who have had an
anaphylactic reaction or who have demonstrated anti-​IgA should
receive cellular products that have been saline-​washed and plasma
from only IgA-​deficient donors. Washed RBCs may also benefit pa-
tients without IgA deficiency, but who have experienced repeated
moderate to severe allergic transfusion reactions.
Septic reactions
Blood products can become contaminated by bacteria if a donor is
bacteraemic at the time of collection or if the donor’s arm is im-
properly cleansed before venepuncture. Transfusing blood products
contaminated by bacteria is particularly dangerous and can result
in profound hypotension and shock. The risk of septic transfusion
reactions is higher for platelet transfusions than other blood com-
ponents because platelets are stored at room temperature. As noted
previously, in an attempt to reduce the risk of transfusion-​associated
bacterial sepsis, blood collection facilities in the United States have
implemented several strategies to detect bacterial contamination of
platelet units. These include culture of the product as well as sur-
rogate methods. The latter comprise (1) visual inspection for loss
of swirling that occurs with a change in platelet shape associated
with the fall in pH; (2) direct visualization of microorganisms using
the Gram stain, Wright stain, or acridine orange; and (3) the use of
Table 22.8.1.5  Symptoms, signs, and management of transfusion reactions
Reaction
Symptoms and signs
Management/​treatment
Acute intravascular
haemolytic reaction
Back pain, fever, hypotension, shock, dyspnoea, haemoglobinuria,
haemoglobinaemia, positive direct Coombs’ test
Stop transfusion, IV fluids, vasopressor support, maintain
diuresis, corticosteroids, dialysis if indicated
Delayed extravascular
haemolytic reaction
Anaemia, jaundice, fever, positive direct Coombs’ test
Stop transfusion, fluid support, follow lab results (haematocrit,
lactate dehydrogenaase, bilirubin)
Febrile reaction
Fever, chills, rigors, mild dyspnoea
Stop transfusion, antipyretics, consider leucoreduced product
for subsequent transfusions
Allergic (mild)
Pruritus, urticaria
Antihistamines, may continue transfusion if symptoms improve
in <30 min, otherwise stop transfusion
Allergic (anaphylactic)
Urticaria, bronchospasm, dyspnoea, nausea, hypotension
Stop transfusion, antihistamines, vasopressor support,
corticosteroids, consider premedication or washed RBCs for
subsequent transfusions
Septic reaction
Rapid onset of chills, fever, hypotension
Stop transfusion, culture sample from product and patient,
vasopressor support, IV fluids, broad spectrum antibiotics
Transfusion-​related
acute lung injury
Dyspnoea, tachypnoea, cyanosis, fever, hypotension
Respiratory support
22.8.1  Blood transfusion
5573
dipsticks for pH and glucose readings. These surrogate methods are
more rapid and less costly, but also much less sensitive than culture
and are being phased out. The sensitivity of culture in detecting bac-
terial contamination of blood products is affected by several factors,
however, such as growth characteristics of the organism, timing of
specimen collection, specimen volume, and the degree of initial bac-
terial contamination.
Organisms commonly implicated in septic transfusion reac-
tions include Gram-​positive (staphylococcus) and Gram-​negative
(enterobacter, yersinia, pseudomonas) bacteria. The literature
demonstrates that the risk of bacterial contamination on the day of
transfusion for apheresis platelets previously screened as negative
by early culture is approximately 1:5000. The risk of septic trans-
fusion reactions caused by contamination of platelets is approxi-
mately 1:107 000. Blood cultures should be obtained from patients
who develop high fevers following or during transfusion, espe-
cially if they become hypotensive. A Gram stain of the suspected
contaminated product may be helpful but is often negative, and the
product should be cultured if possible. Other symptoms attributed
to preformed endotoxin and cytokines include skin flushing, severe
rigors, and rapid-​onset cardiovascular collapse. The symptoms may
occur during or a minute to hours after the transfusion is completed.
Treatment includes fluids, cardiorespiratory support, and broad-​
spectrum antibiotics.
Transfusion-​related acute lung injury
Transfusion-​related acute lung injury is a serious complication of
blood transfusion that presents as noncardiogenic pulmonary oe-
dema. It typically occurs within 6 h of transfusion and is clinic-
ally similar to the acute respiratory distress syndrome. The most
common clinical findings are rapid-​onset dyspnoea, tachypnoea,
cyanosis, fever, and hypotension. Lung auscultation reveals diffuse
crackles and decreased breath sounds. Invasive cardiac monitoring
demonstrates normal cardiac pressures and function with hypox-
aemia and decreased pulmonary compliance. Radiographic findings
include diffuse, fluffy infiltrates typical of pulmonary oedema.
