16.17.3 Secondary hypertension 3778 Morris J. Brow

16.17.3 Secondary hypertension 3778 Morris J. Brown and Fraz A. Mir

section 16  Cardiovascular disorders 3778 Broderick J, et al. (2007). Guidelines for the management of spon- taneous intracerebral hemorrhage in adults:  update. Stroke, 38, 2001–​23. Colhoun DA, et al. (2008). Resistant hypertension: diagnosis, evalu- ation and treatment. A scientific statement from the American heart association professional education committee of the council for high blood pressure research. Hypertension, 117, e510–​26. Conroy RM, et al. (2003). Estimation of ten-​year risk of fatal cardio- vascular disease in Europe:  the SCORE project. Eur Heart J, 24, 987–​1003. Douglas JG, et  al. (2003). The Hypertension in African Americans Working Group. Management of high blood pressure in African Americans. Arch Intern Med, 163, 525–​41. Higgens B, et al. (2007). Pharmacological management of hyperten- sion. Clin Med, 7, 612–​16. James PA, et al. (2014). 2014 Evidence-​Based guidelines for the man- agement of high blood pressure in adults. Report from the panel members appointed to the Eighth Joint National Committee (JNC8). JAMA, 311, 507–​20. Kjeldsen SE, et al. (2016). Treatment of high blood pressure in eld- erly and octagenarians. European Society of Hypertension, Scientific Newsletter, Updated on Hypertension Management, 17, nr 63. Mancia G, et al. (2013). 2013 ESH/​ESC guidelines for the management of arterial hypertension: the task force for the management of ar- terial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). Eur Heart J, 34, 2159–​219. Mendis S, et  al. (2007). World Health Organization (WHO) and International Society of Hypertension (ISH) risk prediction charts: assessment of cardiovascular risk for prevention and control of cardiovascular disease in low and middle-​income countries.
J Hypertens, 25, 1578–​82. National Clinical Guideline Centre (2011). Hypertension—​the clinical management of primary hypertension in adults. Clinical guideline. http://​www.nice.org.uk/​guidance/​cg127. National Clinical Guideline Centre (2019). Hypertension in adults:
diagnosis and management. NICE Guideline 136, https://www.nice. org.uk/ng136. The task force for the management of arterial hypertension of the European Society of Cardiology and the European Society of Hypertension (2018). 2018 ESC/ESH guidelines for the manage- ment of arterial hypertension. J Hypertens, 36, 1953–2041. 16.17.3  Secondary hypertension Morris J. Brown and Fraz A. Mir ESSENTIALS The term ‘secondary hypertension’ is used to describe patients whose blood pressure is elevated by a single, identifiable cause, with an important subdivision being into reversible and irreversible causes: clinically, it is important to exclude the former, but not necessarily to find the latter. In the first two decades of life, the prevalence of secondary hyper- tension is greater than that of essential hypertension; thereafter, a patient is much more likely to have essential hypertension, but inves- tigations for secondary hypertension should still be assiduous in the twenties and thirties because the alternative entails so many years of tablet-​taking. Overall, it is estimated that about 10% of all patients with hypertension may have a secondary cause. All patients with hypertension should have a minimum set of in- vestigations (see Chapter 16.17.2). Common indications for further investigations are (1) any evidence of an underlying cause on his- tory or examination; (2) proteinuria, haematuria, or elevated serum creatinine (eGFR<30; CKD 4/​5); (3) hypokalaemia, even if caused by diuretics; (4) accelerated (malignant) hypertension; (5) documented recent onset or recent worsening of hypertension; (6)  resistant hypertension (not controlled with three antihypertensive drugs); (7) young age—​any hypertension at less than 20 years; any hyperten- sion needing treatment at less than 35 years. The minimum screen in younger patients should include serum electrolytes and bicarbonate, plasma renin and metanephrines to exclude phaeochromocytoma; 24-​h urinary sodium excretion should be measured either in all patients, or in those with ab- normal renin levels. Primary aldosteronism (Conn’s syndrome) Recent studies suggest that aldosterone-​producing adrenal aden- omas, which cause increased sodium retention through the epithe- lial sodium channel (ENaC) in the distal tubule and cortical collecting duct, are the single commonest known cause of hypertension. Diagnosis is usually on the basis of suggestive clinical biochemistry, radiological imaging showing an adenoma, and lateralization on selective adrenal sampling. Medical treatment is preferred for bilat- eral adrenal hyperplasia, control of hypertension and hypokalaemia before surgery for adenoma removal, older patients with adenomas who are well controlled, or where there is any doubt about diag- nosis or lateralization. Spironolactone is the treatment of choice but causes gynaecomastia in men. Elective surgery is indicated for younger patients with adenomas, and older patients intolerant of—​or uncontrolled by—​medical treatment. Renovascular hypertension This is most commonly due to intrinsic disease of the intima (e.g. atherosclerosis) or media (e.g. fibromuscular dysplasia). The main clinical clue is the finding in 50% of cases of a bruit anteriorly or pos- teriorly over a renal area. The diagnosis is made radiologically, most commonly by CT or MR angiography. In fibromuscular dysplasia, angioplasty is usually curative, with about three-​quarters of patients able to discontinue antihypertensive treatment. In atheromatous dis- ease, angioplasty is much less likely to be successful. Complete cure of hypertension is rare, and studies suggest no significant benefit in prevention of decline in renal function. Coarctation of the aorta Coarctation causes less than 1% of all cases of hypertension. The clas- sical clinical finding is radio–​femoral pulse delay or weak lower limb pulses. Diagnosis is confirmed by two-​dimensional echocardiog- raphy, or by CT or MR angiography. Treatment is by surgery, balloon dilatation, or stenting.

16.17.3  Secondary hypertension 3779 Phaeochromocytomas and paragangliomas Phaeochromocytomas are rare tumours of chromaffin tissue that account for 0.1 to 1% of cases of hypertension. About 40% of phaeochromocytomas and paragangliomas are associated with a single driver germline mutation, some with genetic syndromes, including von Hippel–​Lindau, multiple endocrine neoplasia type 2, and neurofibromatosis type 1. Hypertension, usually in associ- ation with one or more symptoms of headache, sweating, anxiety, and palpitations, is the most common presentation. Diagnosis is usually not difficult once the possibility of these conditions has been entertained, but excluding the diagnosis in patients who have clinical and/​or biochemical features of physiological catecholamine excess is problematic. The best screening test is to measure plasma normetanephrine (normetadrenaline) and metanephrine (metadrenaline) levels or, if unavailable, 24-​h urine metanephrines. CT or MRI scanning usually provides excellent imaging of the adrenal gland. Radioisotope scanning with the iodinated analogue of noradrenaline, m-​iodobenzylguanidine (mIBG), is usually helpful in localizing extra-​adrenal tumours. Surgery is the definitive treatment that cures hypertension in most patients. The offer of genetic counselling and screening ought to be considered in all patients. Introduction The term ‘secondary hypertension’ is used to describe patients whose blood pressure is elevated by a single, identifiable cause. Until the last decade there was an optimistic view that description of new causes of hypertension would mean that those regarded as having ‘essential hypertension’ would be an ever-​diminishing group. However, as discussed in Chapter 16.17.1, genome-​wide investiga- tion into the genetic bases of hypertension have shown that there are no common inherited susceptibility alleles that can explain more than 1–​2 mm Hg of a person’s blood pressure, and none has led to the discovery of inherited major gain-​ or loss-​of-​function mutations explaining hypertension in a small number of affected individuals. Hence it is now thought likely that essential hypertension differs from secondary hypertension not only in being unexplained, but in being, within each patient, due to a multiplicity of inherited and ac- quired characteristics. An important subdivision of secondary hypertension is into re- versible and irreversible causes: clinically it is important to exclude the former, but not necessarily to find the latter. Their elucidation may lead to improved medical therapy (e.g. by predicting the best diuretic in the monogenic causes of low-​renin hypertension), or help assess prognosis, as in the patient with proteinuria. However, the resource implications of finding causes, which can be consid- erable, need to be balanced against achievable gains. These in turn are influenced by the patient’s age, usually meaning that a search for secondary causes is easier to justify in young patients in whom small benefits are multiplied over many years. Age-​related prevalence of secondary hypertension Whereas essential hypertension is clearly an age-​related phenom- enon, the same is less true of secondary hypertension, although different causes predominate at different ages. The net likelihood of a given patient with hypertension having a secondary cause is higher at a young age (Fig. 16.17.3.1). In the first two decades of life, essential hypertension used to be so uncommon that even the absolute prevalence of secondary hypertension was greater than that of essential hypertension. However, this picture may now be changing in the second decade as a consequence of childhood obesity. After the teens, a patient is much more likely to have es- sential than secondary hypertension, but investigations for sec- ondary hypertension should still be aggressive in patients in their 0 10 20 30 40 50 60 70 80 90 100 5 75 65 55 45 35 25 15 Age Prevalence (% of all hypertension) Fig. 16.17.3.1  The age-​related prevalence of secondary hypertension. The red line shows the prevalence of essential hypertension by age (years), the dotted line the prevalence of secondary hypertension by age, and the bars show the percentage of all hypertensives with a secondary cause.

