# 15.26.1 Acute pancreatitis 3209

# 15.26.1 Acute pancreatitis 3209

CONTENTS
	15.26.1	 Acute pancreatitis  3209
R. Carter, Euan J. Dickson, and C.J. McKay
	15.26.2	 Chronic pancreatitis  3218
Marco J. Bruno and Djuna L. Cahen
	15.26.3	 Tumours of the pancreas  3227
James R.A. Skipworth and Stephen P. Pereira
15.26.1  Acute pancreatitis
R. Carter, Euan J. Dickson, and C.J. McKay
ESSENTIALS
Acute pancreatitis affects 300 to 600 new patients per million popu-
lation per year and is most commonly caused by gallstones or al-
cohol. Careful imaging reveals that most so-​called idiopathic acute 
pancreatitis is due to small (1–​3-​mm diameter) gallstones.
Diagnosis is made by a combination of a typical presentation 
(upper abdominal pain and vomiting) in conjunction with raised 
serum amylase (more than three times the upper limit of normal) 
and/​or lipase (more than twice the upper limit of normal). Several 
other acute abdominal emergencies can mimic acute pancreatitis 
and may be associated with a raised serum amylase. In equivocal 
cases, a CT scan is indicated to exclude other causes and confirm 
the diagnosis.
Initial management is with (1) analgesia, (2) ensuring adequate 
oxygenation, and (3) intravenous fluid administration. The revision of 
the Atlanta classification separates patients clinically into (1) mild—​
with early resolution without complications, (2)  moderate—​local 
complications without organ failure, and (3) severe—​complications 
associated with organ failure.
Mild acute pancreatitis responds to analgesia and intravenous 
fluids. If gallstones have been identified, then cholecystectomy 
(or endoscopic retrograde cholangiopancreatography (ERCP) 
sphincterotomy where clinically appropriate) should be performed 
during the same admission, or at least within 2 to 4 weeks to pre-
vent recurrent attacks. Severe acute pancreatitis carries a high 
mortality (up to 20%). Management in the early stages is centred 
on organ support (respiratory, circulatory, and renal failure). Later 
management involves surgical or radiological intervention for 
sepsis, usually within a specialist pancreatic unit.
With regard to some specific aspects of management:  (1) 
there is no indication for a nasogastric tube or starvation:  en-
teral feeding of patients with prolonged illness is associated 
with fewer risks and side effects than total parenteral nutrition. 
(2)  Antibiotics should not be given until and unless a specific 
indication arises. (3) ERCP—​occasionally cholangitis may be as-
sociated with hyperamylasaemia, in which case urgent biliary 
decompression at ERCP is indicated. (4) There are no pharmaco-
logical agents that benefit any form of acute pancreatitis. (5) The 
case for decompressive laparotomy for abdominal compartment 
syndrome remains unproven.
Epidemiology
Population studies from Scotland and Finland have shown the in-
cidence of acute pancreatitis is about 400 patients/​million per year. 
Incidence would appear to be rising, with a 100% increase in the 
hospitalization rate in the United States of America over the last 
20 years, a 75% increase in admissions in the Netherlands, and a 
3.1% yearly rise in incidence in the United Kingdom. The mean age 
at presentation is 53 years with a roughly equal sex distribution, al-
though the largest increase in incidence has been among women 
under 35  years, which may reflect increasing obesity within the 
population. The disease was most prevalent in those with poorer 
socioeconomic status, especially where alcohol was the cause. 
Overall mortality is from 2.0 to 7.5%, highest in those who are over 
70 years, obese individuals, and those with comorbidity at the time 
of onset. Prospective and retrospective studies record 45 to 50% of 
deaths as occurring in the initial week of the illness secondary to 
fulminant multiple organ failure.
15.26
Diseases of the pancreas


section 15  Gastroenterological disorders
3210
Clinical features
Sudden onset of severe upper abdominal pain focused in the epigas-
trium with vomiting is the most common presentation. This tends 
to progressively lessen in severity over the first 48 to 72 h, and it 
is not usually a significant factor beyond this time. There may be 
upper abdominal tenderness and guarding, but these signs are often 
less marked than might be suspected from the severity of the pain, 
and bowel sounds are usually absent in the early stages. Vomiting is 
prevalent in the first 12 h of illness, contributing to hypovolaemia 
and hypotension. Clinical jaundice is rare on admission, although 
minor abnormalities of biochemical liver blood tests occur in 80% 
of patients with a biliary aetiology. Occasionally, hyperamylasaemia 
occurs in association with cholangitis, which if overlooked may re-
sult in irreversible multiple organ failure. The presence of jaundice 
and pyrexia on the day of admission is therefore an indication for ur-
gent endoscopic retrograde cholangiopancreatography (ERCP) and 
biliary decompression.
Pathology
The initial phase of acute pancreatitis is characterized by oedema 
and the development of an acute inflammatory infiltrate, rich in 
neutrophils, with tiny spots of fatty tissue necrosis, mainly on the 
surface but also in the intralobular fatty tissue. In mild cases, the 
changes are most marked in the peripancreatic tissue, but because 
the initial pancreatic histological change is within the intralobular 
fat, there is a relationship between extent of necrosis and the amount 
of fat within the pancreas. In severe disease, intravascular throm-
bosis and local enzymic necrosis leads to confluent areas of fat ne-
crosis, which extends beyond the pancreas into the peripancreatic 
fat. Within the pancreas, disseminated ductal and periductal ne-
crosis may be evident.