In most cases, the aetiology is believed to involve an immune-​
mediated reaction between passively transferred donor antileucocyte
antibodies present in a plasma-​containing blood product with the
recipient’s white cells, resulting in leucocyte activation. Much less
frequently, the antibodies present in the recipient may react with
white cells in the transfused products. Granulocytes are first acti-
vated by HLA or other antigen–​antibody complexes and then mi-
grate to the lungs. The activated neutrophils bind to the pulmonary
capillary bed via cell adhesion molecules where they release pro-
teolytic enzymes that destroy tissue, resulting in a capillary leak
syndrome and pulmonary oedema. More recently, reactive lipid
products released from donor cell membranes have been associated
with the development of transfusion-​related lung injury.
Transfusion-​related acute lung injury is a clinical diagnosis and
should be suspected in patients with severe, rapid-​onset respira-
tory distress during or soon after transfusion therapy. Definitive
diagnosis requires identification of HLA and/​or granulocyte anti-
bodies in either the donor’s or recipient’s serum, as well as the cor-
responding antigens on the recipient’s or donor’s leucocytes. This
testing is performed in only a few specialized laboratories. Most
patients with this syndrome will survive with supportive care,
including aggressive respiratory support with supplemental oxygen,
and if necessary, mechanical ventilation. Often, the hypoxia that
develops during or after transfusion is attributed to fluid overload,
and diuretics are empirically administered. Although not absolutely
contraindicated, diuretics may be harmful and should be used with
extreme caution. Corticosteroids have no proven role in the man-
agement of transfusion-​related acute lung injury. As discussed pre-
viously, in order to reduce its incidence, plasma from female donors
is no longer used as a source of fresh frozen plasma in the United
Kingdom, with the United States of America beginning to follow this
course as of 2007.
Transfusion-​associated circulatory overload
Transfusions contribute to fluid overload that result in pulmonary
oedema and hypertension. Transfusion-​associated circulatory
overload is often unrecognized and underreported; the docu-
mented incidence is between 1 and 6%. The reaction should occur
during or within 6 h of transfusion. Patient develops respiratory
symptoms, hypoxaemia with a reduction in O2 saturation, tachy-
cardia, hypertension, and pulmonary/​pedal oedema. Chest radiog-
raphy will show bilateral infiltrates and patient will have an elevated
brain natriuretic peptide. Risk factors include extremes of age, his-
tory of congestive heart failure, and renal failure. It is prevented by
a slow transfusion rate and closely monitoring patient fluid status
and it is managed by stopping the transfusion, respiratory support,
and diuretics.
Transfusion-​associated dyspnoea
The diagnosis of transfusion-​associated dyspnoea is considered if
the patient develops respiratory distress and did not meet the criteria
for transfusion-​related acute lung injury, transfusion-​associated cir-
culatory overload, or an allergic reaction and the symptoms are not
explained by the patient’s underlying condition.
Transfusion-​associated graft-​versus-​host disease
Acute graft-​versus-​host disease (GVHD) is a rare complication of
blood transfusion, but is fatal in approximately 90% of patients.
Transfusion-​associated graft-​versus-​host disease (TA-​GVHD) oc-
curs when donor immunocompetent T and NK cells attack im-
munocompromised recipient cells because these recipient cells
appear foreign due to differences in major or minor histocompati-
bility antigens. The risk of TA-​GVHD is related to the number of
viable T lymphocytes transfused, the recipient’s immune status, and
the HLA disparity between donor and host. Therefore, multiply
transfused patients who receive cells from donors who share HLA
haplotypes with the recipient (i.e. blood relatives) are at greatest risk.
Clinically, TA-​GVHD is characterized by the acute onset of rash, ab-
dominal pain, diarrhoea, liver abnormalities (elevated liver enzymes,
hyperbilirubinaemia), and bone marrow suppression 2 to 30 days
following transfusion. The maculopapular rash seen is similar to that
observed in acute GVHD following bone marrow transplant, and
biopsy of the skin may help confirm the diagnosis. Pancytopenia
may be severe and is attributed to destruction of recipient marrow
stem cells by donor lymphocytes. Immunosuppressive therapy
with prednisone and ciclosporin has had little effect on TA-​GVHD.
Fortunately, the development of this condition can be prevented by
irradiating cellular blood products before transfusion.
section 22  Haematological disorders
5574
Acute pain transfusion reaction
This reaction occurs shortly after the initiation of transfusion, the
mechanism is currently unknown. Patients complain of chest, ab-
dominal, and back pain, and pain in the proximal extremities that
resolves within an hour of stopping the transfusion. It can be associ-
ated with other symptoms including dyspnoea, hypertension chills,
and headache. These symptoms can be manifested in other transfu-
sion reactions and they should be ruled out prior to establishing a
diagnosis of acute pain reaction.
Hypotensive transfusion reaction
Hypotensive transfusion reaction is a newly recognized
complication—​if not recognized early it can be life-​threatening.