section 16  Cardiovascular disorders 3780 twenties and thirties because the alternative entails so many years of tablet-​taking. In the first decade of life the main causes of secondary hyper- tension are (1)  the monogenic syndromes causing low-​renin (Na+-​dependent) hypertension, and (2)  congenital causes (e.g. coarctation). However, the rarity of blood pressure measure- ment or of complications in the first decade means diagnosis is often later, hence these are also the main causes of hypertension diagnosed in the second decade. Additional causes by this time are some acquired renal diseases, and the familial phaeochromo- cytoma syndromes. Primary aldosteronism (Conn’s syndrome) is probably the commonest reversible cause of hypertension in adults. From the fifth decade onwards, atheromatous renal artery stenosis is also an important cause of hypertension. The clinical approach to secondary hypertension All patients with hypertension should have a minimum set of inves- tigations, as described in Chapter 16.17.2, but ought to be guided by the clinical history and thorough examination. Common indi- cations for further investigations are shown in Box 16.17.3.1. If possible, patients with blood pressure requiring treatment in their twenties or thirties should be investigated before initiation of treatment because this is rarely pressing at a young age, and some of the tests are easier to interpret off treatment. The minimum screen in younger patients should include serum electrolytes, plasma bicarbonate, plasma renin, and metanephrines to exclude a phaeochromocytoma; 24-​h electro- lyte (sodium) excretion should be measured, either in all patients, or at least in those with suppressed renin levels. Sodium intake is most readily estimated at steady state (i.e. no recent change in diet or drugs) by measuring sodium excretion:  intakes be- tween 100 and 200 mmol (c.6–​12 g)/​day have little effect upon plasma renin, whereas outside this range there is a steep inverse relationship. Further investigations pursuing specific diagnoses that might be considered in particular cases (Table 16.17.3.1) are described in the following sections. Table 16.17.3.1  Evidence in history, examination, or routine investigations suggesting a secondary cause for hypertension Clinical Evidence Condition to consider History Paroxysmal features—​palpitations, sweating, pallor, panic, headache or chest pain, cool peripheries Phaeochromocytoma, paraganglioma Flushing, labile blood pressure Carcinoid syndrome Personal or family history of renal disease Renal hypertension Pregnancy Pre-​eclampsia, HELLP syndrome Drug history—​oestrogen-​containing oral contraceptives; corticosteroids; nonsteroidal anti-​inflammatory drugs; sympathomimetics (amphetamines, cocaine, cold cures, nasal decongestants); corticosteroids; ciclosporin; angiogenesis inhibitors, leflunomide; liquorice; caffeine; carbenoxolone; sodium bicarbonate (often found in excessive amounts in effervescent medications, antacids); ergotamine; triptans; monoamine oxidase inhibitors (with tyramine-​containing foods); erythropoietin; chronic arsenic exposure; long-​term alcohol use; smoking cessation therapies; tramadol (enhances serotonergic and adrenergic transmission) Drug-​induced hypertension Tetany, muscle weakness, fatigue Primary aldosteronism (Conn’s syndrome) Examination General appearance Cushing’s syndrome, acromegaly, thyroid disorders, obstructive sleep apnoea Palpable kidney(s) Adult polycystic kidney disease, tuberous sclerosis Abdominal or loin bruits Renovascular disease Delayed or weak femoral pulses Coarctation of the aorta Investigations (basic) Proteinuria, haematuria, or elevated serum creatinine (eGFR<30; CKD 4/​5) Renal or renovascular disease Hypokalaemia, metabolic alkalosis Primary aldosteronism (Conn’s syndrome) Hypercalcaemia Hyperparathyroidism Hyperglycaemia Phaeochromocytoma Box 16.17.3.1  Indications for investigation for secondary causes of hypertension • Any evidence of an underlying cause in the history or examination (Table 16.17.3.1) • Proteinuria, haematuria, or elevated serum creatinine (eGFR<30; CKD 4/​5) • Hypokalaemia, even if caused by diuretics • Hypercalcaemia • ‘Incidentalomas’ of the adrenal glands seen on radiological imaging • Accelerated (malignant) hypertension • Documented recent onset or recent worsening of hypertension • Resistant hypertension (not controlled with three antihypertensive drugs) • Young age—​any hypertension <20 years; any hypertension needing treatment <35 years

16.17.3  Secondary hypertension 3781 Primary aldosteronism (Conn’s syndrome) History In 1955, with the words ‘to our surprise and delight, a cortical ad- enoma was observed to be arising from the right adrenal gland’, Jerome Conn reported the first observation of the benign aldosterone-​secreting tumour that now bears his name. The patient had presented with severe hypertension and hypokalaemia, shortly after the discovery of aldosterone (‘electrocortin’) in London by the Taits in 1953. On detecting a high level of aldosterone in the patient’s urine, and in the absence of any easy method for imaging adrenals in the 1950s, Conn decided to remove both glands. There is an his- torical irony in this decision: not so much because the patient even- tually retained her left adrenal, but because the finding of unilateral disease in this patient has largely pre-​empted the same decision being made in patients with truly bilateral disease. Conn’s report led to a flurry of similar diagnoses and optimism that as much as 20% of hypertension might be due to his tumour. However, it soon became apparent that no adenoma could be found in perhaps 50% of patients with the clinical and biochemical fea- tures of primary aldosteronism, some being diagnosed instead as having bilateral nodular hyperplasia. With waning enthusiasm for finding a curable cause of hypertension, estimated prevalence fell to less than 1% of hypertension, but the picture again reversed with the recognition that not all patients with primary aldosteronism have an elevated plasma aldosterone concentration; indeed, it is now estimated that 5–​10% of hypertensive patients have a poten- tially curable cause. Whereas previously low-​renin hypertension was often considered a separate diagnosis, increasingly it is felt that even in such patients aldosterone drives the suppression of renin and that smaller adenomas are found when newer radiotracer imaging modalities are employed. Aetiology and pathology A Conn’s adenoma is a small (0.5–​3.5 cm), benign tumour. Although aldosterone is normally secreted selectively by the (outer) zona glomerulosa of the adrenal, classical Conn’s adenomas resemble the cortisol-​secreting zona fasciculata, and secrete more cortisol than aldosterone; occasionally, the extra cortisol may be sufficient to cause suppression of the contralateral adrenal. In recent years, it has become apparent that aldosterone-​producing adenomas of the zona glomerulosa are also common, but are often missed because of their smaller size. Adenomas arising in the two zones are character- ized by somatic mutations in different genes (KCNJ5, encoding a K+ channel, in zona fasciculata tumours; CACNA1D encoding Cav1.3, an L-​type Ca2+ channel; ATP2B3 and ATP1A1, encoding a Ca2+-​ and Na+/​K+-​ATPase, respectively, in zona glomerulosa tumours). Presentation in pregnancy and after the menopause of aldosterone-​ producing adrenal adenomas harbouring activating mutations of CTNNB1, encoding β-​catenin in the Wnt cell-​differentiation pathway, has also been described. Bilateral adrenal hyperplasia is a distinct condition in which either radiologically or histologically there are macro-​or micro-​ nodules in the adrenal cortex where the monolayered arcades of the normal zona glomerulosa are replaced by bi-​or multicel- lular layered arcades. In the one type of primary aldosteronism of known cause—​glucocorticoid-​remediable aldosteronism (see Chapter  16.17.4)—​there is no anatomical lesion in the adrenals other than expansion of the zona glomerulosa. It remains unknown whether some patients develop single adenomas on the background of nodular hyperplasia, with suppression of all but the dominant nodule, or whether unilateral adenomas are usually a different condition from hyperplasia. In favour of the latter are several bio- chemical and pharmacological differences, and the fact that patients with glucocorticoid-​remediable hyperaldosteronism never develop a super-​added adenoma. Patients with classical Conn’s adenomas show an exaggerated diurnal rhythm in aldosterone secretion, con- sistent with an enhanced ACTH-​dependent cAMP pathway. By contrast, patients with hyperplasia, and those with the small zona glomerulosa tumours, show exaggerated aldosterone response to stimulation by angiotensin II and therefore have higher erect than supine aldosterone levels. Primary aldosteronism causes increased sodium retention through the epithelial sodium channel (ENaC) in the distal tubule and cortical collecting duct. The chronic sodium retention leads to hypertension, which is an essential feature of Conn’s syndrome. Electroneutrality in the tubular cell is retained by secreting K+ and/​or H+ ions in exchange for the Na+ with consequent hypokal- aemic alkalosis. Epidemiology Adenomas are slightly commoner in women, bilateral hyper- plasia commoner in men. Conn’s syndrome is not a cause of child- hood hypertension, except for the rare monogenic syndrome of glucocorticoid-​remediable hyperaldosteronism. Hyperplasia is said to be commoner among older patients with hypertension. However, it is difficult clinically to distinguish hyperplasia from small zona glomerulosa adenomas. Since it is likely that the latter have been pre- sent for many years or decades before presenting with often resistant hypertension, the best and easiest time to look for them is in younger patients, where hyperplasia is less likely and surgical treatment is most rewarding. Another condition which needs distinguishing from aldosterone-​producing adenomas in older patients is non-​ functioning adrenal adenomas (‘incidentalomas’), which are present in at least 4% of people over 50. Overall prevalence remains contentious because of the detailed investigations required to establish presence or absence of func- tioning adenomas. In younger patients, where non​functioning adenomas and low-​renin hypertension are both uncommon, and response to surgery is more clear-​cut, a conservative estimate would be 2% of those with hypertension, but the discovery of the smaller zona glomerulosa tumours may double this number. The prevalence among older patients with hypertension is probably similar. At present, a higher proportion of the older age group are likely to be investigated, having presented with either resistant hypertension or an adrenal incidentaloma, but in reality a smaller proportion are likely to benefit from surgery. Whether investi- gations reduce or increase in coming years may depend on the success of less invasive modalities than in current use for both in- vestigation and treatment of adenomas, and extension of the latter into bilateral disease. Clinical features Patients with primary aldosteronism ‘escape’ from the effects of al- dosterone before sufficient Na+ is retained to cause overt oedema,