Diagnosis and assessment of severity
The diagnosis is usually made from the clinical presentation of 
upper abdominal pain and vomiting associated with an elevation 
of serum amylase or lipase. A CT scan should be performed when 
the diagnosis is not clear. This often demonstrates pancreatic oe-
dema and peripancreatic inflammatory stranding, fluid collec-
tions, and poor contrast enhancement of the gland (Fig. 15.26.1.1). 
Occasionally, the diagnosis may first be made at laparotomy, when 
simple washout and closure is all that should be done.
The differential diagnosis is that of an acute abdomen 
(Box 15.26.1.1). A raised amylase may be associated with several 
of these conditions, but the (near) universal availability of CT has 
simplified diagnosis when doubt exists.
Biochemical abnormalities
A multitude of biochemical phenomena are found in acute pan-
creatitis. Various pancreatic enzymes are released that are useful 
as diagnostic markers. Acinar cell disruption leads to high serum 
levels of amylase, lipase, trypsin, chymotrypsin, phospholipase, elas-
tase, trypsinogen activation peptide, and phospholipase activation 
peptide. These are also elevated in peritoneal and retroperitoneal 
tissues as well as lymphatic fluid. C-​reactive protein, an acute-​phase 
reactant, is of most use for longitudinal monitoring of progress. Very 
high concentrations of circulating cytokines occur in the blood at 
an early stage in the disease, including tumour necrosis factor-​α, 
platelet activating factor, and interleukin 6 with maximal levels in 
those with severe pancreatitis: these are of research interest rather 
than clinical value.
Grading disease severity
Many biochemical scoring systems attempting to objectively grade 
severity of an attack of acute pancreatitis have been developed, 
including the Glasgow and Ranson scoring systems. While there 
is value in directing less experienced clinicians to the multisystem 
organ dysfunction associated with a severe attack, none are suffi-
ciently accurate to direct treatment, and the utility lies in assessing 
equipoise within clinical trials, with regard to which the APACHE 
II system has been shown to be useful in the stratification of severity 
of acute pancreatitis.
CT scanning can be useful in demarcating location and ex-
tent of pancreatic injury, and a CT severity index has been de-
veloped, although CT is more commonly used to monitor the 
development of evolving complications. In practice, sequen-
tial physiological scoring systems (e.g. NEWS) can assist the 
Fig. 15.26.1.1  CT scan 36 h from onset of pain showing reduced 
enhancement of neck and body, a perfused pancreatic tail, and 
peripancreatic stranding.
Box 15.26.1.1  Differential diagnosis of acute pancreatitis
	•	 Mesenteric ischaemia/​infarction
	•	 Small-​bowel obstruction/​perforationa
	•	 Renal failure
	•	 Macroamylasaemia
	•	 Dissecting aortic aneurysm
	•	 Diabetic ketoacidosis
	•	 Perforated duodenal ulcera
	•	 Acute cholangitisa
	•	 Acute cholecystitisa
	•	 Atypical myocardial infarctiona
	•	 Ectopic pregnancya
a Amylase is usually normal.


15.26.1  Acute pancreatitis
3211
identification of clinical deterioration and are in common use in 
most surgical units.
Other factors affecting prognosis
Other factors affecting prognosis include age and obesity. Many 
studies have shown that those aged over 70 years have a higher mor-
tality. Chronic cardiorespiratory or renal impairment is common in 
this age group and further increases the risk of death. Acute pan-
creatitis carries a significantly higher mortality and morbidity in 
patients with a body mass index of more than 30 kg/​m2, mainly 
because of an increased risk of hypoxaemia, but also from other 
associated factors.
Aetiology
Aetiological factors and rare associations of acute pancreatitis are 
listed in Boxes 15.26.1.2 and 15.26.1.3.
Major factors
Biliary disease and alcohol abuse together account for over 80% of 
patients with acute pancreatitis in most prospective studies.
Gallstones
Gallstones are the predominant cause of acute pancreatitis and 
all patients should have an abdominal ultrasound examination 
to identify these, even where there is a history of alcohol excess. 
Smaller stones pass through the cystic duct more easily and are 
at increased risk of precipitating acute pancreatitis. Endoscopic 
ultrasonography (EUS) is more sensitive than transabdominal 
ultrasonography in identifying small gallstones (microlithiasis). 
This may be helpful in demonstrating gallstone aetiology in those 
patients previously labelled ‘idiopathic’.
The mechanism by which gallstones induce acute pancreatitis is 
not certain, but increased back pressure in the pancreatic duct fol-
lowing transient impaction of a migrating gallstone at the ampulla 
of Vater is considered to be the likely initiating event. Subsequent 
intracellular events lead to activation of proteases within acinar cells, 
acinar cell injury, and a local inflammatory response.
Alcohol
The proportion of patients in whom pancreatitis is due to alcohol 
abuse is dependent on the population under study: rates may be 
as high as 70 to 80% (in New York (United States of America) and 
Helsinki (Finland)). The risk is highest in young males who drink in 
excess of 80 g of alcohol per day. Smoking is a cofactor in the devel-
opment of both acute and chronic pancreatitis. Many patients with 
a possible alcohol history also have gallstones and the diagnosis of 
alcohol-​induced acute pancreatitis should be one of exclusion.
Alcohol probably causes acute pancreatitis through intracellular 
lysosomal release modulated by elevations in cytosolic and mito-
chondrial ionized Ca2+ concentration.