It is characterized by the onset of clinically significant hypoten-
sion (systolic <89 and 30 mmHg reduction in systolic blood pres-
sure) during or within 1 h of transfusion. It can be associated with
other symptoms, such as facial flushing, dyspnoea, and abdominal
pain. Other transfusion reactions, such as allergic, acute haemo-
lytic, and septic reactions should be ruled out. It has been asso-
ciated with the use of negatively charged bedside leucoreduction
filters and it is reported in patients using an angiotensin converting
enzyme inhibitor, indicating that bradykinin that is produced by
the activation of the contact system is a possible causative agent.
Hypotensive transfusion reaction is managed with fluids, placing
the patient in the Trendelenburg position and vasopressors. Patient
should improve with supportive measures and after the cessation
of transfusion.
Post-​transfusion purpura
Post-​transfusion purpura patients will develop thrombocyto-
penia with counts decreased to at least 20% of pretransfusion
count. This occurs 5 to 12 days following transfusion of platelets,
RBCs, or plasma. Patient will present with purpura, mucosal
bleed, epistaxis, urinary, and gastrointestinal bleeding. Post-​
transfusion purpura occurs due to the presence of platelet-​specific
alloantibodies present in patients due to previous sensitizations
by transfusion or pregnancy, most commonly antihuman platelet
antigen 1a (anti-​HPA1a). A subsequent exposure will trigger the
destruction of both the transfused and, importantly, autologous
platelets, as well. Antigen negative or antigen positive platelet
transfusions do not effectively increase the platelet count. Patients
are managed with corticosteroids, plasmapheresis, and intravenous
immunoglobulin.
Transfusion-​transmitted disease
Despite major improvements in blood safety during the past two
decades, a small risk of transfusion-​transmitted disease still remains.
The use of volunteer donors and predonation screening question-
naires were the first steps taken to reduce the risk of transfusion-​
related hepatitis and HIV. These risks continue to drive mandated
pretransfusion testing requirements in developed countries. The
advent of enzyme immunoassays in the 1970s and more recently,
nucleic acid amplification testing, have further decreased the risk
of transfusion-​transmitted disease (Table 22.8.1.6). Transfusion-​
transmitted disease is a persistent problem in parts of the world that
do not have access to screening tests.
At present, pretransfusion testing in the United States of America
and Europe includes screening for hepatitis B (HBsAg, anti-​HBc,
nucleic acid amplification), hepatitis C (anti-​hepatitis C virus
(HCV), nucleic acid amplification), HIV (anti-​HIV-​1/​2, nucleic
acid amplification), human T-​cell lymphotropic virus (anti-​HTLV-​
I/​II), and syphilis (RPR). Additionally, in the United States, donated
blood is screened for West Nile virus (nucleic acid amplification)
and Trypanosoma cruzi as a one-​time donor testing. Nucleic acid
amplification testing for HCV and HIV is typically performed on
small pools of donor samples.
The current estimate of the risk of transfusion-​related HIV is ap-
proximately 1 in 2 million units transfused. With the introduction
of screening by amplification of nucleic acid templates, the ‘window
period’, in which the virus could be transmitted by an HIV-​infected
Table 22.8.1.6  Organisms potentially transmitted by blood transfusion
Agent/​organism
Estimated risk per unit transfuseda
Pretransfusion testing
Hepatitis B virus
1:280 000
HBsAg, anti-​HBc, ALT, NAT
Hepatitis C virus
1:1.9 million
Anti-​HCV, NAT
HIV-​1/​2
1:2.1 million
Anti-​HIV-​1/​2, NAT
HTLV-​I/​II virus
1:650 000
Anti-​HTLV-​I/​II
West Nile virus
Unknown
NAT
CMV
1:10–​1:20 (see text)
Some units tested for anti-​CMV antibodies
Parvovirus B19
Unknown
None
Bacterial contamination
1:1500
None
Treponema pallidum
Rare
RPR
Parasites (plasmodium, ehrlichia, Babesia microti)
Rare
None
Trypanosoma cruzi
1:25 000 (seroprevalence)
Anti-​Trypanosoma cruzi
vCJD
Unknown
Deferral based on history
CMV, cytomegalovirus; NAT, nucleic acid testing; vCJD, variant Creutzfeldt–​Jakob disease;
a USA figures.
22.8.1  Blood transfusion
5575
but seronegative donor, has decreased to approximately 10  days.
Genomic testing for HCV RNA has also been implemented in
the United States and Europe to detect seronegative yet infectious
units. Nucleic acid amplification testing screening has decreased
the transfusion-​related hepatitis C risk by decreasing the window
period to 10 to 20 days.
The first cases of transfusion-​transmitted West Nile virus
infection were documented in the United States of America in
2002; the following year, national blood donation screening for
West Nile virus was initiated using nucleic acid amplification
testing technology. The risk of transfusion-​related transmission
of this virus since instituting this screening test has not been
established.
Several techniques have been developed to inactivate viruses in
blood products; see ‘Pathogen reduction’ for more details.