section 16  Cardiovascular disorders 3782 hence the clues and confirmation of the diagnosis are largely bio- chemical. The classic picture in Conn’s syndrome is hypertension in which the plasma electrolytes show a low K+, elevated bicarbonate, and a Na+ typically at the upper end of the normal range, but some- times above this. The hypertension is often resistant to treatment with conventional treatment for the patient’s age group; for example, angiotensin converting enzyme (ACE) inhibition in a younger pa- tient, or to multiple drugs including a thiazide diuretic in the older age groups. It is important to mention, however, that the classical hallmark—​hypokalaemia—​is not always present, and yet the con- sequences of K+ depletion—​weakness, tiredness, U wave on ECG (Fig. 16.17.3.2)—​might still be manifest. The severity of hypo- kalaemia varies steeply with the Na+ load presented to the ENaC, this depending partly on dietary Na+ intake and partly on drugs—​ principally thiazide diuretics—​which affect the proportion of the filtered Na+ load reaching the distal tubule. The commonest reason for the biochemical features of Conn’s to be masked is concurrent treatment with a calcium channel blocker. Hence, when considering the possibility of Conn’s in a patient with hypertension apparently resistant to conventional treatment, it is important to look not just at the current plasma electrolytes but at an historical set of results for any finding of hypokalaemia or alkalosis, and also to reflect that hypokalaemia on a low dose of thiazide is a reason for pursuing (ra- ther than dismissing) the diagnosis of primary aldosteronism. Differential diagnosis The hypokalaemic hypertensive is an interesting diagnostic challenge that can usually be solved by a series of logical moves. The finding of a Conn’s adenoma is the most satisfying outcome because surgical ex- cision is most likely to lead to long-​term cure if the patient is young. The other curable cause is liquorice consumption which, taken in excess, inhibits the enzyme 11β-​hydroxysteroid dehydrogenase (11HSD) and permits cortisol access to the mineralocorticoid re- ceptor (see ‘Apparent mineralocorticoid excess’ in Chapter 16.17.4). Excess production of cortisol in Cushing’s syndrome can also mimic Conn’s. This is most likely to happen when there is ectopic ACTH production or with a malignant adrenocortical tumour, resulting in gross excess of cortisol and consequent saturation of the 11HSD en- zyme (Fig. 16.17.3.3). Cosecreting, or coexisting, aldosterone-​ and cortisol-​producing adenomas should also be considered, although generally the clinical picture is predominantly of one or the other. Clinical investigation Electrolytes The critical tests in the investigation of hypokalaemic hyperten- sion are plasma and urine electrolytes, and plasma renin and al- dosterone. If the recommendations described earlier for screening tests in young patients with hypertension have been observed, all but the plasma aldosterone should already have been performed. The urine K+ (which can be performed on a spot sample) is usu- ally in excess of 40 mmol/​litre if hypokalaemia is due to increased urinary loss, but this test is valuable only when performed when plasma K+ is low. Transient hypokalaemia is common, and hypo- kalaemia transient commonly even in Conn’s syndrome, hence it is important not to postpone urine K+ estimation and risk missing a one-​off opportunity for sparing a patient the further investigations required for renal K+ loss. Renin Of the triad of hypokalaemia, suppressed plasma renin, and ele- vated aldosterone, the renin is of most importance in the diagnosis of Conn’s—​although renin suppression is not invariable, even in un- treated patients. The diagnosis should be entertained in the absence of an elevated aldosterone, especially in patients where a suppressed renin is unexpected: the younger patient (aged <45 years), particu- larly if already on an ACE inhibitor or angiotensin receptor blocker (ARB); and the older patient with resistant hypertension, receiving multiple drugs which normally elevate renin. Fig. 16.17.3.2  A 12-​lead electrocardiogram (ECG) showing ‘u’ waves (arrowed) which may be indicative of hypokalaemia.

16.17.3  Secondary hypertension 3783 The main confounders in interpretation of the plasma renin level are excessive dietary salt intake and drugs. A low renin in the pres- ence of β-​blockade is of no significance, and a β-​blocker (which is unlikely to help with blood pressure control in Conn’s anyway) should be discontinued or substituted by an ACE inhibitor or ARB two weeks before renin measurement. Conversely, spironolactone or amiloride will ‘desuppress’ renin in most patients. Guidelines require these drugs to be stopped prior to hormone measure- ments, in the case of spironolactone for 6 weeks because of its long-​lived active metabolite. This can be clinically challenging in the common group of patients with primary aldosteronism with resistant hypertension. Empirically, a plasma renin found to be suppressed despite low-​dose spironolactone may be sufficient in- dication to investigate adrenal anatomy. Renin itself is very stable in blood, providing this is not chilled (which cryoactivates the renin precursor, prorenin). Although changes in posture and activity cause two-​to threefold changes in renin, the range of renin between high-​and low-​renin patients is some 1000-​fold, hence it is simple to interpret results taken in rou- tine outpatient clinics or surgeries, providing the blood sample (taken into an EDTA tube) reaches the laboratory for plasma separ- ation on the same day as the blood is taken. Aldosterone Plasma aldosterone is often elevated above the normal range (100–​400 pmol/​litre), and is generally higher in patients with macroadenomas (>1 cm) than in those with microadenomas or hyperplasia. In patients with adenomas there is an exaggerated influence of ACTH leading to pronounced diurnal variation in aldosterone levels, which are more likely to be normal when sampled in the afternoon. By contrast, patients with hyper- plasia have an exaggerated response to angiotensin II, so that levels may actually rise during the day in response to activity and be normalized by drugs blocking the renin system, particu- larly angiotensin receptor blockade. However, the most pro- found influences are serum K+ and the use of calcium channel blocker treatment, which (as already stated) is probably now the commonest reason for the diagnosis of Conn’s syndrome to be missed. Aldosterone/​renin ratio The recognition that aldosterone is often normal, despite correc- tion of hypokalaemia and withdrawal of confounding medica- tions, led to the concept of the aldosterone/​renin ratio. However, in practice, because renin is log-​normally distributed and aldos- terone distribution is normal, the aldosterone/​renin ratio is high in most low-​renin patients, except in the low-​renin, low-​aldosterone differential diagnoses considered earlier for hypokalaemic hyper- tension. Indeed, the suppression of plasma aldosterone in patients with liquorice-​induced hypertension is an important observation, showing that the appropriate aldosterone response to suppression of renin is suppression of aldosterone, unless aldosterone is itself the cause of Na+ retention and renin suppression. Because patients with a normal aldosterone but elevated aldosterone/​renin ratio are so numerous, the key question is how to avoid unnecessary inves- tigations in these cases. Until investigation and cure of primary al- dosteronism becomes simpler, the empirical answer is that in the absence of other clues—​hypokalaemia, high/​high-​normal plasma Na+, alkalosis—​investigation be undertaken only in young patients (aged <35) and those with resistant hypertension. Confirmatory tests In patients lacking the triad of plasma aldosterone more than 550 pmol/​litre, undetectable plasma renin, and spontaneous hypokal- aemia, a dynamic test is recommended before proceeding to radio- logical investigations, although the accuracy of such tests is disputed. The most commonly employed is a 4-​hour infusion of saline, after which plasma aldosterone should be more than 190 pmol/​litre in those with primary aldosteronism. An alternative is the captopril suppression test, requiring plasma aldosterone to be similarly unsup- pressed (>190 pmol/​litre) 4 hours after captopril 25 mg. Since patients with resistant hypertension are, by definition, already receiving a drug equivalent to captopril, the decision whether to look for an adrenal adenoma in this group need not be dependent on a confirmatory test Fig. 16.17.3.3  Coronal (a) and axial (b) CT scan images of a large right-​sided malignant adrenocortical tumour (horizontal arrow) that is invading the liver (vertical arrows).