Minor factors
Drugs
The drugs most commonly implicated in causing acute pancreatitis 
are valproic acid, azathioprine, l-​asparaginase, and corticosteroids. 
There is equivocal evidence regarding thiazide and other diuretics. 
However, unless viral titres have been determined, together with 
adequate biliary investigations including endoscopic examination 
of the ampulla of Vater, it is unwise to ascribe acute pancreatitis to 
a particular drug. Repeat exposure to the same drug again causing 
acute pancreatitis is the strongest evidence of a direct association.
Viral infection
Viral infection, particularly mumps, Coxsackie B, and viral hepa-
titis, can cause acute pancreatitis. One clinical feature that may prove 
useful is prodromal diarrhoea, which is rare in all other types of 
acute pancreatitis. Of increasing importance are the effects of HIV 
infection, where acute pancreatitis may be associated with both the 
primary viral infection and treatment (antiretroviral drugs). Single 
combination agent therapy (tenofovir, lamivudine, and efavirenz) 
is considered the most pancreas friendly, but the evidence for 
favouring one agent over another is weak. Alcohol abuse is common 
in many HIV-​positive patients, particularly in Africa, and it may be 
difficult to define a specific causal agent.
Box 15.26.1.2  Aetiological factors in acute pancreatitis 
(according to frequency)
Major
	•	 Biliary disease
	•	 Alcohol abuse
Minor
	•	 After ERCP
	•	 Sphincter of Oddi dysfunction
	•	 Hyperparathyroidism
	•	 Hyperlipoproteinaemia
	•	 Blunt or surgical trauma
	•	 Autoimmune pancreatitis
	•	 Drugs
	•	 HIV-​associated pancreatitis
	•	 Hereditary (trypsinogen gene defects)
	•	 Ampullary or pancreatic tumour
	•	 Cancers metastatic to pancreas:
	—	Renal
	—	Stomach
	—	Breast
	—	Ovarian
	—	Lung
Box 15.26.1.3  Rare associations with acute pancreatitis
	•	 Hypothermia
	•	 Coxsackie B virus
	•	 Mumps virus
	•	 Sclerosing cholangitis
	•	 α1-​Antitrypsin deficiency
	•	 Virus infection (non-​HIV)
	•	 Worm infestationa
	•	 Scorpion biteb
	•	 Duodenal duplication
a In South-​East Asia. b In Trinidad.


section 15  Gastroenterological disorders
3212
Benign pancreatic duct stricture
A focal area of pancreatic necrosis in a primary attack of acute pan-
creatitis can cause secondary fibrosis with main duct stricture for-
mation and segmental ‘upstream’ recurrent attacks of pancreatitis 
as a consequence. Stricture dilatation or occasionally surgical de-
compression or distal pancreatectomy may be required. Congenital 
or developmental anatomical abnormalities can present with 
pancreatitis (choledochal cyst, duodenal duplication, anomalous 
pancreaticobiliary junction). Pancreas divisum (nonunion of main 
and accessory ducts) occurs in 3 to 5% of people and is not con-
sidered to be a primary cause of pancreatitis.
Periampullary or obstructive pancreatic tumours
Periampullary adenoma or carcinoma resulting in upstream ob-
struction of the main pancreatic duct is an important association. 
Ampullary tumours are best diagnosed with side-​viewing endo-
scopic biopsy. With the increase in this approach to diagnosis, tu-
mours at or close to the ampulla have been shown to cause 0.4% of 
cases of acute pancreatitis. Effective treatment of the tumour abol-
ishes recurrent attacks. This usually involves surgical resection, but 
endoscopic laser therapy or endoscopic papillectomy can be effective 
in older and less fit patients. Carcinoma of pancreas can occasionally 
present with clinical acute pancreatitis and other primary tumours 
(such as neuroendocrine tumours) or tumours metastasizing to the 
pancreas (such as renal carcinoma) may present in this way, prob-
ably by causing pancreatic duct obstruction.
Pancreatitis in association with a side branch intraductal papillary 
mucinous neoplasm is considered a risk factor for malignant trans-
formation and requires assessment for potential resection following 
resolution.
Hyperparathyroidism
Hypercalcaemia secondary to hyperparathyroidism is now recog-
nized to be an uncommon cause of acute pancreatitis. Removal of 
a parathyroid adenoma usually prevents further acute pancreatitis 
since persistent hypercalcaemia appears to be the provoking factor.
Hyperlipidaemia
Patients with type I and type V hyperlipoproteinaemia may de-
velop acute pancreatitis in the absence of alcohol ingestion when 
triglyceride levels exceed 11 mmol/​litre. Both subtypes are asso-
ciated with chylomicrons, of which greater than 90% are trigly-
cerides. Dietary restriction of lipids and various lipid-​lowering 
drugs are valuable in therapy. Hyperlipidaemia of any cause where 
triglyceride levels reach in excess of 2000 mg/​dl may cause acute 
pancreatitis, and acute pancreatitis may rarely complicate hyper-
lipidaemia of pregnancy or diabetic ketoacidosis. High triglyceride 
levels may be present during an attack of acute pancreatitis and 
identifying high fasting lipid levels following resolution is key to 
confirmation.
Hypothermia
This is an important association. In younger patients, this may be 
associated with alcohol abuse, particularly if patients fall asleep 
out of doors, or in the elderly in an unheated house. Management 
is directed at gradual warming and supportive measures for organ 
compromise.