Cytomegalovirus (CMV) and parvovirus B19 are common in
the general donor population, and may pose a serious threat to im-
munocompromised patients. Approximately 40 to 60% of blood
donors have been exposed to CMV during their lifetime and sub-
sequently are CMV seropositive. However, only about 2% of CMV-​
seropositive donors are actively infected and transfusing their blood
to an immunocompromised recipient could cause acute CMV infec-
tion. The actual risk of post-​transfusion seroconversion of a CMV-​
negative recipient who receives CMV-​untested blood depends on
the prevalence of CMV seropositivity in the donor population.
As for parvovirus B19, only a few cases of transmission by blood
components have been reported in immunocompetent recipients.
Thus, blood donor screening tests for parvovirus have not been
recommended.
Since its initial description in the United Kingdom in 1996, variant
Creutzfeldt–​Jakob disease (vCJD) has raised additional transfusion
safety concerns. The observations from the study conducted in the
United Kingdom, the Transfusion Medicine Epidemiology Review,
have provided evidence that vCJD can be transmitted through blood
transfusion. As of 2006, of the 66 British patients identified as having
received blood from donors who went on to develop vCJD, three
or four probable cases of transfusion-​transmitted vCJD have been
documented. These numbers may, however, be an underestimate of
the overall risk of transfusion transmission of this disease. Given the
long incubation period of vCJD, some surviving recipients of blood
products derived from ‘vCJD donors’ may still develop the disease.
Moreover, a significant number of deceased blood recipients may
not have survived long enough to manifest clinical disease even if
infected. The introduction of universal leucocyte depletion of the
United Kingdom blood supply in 1999 may have reduced the risk to
blood recipients.
A number of parasitic diseases are known or suspected to be
transmitted by blood transfusion. These include malaria, Chagas’
disease, babesiosis, anaplasmosis, leishmaniasis, and toxoplas-
mosis. Transmission of Lyme disease (Borrelia burgdorferi) by
transfusion has not been documented. Infection with babesia, if
untreated, can be dangerous in at-​risk populations such as asplenic
patients. A screening test for babesiosis is available in the United
States of America. Testing for T. cruzi was not initially mandated,
however, blood centres started testing donors with an EIA test ap-
proved by the FDA in 2007. In 2010, one-​time screening of every
donor was recommended by the FDA. At this time, blood centres
maintain varying algorithms regarding the testing of T. cruzi. These
algorithms range from testing first time donors only, to testing first
time donors and retesting donors who report having travelled to en-
demic areas of Central and South America.
Use of special blood products
Leucoreduction
Leucocytes contained in blood components can provoke febrile
nonhaemolytic reactions, induce HLA alloimmunization, and
transmit CMV to at-​risk recipients. Leucocytes are effectively
removed from red cell and platelet concentrates by leucocyte re-
duction filters. American standards require that units labelled
‘leucoreduced’ contain less than 5 × 106 white blood cells, whereas
the European standard is less than 1 × 106 white blood cells per
unit. Red cells are most commonly leucoreduced shortly after
blood collection (prestorage leucodepletion). Filters are similarly
used to leucoreduce platelet concentrates. Apheresis devices have
been designed to collect leucoreduced platelets directly (process
leucoreduction).
Leucoreduction has been shown to decrease the prevalence
and severity of febrile transfusion reactions and the risk of HLA
alloimmunization. Other generally accepted benefits of white blood
cell reduction include reducing platelet refractoriness and decreasing
the risk of transmitting white blood cell-​related infectious agents
including CMV, HTLV-​I/​II, ehrlichia, and anaplasma. Prestorage
leucoreduced products are preferable because they contain less cyto-
kines and other biological response modifiers produced by white
blood cells. With the dramatic decrease in the risk of viral transmis-
sion, investigators are focusing on the immunomodulatory effects of
blood transfusion. These effects specifically deal with associations
between allogeneic transfusion and bacterial infection, tumour pro-
gression, and tumour recurrence. Universal leucoreduction of both
RBCs and platelets has been required and/​or is being implemented
in a number of European countries and in parts of North America.
Universal leucoreduction is not FDA mandated in the United States
of America.
Irradiation
Blood components are irradiated to prevent potentially lethal
TA-​GVHD by interfering with the ability of donor lymphocytes
to proliferate. GVHD occurs in immunocompromised recipi-
ents when an immunocompetent donor is homozygous for an
HLA haplotype and the recipient is heterozygous for that haplo-
type. The immunocompetent donor lymphocytes will recognize
the recipient as foreign and mount an immune response leading
to GVHD. Irradiation of blood components is indicated for bone
marrow or peripheral blood stem cell transplant recipients, pa-
tients with congenital immunodeficiency states, neonates, pre-
mature infants, during intrauterine exchange transfusion, when
transfusing (seemingly) HLA-​compatible platelet units and blood
products from a blood relative. Patients with AIDS commonly re-
ceive irradiated components, although no clear increased risk of
TA-​GVHD exists in this population. Standard guidelines recom-
mend irradiating RBCs, platelets, and granulocytes with a min-
imum dose of 2500 cGy. Platelets are not adversely affected by this
exposure. Red cells’ shelf life, however, is shortened to 28 days after
irradiation. This is due to the irradiation causing changes in the red
section 22  Haematological disorders
5576
cell membrane that induces a leak of intracellular potassium. It is
not necessary to irradiate frozen noncellular blood products such
as fresh frozen plasma or cryoprecipitate because they contain very
few viable leucocytes.
