Diagnosis of Cushings Syndrome

Determination of 24-h UFC excretion is the best screening test available for documentation of endogenous hypercortisolism (52,53). Values consistently in excess of200 ^g/d (550 nmol) are virtually diagnostic of CS (Fig. 2). Medications or synthetic steroids other than cortisol do not interfere with the performance of high-performance liquid chromatography (HPLC) in measuring urinary cortisol. Assuming complete collections have been performed, there are virtually no false-negative results. In patients with episodic CD (22-25,27,29,30), repeated intermittent UFC collections are necessary, usually in 3-monthly intervals. False-positive results, however, may be obtained in several non-Cushing's hypercortisolemic states (Table 1). Generally, in these states, the UFC levels rarely are higher than 200 ^g/d (552 nmol). UFC remains constant throughout life when normalized per square meter of body surface area, obviating the need of using age-specific normal values in children or subjects who are obese (53).

When assays for UFC excretion are not available, measurement of urinary 24-h 17-hydroxysteroids can be helpful. These compounds include all cortisol metabolites with a 17-dihydroxyacetone side chain and, thus, give an indirect measure of cortisol secretion rate. Correction is required, however, for urinary creatinine excretion because size and adiposity influence their daily production.

The overnight 1-mg dexamethasone suppression test is a simple screening procedure for hypercortisolism (14,54,55). The test has a low incidence of false normal suppression (less than 3%). However, the incidence of false-positive results is significantly high (approx 20% to 30%). In children, the dose of dexamethasone that should be employed is 15 ^g/kg body weight. In the United States, dexamethasone (in adults, 1 mg) is administered orally at 11 PM and plasma cortisol measured the next morning at 8 AM. The typical plasma cortisol cutoff level (from 5 ^g to 2.5 ^g/dL; 69 nmol) has been readjusted over the years to improve sensitivity. With these more stringent cortisol cutoff criteria, the

Fig. 1. (continued) (From Oldfield EH, Doppman JL, Nieman, LK, et al. Petrosal sinus sampling with and without corticotropin releasing hormone for the differential diagnosis of Cushing's syndrome. N Engl J Med 1991;325:897.)

Cushing Algorithm

Fig. 2. Algorithm for the diagnosis of Cushing's disease. To convert serum cortisol values to nmol/L and to convert UFC values to nmol/d, multiply by 27.59 (serum) and 2.759 (urine). ACTH, 5 ng/L = 1.1 nmol/L; BIPPS, bilateral inferior petrosal sinus sampling; CRH, corticotropin-releasing hormone; CT, computed tomography; MRI, magnetic resonance imaging.

Fig. 2. Algorithm for the diagnosis of Cushing's disease. To convert serum cortisol values to nmol/L and to convert UFC values to nmol/d, multiply by 27.59 (serum) and 2.759 (urine). ACTH, 5 ng/L = 1.1 nmol/L; BIPPS, bilateral inferior petrosal sinus sampling; CRH, corticotropin-releasing hormone; CT, computed tomography; MRI, magnetic resonance imaging.

sensitivity of the overnight low-dose dexamethasone suppression test in CS diagnosis is greater than 95%, whereas the specificity can be as low as 88%.

False-positive results are seen in stress conditions, such as acute illness, extreme obesity, alcoholism, and depression; in high estrogen states; and with medications that accelerate dexamethasone catabolism, such as primidon, phenobarbital, phenytoin, and rifampin.

CS is generally excluded if the response to the single-dose dexamethasone suppression test and the 24-h UFC or 17-hydroxysteroid excretion are normal. One should bear in mind, however, that cortisol hypersecretion may be intermittent and/or periodic in approx 10% of patients with CS of any etiology. Documenting loss of diurnal variation of plasma cortisol (taking into account time zones of travel or shift work) would support the CS diagnosis and vice versa. Multiple blood samples taken in both morning and evening increase the value of the test, because a significant variability of cortisol levels may be present (7,56). Similarly, isolated plasma ACTH determinations are of limited value especially because there is significant overlap between the ACTH levels in patients with Cushing's disease and normal subjects.

On the other hand, plasma ACTH measurements can be useful in providing an early distinction between ACTH-dependent and ACTH-independent sources, because ACTH concentrations higher than 200 pg/mL (44 pmol/L) strongly suggest ectopic (nonpituitary) production. In borderline cases (those with only mild elevation of UFC and/or positive dexamethasone suppression test), the 2-d Liddle test can be conducted (57). In this test, patients receive dexamethasone 0.5 mg orally every 6 h for 48 h, with measurement of UFC and 17-hydroxy-corticosteroids (levels of > 25 ^g/d [69 nmol] or 4 mg/d [11 ^mol], respectively, are abnormal). In addition, a plasma cortisol level 6 h after the last dexamethasone dose can be measured: a level > 1.8 ^g/dL (50 nmol) is abnormal, and when using this criterion, the Liddle test has a sensitivity and specificity of 95% (7).