Hereditary
This condition is increasingly being studied since the discovery of 
genetic mutations of the cationic trypsinogen gene (PRSS1), which 
shed light on the mechanism of acute pancreatitis. A Europe-​wide 
study (EUROPAC) has tracked multiple families in the United 
Kingdom and Europe, and similar work in Japan and the United 
States of America is ongoing. The two most common mutations are 
R122H and N29I. An autosomal dominant pattern of inheritance is 
seen. Severe acute inflammatory changes are rare and diagnosis is 
often delayed. Patients usually have a long history of recurrent ab-
dominal pain from childhood or adolescence. Changes of chronic 
fibrosis may be present at diagnosis. Typically, chronic pancreatitis 
is evident by the age of 20 to 40 years, and the risk of pancreatic car-
cinoma in those aged over 60 years is significantly increased.
Mutations in the cystic fibrosis transmembrane conductance 
regulator gene (CTFR) and the pancreatic secretory trypsinogen in-
hibitor gene (SPINK1) are also linked to pancreatitis.
Trauma
Hyperamylasaemia may occur after blunt abdominal trauma, usu-
ally from a crush injury to the body of the pancreas against the ver-
tebral column. The risk of associated injuries to surrounding organs 
is high, and in the acute phase these are usually more significant 
than the pancreatic injury. The identification of a pancreatic injury 
during a trauma laparotomy should be managed by simple drainage 
in most cases. When there is transection of the main pancreatic duct, 
therapeutic options include endoscopic transpapillary stenting and 
distal pancreatectomy. Late presentation is associated with pseudo-
cyst formation due to leakage from the damaged pancreatic duct. 
Pancreatic manipulation during surgical mobilization for colonic, 
gastric, or splenic surgery may result in inflammation.
Iatrogenic
Surgical or endoscopic procedures involving the ampulla of 
Vater can induce pancreatitis. In recent years, diagnostic ERCP 
(2% risk of acute pancreatitis) has largely been replaced by 
noninvasive imaging modalities (EUS and magnetic resonance 
cholangiopancreatography). This has reduced the overall incidence 
of postprocedural acute pancreatitis, but EUS fine needle aspiration 
of pancreatic lesions, particularly close to the main duct, can itself 
result in pancreatitis. The risk of acute pancreatitis increases to 4 to 
6% where a therapeutic endoscopic sphincterotomy has been per-
formed and may be as high as 20% in high-​risk patients (sphincter 
of Oddi dysfunction). Iatrogenic perforation should be considered 
in all patients who develop pancreatitis after therapeutic ERCP, and 
patients with early organ dysfunction or abdominal signs should 
undergo a CT without delay if any doubt exists as this will affect 
management (antibiotics).
Autoimmune pancreatitis
This is a rare condition, which is considered part of the systemic 
IgG4-​related autoimmune disease spectrum and is associated with 
other autoimmune diseases (polyarteritis nodosa, systemic lupus 
erythematosus, other vasculitides) and inflammatory bowel disease 
(Crohn’s disease and ulcerative colitis). It may present as abdominal 
pain with obstructive jaundice more typical of chronic than acute 
pancreatitis. Other features may include (1)  an increased IgG4/​


15.26.1  Acute pancreatitis
3213
IgG ratio in serum, (2)  homogeneous gland enlargement with a 
well-​defined halo on CT, (3) characteristic diffuse abnormality on 
EUS, and (4) periductal lymphoplasmacytic infiltrate on biopsy. The 
latter may also be associated with abnormalities in the extrahepatic 
biliary tree resembling sclerosing cholangitis (IgG4-​associated 
cholangiopancreatopathy). Focal autoimmune pancreatitis may 
prove difficult to differentiate from carcinoma. A good response to 
steroids is diagnostic (HISORt criteria).
Worm infestation
Ascaris lumbricoides within the ampullary area may manifest as 
acute pancreatitis clinically, and other worms lodged in this area can 
produce the same effect.
Sphincter of Oddi dyskinesia
Rarely, sphincter of Oddi dyskinesia can present with acute ab-
dominal pain, although the more common presentation is one of 
chronic relapsing abdominal discomfort. The attacks associated 
with hyperamylasaemia are usually mild (except where pancrea-
titis follows an ERCP). With regard to functional gallbladder and 
sphincter of Oddi disorders (SODs), the Rome III diagnostic criteria 
for functional gastrointestinal disorders require episodes of pain in 
the epigastrium or right upper quadrant and all of the following: epi-
sodes lasting 30 min or longer; recurrent symptoms occurring at 
different intervals (not daily); the pain builds up to a steady level 
and is moderate/​severe enough to interrupt the patient’s daily ac-
tivities or lead to a visit to the emergency department; the pain is 
not relieved by bowel movements, postural change, or antacids; and 
exclusion of other structural disease that would explain the symp-
toms. If these criteria are satisfied, then Rome III allowed further 
classification into functional gallbladder disorder (SOD I), where 
the gallbladder is present and biochemical tests are normal; func-
tional biliary SOD (SOD II), where amylase and lipase are normal, 
but association of at least two episodes of pain with elevation of 
serum transaminases, alkaline phosphatase, or conjugated bilirubin 
is a supportive criterion; and functional pancreatic SOD (SOD III), 
characterized by elevation of serum amylase and/​or lipase. More re-
cently, Rome IV, recognizing good evidence that sphincterotomy is 
no better than sham treatment in patients previously classified as 
having SOD III, discarded the use of this term and renamed SOD II 
‘Functional Biliary Sphincter Disorder’. However, whether this exists 
as a clinical entity remains a matter of debate, and whether it is a 
potential cause of acute pancreatitis is controversial. Most clinicians 
would not recommend pancreatic sphincterotomy in the absence of 
ampullary stenosis.