Bone marrow or peripheral blood stem cells must never be irradi-
ated prior to transplant.
Cytomegalovirus-​safe components
CMV infection is a leading cause of morbidity and mortality in
marrow and solid-​organ transplant patients. Most serious CMV in-
fections that develop in these populations are a result of latent re-
activation of recipient CMV, but CMV can also be transmitted by
blood transfusion. Therefore, blood banks supply products that have
a low potential of transmitting CMV. The available products include
CMV-​seronegative units prepared from donors who are CMV IgG
antibody negative. However, seroprevalence of CMV in the popu-
lation ranges from 40–​80% and it is logistically difficult to provide
a sufficient quantity of this product. The other available products
are leucodepleted components. The latter refers to blood compo-
nents leucoreduced in a blood centre or laboratory using ‘good
manufacturing practice’ techniques. Studies suggest that CMV-​
seronegative and leucodepleted filtered products are equivalent in
preventing CMV transmission. Many transfusion specialists con-
sider leucodepleted units produced under conditions of good manu-
facturing practice as CMV ‘safe’ in that they are unlikely to transmit
CMV disease. In addition to CMV-​seronegative marrow and solid-​
organ transplant recipients, CMV-​seronegative or safe components
are generally indicated for premature infants, during intrauterine
transfusions, for patients with congenital immunodeficiencies,
CMV-​seronegative pregnant women, and seronegative patients with
HIV. The British Committee for Standards in Haematology has con-
cluded that leucoreduced components are an ‘effective alternative’ to
seronegative products for preventing CMV transmission by transfu-
sion. In addition, pathogen reduction technology, recently approved
by FDA for use in the United States of America, also reduces the risk
of CMV transmission.
Washed blood products
RBCs and platelets can be washed to remove the plasma that con-
tains proteins and cytokines, and replace it with saline. It is per-
formed for patients with repeated severe allergic/​anaphylactic
reactions and in neonatal alloimmune thrombocytopenia patients
receiving maternal platelets, to remove the maternal alloantibodies.
Approximately 20 to 30% of the red bloods cells or platelets are lost
during this process. RBCs must be transfused within 24 h and plate-
lets within 4 h of washing. Considering the loss of RBCs and/​or
platelets during the washing process, it is important to evaluate the
clinical necessity of washing.
Volume reduction
This process is performed by a centrifugation step and the re-
moval of the supernatant to concentrate the product. Volume re-
ducing of RBCs is performed to transfuse patients who are prone to
volume overload and to prevent hyperkalaemia when older stored
red cells are transfused to neonates or young children. Volume-​
reduced platelets can be resuspended in saline and should be trans-
fused within 4 h. The indication for platelet volume reduction is
transfusion to patients who are prone to circulatory overload, for
out-​of-​ABO-​group platelet transfusions and to reduce the incidence
of febrile nonhaemolytic transfusion reactions by decreasing the
cytokines that accumulate in plasma during storage.
Frozen products
Both RBCs and platelets can be cryopreserved and frozen. The RBC
preservation process is widely used in the United States of America
to store rare phenotyped red cell units and autologous blood collec-
tions. RBCs are cryopreserved with glycerol to prevent dehydration
and frozen at less than −65°C. This process should be performed
within 6 days of collection. Once the RBC products are frozen, they
can be stored for up to 10 years. Prior to transfusion the product
need to be thawed at 37°C, deglycerolized-​washed. Platelet cryo-
preservation, using DMSO, is investigational only and not licensed
for use in the United States of America.
Pathogen reduction
Transfusion-​transmitted infections from known and emerging
pathogens pose a risk to the blood supply. With donor screening
and use of a donor history questionnaire, coupled with serological
testing, and nucleic acid testing, the risk of transfusion-​transmitted
infections has decreased, but has not been completely eliminated.
Newer technologies, such as pathogen reduction, will reduce this
risk further by inactivating both intracellular and extracellular
agents including viruses, parasites, and bacteria. Advantages of
pathogen reduction include the potential to eliminate the need for
future additional donor infectious disease testing, decreasing donor
deferrals and extending the limited shelf life of platelets to 7 days due
to the decreased risk of bacterial contamination with treated plate-
lets stored at room temperature. There are several available methods
of pathogen reduction which have been approved for clinical use
by the FDA. Other pathogen-​reduction technologies are in various
stages of development for use with either whole blood plasma or cel-
lular blood components.