Salivary cortisol is in equilibrium with free plasma cortisol and has been recently used in the diagnosis of CS (58). Measurement of a single midnight plasma cortisol under nonstressed conditions (abnormal if >1.8 ^g/dL) is highly sensitive (99%) but not practical for outpatients (56,59).

Distinguishing Mild Cushing's Syndrome From Pseudo-Cushing States

The clinical and biochemical presentation of mild hypercortisolism in CS is often indistinguishable from that seen in pseudo-Cushing states, such as depression or chronic active alcoholism (Table 1) (60). A hyperactive/hyper-responsive hypothalamic CRH neuron is central to the hypercortisolism of pseudo-Cushing states in the context of a pituitary-adrenal axis that is otherwise appropriately, albeit not fully, restrained by negative cortisol feedback (61). In contrast, the hypercortisolism of CS, regardless of the classification, feeds back negatively and completely suppresses hypothalamic CRH secretion. These concepts form the basis for the tests employed in the differential diagnosis of mild hypercortisolism. Thus, most patients with CS (80%-90%) show inadequate suppression to low-dose

Table 1

Classification of Hypercortisolism

Physiologic states Stress Pregnancy

Chronic strenuous excercise Malnutrition Pathophysiologic states Cushing's syndrome

ACTH-dependent (85%) Pituitary adenoma (80%) Ectopic ACTH (20%) Ectopic CRH (rare) ACTH-independent (15%) Adrenal adenoma Adrenal carcinoma Micronodular adrenal disease (rare) Macronodular adrenal disease (rare) Psychiatric states

Melancholic depression Obsessive-compulsive disorder Chronic active alcoholism Panic disorder Anorexia nervosa Narcotic withdrawal Glucocorticoid resistance

ACTH, adenocorticotropic hormone; CRH, corticotropin-releasing hormone

(0.5 mg every 6 h for 2 d) dexamethasone and do not respond to insulin-induced hypoglycemia, contrasting the normal responses of depressed and other pseudo-Cushing patients. In addition, patients with CD (85%) have a "normal" or exaggerated ACTH response to CRH, whereas patients with depression (75%) show a blunted response. Whether these three tests, however, are considered individually or evaluated in combination, their diagnostic accuracy in the differential diagnosis of mild hypercortisolism does not exceed 80%.

The combined dexamethasone suppression (0.5 mg every 6 h for 2 d) and ovine CRH stimulation test optimize the ability of oCRH to distinguish between the hypercortisolism of pseudo-Cushing's states and CD (62). In the former, the pituitary corticotroph is appropriately restrained by glucocorticoid feedback and does not respond to CRH, whereas in the latter, the corticotroph tumor is generally resistant to this dose of dexamethasone and responds to CRH. Thus, the dexamethasone/CRH test achieves nearly 100% specificity, sensitivity, and diagnostic accuracy. This test should be reserved, however, for those patients who are borderline/mildly hypercortisolemic who have already failed to suppress to 1 mg of overnight dexamethasone and in whom the clinician suspects CD. The criterion used for the diagnosis of CD is a 15-min cortisol level of >38 nmol per liter (1.4 ^g/dL) after the CRH injection.

Another strategy, which is always helpful in diagnosing or ruling out CS, is closely monitoring the patient for a few months. Although true hypercortisolism will persist and cause further symptomatology, the hypercortisolism of pseudo-Cushing's states will frequently subside spontaneously, with effective antidepressant treatment or abstinence from alcohol.

Differential Diagnosis of Cushing's Syndrome

Once the diagnosis of endogenous CS has been established, the next challenge is to establish its specific cause (7,14). Generally, a relatively acute onset of symptoms with rapid progression and associated hypokalemic alkalosis point toward an ectopic ACTH source. Accurate differential diagnosis, however, can only be achieved by the combination of dynamic endocrine testing of the integrity of the feedback regulation of the hypothalamic-pituitary-adrenal (HPA) axis, and imaging techniques used mainly to examine the size and shape of the pituitary and adrenal glands and to localize ectopic ACTH- or CRH-secreting tumors (Fig. 2). It is essential that dynamic adrenal testing is performed when the patient is clearly hypercortisolemic. Hypercortisolemia needs to be documented always at the time of testing to avoid mistakes. To that purpose, all adrenal blocking agents should be discontinued for at least 4 wk before testing.

