Cushing's syndrome is a hormonal disorder caused by an abnormally high level of corticosteroid hormones in the blood. Cushing's syndrome is characterized by a reddened, moon-shaped face, wasting of the limbs, thickening of the trunk, and a humped upper back.
Other symptoms include acne; stretch marks on the skin; bruising; osteoporosis (loss of bone density); susceptibility to infection and peptic ulcers; and, in women, increased hairiness. Mental changes frequently also occur, causing depression, insomnia, paranoia,or, euphoria. Oedema, hypertension, and diabetes mellitus may develop. In children, growth may be suppressed.
The excess of hormones is most commonly due to prolonged treatment with corticosteroid drugs. Such cases of Cushing's syndrome are usually mild. In other cases, high hormone levels are due to overactivity of the adrenal glands because of an adrenal tumour, or due to a pituitary tumour affecting production of ACTH (adrenocortocotrophic hormone), which in turn stimulates the adrenal glands.
Cushing's syndrome due to cortico-steroid drugs usually disappears when the dose of the drug is gradually reduced. In cases of Cushing's syndrome that are caused by an adrenal gland tumour, the tumour will be removed surgically. If the cause of the disease is a pituitary tumour, it may be removed surgically or shrunk by irradiation and drug treatment. In both of these cases, surgery is followed by hormone replacement therapy.
Cushing's syndrome in more detail - technical
Cushing’s syndrome may be (1) ACTH-dependent—usually due to a pituitary adenoma (Cushing’s disease), but sometimes to nonpituitary tumours producing ACTH (most commonly small-cell carcinoma of the bronchus); (2) ACTH-independent—most often adrenal adenoma (rarely carcinoma).
Clinical features—typical presentation is with ‘classical’ manifestations of centripetal obesity, moon face, hirsutism, and plethora, with signs (when present) that best distinguish from simple obesity being bruising and muscle weakness (typically proximal).
Diagnosis of the presence of Cushing’s syndrome—this can be confirmed by finding (1) elevated 24-h urinary free cortisol; and/or (2) raised midnight salivary/plasma cortisol; and/or (3) impaired plasma cortisol suppression (09.00 h sample) in response to a low-dose overnight dexamethasone suppression test.
Diagnosis of the cause of Cushing’s syndrome—ACTH-dependent causes can be distinguished from ACTH-independent causes by measurement of plasma ACTH (09.00 h sample). Determining whether elevated ACTH is coming from the pituitary (Cushing’s disease) or from an ectopic source can be difficult, but may be achieved by consideration of (1) plasma potassium—hypokalaemia is a typical feature of ectopic ACTH but not of Cushing’s disease; (2) high-dose dexamethasone suppression test—which tends to suppress plasma cortisol in Cushing’s disease but not ectopic ACTH; (3) CRF test—producing an exaggerated rise in ACTH and cortisol in Cushing’s disease but not in ectopic ACTH; (4) inferior petrosal sinus sampling/selective venous catheterization—the most robust test to distinguish Cushing’s disease from ectopic ACTH syndrome.
Imaging—pituitary MRI is the investigation of choice if biochemical testing suggests Cushing’s disease, and abdominal CT scanning if biochemical testing suggests ACTH-independent Cushing’s syndrome.
Management—drugs that interfere with cortisol synthesis (e.g. metyrapone, ketoconazole) can lower cortisol levels, but definitive treatment depends on the cause: (1) adrenal adenomas—unilateral adrenalectomy; (2) Cushing’s disease—trans-sphenoidal removal of the pituitary tumour; (3) ectopic ACTH—surgical removal of the tumour is rarely possible but can lead to cure.
Glucocorticoid excess: Cushing's syndrome in great detail
Harvey Cushing first described a case of polyglandular syndrome secondary to pituitary basophilia in 1912, and several years later linked this to bilateral adrenal hyperplasia. The first case of an adrenal adenoma was probably reported by H G Turney in 1913.
Cushing’s syndrome comprises the symptoms and signs associated with prolonged exposure to inappropriately elevated levels of free plasma glucocorticoids. This definition thus takes into account the elevated corticosteroid levels that may be found in severely depressed patients, but which appear to be appropriate to the condition, and also the increased total (but normal free) glucocorticoid levels found when there is an increase in circulating cortisol-binding globulin (e.g. in patients on oestrogen therapy). The use of the term glucocorticoid in the definition covers both endogenous (cortisol) and exogenous steroid excess (e.g. prednisolone, dexamethasone).