section 16  Cardiovascular disorders 3784 of their diagnosis. Its role in diagnosis can be reserved for a later stage in the diagnostic algorithm, in patients where the case for surgery is borderline, and when extra certainty about diagnosis is required. Genetic testing This is rarely required, but if there is a family history of early-​onset hypertension, and particularly of strokes at a young age, the patient should be screened for glucocorticoid-​remediable aldosteronism (see Chapter 16.17.4), of which there are only a few known families in the United Kingdom. Interestingly, research is increasingly providing fascinating insights into the genetics of aldosterone-​producing ad- enomas, which in turn may lead to improved diagnosis in future, without the need for adrenal venous sampling. Further discussion can be found in the ‘Aetiology and pathology’ section of this chapter. Scanning The adrenals are easily imaged by either CT or MRI, except when there is a dearth of intra-​abdominal fat (Fig. 16.17.3.4). There is no proven advantage of one of these two modalities over the other. MRI may be preferred in younger patients, to spare radiation, and the fat-​suppression sequence is useful for differentiating aden- omas both from other adrenal masses, and sometimes the adjacent normal adrenal. Resolution may be higher with CT, but the limit for both modalities is not so much inherent resolution as the existence of 0.3–​0.6-​cm adenomas that do not create a discrete bulge within an adrenal limb. Even 1-​cm adenomas at the bifurcation of the two limbs can be difficult to distinguish from a normal gland. It is valu- able to request coronal reconstructions, which may show or confirm adenomas less evident on the axial views. Neither MRI nor CT can differentiate functional from incidental adenomas. Functional lateralization This is the key but most difficult stage of diagnosis. Lateralization is essential in predicting that removal of one adrenal will have a substantial benefit, as well as indicating which adrenal to remove, although it might occasionally be omitted in younger patients (aged <35 years) with macroadenomas, or where the tumour is more than 3.5 cm in diameter and needs to be removed to exclude a mixed ad- renal carcinoma. At present, the only reliable form of lateralization available at most specialist centres is adrenal vein sampling. This is technically demanding and should be undertaken only by experienced radiolo- gists (Fig. 16.17.3.5). On the left side, the adrenal vein is the only vein to enter the renal vein superiorly, and cannulation is relatively straightforward. On the right, however, the adrenal vein is one of sev- eral small veins (<1 mm diameter) entering the inferior vena cava posteriorly. A  fish-​hooked ‘Cobra’ catheter with side-​holes maxi- mizes the chances of success at 80%, providing several veins are sam- pled, with reference samples also taken in the inferior vena cava above and below the adrenal veins. Centres vary in whether or not they ad- minister ACTH 250 micrograms as a 2-​hour infusion prior to the procedure. Confirmation of adrenal vein cannulation requires that the cortisol measurement is at least threefold higher than in the IVC. All samples need to be assayed for aldosterone and cortisol, with the ratio compared between the two sides: a ratio above 4 is considered diagnostic. Ratios of two-​to fourfold can be compatible with lateral- ization, but are best confirmed on repeat sampling. In such cases ac- curacy might be enhanced by simultaneous sampling from both veins, or—​if aldosterone levels were low on the first occasion—​by prior ACTH stimulation. When an adrenal vein (usually the right) cannot be cannulated, it is very risky to draw conclusions from the contra- lateral sample alone: concentrations of aldosterone can be very high, even from a normal gland, because adrenal vein blood flow is so low. Isotope scans can also be used for lateralization. 131I-​cholesterol (or 75Se-​methyl-​19-​norcholesterol) can be bought or generated for scanning in any nuclear medicine department, but 11C-​metomidate (Fig. 16.17.3.6) must be synthesized on site in centres with a cyclotron and positron emission tomography (PET) scanner. The cholesterol scans rely on its role as precursor of steroid synthesis, and the scan is performed 1 week after isotope administration to Fig. 16.17.3.4  Conn’s adenoma (arrow): CT transverse view (left), coronal view (middle), surgical specimen showing classical yellow, lipid-​rich adenoma (right).

16.17.3  Secondary hypertension 3785 permit cholesterol turnover and elimination from non​adrenal sites. However, the technique has a generally unreliable record, possibly because the dexamethasone taken during the week of in- vestigation has variable influence on zona glomerulosa as well as zona fasciculata uptake. Metomidate binds to synthetic enzymes in both the aldosterone and cortisol pathway, but its uptake is in- creased in aldosterone-​producing adenomas, compared to adjacent or contralateral normal adrenal cortex. This in vivo selectivity for aldosterone synthase probably reflects the higher concentration of the enzyme in adenomas—​especially those of the small zona-​ glomerulosa subtype—​due to their high density of mitochondria and inner membrane reduplication, as seen on electron microscopy. Treatment Medical Medical treatment is preferred for bilateral adrenal hyperplasia, before surgery for adenoma, in older patients with adenoma who are well controlled, or where there is any doubt about diagnosis or lateralization. Chronic medical treatment is by K+-​sparing diuretic, preferably spironolactone or amiloride. Spironolactone is a competitive an- tagonist of aldosterone, hence patients with very high aldosterone levels may require higher doses than used in resistant hypertension. While this is possible for preoperative use, long-​term administra- tion causes gynaecomastia. High-​dose amiloride (20–​40 mg daily) is better tolerated but may be less effective. Eplerenone also avoids the gynaecomastia of spironolactone, but again is less effective and more expensive. A possible strategy is to combine eplerenone or a low dose of spironolactone (≤25 mg daily) with amiloride, but vigi- lant monitoring of plasma electrolytes is required. It may not be possible to control blood pressure entirely by diur- esis, especially in older patients, where calcium channel blockers or α-​blockers can usefully be added. In patients who are difficult to con- trol, the maximum useful dose of diuretic can be found by titrating dose against plasma renin: once this is de-​suppressed it becomes lo- gical to add ACE inhibition or an ARB. In patients with bilateral hyperplasia, one of these classes is often required, even when renin is suppressed. This may reflect either the resistant nature of hyper- tension that often ensues with long-​standing hyperaldosteronism, or the increased sensitivity to angiotensin in salt-​loaded patients. Surgical Elective surgery is indicated for younger patients with adenomas, and older patients intolerant of medical treatment, or uncontrolled by it. A good blood pressure response to spironolactone may augur well for the blood pressure response to surgery, indicating that hypertension is largely due to excess aldosterone (rather than sec- ondary consequences of this). If lateralization has been correctly performed, most patients can expect cure of hypokalaemia, if pre- sent, and a substantial reduction in number of medicines required to control blood pressure. A bonus in many patients is alleviation of chronic fatigue, presumably attributable to total body K+ depletion. Surgery should be undertaken by a surgeon experienced in laparo- scopic adrenalectomy, but patients warned that anatomical anom- alies, or perioperative eventualities such as tear of the inferior vena Vein Aldosterone (pmol/litre) Cortisol (nmol/litre) Ratio 1 Left adrenal 3520 3440 1.02 2 Right adrenal 7520 4310 644 552 11.67 7.80 3 IVC 254 187 1.35 Fig. 16.17.3.5  Adrenal vein sampling for a right adrenal adenoma. CT PET PET/CT overlay Fig. 16.17.3.6  11C-​metomidate PET/​CT scan of a right adrenal aldosteronoma. Uptake correctly differentiated hot and cold nodules, as confirmed by presence and absence of aldosterone secretion from the nodules when cultured postoperatively.

section 16  Cardiovascular disorders 3786 cava, may necessitate conversion to open adrenalectomy in about 1/​40 procedures. No special preoperative care is required, although it is sensible to undertake assessment to exclude hypercortisolism in those patients with large adenomas. Diuretics should be withdrawn from the time of surgery, but any additional antihypertensive treat- ment continued until any change in blood pressure becomes clear over the following weeks. Radiofrequency ablation of adenomas is starting to be reported, preserving the adjacent normal adrenal gland. To date, these have been undertaken percutaneously or via a laparoscope. However, left-​ side adenomas, sitting close to the stomach, are accessible for abla- tion, delivered using endoscopic ultrasound. A prospective study of this approach is in progress, which—​if successful—​should lower the bar at which intervention for a benign tumour is considered. Ablation also opens up the possibility of cure for the increasing number of pa- tients found to have bilateral aldosterone-​producing adenomas but in whom bilateral adrenalectomy would never be considered an option. Prognosis The average cure rate across reported studies is 30–​60%. Younger patients have a higher likelihood of cure than older, which probably reflects a shorter time of exposure to aldosterone excess. However, younger patients are also more likely to have larger KCNJ5-​mutant adenomas than smaller adenomas with other mutations. The latter appear more often on the background of multiple adenomas or hyperplasia and are perhaps less likely to be completely unilateral than binary clinical tests suggest. Hypokalaemia is usually cured by surgery, and either the number of medicines required is reduced and/​or there is improved blood pressure control. Renal hypertension The principal curable cause is renovascular hypertension. This is usu- ally due to a stenosis in one or both renal arteries, but can be due to a suprarenal aortic stenosis. Other curable causes include renal tumours (hypernephroma and, the rarest of all secondary causes, a juxtaglo- merular renin-​secreting tumour or reninoma); a unilateral, poorly functioning scarred or hydronephrotic kidney which hypersecretes renin, and can be removed without unacceptable loss of renal function; kidneys that are subject to long-​standing compression (e.g. by post-​ traumatic subcapsular haematoma or extrinsic mass; so-​called Page kidneys; Fig. 16.17.3.7); and various causes of acute/​subacute glom- erulonephritis, some associated with systemic disorders whose treat- ment by immunosuppression cures the hypertension and underlying disorder. Interestingly, aortic dissection, often itself a complication of arterial hypertension, may extend into the renal arteries and thereby ex- acerbate hypertension by causing increased secretion of renin. However, whatever the cause of renal hypertension, there is no evidence to sup- port renal denervation therapy, even in seemingly resistant cases. Renovascular hypertension This is most commonly due to intrinsic disease of the intima (ac- quired, as in atherosclerosis) or media (congenital, as in fibromuscular dysplasia). Extrinsic narrowing can be caused by ligamentous bands or by tumours (e.g. neurofibromas). Fibromuscular dysplasia (FMD) accounts for only 10–​20% of all patients with renovascular hypertension, but is the commonest cause under the age of 40. It is a non​atherosclerotic and non​inflammatory disease of small and medium arteries, usually affecting the media, less commonly the adventitia (<25%), and rarely the intima. The classical ‘string of beads’ appearance seen at arteriography results from pro- liferation of the extracellular matrix and disruption of the internal elastic lamina, causing multiple stenoses and poststenotic saccular aneurysms. The condition affects women more often than men, and there is usually no family history of hypertension. FMD involves extrarenal arteries in about one-​quarter of patients, with cerebral in- farction recorded due to relative hypotension and hypoperfusion of FMD-​affected carotid arteries following successful renal angioplasty. The typical medial form of FMD does not affect the proximal part of the renal arteries (Fig. 16.17.3.8) and is bilateral in about one-​ third of cases. Other vascular beds (e.g. the cerebral arteries), can be affected. Complications (other than renal ischaemia) are rare, whereas dissection or thrombosis can ensue in the rarer intimal or adventitial form of FMD. Rupture of renal artery aneurysms is rare. Fig. 16.17.3.7  A Page kidney. CT scan image showing a left-​sided neuroblastoma compressing the kidney (arrow). The compression impedes renal blood flow, resulting in excess secretion of the hormone renin and consequently hypertension. Named after Irvine Page (1901–​89) who demonstrated that wrapping cellophane tightly around an animal’s kidney caused arterial blood pressure to rise. The patient’s hypertension was cured after surgery to remove the tumour. Fig. 16.17.3.8  MR angiogram demonstrating fibromuscular dysplasia of the right renal artery causing stenosis (arrow).

