1: JAMA 1993 May 5;269(17):2232-8 Comment in: JAMA. 1993 Nov 17;270(19):2301-2. Corticotropin-releasing hormone stimulation following low-dose dexamethasone administration. A new test to distinguish Cushing's syndrome from pseudo-Cushing's states. Yanovski JA, Cutler GB Jr, Chrousos GP, Nieman LK. Developmental Endocrinology Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD 20892. OBJECTIVE--The biochemical and phenotypic presentation of mild hypercortisolism in Cushing's syndrome is often indistinguishable from that seen in pseudo-Cushing's states such as depression. Both dexamethasone suppression and corticotropin-releasing hormone (CRH) stimulation tests have been used individually to distinguish these conditions, but neither approach has achieved a diagnostic accuracy greater than 85%. Therefore, we sought to develop a combined dexamethasone-CRH test that would take advantage of the altered sensitivity of patients with Cushing's syndrome to both dexamethasone and CRH and would achieve greater accuracy in the diagnosis of Cushing's syndrome. DESIGN--Prospective cohort study. SETTING--Tertiary care research hospital. PATIENTS--A total of 58 adults referred for evaluation of mild hypercortisolism (urine free cortisol level < 1000 nmol/d). The diagnosis of Cushing's syndrome was confirmed at surgery in 39 patients. The diagnosis of a pseudo-Cushing's state was made in 19 patients on the basis of extended follow-up (mean, 28 months) without progression of cushingoid features. INTERVENTION--The low-dose dexamethasone suppression test, the CRH stimulation test, and the CRH stimulation test started 2 hours after completion of low-dose dexamethasone suppression (the dexamethasone-CRH test) were performed in all patients. MAIN OUTCOME MEASURES--Sensitivity, specificity, and accuracy of the three procedures for diagnosis of Cushing's syndrome were calculated from plasma corticotropin, plasma cortisol, urine free cortisol, and urine 17-hydroxycorticosteroid values. RESULTS--The low-dose dexamethasone suppression test had 74% specificity, 69% sensitivity, and 71% diagnostic accuracy, using the standard criterion (17-hydroxycorticosteroid excretion level > 11.0 mumol/d on the second day of dexamethasone administration). With a urine free cortisol criterion for Cushing's syndrome of greater than 100 nmol/d, the low-dose dexamethasone suppression test had 100% specificity, 56% sensitivity, and 71% diagnostic accuracy. The CRH stimulation test without dexamethasone pretreatment had 100% specificity, 64% sensitivity, and 76% diagnostic accuracy. The diagnostic accuracy of the dexamethasone-CRH test for Cushing's syndrome was significantly greater than the accuracy of either the low-dose dexamethasone test or the CRH test alone (P < .01). A plasma cortisol concentration greater than 38 nmol/L measured 15 minutes after the administration of CRH correctly identified all cases of Cushing's syndrome and all cases of pseudo-Cushing's states (100% specificity, sensitivity, and diagnostic accuracy). CONCLUSION--The dexamethasone-CRH test is a more accurate test to distinguish Cushing's syndrome from pseudo-Cushing's states in patients with mild hypercortisolism. PMID: 8386285 [PubMed - indexed for MEDLINE] Introduction. Many of the laboratory abnormalities that are characteristic of Cushing's syndrome [1], such as elevated urine free cortisol and 17-hydroxycorticosteroid excretion, disruptions in the normal diurnal pattern of plasma cortisol secretion, and lack of suppression of plasma cortisol after administration of 1 mg of dexamethasone, may be seen in other conditions. Situations that cause these pseudo-Cushing's states include long-term, active alcoholism and withdrawal from ethanol intoxication [2,3], stressful conditions such as surgery or severe illness [4], renal failure [5], anorexia and bulimia nervosa [6,7,8], the depressed phase of affective disorders [9,10], primary glucocorticoid receptor resistance [11], and severe obesity [9,12]. While the context and physical examination may provide important clues that a particular hypercortisolemic patient does not have Cushing's syndrome, definitive diagnosis may be quite difficult in obese patients with mild hypercortisolism, hirsutism, hypertension, or depression. Thus, such patients may visit their internist, family physician, gynecologist, psychiatrist, or other subspecialist, complaining of symptoms and exhibiting signs consistent with Cushing's syndrome. Further evaluation of these patients may then reveal 24-hour urine excretion of free cortisol or 17-hydroxycorticosteroid compatible with the diagnosis of mild Cushing's syndrome or a pseudo-Cushing's state. The hypercortisolism of pseudo-Cushing's states is believed to be mediated through increased hypothalamic secretion of corticotropin-releasing hormone (CRH) in the context of a hypothalamic-pituitary-adrenal axis that is otherwise appropriately restrained by negative cortisol feedback. In contrast, hypothalamic CRH secretion is suppressed by the hypercortisolism of true Cushing's syndrome, which is less responsive to the negative feedback of exogenous glucocorticoid [8]. Thus, in comparison with true Cushing's disease, patients with pseudo-Cushing's states show a diminished response to administration of exogenous CRH and a greater inhibition