The condition is most readily classified into ACTH-dependent and ACTH-independent causes (Table 1). The term ‘Cushing’s syndrome’ is used to describe all causes, whereas ‘Cushing’s disease’ is reserved for cases of pituitary-dependent Cushing’s syndrome.
When iatrogenic causes are excluded, the most frequent cause of Cushing’s syndrome is Cushing’s disease, which accounts for approximately 70% of cases. The adrenal glands show bilateral adrenocortical hyperplasia, with widening of the zona fasciculata and zona reticularis. Nodules may form within the hyperplastic glands.
|Table 1 Classification of causes of Cushing’s syndrome|
|Iatrogenic (treatment with ACTH1–39 or Synacthen®, ACTH1–24)|
|Cushing’s disease (pituitary-dependent)|
|Ectopic ACTH syndrome|
|Ectopic corticotrophin-releasing factor syndrome|
|Iatrogenic (such as pharmacological doses of prednisolone or dexamethasone)|
|Aberrant receptor expression (gastric inhibitory polypeptide, interleukin 1β).|
ACTH, adrenocorticotrophic hormone.
Cushing himself raised the question of whether his disease was a primary pituitary condition or secondary to an abnormality of the hypothalamus. There is abundant evidence to indicate that the condition is related to the pituitary rather than the hypothalamus. In over 90% of cases the disease is caused by a pituitary adenoma of monoclonal origin; basophilic hyperplasia is very uncommon, and selective surgical removal of the microadenoma usually results in cure, with a low recurrence rate.
Ectopic production of corticotropin-releasing factor (CRF)
This is a very rare cause of pituitary-dependent Cushing’s disease. However, cases have been described in which a tumour (e.g. medullary thyroid, prostate carcinoma) has been shown to produce CRF, but not ACTH.
Ectopic ACTH syndrome
Cushing’s syndrome may be caused by nonpituitary tumours producing ACTH, most commonly a malignant small-cell carcinoma of the bronchus (Table 2). However, the most challenging diagnostic problems relate to ACTH secretion from more benign and indolent carcinoid tumours, which may present with Cushing’s syndrome many years before the occult tumour manifests.
Adrenal adenoma and carcinoma
With the exclusion of iatrogenic Cushing’s syndrome, a solitary cortisol-secreting adrenal adenoma is the cause in about 10% of cases. Carcinomas are rarer, have a poor prognosis, and may be associated with the secretion of other hormones in addition to cortisol (usually adrenal androgens). The aetiology of these tumours is unknown.
|Table 2 Tumours associated with the ectopic ACTH syndrome|
|Tumour type||Approximate incidence (%)|
|Small-cell lung carcinoma||50|
|Non-small-cell lung carcinoma||5|
|Pancreatic tumours (including carcinoids)||10|
|Thymic tumours (including carcinoids)||5|
|Medullary carcinoma of thyroid||5|
|Phaeochromocytoma and related tumours||3|
|Rare carcinomas of prostate, breast, ovary, gallbladder, colon||10|
Carney’s syndrome (OMIM 160980)
This is an autosomal dominant condition involving mesenchymal tumours (especially atrial myxomas), spotty skin pigmentation, peripheral nerve tumours, and various endocrine tumours, one of which may lead to Cushing’s syndrome. The adrenals then contain multiple small, pigmented nodules. The condition has been described as pigmented multinodular adrenocortical dysplasia, and results from mutations in the regulatory subunit R1A of protein kinase A, causing adrenal hyperfunction.
McCune–Albright syndrome (OMIM 174800)
In this condition, fibrous dysplasia and cutaneous pigmentation may be associated with pituitary, thyroid, adrenal, and gonadal hyperfunction. The adrenal hypersecretion may produce Cushing’s syndrome. The underlying abnormality is a somatic mutation in the α subunit of the stimulatory guanine nucleotide-binding protein (G protein) that is linked to adenylate cyclase. The mutation results in the G protein being constitutively activated, which, in the adrenal gland, mimics constant ACTH stimulation. Adrenal nodular formation may occur.
Aberrant receptor expression (OMIM 219080)
Patients have been described with nodular hyperplasia, ACTH-independent Cushing’s syndrome, and enhanced adrenal responsiveness to gastric inhibitory polypeptide (GIP). The biochemical clues are the presence of subnormal morning levels of plasma cortisol and a rise in cortisol after food. This food-dependent form of Cushing’s syndrome results from the normal increase in GIP after eating. Not surprisingly, the clinical syndrome is related to food intake; fasting can produce adrenal insufficiency. It is now appreciated that a similar form of Cushing’s syndrome can result from the aberrant expression of other receptors, including those for interleukin 1, luteinizing hormone, and serotonin.