Classification of severity
The original Atlanta classification of acute pancreatitis dichotom-
ized clinical behaviour into mild or severe acute pancreatitis, and 
intervention for necrosis was often focused on early removal of 
sterile or infected necrosis, usually by open necrosectomy. This 
oversimplification proved inadequate in clinical practice and the 
Atlanta criteria were revised in 2013 to address the importance of 
early systemic organ dysfunction in determining disease severity 
and outcome (Table 15.26.1.1). The outcome following interven-
tion for the management of local fluid or necrotic collections is also 
heavily influenced by the degree of systemic disturbance, and this is 
reflected in an additional category of ‘moderately severe’ pancrea-
titis, where collections are present in the absence of organ dysfunc-
tion. In addition to disease severity, mortality is strongly associated 
with age and comorbidity. The significance of infection has been 
recognized in an addendum adding a category of ‘critical’, recog-
nizing that those patients with sepsis and organ failure have the 
highest mortality.
This classification further separates local complications/​collec-
tions on the basis of time from presentation (<4 or >4 weeks) and 
on the presence of necrosis, leading to definitions aimed at permit-
ting comparison of case series (Table 15.26.1.2). The ‘early’ phase 
is characterized by the initial host response to the pancreatitis, the 
severity being determined by the magnitude of organ disturbance/​
failure, and a ‘late’ phase typified by the persistence of organ dys-
function and the management of local or systemic complications. 
The vast majority of acute fluid collections without necrosis will re-
solve within 4 weeks, and a persistent fluid collection with minimal 
or no necrotic component (‘pseudocyst’) is very rare. Collections 
may be sterile or infected. Most clinically significant peripancreatic 
complications are therefore related to either acute necrotic collec-
tions (<4 weeks) or walled-​off pancreatic necrosis (>4 weeks). This 
temporal separation is somewhat arbitrary as the clinical manage-
ment and surgical approach is determined by multifactorial indi-
vidual patient factors. However, this does serve to provide a timeline 
beyond which, if appropriate, intervention should be delayed.
Management
General aspects
The principles of organ support in critical illness should be followed, 
ensuring reversal of hypoxaemia, restoration of circulating volume, 
and maintenance of tissue perfusion ideally within a critical care en-
vironment. Early restoration of circulating blood volume is associ-
ated with an improved outcome. The combination of fluid lost from 
vomiting and loss of capillary integrity can be very substantial. The 
introduction of vasoconstrictor therapy should only be considered 
after establishing adequate volume resuscitation. Hypoxaemia re-
flects disease severity in acute pancreatitis, and while most patients 
Table 15.26.1.1  Grades of severity of pancreatitis (based on the 
clinical parameters of the presence or absence of organ failure 
and/​or complications
Mild acute 
pancreatitis
•	 No organ failure
•	 No local or systemic complications
Moderately 
severe acute 
pancreatitis
•	 Organ failure that resolves within 48 h (transient 
organ failure) and/​or
•	 Local or systemic complications without persistent 
organ failure
Severe acute 
pancreatitis
•	 Persistent organ failure (>48 h):
•	 Single organ failure
•	 Multiple organ failure
Critical acute 
pancreatitis
In addition to the other three categories taken from 
the Atlanta revision, an additional ‘critical’ category 
(persistent organs failure and sepsis) was proposed by 
Petrov et al.


section 15  Gastroenterological disorders
3214
can be managed with supplemental humidified oxygen, assisted ven-
tilation may be required in more severe cases.
For initial resuscitation, Ringer’s lactate is recommended, with 
goal-​directed intravenous fluid (5–​10 ml/​kg per hour) to produce 
0.5 ml/​kg of urine per hour. Renal replacement therapy by haemo-
dialysis or haemofiltration may be required if renal failure becomes 
established. The more severe cases of acute pancreatitis are charac-
terized by the development and persistence of a systemic inflam-
matory response syndrome, one aspect being the development of 
a swinging pyrexia. This is often inappropriately taken as evidence 
of sepsis but is usually a reflection of the inflammatory cascade and 
the presence of devitalized tissue in the retroperitoneum rather than 
bacterial infection.
Specific aspects
ERCP
There is no role for early ERCP in mild disease. High fever within 
the first 24 h is rare and, if associated with jaundice, is suspi-
cious of ascending cholangitis where the cholangitis, rather than 
any pancreatic inflammation, is driving organ dysfunction. This 
is the only unequivocal indication for ERCP and biliary decom-
pression. The role of early ERCP remains controversial in all other 
circumstances. Several randomized studies and cohort series have 
looked at the role of early ERCP with endoscopic sphincterotomy 
compared with conservative management in acute gallstone pan-
creatitis. A number of meta-​analyses have failed to demonstrate 
a definite benefit, and early ERCP is difficult to justify in patients 
who are not jaundiced.
Antibiotics
The dual peak in mortality in acute pancreatitis is well recognized. 