16.17.3  Secondary hypertension 3787 Atheromatous renal artery stenosis has the same risk factors as atheromatous disease of other arteries, which often coexists. It is thus commoner in older men, and whereas FMD rarely causes renal impair- ment, atheromatous disease is often discovered in the context of inves- tigation of hypertension with chronic kidney disease. Apart from the obvious difference in biology of FMD and atheromatous renovascular hypertension, there is a difference in location of the stenosis, which is more likely to be proximal in atheromatous disease (Fig. 16.17.3.9). Mechanism of hypertension Unilateral renal artery stenosis gives rise to an endocrine disorder, because reduced pressure in the afferent arteriole causes juxtaglo- merular hyperplasia and increased renin secretion. The consequent increase in angiotensin II formation causes hypertension, partly by vasoconstriction, and partly through increased aldosterone se- cretion. Although secondary hyperaldosteronism is not usually a marked feature of renal artery stenosis, the combination of hypokal- aemia and hyponatraemia should raise suspicion of the diagnosis, the latter being dilutional and due to the inhibition by angiotensin II of free water clearance. The effect on renin secretion is less predict- able when renal artery stenosis affects both renal arteries: sometimes it is high, but sometimes bilateral reduction in GFR leads to suffi- cient sodium retention that renin is suppressed. Diagnosis Most cases of renovascular hypertension are probably not diagnosed because of the absence of sensitive clinical or biochemical markers. Lack of a family history of hypertension in younger patients, or re- cent onset (or exacerbation) of hypertension in older patients is more likely than in essential hypertension. Acute shortness of breath, due to flash pulmonary oedema, can be the presenting feature of bilateral renal artery stenosis. However, the main clinical clue is the finding, in about one-​half of the patients, of a bruit anteriorly or posteriorly over a renal area. It is important to remember, however, that such a bruit is never diagnostic: normal abdominal arteries can give rise to innocent flow murmurs in younger patients and in older pa- tients a bruit could arise from any of a number of arteries within the abdomen. The response to antihypertensive drugs can also give clues: in particular, poor response to β-​blockade in younger patients, or rapid worsening of renal function in older patients. The diagnosis of renal artery stenosis is made radiologically. The cheapest investigation is a nuclear medicine scan using technetium-​ labelled MAG3, both the uptake and elimination of this being delayed on the ischaemic side, with the difference in excretion rate between sides greatly increased following a single dose of captopril (25 mg) because of dilatation of the efferent arterioles in glomeruli and conse- quent reduction in filtration fraction. For this reason the scan is best performed initially with captopril; if abnormal, it is repeated on a sub- sequent visit without captopril, with partial or complete normaliza- tion being evidence that the previous abnormality was due to vascular rather than renal parenchymal disease. However, the MAG3 scan is not always positive, with chronic use of ACE inhibitors being a cause of some false negatives, and it may also miss bilateral renal artery sten- oses that do not cause significant asymmetry between the kidneys. Partly for these reasons, nuclear imaging is not performed for suspected renal artery stenosis in most centres, with investigation proceeding to direct imaging of the renal arteries by CT or MR angiography (Fig. 16.17.3.8 and Fig. 16.17.3.9). In patients under 20 years of age some form of angiography should always be under- taken, except in those with low-​renin syndromes, because of the high likelihood of a secondary cause being present, and that this will be a vascular abnormality. As well as providing an accurate estimate in most patients of the severity of any stenosis, angiography will also detect suprarenal aortic stenoses. False-​positive and false-​negative diagnoses still occur; for example, the poststenotic dilatations of FMD can—​if they expand proximally around the artery—​be a cause of stenoses being missed. However, the risk of diagnostic error can be reduced by careful review of images taken in more than one projec- tion, and it is useful to remember that stenoses are not usually isolated lesions in both FMD and atheromatous disease (Fig. 16.17.3.9). Some centres use Doppler flow measurements for diagnosis, but these are more user-​dependent than angiography, which is still re- quired subsequently for anatomical diagnosis. On the other hand, there are some patients in whom an anatomical diagnosis is made first, but the severity is in question. Here it can be useful to per- form Doppler or MAG3 scan as the second investigation before proceeding to treatment. Another investigation that is sometimes helpful at this stage is renal vein sampling for renin determination, the main use for which is before removing a kidney thought respon- sible for causing hypertension through elevated renin secretion. The contralateral—​anatomically normal—​kidney has often sustained microvascular damage as a consequence of prolonged hypertension and renin excess, and is found to secrete as much renin as (or more than) the diseased kidney. Nephrectomy should not normally be contemplated where significant renal function remains, but in any circumstance there would rarely be an indication for removing a Fig. 16.17.3.9  MRI scan image showing a left renal artery stenosis (yellow arrow at bottom of image) with concomitant left subclavian artery stenosis (yellow arrow at top of image) and also an infrarenal aortic stenosis (blue arrow). This patient presented with a subclavian steal syndrome and lower blood pressure readings in the left arm.

section 16  Cardiovascular disorders 3788 kidney showing less than 25% excess renin secretion compared to the contralateral side. Treatment There are several options, one of which is simply to continue optimal drug treatment if for any reason the risks of other intervention appear excessive. Among interventions, the options are as for any other arterial stenosis, namely angioplasty, stenting, or surgery. For FMD, angio- plasty is usually curative, and about three-​quarters can discontinue antihypertensive treatment (Fig. 16.17.3.10). In atheromatous disease, angioplasty is much less likely to be successful, especially for lesions at the origin of the artery, and restenosis can occur. It is reasonable to rec- ommend stenting as a backup procedure when angioplasty has failed. Complete cure of hypertension is very much less likely than in FMD. In the past, the purpose of intervention was often to try to pro- tect or improve renal function. The ASTRAL and CORAL trials have largely rebutted this objective, although some argue the meth- odologies were flawed and that patients were excluded from partici- pation where clinicians were certain of benefit from intervention. Few nephrologists now pursue renal revascularization with the same vigour that was common ten or so years ago, although many still sup- port intervention in selected patients, such as those with renal artery stenosis of more than 80% with a significant translesional pressure gradient; younger patients with difficult to control blood pressure on more than three antihypertensives; and those with truncal rather than ostial stenosis, rapidly deteriorating renal function, flash pul- monary oedema, or post-​transplant renal artery stenosis. Sometimes angioplasty is unsuccessful because balloon infla- tion fails to dent the stenosis. Surgery is required in this situation, or when failure can be predicted because stenosis is due to external compression or there is complete occlusion. A  favoured surgical procedure is autotransplantation to the pelvis. Coarctation of the aorta Coarctation of the aorta, a congenital cause of hypertension, was described pathologically in the 1700s and recognized clinically in the early 1900s. The term describes a constriction of the aorta at the point where the fetal arterial duct originates, and the condi- tion should ideally be diagnosed in early childhood, with most cases treated before hypertension develops. Coarctation represents 5 to 8% of all causes of congenital heart disease, and 25% of patients with this condition will also have a bi- cuspid aortic valve, but coarctation causes less than 1% of all cases of hypertension. Diagnosis is often delayed until the patient presents Fig. 16.17.3.10  Renal angiography (Left panel), followed by balloon angioplasty (Right panel) of right-​sided renal artery stenosis secondary to fibromuscular dysplasia. (a) (b) Fig. 16.17.3.11  (a) Cross-​sectional CT image of a 52-​year-​old woman with aortitis. Note the increased inflammatory ‘cuffing’ around the aorta (arrow). (b) Cross-​sectional and sagittal PET CT images of the same patient. Note how the aorta ‘lights up’.

16.17.3  Secondary hypertension 3789 in adulthood with hypertension, and high blood pressure can some- times develop even after surgical cure of the coarctation. Stenosis may also develop at a lower level of the aorta as a consequence of aortitis (Fig. 16.17.3.11). The mechanism of hypertension is pre- dominantly the relative renal ischaemia consequent on low perfu- sion pressure in the aorta beyond the coarctation or stenosis. The classical clinical finding in coarctation is radiofemoral pulse delay or weak lower limb pulses, confirmed by measurement of reduced blood pressure in the legs. Of greater sensitivity and spe- cificity in the clinic is a bruit—​systolic or continuous—​over the front and back of the precordium, which arises in the intercostal collaterals. Rarely, subtle clues may be seen on plain chest radiog- raphy (Fig. 16.17.3.12). The diagnosis should be confirmed by two-​dimensional (2D) echocardiography (suprasternal view) or by CT or MR angiog- raphy (Figs. 16.17.3.13 and 16.17.3.14). Treatment is by surgery, balloon dilatation, or stenting. Surgery or balloon dilation are the preferred approaches in childhood, balloon dilation and stent im- plantation in adolescents and adults. Although upper limb hyper- tension is usually cured, recurrence has been attributed to a variety of unproven factors, including a systemic vasculopathy. Phaeochromocytoma and paragangliomas Aetiology and pathology Catecholamine biochemistry Catecholamine biochemistry is summarized in Fig. 16.17.3.15. The final step in the biosynthetic pathway is the N-​methylation of nor- adrenaline (norepinephrine) to adrenaline (epinephrine), which outside the brain occurs only in the adrenal medulla because the enzyme phenylethanolamine N-​methyltransferase in the adrenal is dependent for induction on glucocorticoids, secreted at high con- centration into the adrenal portocapillary circulation. The clinical importance of this is that extra-​adrenal phaeochromocytomas, or paragangliomas, rarely produce adrenaline. Fig. 16.17.3.12  Plain chest radiographs showing ‘rib-​notching’ in a patient with coarctation (Left panel, arrow) due to the formation of significant collateral intercostal arteries. The ‘3 sign’ in a different patient with coarctation (Right panel, arrows) is caused by pre and post-​stenotic dilatation of the aorta. Fig. 16.17.3.13  MR angiogram showing coarctation of the aorta (arrow).