In the original description of this syndrome, urinary and plasma cortisol levels were elevated, but were not suppressed with dexamethasone. Plasma ACTH may be normal or suppressed. The frequency and pathogenesis of this condition remain unknown, but a two-hit hypothesis has been put forward to explain its aetiology. Chronic liver disease, irrespective of the cause, is associated with impaired cortisol metabolism, but in alcoholics this is associated with an increase in the cortisol secretion rate, rather than concomitant suppression in the face of impaired metabolism. With abstinence from alcohol the biochemical abnormalities rapidly revert to normal.
Clinical features of Cushing’s syndrome
The classical features of Cushing’s syndrome—centripetal obesity, moon face, hirsutism, and plethora—are well known following Cushing’s initial description in 1912.However, this gross clinical picture is not always present. The signs and symptoms in patients with Cushing’s syndrome are listed in Table 3, together with the most discriminatory features distinguishing Cushing’s syndrome from simple obesity. Weight gain and obesity are the most common symptom and sign, but the distribution of fat is not invariably centripetal—a ‘buffalo hump’ is present in about one-half of patients.
Gonadal dysfunction is very common, with menstrual irregularity in females and loss of libido in males, resulting from a suppressive effect of cortisol on gonadotropin secretion. Hirsutism is frequently found in female patients, as is acne, and reflects ACTH-stimulated hyperandrogenism.
Psychiatric abnormalities have been reported in all series of patients with Cushing’s syndrome, regardless of cause. Depression and lethargy are among the most common problems, but poor concentration, paranoia, and overt psychosis are also well recognized. Lowering of plasma cortisol by medical or surgical therapy usually results in a rapid improvement in the psychiatric state.
Many patients with long-standing Cushing’s syndrome have lost height because of osteoporotic vertebral collapse. Pathological fractures, either spontaneous or after minor trauma, are not uncommon. Rib fractures, by contrast with those of the vertebrae, are often painless. The radiographic appearance is typical, with exuberant callus formation at the site of the healing fracture.
The plethoric appearance of the patient with Cushing’s syndrome results from thinning of the skin, not true polycythaemia. The typical red-purple livid striae of the syndrome are found most frequently on the abdomen, but may also be present on the upper thighs and arms. They are very common in younger patients, and less so in those over 50 years of age.
Myopathy and bruising are two of the most discriminatory features of the syndrome. The myopathy involves the proximal muscles of the lower limbs and the shoulder girdle. Complaints of weakness, such as an inability to climb stairs or get up from a deep chair, are relatively uncommon, but observation of whether the patient can rise from a crouching position often reveals the problem. Bruising of the skin is often extensive and occurs with unknown or trivial trauma.
|Table 3 Prevalence of symptoms and signs in Cushing’s syndrome and discriminant index compared with prevalence of features in patients with simple obesity|
|Loss of scalp hair||13|
Data from Ross and Linch (1982).
GTT, glucose tolerance test.
Hypertension is another prominent feature. Even though epidemiological data show a strong association between blood pressure and obesity, hypertension is much more common in patients with Cushing’s syndrome than in those with simple obesity.
Pigmentation is rare in Cushing’s disease, but common in ectopic ACTH syndrome. However, in some pituitary tumours there is abnormal processing of the pro-opiomelanocortin (POMC) precursor molecule, with resulting pigmentation.
Infections are more common in patients with Cushing’s syndrome than in unaffected individuals. In many instances these are asymptomatic, as the normal inflammatory response may be suppressed. Reactivation of tuberculosis has been reported. Fungal infection of the skin is frequently found. Glucose intolerance may be a predisposing factor, with overt diabetes being present in up to one-third of patients in some series.
Ocular effects may include raised intraocular pressure, chemosis, and exophthalmos (present in up to one-third of patients in Cushing’s original series). Cataracts, a well recognized complication of exogenous corticosteroid therapy, seem to be uncommon, except as a complication of diabetes.