One licensed pathogen-​reduction method targets cell membranes
by using a solvent and a detergent to attack and damage the cell
membrane of pathogens. This technology is thus applicable only to
plasma products and not to cellular blood components. Methods
used for pathogen reduction of plasma, besides solvent/​detergent
treatment, include methylene blue light treatment, and an FDA-​
approved psoralen–​ultraviolet (UV)-​A light treatment technology.
In addition, plasma can be inactivated using riboflavin (vitamin B2)
light treatment.
Most technologies for pathogen reduction of cellular blood com-
ponents target nucleic acids, preventing the replication of pathogens
and leucocytes. Several techniques have been developed to inactivate
pathogens in platelets. One FDA-​approved methodology, uses psor-
alen and UV-​A light to inactivate pathogens in units of single donor
platelets. This is the same FDA-​approved technology described
previously for use in pathogen reduction of plasma. For platelets,
riboflavin–​light treatments, currently under investigation, have also
been shown to inactivate many pathogens known to be transmitted
by transfusion. The currently approved method in the United States
of America, however, only utilizes the psoralen–​UV-​A technology.
Methods to inactivate infectious pathogens in red cells are currently
under development. A solvent/​detergent approach cannot be used as
it would adversely affect RBC membranes, nor can a psoralen–​UV-​
A approach be used as the haemoglobin in red cells blocks the UV-​A
22.8.1  Blood transfusion
5577
light from activating the psoralen compound drastically reducing
the effectiveness of the system to inactivate pathogen nucleic acids.
Albumin, immune globulin, factor concentrates, and other plasma
derivatives are treated using solvent/​detergent and other protocols
that essentially eliminate the risk of viral transmission. Pathogen-​
reduction technologies are safe, nontoxic, and achieve an adequate
level of pathogen inactivation while maintaining cellular quality and
adequate levels of functional clotting factors. It should be noted that
current technologies and most pathogen-​reduction technologies
currently available and under development are ineffective against
spores or prions.
Alternatives to blood component therapy
Autologous transfusion
Commonly used forms of autologous transfusion include preopera-
tive blood donation, acute normovolaemic haemodilution, and
autologous blood salvage. Many blood centres provide autologous
preoperative blood donation services in which a patient’s blood is
drawn and stored for later use, usually during a surgical procedure.
The criteria for autologous donations are less stringent than those
for allogeneic donors. Preoperative blood donation can be utilized
in elderly patients, although there is a higher risk of anaemia and
more serious cardiovascular complications associated with the do-
nation. Although the use of autologous blood decreases the risk of
viral infection, the risk of bacterial contamination remains. Acute
normovolaemic haemodilution is performed by removing blood
from a patient immediately before surgery and replacing the blood
volume with crystalloid or colloid solutions to maintain haemo-
dynamic stability. The withdrawn blood is then later reinfused.
Autologous blood salvage is performed by collecting and then re-
turning blood lost during or shortly following operative procedures
using intraoperative salvage devices. This technique is primarily
used in cardiac and orthopaedic surgery.
Growth factors
Haematopoietic growth factors used in transfusion therapy are
designed to limit the exposure of patients to allogeneic blood.
The isolation, characterization, and subsequent synthesis of
erythropoietin by recombinant technology were important ad-
vances in decreasing red cell transfusions. The use of recom-
binant human erythropoietin has reduced the transfusion needs
of some patients with renal failure and various anaemias. In the
United States of America, use of recombinant human erythro-
poietin is being restricted due to reported adverse vascular and
other events. Granulocyte colony-​stimulating factor has been
shown to decrease infection rates in neutropenic patients under-
going chemotherapy, replacing marginally effective granulocyte
transfusions. Thrombopoietic growth factors, such as recombinant
thrombopoietin, as well as small molecules with thrombomimetic
activity, are currently being evaluated and some formulations are
licensed in the United States of America.
Blood substitutes
For over a century, ongoing research has sought to develop
haemoglobin-​based oxygen-​carrying compounds that can serve
as an alternative to allogeneic red cell transfusion. The earliest
products consisted of stroma-​free haemoglobin, which was aban-
doned because of its renal toxicity, and polymerized haemoglobin.
Most of these agents are not used clinically because of vasoactivity
and other untoward effects; other formulations are in various
phases of clinical trials. No formulation is currently licensed in the
United States of America.