Basal Plasma Adenocorticotropin Hormone and Corticotropin-Releasing Hormone Concentrations

Morning measurement of plasma ACTH concentrations simultaneously with plasma cortisol would distinguish ACTH-dependent from ACTH-independent CS. Circulating ACTH is typically suppressed/undetectable (<1.1 pmol/L; 5 pg/mL measured by immunoradiometric assay [IRMA]) in adrenal cortisol-secreting tumors, micronodular adrenal disease, and autonomously functioning massive macronodular adrenals. In contrast, plasma ACTH concentrations are normal or elevated in CD and ectopic ACTH and CRH syndrome. Patients with ectopic ACTH syndrome frequently have greater plasma ACTH levels than do those with CD. Interestingly, in many of these patients, ACTH immunoreactivity consists primarily of larger precursor molecules (63). Thus, specific measurement ofACTH precursors, if available, may provide a better marker of ectopic ACTH syndrome.

If ectopic CRH secretion is suspected to be the cause of CS, detection of elevated CRH concentrations in the circulation (>50 pg/mL; 11 pmol/L) is helpful.

CRH Stimulation Test

Most patients (80-90%) with CD respond to CRH with increases in plasma ACTH and cortisol, whereas patients with ectopic ACTH production usually do not (64). New criteria were developed for the interpretation of the morning CRH test that maximize simplicity and cost-effectiveness without compromising diagnostic accuracy (65). The diagnostic accuracy of the CRH test is approx 80-90%. Approximately 10-20% of patients with CD do not respond to CRH stimulation, whereas a few patients with ectopic ACTH syndrome might respond. The best cortisol criterion suggestive of CD is a mean increase at 30 and 45 min of greater than 20% above mean basal values at -5 and 0 min (91% sensitivity and 88% specificity). Similarly, an increase of mean ACTH concentrations at 15 and 30 min after CRH by at least 35% above the mean basal values achieves sensitivity of 91% and specificity of nearly 100%. Indeed, although all patients with ectopic ACTH secretion have less than a 35% increase in ACTH, the probability of CD remains high at all levels of responses, suggesting that in the absence of a discrete lesion on pituitary imaging, it is prudent to perform a second test (e.g., the high-dose dexam-ethasone suppression test or bilateral inferior petrosal or cavernous sinus sampling) to confirm the diagnosis.

Liddle Dexamethasone Suppression Test

The standard low-dose, high-dose dexamethasone suppression test, developed by Grant Liddle (57) has been used extensively for differentiating CD from ectopic ACTH syndrome. In patients with CD, the abnormal corticotrophs are sensitive to glucocorticoid inhibition only at the high dose of dexamethasone (2.0 mg every 6 h for 2 d). In contrast, patients with ectopic ACTH syndrome or cortisol-secreting adrenal tumors usually fail to respond to the dose of 8 mg/d. The classic Liddle criterion for a positive response consistent with CD is a greater than 50% drop in 17-hydroxysteroid excretion on d 2 of high-dose dexamethasone (80% diagnostic accuracy). The diagnostic accuracy of the test, however, increases to 86% by measuring both UFC and 17-hydroxysteroid excretion and by requiring greater suppression of both steroids (greater than 64% and 90%, respectively, for 100% specificity) (66).

Overnight 8-mg Dexamethasone Suppression Test

A simple, reliable, and inexpensive alternative to the Liddle dexamethasone suppression test is the overnight 8-mg dexamethasone suppression test. The advantages are its outpatient administration and the avoidance of errors attributable to incomplete urine collections. The diagnostic accuracy of this overnight test may be similar to that of the standard Liddle dexamethasone suppression test (67,68). Dexamethasone 8 mg orally is administered at 11 PM with measurement of plasma cortisol the next morning at 8 AM. Patients with CD usually show plasma cortisol suppression to <50% (88% sensitivity, 57% specificity but 71% sensitivity, 100% specificity if cutoff <69%) of baseline values (68). Approximately 15% of patients with CD do not show suppression with oral high-dose dexamethasone, whereas approx 30% of patients with ectopic ACTH secretion do show suppression with an iv infusion of dexamethasone for 7 h (69).

Metyrapone Testing

Metyrapone inhibits 11-P-hydroxylase and subsequent cortisol production, thereby stimulating endogenous CRH. This is a relatively simple test but not as reliable as the dexamethasone suppression test. Its overall accuracy ranges from 40% to 70% (70).

Alcohol No More

Alcohol No More

Do you love a drink from time to time? A lot of us do, often when socializing with acquaintances and loved ones. Drinking may be beneficial or harmful, depending upon your age and health status, and, naturally, how much you drink.

Get My Free Ebook

Post a comment