In patients with ectopic ACTH syndrome caused by small-cell lung carcinoma, the clinical presentation more commonly resembles Addison’s disease than Cushing’s syndrome. The patients are very commonly pigmented and have lost weight, but the association of these with hypokalaemic alkalosis and glucose intolerance should alert the clinician to the diagnosis. Patients with more indolent causes, such as bronchial carcinoids, present with the more typical features of Cushing’s syndrome.
Special features of Cushing’s syndrome
Cyclical Cushing’s syndrome
Of particular clinical interest has been a group of patients with cyclical Cushing’s syndrome, characterized by periods of excess cortisol production (e.g. 40 days), followed by intervals of normal cortisol production (e.g. 60–70 days). Some of these patients demonstrate a paradoxical rise in plasma ACTH and cortisol when treated with dexamethasone. Most patients have been thought to have pituitary-dependent disease, and in many of these patients basophil adenomas have been removed, some with long-term cure. However, cortisol secretion may show some evidence of cyclicity in other causes of Cushing’s syndrome, notably ectopic ACTH syndrome.
All the above features occur in children, but growth arrest is almost invariable. The dissociation between height and weight on the growth chart is obvious. If the child is growing along the same centile line then the diagnosis of Cushing’s syndrome is highly unlikely. In addition to glucocorticoid-induced growth arrest, androgen excess may result in precocious puberty.
Pregnancy is rare in women with Cushing’s syndrome because of associated amenorrhoea resulting from androgen excess or hypercortisolism. However, approximately 100 such cases have been reported, 50% of which resulted from adrenal adenomas.
A few cases of true pregnancy-induced Cushing’s syndrome have been reported, with postpartum regression. In these cases the aetiology is unknown. Establishing a diagnosis and cause can be difficult; normal pregnancy is associated with a threefold increase in plasma cortisol caused by increased production rates and increases in cortisol-binding globulin. Urinary free cortisol also rises, and dexamethasone does not suppress plasma cortisol to the same degree as in the nonpregnant state. Untreated, the condition has high maternal and fetal morbidity and mortality. Adrenal and/or pituitary adenomas should be excised. Metyrapone, which is not teratogenic, has been effective in controlling the hypercortisolism in many cases.
In addition to the normal features of glucocorticoid excess the patient may present with other problems relating to: (1), the tumour, e.g. abdominal pain from the primary tumour or secondary deposits, or (2), the secretion of other steroids such as androgens or mineralocorticoids. Thus, in addition to hirsutism, there may be other features of virilization in females, including clitoromegaly, breast atrophy, deepening of the voice, temporal recession, and severe acne.
Investigation of patients with suspected Cushing’s syndrome
There are two stages in the investigation of a patient with suspected Cushing’s syndrome: (1), does the patient have Cushing’s syndrome? (2), if the answer to (1) is yes, what is the cause? Unfortunately many investigators fail to make this distinction and ill-advisedly use tests that are relevant to question (2) to try to answer question (1). In particular, it is essential that appropriate radiological investigations are not undertaken until Cushing’s syndrome has been confirmed biochemically. The principal diagnostic tests are listed in Table 5 below.
Practically, three screening tests should be used to confirm Cushing’s syndrome. Depending on the index of clinical suspicion these can be performed in isolation or combination.
Urinary free cortisol
For many years the diagnosis of Cushing’s syndrome was based on the measurement of urinary metabolites of cortisol (24-h urinary 17-hydroxycorticosteroid or 17-oxogenic steroid excretion, depending on the method used). However, the sensitivity and specificity of these methods is poor and these assays have now been replaced with the much more sensitive measurement of urinary free cortisol excretion. Urinary free cortisol is an integrated measure of plasma free cortisol. As cortisol secretion increases, the binding capacity of cortisol-binding globulin is exceeded, resulting in a disproportionate rise in urinary free cortisol. This is a useful screening test, but even so, it is accepted that urinary free cortisol may be normal in 7 to 10% of patients with Cushing’s syndrome.
|Table 5 Tests used in the diagnosis and differential diagnosis of Cushing’s syndrome|
|—does the patient have Cushing’s syndrome?|
|Circadian rhythm of cortisol-late night plasma or salivary cortisol.|
|Urinary free cortisol excretion*|
|Low-dose dexamethasone suppression test*|
|Insulin tolerance test|
|—what is the cause of the Cushing’s syndrome?|
|High-dose dexamethasone suppression test|
|Inferior petrosal sinus ± selective venous sampling for ACTH|
|MRI/CT scanning of pituitary/adrenals|
* Valuable outpatient screening tests (see text).