Molecular testing in the blood bank
The molecular basis of blood group antigens has been extensively
studied and many of the genes that code for antigens on RBCs,
platelets, and leucocytes have sequenced. Molecular-​based typing
in the blood bank is referred to as genotyping, as compared to the
current gold standard serological typing via haemagglutination
known as phenotyping. Genotyping has been more utilized in re-
cent years. However, it will not soon replace serological testing
because serology is inexpensive and less complex compared with
genotyping. In addition, the safety of most transfusions is assured
with current methods by means of ABO typing, screening, and
crossmatching. Genotyping red cells has its advantages in cer-
tain circumstances, for example, to resolve a typing discrepancy,
in positive direct antiglobulin testing, after multiple transfusions,
when typing reagents are not available for certain blood group
antigens, to identify a fetus at risk of haemolytic disease of the
newborn, and with patients in need of chronic transfusions with
an increased risk of alloimmunizations and matching their blood
is problematic. This population of patients includes sickle cell,
thalassemia, and oncology patients. Genotyping assays used in the
blood bank are polymerase chain reaction-​based assays and many
platforms have been developed in recent years. These assays have
their limitations. In addition to the complexity of these assays,
there are approximately 300 antigens identified and more than
1000 alleles coding them. A particular phenotype can result from
multiple genetic variations and one should have a full knowledge
of the different alleles present in a population prior to developing
a molecular-​based assay. At the current time, it is recommended
that DNA testing should be used as an adjunct to serological tests
especially to confirm negativity.
FURTHER READING
American Association of Blood Banks (AABB) (2011). Guidelines for
PBM and blood utilization AABB, Bethesda, MD.
American Association of Blood Banks (AABB) (2014). Technical
manual, 18th edition. AABB, Bethesda, MD.
Anstee DJ (2009). Red cell genotyping and the future of pretransfusion
testing. Blood, 114, 248–​56.
Ballen KK, et al. (2004). Autologous stem-​cell transplantation can be
performed safely without the use of blood-​product support. J Clin
Oncol, 22, 4087–​94.
BCSH Blood Transfusion Task Force (2007). Guidelines on gamma
irradiation of blood components for the prevention of transfusion-​
associated graft-​versus-​host disease. Transfus Med, 6, 261–​71.
Blajchman M (2006). The clinical benefits of the leukoreduction of
blood products. J Trauma, 60 Suppl 6, S83–​90.
section 22  Haematological disorders
5578
Blajchman M (2008). Platelet transfusions: an historical perspective.
Hematology Am Soc Hematol Educ Program, 2008, 197.
Carson JL, et al. (2017). Indications for and Adverse Effects of Red-Cell
Transfusion. N Engl J Med, 377(13), 1261–72.
Chou ST (2013). Transfusion therapy for sickle cell disease: a balancing
act. Hematology Am Soc Hematol Educ Program, 2013, 439–​46.
Cohn E, et al. (1950). A system for separation of components of human
blood. J Am Chem Soc, 72, 465–​74.
Department of Health and Human Services (2011). 2011 national
blood collection and utilization survey report. Department of Health
and Human Services, Washington DC.
Dobroszycki J, et  al. (1999). A cluster of transfusion-​associated
babesiosis cases traced to a single asymptomatic donor. JAMA,
281, 927–​30.
Dorsey KA, et al. (2013). A comparison of human immunodeficiency
virus, hepatitis C virus, hepatitis B virus, and human T-​lymphotropic
virus marker rates for directed versus volunteer blood donations
to the American Red Cross during 2005 to 2010. Transfusion, 53,
1250–​6.
Dzik WH (2007). New technology for transfusion safety. Br J Haematol,
136, 181–​90.
Friedberg RC, Donnelly SF, Mintz PD (1994). Independent roles for
platelet crossmatching and HLA in the selection of platelets for
alloimmunized patients. Transfusion, 34, 215–​20.
Goodnough LT (2003). Risks of blood transfusion. Crit Care Med, 31
Suppl 12, S678–​86.
Goodnough LT, et al. (1999). Transfusion medicine. Second of two
parts–​blood conservation. N Engl J Med, 340, 525–​33.
Goodnough LT, et  al. (2009). Implementation of a two-​specimen
requirement for verification of ABO/​Rh for blood transfusion.
Transfusion, 49, 1321–​8.
Hébert PC, et al. (1999). A multicenter, randomized, controlled clin-
ical trial of transfusion requirements in critical care. Transfusion
Requirements in Critical Care Investigators, Canadian Critical Care
Trials Group. N Eng J Med, 340, 409–​17.
Heddle N, Webert KE (2007). Febrile, allergic, and other noninfectious
transfusion reactions. In:  Hillyer CD, et  al. (eds) Blood banking
and transfusion medicine, 2nd edition, pp. 677−90. Churchill
Livingstone, Philadelphia.
Hewitt P (2006). vCJD and blood transfusion in the United Kingdom.
Transfus Clin Biol, 13, 312–​16.
Hod E, Schwartz J (2008). Platelet transfusion refractoriness. Br J
Haematol, 142, 348–​60.
Issitt PD, Anstee DJ (eds) (1998). Applied blood group serology, 4th edi-
tion. Montgomery Scientific, Durham.
Kaufman RM, et  al. (2015). Platelet transfusion:  a clinical practice
guideline from the AABB. Transfusion, 162, 205–​13.
Klein HG (2006). Transfusion-​associated graft-​versus-​host dis-
ease: less fresh blood and more gray (Gy) for an aging population.
Transfusion, 46, 878–​80.
Klein HG, Strauss RG, Schiffer CA (1996). Granulocyte transfusion
therapy. Semin Hematol, 33, 359–​68.