The late peak is determined by complications associated with ne-
crosis, including the development of pancreatic or peripancreatic 
infection. The consensus of the most recent systemic reviews/​meta-​
analyses is that there is no evidence supporting the use of prophy-
lactic antibiotics in either mild or severe acute pancreatitis, and 
the recommendation is to avoid their use, using targeted antibiotic 
therapy for episodes of proven infection. It is, however, reasonable to 
commence antibiotics in the deteriorating patient with radiological 
and clinical evidence of sepsis while awaiting culture confirmation.
Nutrition support
There is no benefit from enteral feeding in mild pancreatitis, and 
these patients need have no dietary restrictions. Assisted feeding 
may be required in severe acute pancreatitis to provide long-​term 
nutritional support. Randomized studies have shown that enteral 
nutritional support is cheaper and is associated with fewer side ef-
fects than total parenteral nutrition. In addition, there is no differ-
ence in outcome with nasogastric compared to nasojejunal feeding. 
In clinical practice, therefore, the mode of nutritional support does 
not appear to influence the disease process, and the choice of de-
livery relates to tolerance and minimizing morbidity associated with 
the delivery system. Probiotics may be detrimental in acute pancrea-
titis and specifically supplemented nutrition should only be admin-
istered in the context of clinical trials.
It has been suggested that nutritional support may help to pre-
serve mucosal function and limit the stimulus to the inflammatory 
response. The experimental evidence supporting this has not been 
confirmed in clinical practice and there is no evidence that the mode 
of feeding alters disease progression.
Other approaches
Previous randomized studies have failed to provide sufficient evi-
dence to recommend the use of antiproteases, antisecretory, anti-​
inflammatory agents, or antioxidant therapy in patients with acute 
pancreatitis. There is therefore no proven pharmacological therapy 
for the treatment of acute pancreatitis.
Intervention for postacute pancreatitis fluid  
and necrotic collections
The definitions surrounding acute fluid-​predominant and acute nec-
rotic collections have been detailed previously in Table 15.26.1.2. 
Due to the complexity and diversity of an individual patient’s clinical 
course, it is challenging to define specific triggers for intervention. 
Delayed intervention where possible, along with planned minim-
ally invasive sepsis control, has become key to management in most 
pancreatic units. There is no role for early (week 1) intervention 
unless vascular complications are suspected. Intervention is most 
commonly required for suspected or proven sepsis of acute necrotic 
collections or walled-​off necrosis after resolution of the initial in-
flammatory phase. Optimal management may involve a number or 
combination of techniques and require discussion with a specialist 
regional pancreatic centre.
The indication for early (<6 weeks) intervention is for the con-
trol of proven or suspected sepsis. Secondary clinical or biochemical 
deterioration after the initial period of organ compromise should 
lead to a full bacteriological septic screen and contrast-​enhanced CT 
scan. Gas within an acute necrotic collection or walled-​off necrosis 
are indicative of bacteriological contamination but may result from 
spontaneous fistulation into the intestine (particularly where there 
is an air–​fluid level). In general, fistulation into the foregut may be 
associated with clinical improvement despite CT appearances, while 
Table 15.26.1.2  Local complications in acute pancreatitis (2012 Revised Atlanta Classification)
Timescale
Necrosis absent
Necrosis present
<4 weeks
Acute peripancreatic fluid collection (peripancreatic fluid associated 
with interstitial oedematous pancreatitis with no associated 
peripancreatic necrosis)
Acute necrotic collection (a collection containing variable amounts 
of both fluid and necrosis; the necrosis can involve the pancreatic 
parenchyma or the extrapancreatic tissues)
>4 weeks
Pancreatic pseudocyst (an encapsulated collection of fluid with a 
well-​defined inflammatory wall usually outside the pancreas with 
minimal or no necrosis)
Walled-​off necrosis (a mature, encapsulated collection of pancreatic 
or extrapancreatic necrosis that has developed a well-​defined 
inflammatory wall)
Infection
Each collection type may be sterile or infected


15.26.1  Acute pancreatitis
3215
hindgut fistulation often results in clinical deterioration mandating 
intervention. Smaller gas bubbles are more commonly associated 
with bacterial translocation and require drainage by one of the 
methods described in the following subsections if clinically indi-
cated (Fig. 15.26.1.2).
Sepsis associated with acute necrotic collections or early 
walled-​off necrosis
Initial walled-​off necrosis
The trend in the last 20 years has been towards minimizing the mag-
nitude of any intervention while achieving adequate sepsis control. 
Initially most methods were aimed at mirroring the open surgical 
debridement techniques to achieve clearance of any necrotic ma-
terial as this was felt to be key to recovery, but the importance of 
sepsis control over removal of necrosis has increasingly been rec-
ognized. This was perhaps the most significant finding within the 
PANTER trial from the Dutch Pancreatitis Study Group, which 
not only provided randomized data regarding the management of 
infected pancreatic necrosis, but also demonstrated that a third of 
patients (35% within PANTER) with pancreatic or peripancreatic 
necrosis may resolve completely with simple percutaneous drainage. 
In this trial, patients requiring surgical intervention for pancreatic 
necrosis were randomized to either primary open necrosectomy or 
a ‘step-​up’ approach based on percutaneous drainage as the initial 
intervention, with progression to retroperitoneal debridement with 
lavage if no improvement was observed. There is now a consensus 
advocating a principle of early organ support, nutritional optimiza-
tion, followed ideally by delayed and selective minimally invasive 
intervention if required.