Measurement of the cortisol:creatinine ratio in the first urine specimen passed on waking obviates the need for a timed collection, and has been used by some as a sensitive screening test, particularly if cyclical Cushing’s syndrome is suspected. Urine aliquots are stable at room temperature for up to 7 days, and can then be sent by post to the local endocrine laboratory.
Late night plasma/salivary cortisol
In normal subjects, plasma cortisol concentrations are at their highest first thing in the morning and reach a nadir at around midnight (up to100 nmol/litre in the morning and <50 nmol/litre at midnight effectively excluding Cushing’s syndrome). This circadian rhythm is lost in patients with Cushing’s syndrome, such that in most patients the 09.00 level of plasma cortisol is normal, but nocturnal levels are raised. Random morning levels of plasma cortisol are therefore of little value in making the diagnosis. In addition, various factors, such as the stress of venepuncture, intercurrent illness, and admission to hospital, may result in normal subjects losing their circadian rhythm. It is therefore good practice not to measure plasma cortisol until the patient has been in hospital for 48 h. For this reason midnight plasma cortisol is not routinely used as a first-line screening test.
By contrast, midnight salivary cortisol can be collected at home and offers greater accuracy. A salivary cortisol value of more than 5.5 nmol/litre (2.0 ng/ml) has 100% sensitivity and 95% specificity in diagnosing Cushing’s syndrome. Other screening and confirmatory tests may be required to evaluate false positive results.
Low-dose/overnight dexamethasone suppression tests
In normal subjects, administration of a supraphysiological dose of a glucocorticoid results in suppression of ACTH and hence cortisol secretion. In Cushing’s syndrome of whatever cause there is failure of this suppression when low doses of the synthetic glucocorticoid dexamethasone are given. The overnight test is often used as an outpatient screening test. Various doses of dexamethasone have been used, usually given at midnight, but most experience is with a dose of 1 mg. A plasma cortisol of less than 50 nmol/litre between 08.00 and 09.00 the following morning has a sensitivity of 95% and specificity of 80% in excluding Cushing’s syndrome. Thus the outpatient overnight test has high sensitivity but low specificity, and further investigation is often required.
The conventional low-dose 48-h test is more accurate, but usually requires inpatient admission. Here, plasma cortisol is measured at 09.00 on day 0 and 48 h later, following dexamethasone given at a dose of 0.5 mg every 6 h for 48 h. This test is reported as having a 97 to 100% true positive rate and a false positive rate of less than 1%. Certain drugs (phenytoin, rifampicin) may increase the metabolic clearance rate of dexamethasone, thereby giving false positive results. If pseudo-Cushing’s syndrome is suspected, physicians in North America have modified this test slightly by administering CRF at the end of the dexamethasone suppression.Differential diagnostic tests
Once the biochemical diagnosis has been made, other investigations are required to determine the cause of the Cushing’s syndrome.
Plasma ACTH at 09.00
This will differentiate ACTH-dependent from ACTH-independent causes. ACTH is either within the normal reference range (50% of cases) or elevated in patients with Cushing’s disease. ACTH levels in ectopic ACTH syndrome are high, but overlap the values seen in Cushing’s disease in 30% of cases and cannot therefore be used to differentiate these two conditions. The measurement of ACTH precursors (pro-ACTH, POMC) is not routinely available, but may be more useful in detecting an ectopic source of ACTH.
In patients with autonomous adrenal tumours, plasma ACTH is invariably undetectable. This can also occur with degradation of ACTH; consequently, nonhaemolysed blood samples should be taken on ice and immediately separated.
Diagnosis is a problem in those patients whose plasma ACTH levels are in the low normal range or intermittently detectable. This may occur in macronodular hyperplasia. The danger is that in some patients the asymmetry of the nodular hyperplasia may lead to a diagnosis of adrenal adenoma, the plasma ACTH is ignored, and an inappropriate adrenalectomy is performed. Conversely, in some patients with this syndrome an autonomous adrenal tumour develops and, despite detectable ACTH, unilateral adrenalectomy is required.
Hypokalaemic alkalosis is present in more than 95% of patients with ectopic ACTH syndrome, but in fewer than 10% of patients with Cushing’s disease. Patients with the ectopic syndrome usually have higher cortisol secretion rates that saturate the renal protective 11β-HSD2 enzyme, resulting in cortisol-induced mineralocorticoid hypertension (see ‘Apparent mineralocorticoid excess syndrome’ below). In addition, these patients have higher levels of the ACTH-dependent mineralocorticoid deoxycorticosterone.