Kuter DJ, Begley CG (2002). Recombinant human thrombopoietin:
basic biology and evaluation of clinical studies. Blood, 100, 3457–​69.
Liles WC, et al. (1997). A comparative trial of granulocyte-​colony-​
stimulating factor and dexamethasone, separately and in combin-
ation, for the mobilization of neutrophils in the peripheral blood of
normal volunteers. Transfusion, 37, 182–​7.
Lo YM, et al. (1998). Prenatal diagnosis of fetal RhD status by mo-
lecular analysis of maternal plasma. N Engl J Med, 339, 1734–​8.
Looney MR, Gilliss BM, Matthay MA (2010). Pathophysiology of
transfusion-​related acute lung injury. Curr Opin Hematol, 17, 418–​23.
Looney MR, Gropper MA, Matthay MA (2004). Transfusion-​related
acute lung injury. Chest, 126, 249–​58.
Marcus DM (1969). The ABO and Lewis blood-​group system.
Immunochemistry, genetics and relation to human disease. N Engl J
Med, 280, 994–​1006.
Montgomery SP, et al. (2006). Transfusion-​associated transmission of West
Nile virus, United States 2003 through 2005. Transfusion, 46, 2038–​46.
Murphy MF (1999). New variant Creutzfeldt-​Jakob disease (nvCJD):
the risk of transmission by blood transfusion and the potential
benefit of leukocyte-​reduction of blood components. Transfus Med
Rev, 13, 75–​83.
Murphy MF, Yazer MH (2013). Measuring and monitoring blood util-
ization. Transfusion, 53, 3025–​8.
Novotny VM (2004). Prevention and management of platelet transfu-
sion refractoriness. Vox Sang, 76, 1–​13.
Paglino JC, et al. (2004). Reduction of febrile but not allergic reac-
tions to RBCs and platelets after conversion to universal prestorage
leukoreduction. Transfusion, 44, 16–​24.
Panch SR, et al. (2019). Hemolytic Transfusion Reactions. N Engl J
Med, 381(2), 150–62.
Popovsky M (2012). Transfusion reactions, 4th edition. American
Association of Blood Banks, Bethesda, MD.
Procop GW (2013). Strategies for appropriate test utilization:  the
right test for the right patient at the right time. Cleveland Clinic
Laboratories. https://​clevelandcliniclabs.com/​wp-​content/​assets/​
pdfs/​publications/​2014-​test-​utilization.pdf.
Race RR (1944). An ‘incomplete’ antibody in human serum. Nature,
153, 771–​2.
Reesink HW, et al. (2010). Pathogen inactivation of platelet concen-
trates. Vox Sang, 99, 85–​95.
Reid M, Lomas-​Francis C, Olsson M (2012). The blood group antigen.
Elsevier, Burlington, MA.
Reid ME (2009). Transfusion in the age of molecular diagnostics.
Hematol Am Soc Hematol Educ Program, 2009, 171–​7.
Reid ME, Yazdanbakhsh K (2011). Molecular insights into blood groups
and implications for blood transfusion. Curr Opin Hematol, 5, 93–​102.
Silliman CC (1999). Transfusion-​related acute lung injury. Transfus
Med Rev, 13, 177–​86.
Silliman CC, McLaughlin NJD (2006). Transfusion-​related acute lung
injury. Blood Rev, 20, 139–​59.
Slapak C, Fredrich N, Wagner J (2011). Transfusion safety: is this the
business of blood centers? Transfusion, 51, 2767–​71.
Snyder EL (2008). The role of cytokines and adhesive molecules in fe-
brile non-​hemolytic transfusion reactions. Immunol Invest, 24, 333–​9.
Snyder EL, Stramer SL, Benjamin RJ (2015). The safety of the blood
supply –​ time to raise the bar. N Engl J Med, 372, 1882–​5.
Stanworth SJ, et  al. (2013). A no-​prophylaxis platelet-​transfusion
strategy for hematologic cancers. N Engl J Med, 368, 1771–​80.
Stanworth SJ, et al. (2013). Thrombocytopenia and platelet transfusion
in UK critical care: a multicenter observational study. Transfusion,
53, 1050–​8.
Stramer SL, Dodd RY (2013). Transfusion-​transmitted emerging in-
fectious diseases: 30 years of challenges and progress. Transfusion,
53, 2375–​83.
Utter GH, et al. (2006). Leukoreduction of blood transfusions does
not diminish transfusion-​associated microchimerism in trauma pa-
tients. Transfusion, 46, 1863–​9.
Vogelsang GB, Hess AD (1994). Graft-​versus-​host disease: new direc-
tions for a persistent problem. Blood, 84, 2061–​7.
Wiener AS (1943). Genetic theory of the Rh blood types. Proc Soc Exp
Biol Med, 54, 316.
Winslow RM (2006). Red cell substitutes. Semin Hematol, 44, 51–​9.