More recently the role of endoscopic (EUS) transmural cysto­
gastrostomy in the management of postacute walled-​off collections 
has emerged as an alternative to percutaneous or surgical drainage. 
The choice between initial percutaneous or endoscopic drainage 
is determined by several factors including (1) the anatomy of the 
disease, for example, the position of the collection relative to the 
stomach, colon, liver, spleen, and kidney; (2)  patient physiology, 
as the ability to perform EUS-​guided puncture within an intensive 
therapy unit setting, without the need for patient transfer to the 
radiology department for CT-​guided drainage, may influence the 
management decision where a patient is in extremis, and unstable to 
transfer; and (3) institutional capabilities.
Secondary intervention
The ‘step-​up’ concept is based on staged sepsis control as a bridge 
to surgery or as definitive treatment in some patients. The require-
ment for secondary intervention usually arises from blockage of the 
percutaneous drain, or cystogastrostomy with solid necrotic ma-
terial. Simple replacement, upsizing, or insertion of multiple drains 
has been proposed rather than embarking on one of the secondary 
intervention techniques described in ‘Particular interventional 
techniques’.
The Dutch Pancreatitis Study Group compared the success of fur-
ther upsizing of percutaneous drainage versus retroperitoneal de-
bridement as the initial enhanced step-​up procedure if immediate 
resolution did not occur and showed that more that 50% of patients 
in the dilatation-​alone group settled without formal necrosectomy. 
Drawbacks included limited ability to remove necrotic debris, pro-
longed hospitalization, and the need for multiple procedures. The 
use of grasping forceps to extract the debris after sequential tract 
dilatation has been described in a small series, as has the use of as-
sist devices such as stone retrieval baskets, but these techniques are 
seldom performed in clinical practice. Probably more important 
than the method of drainage, a dedicated team of surgeons/​radi-
ologists prepared to proactively address any undrained sepsis is 
critical for successful percutaneous management of necrotizing 
pancreatitis.
Particular interventional techniques
The open lateral approach initially described in the 1980s utilized a 
loin/​subcostal and retrocolic approach to allow debridement of pan-
creatic and peripancreatic necrosis. This open approach was associ-
ated with major morbidity (enteric fistula 45%, haemorrhage 40%, 
and colonic necrosis 15%), and failed to gain popularity.
Minimally invasive pancreatic necrosectomy
For percutaneous necrosectomy, the catheter is exchanged for a 
radiological guidewire and then a low-​compliance balloon dilator 
is inserted into the collection and dilated to 34 FG. Access to the 
cavity is maintained by an Amplatz sheath through which is passed 
an operating necroscope to allow debridement under direct vision. 
The necroscope has an operating channel that permits standard 
(5-​mm) laparoscopic graspers as well as an irrigation/​suction 
channel. The directed, high-​flow lavage promotes rapid evacu-
ation of pus and liquefied necrotic material, revealing black or grey 
devascularized pancreatic tissue and peripancreatic fat, which if 
loose is extracted in a piecemeal fashion until, after several pro-
cedures, a cavity lined by viable tissue or granulating pancreas is 
created. At the end of each procedure an 8 FG catheter sutured to a 
24 FG drain is passed into the cavity (Fig. 15.26.1.3a) to allow con-
tinuous postoperative lavage of warmed fluid, initially at 250 ml/​h. 
Subsequent conversion of the lavage system to simple drainage may 
be all that is required prior to recovery, or the procedure may be 
repeated until sepsis control is achieved and interval CT confirms 
resolution.
Video-​assisted retroperitoneal debridement
A video-​assisted retroperitoneal debridement (VARD) procedure is 
performed with the patient placed in a supine position with the left 
side 30 to 40° elevated. A subcostal incision of 5 cm is placed in the 
Fig. 15.26.1.2  An acute necrotic collection after 17 days illustrating 
small gas bubbles, managed initially by EUS-​guided cystogastrostomy 
within a step-​up framework.


section 15  Gastroenterological disorders
3216
left flank at the midaxillary line, close to the exit point of the percu-
taneous drain. Using the in situ percutaneous drain as a guide, the 
retroperitoneal collection is entered. The cavity is cleared of puru-
lent material using a standard suction device. Visible necrosis is 
carefully removed with the use of long grasping forceps, and deeper 
access under direct vision is facilitated using a 0° laparoscope, and 
further debridement performed with laparoscopic forceps. As with 
a percutaneous necrosectomy, sepsis control rather than complete 
necrosectomy is the aim of this procedure and only loosely adherent 
pieces of necrosis are removed, minimizing the risk of haemorrhage. 
Two large-​bore single-​lumen drains are positioned in the cavity 
and the fascia closed to facilitate a closed continuous postoperative 
lavage system.
Endoscopic necrosectomy
Endoscopic cystogastrostomy was initially reported for the man-
agement of a mature pancreatic abscess with minimal necrosis, but 
the technique has evolved in the last 10 years to become an estab-
lished procedure, with endoscopic transmural exploration and de-
bridement of the retroperitoneum. Single-​step drainage under EUS 
guidance may be carried out by either a transgastric or (less com-
monly) a transduodenal route and is preferred to ‘blind’ drainage as 
EUS allows for identification of the collection where there is no ob-
vious bulge seen within the stomach and helps identify a safe route 
for puncture, free from intervening vessels (Fig. 15.26.1.3b). The 
presence of significant walled-​off necrosis is no longer considered 
a contraindication, but concerns do remain regarding the adequacy 
of endoscopic drainage, particularly in solid predominant or larger 
collections. The major limitation of this technique is that instru-
ments have yet to be developed to facilitate the adequate removal 
of loose solid material from the cavity when this is compromising 
drainage through the cystogastrostomy.