High-dose dexamethasone suppression test
The rationale for this test is that in Cushing’s disease there is negative feedback control of ACTH, but set at a higher level than normal. Thus, in this disease, cortisol levels are not suppressed with a low dose of dexamethasone, but are suppressed with a higher dose. The original test introduced by Liddle was based on giving dexamethasone at a dose of 2 mg every 6 h for 48 h and measuring urinary 17-oxogenic steroids. Suppression was defined as a greater than 50% fall in 24-h urinary 17-oxogenic steroids. In the modern test, plasma cortisol is measured at 0 and 48 h or, less commonly, plasma cortisol is measured at 08.00 (basal sample), 8 mg dexamethasone is given orally at 23.00 on the same day, and plasma cortisol is measured again at 08.00 on the following morning. In both these tests, greater than 50% suppression of plasma cortisol in comparison with the basal sample has been used to define a positive response. In Cushing’s disease about 90% of patients have a positive 48-h test, compared with 10% with ectopic ACTH syndrome. With overnight high-dose testing, 89% sensitivity and 100% specificity has been reported for Cushing’s disease.
Metyrapone is an 11β-hydroxylase inhibitor that blocks the conversion of 11-deoxycortisol to cortisol, and deoxycorticosterone to corticosterone. This lowers plasma cortisol and, via negative feedback control, increases plasma ACTH. This in turn stimulates an increase in the secretion of adrenal steroids proximal to the block. When metyrapone is given in doses of 750 mg every 4 h for 24 h, patients with Cushing’s disease exhibit an exaggerated rise in plasma ACTH, with 11-deoxycortisol levels at 24 h exceeding 1000 nmol/litre. In most patients with ectopic ACTH syndrome there is little or no response, but occasional patients (possibly those producing both ACTH and CRF) have an 11-deoxycortisol response that may be similar to that in Cushing’s disease.
The metyrapone test was originally used to distinguish patients with Cushing’s disease from those with a primary adrenal cause. However, these can be more reliably distinguished by measuring plasma ACTH and CT scanning of the adrenal glands. As indicated, the test does not reliably distinguish between Cushing’s disease and ectopic ACTH syndrome, and the value of this test has been questioned. It should be reserved for patients in whom the results of other tests are equivocal.
CRF is a peptide of 41 amino acids, identified by Vale in 1981 from ovine hypothalami. The ovine sequence differs by seven amino acid residues from that of the human peptide, but despite this stimulates the release of ACTH in humans. The test involves the intravenous injection of either ovine or human CRF at a dose of 1 µg/kg body weight or a single dose of 100 µg. The test can be performed in the morning or afternoon, and after basal sampling, blood samples for ACTH and cortisol are taken every 15 min for 1 to 2 h after administering CRF.
In normal subjects CRF elicits a rise in ACTH and cortisol, and this response is exaggerated in Cushing’s disease. It is typically absent in ectopic ACTH syndrome and patients with adrenal tumours. In distinguishing pituitary-dependent Cushing’s disease from ectopic ACTH syndrome, the response of ACTH to CRF has a specificity of 90%, and with cortisol as the endpoint, 95%. Using as an endpoint an ACTH increase of 100% over basal, or a cortisol rise of 50%, this positive response eliminates a possible diagnosis of ectopic ACTH syndrome.
Inferior petrosal sinus sampling/selective venous catheterization
This is the most robust test for distinguishing Cushing’s disease from ectopic ACTH syndrome, but also the most costly and technically demanding. As blood from each half of the pituitary drains into the ipsilateral inferior petrosal sinus, catheterization of both sinuses with simultaneous sampling of venous blood can distinguish a pituitary from an ectopic source, and aid in the lateralization of a pituitary microadenoma. In patients with ectopic ACTH syndrome there is no ACTH gradient between the inferior petrosal sinus samples and simultaneously drawn peripheral venous levels. In Cushing’s disease the ipsilateral:contralateral ACTH ratio is usually greater than 1.4. However, because of the problem of intermittent ACTH secretion, it is useful to make measurements before and at intervals (e.g. 2, 5, and 15 min) after intravenous injection of 100 µg of synthetic ovine CRF. Using this approach, patients with Cushing’s disease and bilateral inferior petrosal sinus ratios of less than 1.4 can readily be distinguished from those with the ectopic syndrome. The precise ratio that distinguishes Cushing’s disease from the ectopic syndrome has been debated. Some authors use 2 rather than 1.4.