section 16  Cardiovascular disorders 3790 The metabolism of catecholamines is different from normal in phaeochromocytoma in that adrenaline and noradrenaline are lib- erated directly into the bloodstream, rather than mainly into the synaptic gap around sympathetic nerve endings. Noradrenaline re- leased from these is largely recaptured by neuronal and extraneuronal uptake, and metabolized before any free amine escapes into the bloodstream. Consequently, the proportion of parent amine (nor- adrenaline) to metabolite (adrenaline) is usually higher in blood and urine in the presence of a phaeochromocytoma than in any other cause of elevated catecholamine production. Pathology Phaeochromocytomas arise in chromaffin tissue and their anatom- ical distribution closely parallels the sites where this tissue is pre- sent at the time of birth. The term phaeochromocytoma reflects the dusky colour of the cut surface of the tumour, whereas the term chromaffin refers to the brownish colour caused by contact with di- chromate salts, which oxidize the catecholamines. Most phaeochromocytomas are benign, but the pathologist can rarely provide a clear distinction between those that are benign and those that are malignant: benign tumours can appear to be invading the capsule of the tumour, which is often ill-​defined, while malig- nant tumours may show no mitoses because of their slow rate of division. Yearly surveillance, or earlier if indicated clinically, with measurement of plasma metanephrine levels is recommended. Genetics At least 40% of phaeochromocytomas and paragangliomas (PPGLs) are now thought to be caused by a single driver germ line muta- tion, which makes such tumours among the most highly heritable in humans. Mosaic transmission may also occur, as well as a var- iety of well-​defined somatic mutations. As a result, genetic testing is increasingly being recommended in all patients, and over 15 dif- ferent susceptibility genes have been implicated in familial cases. Most are either involved in processes linked to angiogenesis and hypoxia-​induced cell proliferation, or to aberrant activation of kinase signalling pathways. Several mutations also cause syndromes that include PPGLs (Table 16.17.3.2), the clinical and biochemical features of Fig. 16.17.3.14  3D CT reconstruction of coarctation of the aorta (arrow). Tyrosine Hydroxylase HO CH2CHNH2 COOH CH2CHNH2 COOH CH2CH2NH2 CHCH2NH2 CHCH2NH2 OH OH HO HO HO HO HO HO HO HO Tyrosine DOPA Dopamine Noradrenaline Adrenaline DOPA Decarboxylase Dopamine β-hydroxylase Phenylethanolamine N-methyltransferase (PNMT) CHCOOH CH3O HO HO OH CHCH2NH2 HO OH CHCOOH OH CHCH2NH2 CH3O HO Dihydroxymandelic acid Noradrenaline Normetanephrine COMT COMT MAO MAO OH HO HO VMA Fig. 16.17.3.15  The biosynthetic pathway for epinephrine and norepinephrine (upper panel), and for metabolism of norepinephrine (lower panel). COMT, catechol-​O-​methyltransferase; DOPA, dihydroxyphenylalanine; MAO, monoamine oxidase; VMA, vanillylmandelic acid.

16.17.3  Secondary hypertension 3791 which are variable. Only tumours associated with mutations of succinate dehydrogenase (SDH, mainly subunits B or D) com- monly occur outside the adrenal. Paragangliomas in the head or neck are restricted to SDHD (or rarely SDHC) mutations. VHL and RET mutations may cause multiple tumour types, the site of these being determined by the site of mutation in the gene; for example, VHL type 2c missense mutations cause only phaeo- chromocytoma, while the gene deletions of type 1 cause renal cell carcinoma but not phaeochromocytoma. The main value of genotyping has become prediction of multiple (but usually be- nign) paragangliomas in patients with SDHD mutations, while patients with SDHB mutations have the highest incidence of as- sociation with malignancy or metastatic disease. There is also growing recognition of a link between susceptibility to renal cancer and PPGLs. As well as VHL, mutations in SDH, FH, and TMEM127 have been described in both. Gastrointestinal stromal tumours are known to be associated with germline SDH mutations. Epidemiology Phaeochromocytoma is a rare tumour, responsible for probably 0.1 to 1% of hypertensives, although it is possible that some of its non- ​blood-​pressure presentations are overlooked and that we select- ively detect patients in whom pressure natriuresis no longer com- pensates for the effect of vasoconstriction upon blood pressure. However, despite its rarity, phaeochromocytoma justifies the Table 16.17.3.2  Genes associated with phaeochromocytomas and paragangliomas (PPGLs) Gene Chromosome Exons Function % of mutations in PPGLs (mutation type: G—​germ line S—​somatic M—​mosaic) Frequency of malignant disease (if known) (%) SDHB 1p36.13 8 Tumour suppressor, Krebs cycle intermediate 8–​10 (G) 50 NF1 17qll.2 60 Encodes neurofibromin, suppresses cell proliferation by negatively regulating ras signal transduction 3 (G) 20–​25 (S) 11 SDHA 5p15.33 15 Conversion of succinate to fumarate in Krebs cycle <1 (G/​S) 10 TMEM127 2q11.2 4 Restricts mTORC1 activation 1–​2 (G) 10 VHL 3p25.3 3 Regulation of Hypoxia Inducible Factor (HIF) 8–​10 (G/​S) 5 RET 10qll.21 21 Cell growth and differentiation 4–​6 (G/​S) 3 SDHD 11q23.1 4 Electron transfer to ubiquinone in Krebs cycle 5–​7 (G) <3 KIF1B 1p36.22 48 Transport of mitochondria and apoptosis 4–​5 (G/​S) MEN1 11q13 14 Transcriptional regulation and cell proliferation <2 (G) 2–​10 (S) HRAS 11p15.5 7 Involved in ras signal transduction 7–​8 (S) EPAS1 2p21 17 Encodes HIF2α; involved in vasculogenesis and haematopoiesis during embryonic development 6–​12 (M/​S) ATRX Xq21.1 30 Telomere maintenance and chromosome integrity <5 (S) TP53 17p13.1 12 Gatekeeper for cell growth and division <5 (S) SDHC 1q23.3 6 Electron transfer to ubiquinone in Krebs cycle 1–​2 (G) FH 1q43 8 Hydration of fumarate to malate in Krebs cycle 1–​2 (G) BRAF 7q34 18 Involved in cell growth <2 (S) CDKN2A 9p21.3 12 Regulation of p53 and RB1 pathways <2 (S) EGLN1/​PHD2 1q42-​q43 5 Regulation of stability of HIF1 <1 (G/​S) FGFR1 8p11.23 24 Role in angiogenesis, neural and embryonic development 1 (S) H3F3A 1q42.12 4 Role in regulating transcription, DNA repair, and chromosomal stability <2 (M) IDH 2q34 12 Oxidative decarboxylation of isocitrate <1 (S) KMT2D 12q13.12 54 Regulation of accessibility to DNA <2 (G/​S) MAX 14q23.3 5 Regulates cell proliferation, differentiation, and apoptosis 1–​2 (G/​S) MDH2 7q11.23 10 Reversible oxidation of malate to oxaloacetate in Krebs cycle <2 (G) MERTEK 2q13 24 Regulation of cell survival and phagocytosis of apoptotic cells <2 (G) SDHAF2 11q12.2 4 Flavination of SDHA in Krebs cycle <1