Where there is extensive necrosis, delayed endoscopic necro­
sectomy may be required. It is the chapter authors’ practice to defer 
this for a week following the initial drainage procedure to allow the 
fluid component to drain and any associated sepsis to improve. A re-
cent systematic review of 14 studies including 455 patients found an 
overall success rate of 81% and mortality of 6%, but these studies are in 
highly selected patients and all but one was retrospective. One small 
randomized trial compared endoscopic with surgical drainage and 
found a reduction in significant complications with the endoscopic 
approach. Endoscopic necrosectomy is, however, a challenging pro-
cedure and not without risk. Major complications including fatal air 
embolism, bleeding, and perforation occurred in 26% of patients in 
the multicentre GEPARD study. The use of carbon dioxide insuffla-
tion is therefore now recommended. A persistent problem is the lack 
of availability of suitable endoscopic devices to achieve necrosectomy, 
and although endoscopic access to the cyst cavity is now facilitated 
by metallic stents, piecemeal necrosectomy using standard graspers, 
baskets, and snares is a time-​consuming and painstaking process. 
One possible modification is the use of intracavity hydrogen peroxide 
to facilitate necrosectomy, but further experience is required before 
this can be recommended for routine practice.
An early randomized pilot study exploring the outcome of endo-
scopic transmural drainage versus minimally invasive intervention 
(VARD) (the PENGUIN trial) suggested at least equivalence, if not 
benefit, from endoscopic drainage. This study has been criticized 
due to very small numbers and an excessive mortality (40%), com-
pared to historical results, within the VARD arm. The TENSION 
trial demonstrated that the endoscopic step-up approach was not 
superior to the surgical step-up approach in reducing major compli-
cations or death. The rate of pancreatic fistulas and length of hospital 
stay were lower in the endoscopy group and many now favour endo-
scopic drainage as the index intervention.
Open surgical necrosectomy
Worldwide, open necrosectomy is still commonly employed, but 
increasingly is being superseded by the procedures previously de-
scribed, or performed within a ‘step-​up’ format after initial percutan-
eous drainage has produced sepsis control. Three general variations 
of open necrosectomy are currently practised; although the proced-
ures are broadly similar in terms of the necrosectomy, they differ 
in terms of how they prevent recurrence of an infected collection 
within the debridement cavity: (1) open necrosectomy with open 
or closed packing, (2) open necrosectomy with continuous closed 
postoperative lavage, and (3) programmed open necrosectomy.
(a)
(b)
Fig. 15.26.1.3  Initial ‘step-​up’ drainage using (a) minimally invasive retroperitoneal pancreatectomy percutaneous lavage drain and (b) EUS-​guided 
transgastric cystogastrostomy with a self-​expanding metal stent.


15.26.1  Acute pancreatitis
3217
Management of late collections
A significant proportion of patients with postacute collections do 
not develop sepsis-​related complication in the early phase, allowing 
the collection to organize and mature. Failure to thrive, persistent 
discomfort, and nutritional failure are the common indications for 
late intervention for walled-​off necrosis. Nutritional intake may be 
limited by a combination of early satiety and postprandial pain. The 
extent of actual necrosis, as opposed to reduced enhancement of the 
parenchyma on contrast-​enhanced CT, often results in an overesti-
mation of the extent of necrosis. Spontaneous resolution of even 
large collections is not infrequent, hence continued nonintervention 
is often the best approach, particularly where ongoing maturation 
of a collection may be anticipated and where the clinical picture is 
improving. In any individual case, the choice of intervention may be 
guided by factors including the clinical picture, the position of the 
collection in relation to the stomach and duodenum, and available 
expertise.
Laparoscopic cystogastrostomy
For many years, the conventional approach to the management of late 
walled-​off pancreatic necrosis was open pancreatic cystogastrostomy 
with necrosectomy. This procedure can now be safely and effectively 
carried out using a laparoscopic approach, which has the potential 
advantage of allowing a one-​stage removal of any solid component 
and may also be combined with a cholecystectomy if appropriate.
Endoscopic ultrasonography-​guided  
cystogastrostomy/​necrosectomy
Many reports in the literature describe EUS-​guided drainage 
of ‘pseudocysts’, but it is now recognized that true pancreatic 
pseudocysts are rare following acute pancreatitis as some degree 
of necrosis is usually present where collections persist. The revised 
Atlanta criteria define these collections as walled-​off necrosis, 
but there is still a spectrum of clinical presentations. EUS-​guided 
drainage of these collections is now an established technique in 
specialist units and several different modifications to the technique 
have been described. The frequent requirement for repeated endo-
scopic procedures, particularly in the presence of significant ne-
crosis, have led to a former preference to select fluid-​predominant 
walled-​off necrosis collections for this approach, but this assump-
tion is being currently challenged within a randomized trial in the 
chapter authors’ unit.
(a)
(c)
(b)
Fig. 15.26.1.4  Examples of complications following acute pancreatitis. (a) Haemorrhage from the splenic artery into a retroperitoneal 
collection. (b) Acute necrotic collection with fistulation into the colon. (c) Portal vein thrombosis and cavernous malformation.