Some specialists reserve petrosal sinus sampling for those cases where the differential diagnosis is still in doubt after high-dose dexamethasone, pituitary imaging, and peripheral CRF testing.
Rarely, selective catheterization of vascular beds may be required to identify the source of ectopic ACTH secretion, e.g. from a small pulmonary carcinoid or thymic tumour.
Many tumours responsible for ectopic ACTH syndrome also produce peptide hormones other than ACTH or its precursors. Calcitonin, chromogranin A, and gut hormones such as gastrin and vasoactive intestinal polypeptide should be measured.
There is no doubt that high-resolution contrast-enhanced imaging of thin sections of the pituitary and adrenals by either CT or MRI has revolutionized the investigation of Cushing’s syndrome. However, if mistakes are to be avoided it is essential that the results of any imaging technique always be interpreted in the light of the biochemical results. In imaging the adrenals, asymmetrical nodular hyperplasia may lead to a false diagnosis of adrenal adenoma. Owing to the presence of pituitary incidentalomas, pituitary MRI/CT scanning may produce false-positive results, particularly for lesions of less than 5 mm diameter (see ‘Adrenal incidentalomas’ below).
Pituitary MRI is the investigation of choice if the biochemical tests suggest Cushing’s disease, and has a sensitivity of 70% and specificity of 87%. About 90% of ACTH-secreting pituitary tumours are microadenomas, i.e. less than 10 mm in diameter. The classical features of a pituitary microadenoma are a hypodense lesion after contrast, associated with deviation of the pituitary stalk and a convex upper surface of the pituitary gland. With such small tumours it is not surprising that the sensitivity of CT scanning is relatively low (20–60%), with a similar specificity.
By contrast, CT scanning rather than MRI is the investigation of choice for adrenal imaging, offering better spatial resolution. Once again it is stressed that adrenal incidentalomas are present in up to 5% of normal subjects, and thus adrenal imaging should not be performed unless biochemical investigation suggests a primary adrenal cause. Adrenal carcinomas are large and often associated with metastatic spread at presentation.
In patients with occult ectopic ACTH syndrome, high-definition MRI/CT scanning of the neck, thorax, and abdomen/pelvis, with images every 0.5 cm, may be required to detect small ACTH-secreting carcinoid tumours.
Adrenal scintigraphy is of value in certain patients with primary adrenal pathology. The most commonly used agent is [131I]-6β-iodomethyl-19-norcholesterol, a marker of adrenocortical cholesterol uptake. In patients with adrenal adenomas the isotope is taken up by the adenoma, but not by the contralateral suppressed adrenal. Adrenal scintigraphy is useful in patients with suspected adrenocortical macronodular hyperplasia, in which CT scanning may mislead by suggesting unilateral pathology, whereas with isotope scanning the bilateral adrenal involvement is identified.
Treatment of Cushing’s syndrome
Studies carried out before the introduction of effective therapy suggested that 50% of patients with untreated Cushing’s syndrome died within 5 years, causing some physicians to label this the ‘killing disease’. Even with modern management, an increased prevalence of cardiovascular risk factors persists for many years after an apparent cure. Close follow-up of all patients is recommended.
Adrenal adenomas should be removed by unilateral adrenalectomy, which has a 100% cure rate. With the increasing experience of laparoscopic adrenalectomy in most tertiary centres, this has now become the surgical treatment of choice for unilateral tumours, reducing surgical morbidity and postoperative hospital stay compared with traditional open approaches. After surgery it may take many months or even years for the suppressed adrenal to recover. It is wise therefore to give slightly suboptimal replacement therapy (<25–30 mg hydrocortisone/day or equivalent), with intermittent measurement of the 08.00 level of plasma cortisol after having omitted therapy for 24 h. When the morning plasma cortisol is above 180 nmol/litre a stimulation test such as an insulin tolerance test may then demonstrate whether or not the hypothalamic–pituitary–adrenal axis has recovered.
Adrenal carcinomas have a very poor prognosis and most patients are dead within 2 years. It is usual practice to try to remove the primary tumour, even though metastases may be present, so as to enhance the response to the adrenolytic agent mitotane (see ‘Medical treatment of Cushing’s syndrome’ below). Radiotherapy to the tumour bed and to some metastases, such as those in the spine, may be of limited value.