section 16  Cardiovascular disorders 3792 disproportionate interest that it commands among physicians, combining the potential for being lethal if not diagnosed and treated, and for cure in most patients if diagnosed. The need for maintaining a high awareness of the condition is emphasized by the small number of deaths each year due to undiagnosed phaeo- chromocytoma in both anaesthetic and obstetric practice. Clinical features Hypertension is the most common presentation of phaeochromo- cytoma in clinical practice, but other rare presentations include unexplained heart failure or paroxysmal arrhythmias. Patients with large tumours occasionally remain asymptomatic, and this is the norm for small phaeochromocytomas detected through regular screening of patients with a genetic diagnosis. This is also true of many so-​called ‘incidentalomas’ seen on radiological imaging of the adrenal glands which subsequently turn out to be phaeochromocytomas. In hypertensive patients a spontaneous history or direct en- quiry will usually reveal at least one of a group of characteristic symptoms. The classical triad comprises headache, sweating, and palpitations; less frequent are episodes of pallor, a feeling of ‘impending doom’, and paraesthesiae. Spontaneous haemorrhage and infarction in the tumour can be associated with local pain and (on occasion) systemic features of tissue necrosis, and rarely the patient can present with the features of full-​blown retroperi- toneal haemorrhage, coupled to a pathognomonic swinging blood pressure. Most of the symptoms of phaeochromocytoma can be readily ascribed to the expected effects of catecholamine excess and dis- appear rapidly on initiation of appropriate treatment. Because large tumours principally secrete noradrenaline, even when arising within the adrenal gland, tachycardia is usually only modest, and can be replaced altogether by reflex bradycardia when episodes of hypertension are triggered by release of noradrenaline alone. The bradycardia can be severe enough—​if the hypertension is high enough—​to be misdiagnosed as asystolic cardiac arrest, and the correct treatment is not atropine but phentolamine to reduce the blood pressure. Severe bradycardia is also recorded in response to the paradoxical rise in blood pressure when patients with phaeo- chromocytoma are inadvertently given a non​selective β-​blocker such as propranolol. Often, however, the clinical features are less im- pressive than might be expected, possibly because the adrenoceptors have been down-​regulated by years of exposure before the diagnosis is first entertained. Examination may reveal a bruit over the tumour. A Raynaud’s type of discolouration over the extremities and the larger joints in the limbs can be caused by ischaemia. Clinical investigation The diagnosis of phaeochromocytoma is usually not difficult once the possibility has been entertained; often more difficult is excluding the diagnosis in patients who have clinical and/​or bio- chemical features of physiological catecholamine excess. There are two distinct questions to ask in order. ‘Does the patient have a phaeochromocytoma?’, and ‘Where is it?’. It is unwise to use radio- logical tests to answer the first question because of the risk of false positives and negatives. Biochemical analyses of catecholamines
and their metabolites Twenty-​four-​hour urine samples measure integrated catecholamine release and provide a useful screening test, with catecholamine me- tabolites less temperamental to assay than the more unstable cat- echolamines themselves. Vanillylmandelic acid (VMA) measured by high-​performance liquid chromatography (HPLC) is least prone to interference, L-​DOPA and paracetamol being the main concerns, and although now regarded as less sensitive than some alternatives it is still the exception for VMA to be entirely normal in a patient with hypertension due to a phaeochromocytoma. Metanephrines (sometimes called metadrenalines) measured by radioimmuno- assay or gas chromatography–​mass spectrometry (GCMS) are more sensitive and more specific than VMA, with the assay of ‘frac- tionated metanephrines’ permitting separate evaluation of nor- adrenaline and adrenaline secretion. The ability to differentiate physiological release of noradrenaline from sympathetic nerve endings from pathological secretion from a phaeochromocytoma arises because of the presence of two different enzymes in the two locations: monoamine oxidase (MAO) in sympathetic nerves, and catechol-​O-​methyltransferase (COMT) in the adrenal medulla and phaeochromocytoma (Fig. 16.17.3.15). The measurement of free catecholamines in plasma (which have a very short half-​life) by HPLC with electrochemical detection allows short bursts of secretion during a possible phaeochromocytoma crisis to be detected. However, the technique is susceptible to interfer- ence, especially in the adrenaline peak, and the finding of an adren- aline concentration that is higher than that of noradrenaline should be regarded as suspect. Dopamine levels are usually undetectable in plasma, whereas it is the major catecholamine in urine as a product of renal decarboxylation of plasma dihydroxyphenylalanine. Only several-​fold increases in urinary dopamine are of diagnostic value, and are more likely to indicate neuroblastoma (in a child) or mel- anoma (which secretes dopamine as a by-​product of melanin syn- thesis) than phaeochromocytoma. Paragangliomas, particularly those of the head and neck area, tend not to secrete catecholamines. Up to 70% of those secondary to SDHB and SDHD mutations may have elevated levels of methoxytyramine, a dopamine metabolite. Most adrenal phaeochromocytomas secrete adrenaline (and there- fore metadrenaline), the exceptions being patients with very large tumours, which completely disrupt the portocapillary supply of cor- tisol required to induce phenylethanolamine N-​methyltransferase, and patients with von Hippel–​Lindau syndrome, who tend to secrete noradrenaline and often have normal adrenaline levels even when the tumour is small. By contrast, in patients with multiple endocrine neoplasia (MEN) an elevated plasma adrenaline concentration is the first biochemical abnormality. Usually the normal adrenal predom- inance of adrenaline over noradrenaline is reversed as the tumour enlarges. Occasionally even large tumours secrete mainly adrenaline if either the tumour’s centre is infarcted, leaving a rim still exposed to cortical cortisol supply, or the tumour itself is secreting ACTH or corticotrophin releasing factor. This secretion may be triggered by α-​blocker therapy, and typical Cushing’s features are then absent (as with any ectopic ACTH tumour). Phaeochromocytomas often secrete one or more neuropeptides: somatostatin may exaggerate the episodic nature of catecholamine discharge by inhibiting catecholamine release as soon as a discharge

16.17.3  Secondary hypertension 3793 starts, and it may also contribute to a reversible form of diabetes in phaeochromocytoma. Suppression tests The use of plasma or urine metanephrine measurements, in a re- liable laboratory, has reduced the number of patients with am- biguous results. In deciding which of the ‘grey zone’ patients need further investigations, it is also helpful to remember that modest increases in noradrenaline secretion are usually insufficient to cause severe hypertension. This is partly because of receptor (and postreceptor) desensitization, and partly volume depletion conse- quent on pressure natriuresis. Where doubt about the diagnosis re- mains, a pharmacological suppression test may be performed prior to imaging. Whereas physiological elevations of noradrenaline re- lease are temporarily suppressed by administration of the ganglion-​ blocking drug pentolinium, or the centrally acting α2-​agonist clonidine, these drugs do not suppress autonomous secretion by tumour. However, such pharmacological suppression tests have limited use nowadays given the increased diagnostic sensitivity of plasma and urinary assays for metanephrines. Localization of phaeochromocytomas A substantial clue to localization is provided by measurement of plasma adrenaline (or metadrenaline) or fractionated urinary metanephrines (as stated previously, extra-​adrenal tumours rarely produce adrenaline). CT or MRI scanning provides excellent imaging of the adrenal, where 90% of phaeochromocytomas are found (Fig. 16.17.3.16). They are usually much larger than Conn’s tumours, and may appear heterogeneous. It is best to withhold CT/​MRI scanning for extra-​adrenal phaeochromocytomas until the radiologist can be given some clue as to where to concentrate. This can be achieved by radio- isotope scanning with the iodinated analogue of noradrenaline, m-​iodobenzylguanidine (mIBG), in about 85% of patients. This may carry either an [123I] or [131I] label, the former being more sensi- tive but also more expensive, and may be misinterpreted if users are unaware that normal adrenal glands also accumulate mIBG. There is a case for undertaking mIBG scanning in addition to CT, even for patients found to have an adrenal phaeochromocytoma, to iden- tify extra-​adrenal secondary deposits when tumours are malignant, and because there may be coexisting adrenal and extra-​adrenal phaeochromocytomas. PET scans have been used and may be posi- tive when mIBG is unhelpful. 18F-​DOPA appears to be the most accurate of these, but available only when there is a centre doing neurological research for which a routine supply of this radiotracer is required. 68Ga-​DOTANOC, a somatostatin receptor analogue, is acquiring a reputation for higher sensitivity than mIBG, with good specificity. Selective venous sampling remains of occasional value when diagnostic problems persist. About 25 samples of blood are col- lected under fluoroscopic guidance from various sites, with an arterial sample invaluable for interpreting the results because it enables sites with a positive venoarterial difference to be readily detected. Because of the short half-​life of catecholamines in the circulation (c.1 min), the concentration at the tumour site is usu- ally several-​fold greater than elsewhere. This procedure should not usually be used for adrenal phaeochromocytomas, an exception being patients with von Hippel–​Lindau syndrome with small ad- renal masses, in whom all other biochemical tests may be normal, and the diagnosis of phaeochromocytoma is suggested by a re- versal of the normal excess of adrenaline to noradrenaline in the adrenal vein. Because phaeochromocytomas are vascular tumours, they pro- vide a good tumour blush, and occasionally angiography is required to resolve equivocal scans. Patients must be fully α-​blocked and preferably also β-​blocked prior to angiography. Other investigations It is important to check blood glucose in every patient as there may be α-​mediated inhibition of insulin release prior to effective treat- ment, and all patients should be screened for an associated medul- lary carcinoma of the thyroid. Routine slit lamp examination of the fundi has resulted in more frequent diagnosis of von Hippel–​Lindau syndrome, sometimes as a de novo occurrence. (c) (a) (b) Fig. 16.17.3.16  CT (Panel (a)) and m-​iodobenzylguanidine (mIBG) scan (Panel (b)) of a patient with a left adrenal phaeochromocytoma. Both scans illustrate typical non​homogeneous appearance due to large area of haemorrhage/​infarction at the centre of the tumour. Panel (c) shows the cut surface of a phaeochromocytoma, approximately 6 cm in diameter. Note the heterogeneous texture and dusky brown colour.