Pituitary-dependent Cushing’s disease
The treatment of Cushing’s disease has been improved by trans-sphenoidal surgery conducted by an experienced surgeon. Before the selective removal of a pituitary microadenoma the treatment of choice was bilateral adrenalectomy. This had an appreciable mortality, even in the best centres (c.4%), as well as morbidity. The main risk was the subsequent development of Nelson’s syndrome (postadrenalectomy hyperpigmentation with locally aggressive pituitary tumour). To avoid this, pituitary irradiation was often carried out following bilateral adrenalectomy. These patients required lifelong replacement therapy with hydrocortisone and fludrocortisone. Today, bilateral adrenalectomy is reserved for the occasional patient with Cushing’s disease in whom no pituitary tumour can be found, or when pituitary surgery has failed or the condition has recurred.
After selective removal of a microadenoma, the surrounding corticotrophs are normally suppressed. In these cases plasma cortisol concentrations are also suppressed postoperatively, and glucocorticoid replacement therapy is required, but gradual recovery of the hypothalamic–pituitary–adrenal axis can be anticipated, particularly in patients with normal pituitary function as it relates to other endocrine axes. A nonsuppressed postoperative plasma cortisol suggests that the patient is not cured, even though cortisol secretion may have fallen to normal or subnormal values. Close follow-up of such individuals is required.
In the past, pituitary irradiation was often used in the treatment of Cushing’s disease. However, improvements in pituitary surgery have resulted in far fewer patients being so treated. In children, pituitary irradiation appears to be effective. Radiotherapy is not recommended as a primary treatment, but is reserved for patients not responding to pituitary microsurgery, when bilateral adrenalectomy has been performed, or in those with established Nelson’s syndrome.
Ectopic ACTH syndrome
Treatment of ectopic ACTH syndrome depends on the cause. If the tumour can be found and has not spread, then its removal can lead to cure (e.g. bronchial carcinoid tumours, or thymomas). However, the prognosis for small-cell lung cancer associated with ectopic ACTH syndrome is poor. The cortisol excess and associated hypokalaemic alkalosis and diabetes mellitus can be ameliorated by medical therapy (see below). Treatment of the small-cell tumour itself will also, at least initially, produce improvement. Sometimes, if the ectopic source of ACTH cannot be found, it may be necessary to perform bilateral adrenalectomy and then follow the patient carefully (sometimes for several years) to find the primary tumour.
Medical treatment of Cushing’s syndrome
Several drugs have been used in the treatment of Cushing’s syndrome. Most commonly, metyrapone in Europe or ketoconazole in the United States of America has been given, often to lower cortisol concentrations before definitive therapy, or while awaiting benefit from pituitary irradiation. The daily dose has to be determined by measuring either plasma or urinary free cortisol. The aim should be to achieve a mean plasma cortisol of about 300 nmol/litre during the day, or a normal urinary free cortisol. Metyrapone is usually given in doses ranging from 250 mg twice daily to 1.5 g every 6 h. Nausea may be produced and can be alleviated (if not caused by adrenal insufficiency) by giving the drug with milk. Ketoconazole is an imidazole that has been widely used as an antifungal agent; it produces abnormal liver function tests signifying hepatitis in about 14% of patients. Ketoconazole blocks a variety of steroidogenic cytochrome P450-dependent enzymes and thus lowers plasma cortisol levels. For effective control of Cushing’s syndrome, 400 to 800 mg ketoconazole daily is required.
Aminoglutethimide is a more toxic drug that in high doses blocks the initial steps in the biosynthetic pathway, and thus affects the secretion of steroids other than cortisol. In doses of 1.5 to 3 g daily (starting with 250 mg every 8 h) it commonly produces nausea, marked lethargy, and a skin rash. Trilostane, a 3β-hydroxysteroid dehydrogenase inhibitor, is ineffective in Cushing’s disease since the block in steroidogenesis is overcome by the rise in ACTH. However, it can be effective in patients with adrenal adenomas.
Mitotane is an adrenolytic drug that is taken up by both normal and malignant adrenal tissue, causing adrenal atrophy and necrosis. Because of its toxicity, mitotane has been used mainly in the management of adrenal carcinoma. Doses of up to 8 g/day are required to control glucocorticoid excess, although evidence that it causes tumour shrinkage or improves long-term survival is scant. The drug will also produce mineralocorticoid deficiency, and both glucocorticoid and mineralocorticoid replacement therapy may be required. Side effects are common and include fatigue, skin rashes, and gastrointestinal disturbance.