section 16  Cardiovascular disorders 3794 Treatment Medical management before surgery The definitive treatment for phaeochromocytoma is surgical, with laparoscopic surgery possible for most adrenal tumours. Even the small number of phaeochromocytomas that are recognized to be malignant preoperatively (e.g. by the presence of bone or liver me- tastases) may still benefit from resection of the primary tumour. The task for the physician is to make surgery safe, for which the mainstay of medical treatment is α-​blockade, but not all patients—​especially those without elevated plasma adrenaline levels—​require β-​ blockade. Indeed, the inadvertent use of β-​blockers in patients with undiagnosed phaeochromocytomas may lead to a paradoxical rise in blood pressure due to unopposed α stimulation (Fig. 16.17.3.17). The objective of treatment is not solely control of blood pressure, but also the expansion of blood volume, which is invariably reduced. Without doubt, phaeochromocytomas represents the pure vaso- constriction end of the vasoconstriction-​volume spectrum, and the hypertensive patient is best seen as the exception where pressure natriuresis has failed to compensate adequately for vasoconstriction. Normotension is an indication, not contraindication, for the use of α-​blockade to restore volume preoperatively. The α-​blocker of choice is phenoxybenzamine, which is an irre- versible blocker that actually destroys the α-​receptor by alkylation. More modern α-​blockers, such as doxazosin, and the mixed α-​and β-​blocker labetalol (a much stronger β-​blocker than it is α-​blocker), cause competitive blockade, which can be overcome by a surge of noradrenaline release from the tumour. An additional advantage of phenoxybenzamine is that it will block both α1-​and α2-​receptors, with blockade of the latter possibly advantageous because extra-​synaptic α2-​receptors mediate some of the direct vasoconstriction caused by circulating (non​neuronal) catecholamines. The diabetogenic effect of catecholamines is also an α2-​mediated response. The starting dose of phenoxybenzamine is 10 mg once or twice daily, with increases titrated against blood pressure up to a maximum of 90 mg daily. The effect of irreversible antagonists is cumulative, with the effect of the drug—​and each subsequent dose increment—​taking several days to reach maximum. There is rarely any urgency for surgery, which should not nor- mally be considered in less than 1 month after initiation of treat- ment in patients with symptomatic phaeochromocytomas. Indeed, the more severe the initial clinical picture, the greater the need for prolonged α-​blockade to expand intravascular volume. In most pa- tients there is a low filling pressure at presentation, evident clin- ically as a jugular venous pressure visible only when the patient lies flat, and any postural hypotension should be assumed to re- flect continuing hypovolaemia, not excessive α-​blockade, until the venous pressure is normalized. Usually volume expansion will occur spontaneously with phenoxybenzamine treatment, but ex- pansion should be achieved with intravenous saline if there is per- sistent hypovolaemia. Preoperatively, the aim should be uptitration of phenoxybenzamine dose, concurrent with active volume re- placement, until there is at least a persistent 10 mm Hg postural drop in blood pressure. The need for β-​blockade is indicated by tachycardia, which may become apparent only after treatment with phenoxybenzamine, and the dose of β-​blocking drug necessary is generally lower than that used in the treatment of hypertension. It is usually better to use a β1-​selective agent so that the peripheral vasodilatation mediated by β2-​receptors is not affected. Occasionally, pronounced β2-​receptor mediated effects, including tachycardia or tremor, can oblige use of a non​selective β-​blocker, although blood pressure control may then be more difficult and require addition of a calcium blocker. The reason for using as low a dose of β-​blocker as possible is that there may be a period of hypotension immediately upon removal of the phaeochromocytoma, despite the preoperative preparation that has been outlined. This hypotension should normally be offset by the ability to mount a tachycardia. Otherwise, further volume replace- ment may be required, supplemented if necessary by β-​agonists, Facilitation Inhibition Presynaptic Extrasynaptic Norepinephrine Epinephrine α2 β2 β1 α1 Norepinephrine β–blockers β2 α2 VASODILATATION VASCOCONSTRICTION Postsynaptic Norepinephrine (a) (b) Facilitation α2 β2 β1 α1 β2 α2 Inhibition Presynaptic Postsynaptic Extrasynaptic VASCOCONSTRICTION VASODILATATION Norepinephrine Epinephrine Fig. 16.17.3.17  Paradoxical worsening of blood pressure by β-​blockers in the presence of excess catecholamines. Panel (a) shows the effects of epinephrine and norepinephrine on α and β receptors. Panel (b) shows that β-​blockers, particularly non​selective ones, negate vasodilatation mediated via β2 receptors and inhibit catecholamine reuptake, thereby augmenting α-​mediated vasoconstriction.

16.17.3  Secondary hypertension 3795 most vasoconstrictor drugs being ineffective because of the slow washout of phenoxybenzamine. Malignant phaeochromocytomas The treatment of malignant phaeochromocytomas remains uncer- tain and unsatisfactory. The rate of growth is usually slow, but the prognosis for affected individuals can vary between the extremes of local recurrence at intervals of many years, and rapid demise sometimes precipitated by surgery. The tumours are not particularly sensitive either to chemotherapy or to radiotherapy, although the variability of response may still make them worth trying. There has been interest in the use of therapeutic doses of mIBG as a means of targeting high doses of radioactivity to the tumour: some patients show considerable regression after such treatment, but long-​term results are less certain. It is rare for the pharmacological effects of the tumour to be the principal problem if the primary tumour has been removed or debulked. High doses of phenoxybenzamine are preferable to α-​methyltyrosine, used as an inhibitor of noradrenaline synthesis. There is anecdotal evidence that therapy with high doses of an ARB might slow progression through reflex activation of renin and hence AT2-​receptor mediated apoptosis. Prognosis and genetic screening The removal of a phaeochromocytoma cures hypertension in most patients, especially those that are young. Most (90%) phaeochromocytomas are benign, and the proportion is prob- ably even higher for adrenal tumours, whereas paragangliomas have a greater than 10% likelihood of proving malignant. The latter should all be screened for mutations in the SDHB gene, which carry greater than 50% risk of malignancy. Other genetic screening will be influenced by a mixture of clinical features and cost considerations. A history (or family history) of other relevant tumours will lead to a search for von Hippel–​Lindau syndrome or MEN type 2. Even so, there is increasing consensus that all patients should at least be considered for structured genetic counselling and screening, and this is particularly important in much younger patients. All patients should be followed indefinitely with at least an annual measurement of arterial blood pressure and analysis of one of the indices of catecholamine secretion. In patients at par- ticularly high risk of malignancy, annual (or earlier if indicated clinically) whole-​body MRI/​CT imaging and focal neck scanning ought to be considered. Other endocrine causes of hypertension Conn’s syndrome and phaeochromocytoma have been singled out for attention in this chapter as the two endocrine conditions most likely to present as hypertension. However, hypertension is a fea- ture of several other endocrinopathies: Cushing’s syndrome, acro- megaly, hyperparathyroidism, and is a common complication of type 2 diabetes. The hypertension of Cushing’s syndrome is usually modest, except in ectopic ACTH where there is saturation by high cortisol levels of 11β-​hydroxysteroid dehydrogenase 2 (which nor- mally converts cortisol to the inactive cortisone). The cause of the hypertension in other syndromes is less clear cut and often not cor- rected by surgical cure of the primary problem. FURTHER READING Primary aldosteronism Brown MJ, Hopper RV (1999). Calcium-​channel blockade can mask the diagnosis of Conn’s syndrome. Postgrad Med J, 75, 235–​6. Choi M, et al. (2011). K+ channel mutations in adrenal aldosterone-​ producing adenomas and hereditary hypertension. Science, 331(6018), 768–​72. De Sousa K, et al. (2019). Molecular mechanisms in primary aldoster- onism. J Mol Endocrinol, 242, R67–R79. Dluhy RG, Lifton RP (1999). Glucocorticoid-​remediable aldoster- onism. J Clin Endocrinol Metab, 84, 4341–​4. Gordon RD, et al. (1994). High incidence of primary aldosteronism in 199 patients referred with hypertension. Clin Exp Pharmacol Physiol, 21, 315–​18. Hood SJ, et  al. (2007). The Spironolactone, Amiloride, Losartan, and Thiazide (SALT) double-​blind crossover trial in patients with low-​renin hypertension and elevated aldosterone-​renin ratio. Circulation, 116, 268–​75. Mir FA, Brown MJ, Appleton DS (2007). Lessons in the diagnosis and management of Conn’s syndrome. Clin Med, 7, 530–​2. Mulatero P, et al. (2006). Comparison of confirmatory tests for the diagnosis of primary aldosteronism. J Clin Endocrinol Metab, 91, 2618–​23. Padmanabhan S, et al. (2015). Genetic and molecular aspects of hyper- tension. Circ Res, 116, 937–​59. Rossi GP, et al. (2006). A prospective study of the prevalence of primary aldosteronismin 1,125 hypertensive patients. J Am Coll Cardiol, 48, 2293–​300. Stewart PM (1999). Mineralocorticoid hypertension. Lancet, 353, 1341–​7. Stowasser M (2015). Update in primary aldosteronism. J Clin Endocrinol Metab, 100, 1–​10. Stowasser M, et al. (2003). High rate of detection of primary aldos- teronism, including surgically treatable forms, after ‘non-​selective’ screening of hypertensive patients. J Hypertens, 21, 2149–​57. Young WF Jr (2007). The incidentally discovered adrenal mass. N Engl J Med, 356, 601–​10. Renovascular hypertension and coarctation Caliezi C, Reber P (2006). Images in clinical medicine. Fibromuscular dysplasia of the renal artery. N Engl J Med, 355, 2131. Cooper CJ, et al. for the CORAL Investigators (2014). Stenting and medical therapy for atherosclerotic renal-​artery stenosis. N Engl J Med, 370, 13–​22. Rosenthal E (2005). Coarctation of the aorta from fetus to adult: cur- able condition or lifelong disease process? Heart, 91, 1495–​502. Safian RD, Textor SC (2001). Renal-​artery stenosis. N Engl J Med, 344, 431–​42. The ASTRAL Trial Investigators (2009). Revascularization versus medical therapy for renal-​artery stenosis. N Engl J Med, 361, 1953–​62. White CJ (2006). Catheter-​based therapy for atherosclerotic renal ar- tery stenosis. Circulation, 113, 1464–​73. Phaeochromocytoma Allison DJ, et al. (1983). Role of venous sampling in locating a phaeo- chromocytoma. BMJ, 286, 1122–​4. Brown MJ, et  al. (2009). Pheochromocytoma. Horm Metab Res, 41, 655–​7.