of cortisol production by glucocorticoids. These concepts form the basis for tests used in the differential diagnosis of hypercortisolism. The low-dose dexamethasone suppression test, in which 17-hydroxycorticosteroid excretion is measured during administration of dexamethasone at a rate of 0.5 mg every 6 hours for 2 days, has been used since its description by Liddle in 1960 [13]. A 17-hydroxycorticosteroid excretion value greater than 11.0 micromole/d (4 mg/24 h) on the second day of dexamethasone administration is considered evidence for Cushing's syndrome [14,15]. However, this test may misclassify as many as 6% of patients with Cushing's syndrome and up to 15% of patients with pseudo-Cushing's states [9,10,11,12,13,14,15,16,17,18]. The CRH stimulation test, while useful in distinguishing between pituitary and ectopic secretion of adrenocorticotropin [19], is of limited usefulness in the differential diagnosis between pseudo-Cushing's states and Cushing's disease. Although individuals with depression have, on average, significantly blunted adrenocorticotropin and cortisol responses to CRH, there is a large overlap with the responses of patients with Cushing's disease [20,21,22]. Inability to distinguish patients with a pseudo-Cushing's state from patients with Cushing's disease exposes such pseudo-Cushing's patients to the risk of unnecessary transsphenoidal exploration and, when the hypercortisolism is not cured by surgery, to pituitary irradiation. Conversely, if mild Cushing's disease is diagnosed incorrectly as a pseudo-Cushing's state, delay in treatment may have serious adverse consequences for the patient's mental and physical health. To address the need for an improved laboratory test to differentiate Cushing's disease from pseudo-Cushing's states, we evaluated the diagnostic efficiency of a new test that combined CRH stimulation with low-dose dexamethasone suppression (the dexamethasone-CRH test) by examining the effects of dexamethasone pretreatment on CRH-stimulated cortisol and adrenocorticotropin levels. We hypothesized that, at a dexamethasone dose sufficient to suppress normal cortisol production, patients with pseudo-Cushing's states would exhibit low basal plasma cortisol and adrenocorticotropin levels and a diminished response to exogenous CRH. By contrast, patients with Cushing's disease would have higher basal cortisol and adrenocorticotropin levels after administration of dexamethasone and have greater peak values after CRH when compared to patients with pseudo-Cushing's states. METHODS Patients A total of 58 patients admitted to the National Institutes of Health, Bethesda, Md, from June 1987 through June 1991 for evaluation of hypercortisolism were studied prospectively. Each had a history of mild hypercortisolism documented by urine free cortisol levels between 250 and 1000 nmol/d (normal, 50 to 250 nmol/d). Each patient who underwent a 1-mg overnight dexamethasone suppression test had an 8 AM cortisol concentration greater than 138 nmol/L (5 micrograms/dL). None had renal or hepatic disease. Each subject gave written consent for the administration of ovine CRH as approved by the institutional review board of the National Institute of Child Health and Human Development. A presumptive diagnosis of Cushing's syndrome was made on the basis of the combination of physical examination, endocrine studies, and radiologic evaluations [1,9,15,23]. The diagnosis of Cushing's syndrome was considered confirmed only when verified by a pathological specimen or, in those cases of Cushing's disease where no tumor was found, when clinical and laboratory remission resulted after transsphenoidal surgery. The diagnosis of a pseudo-Cushing's state was based on lack of features of severe Cushing's syndrome at initial presentation (such as marked central adiposity, cutaneous atrophy, proximal myopathy, and large purple striae), preservation of some diurnal variation of plasma cortisol, and a lack of progression of the laboratory values, the clinical signs, and the symptoms of Cushing's syndrome over an observation period of at least 17 months. Of the 58 patients, 35 had Cushing's disease, two had ectopic production of adrenocorticotropin, two had primary adrenal disease, and 19 had pseudo-Cushing's states Table 1. ---------------------------------------------- Table 1. Clinical Characteristics of Patients With Mild Hypercortisolism ---------------------------------------------- All 19 patients with pseudo-Cushing's states underwent pituitary magnetic resonance imaging; however, none had an abnormal examination. Most were obese, and their weight and body mass index were not significantly different from that of patients with surgically confirmed Cushing's disease Table 1. Psychiatric consultation was also obtained in all but one of the patients diagnosed with a pseudo-Cushing's state. Based on the psychiatric Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition [24] criteria, one was diagnosed with an eating disorder not-otherwise-specified, one was found to have withdrawal from substance abuse, nine had a current depressive episode, major depression, or a history of a major depressive episode, one had bipolar illness, three met criteria for an undifferentiated somatoform disorder, one had an adjustment disorder with mixed emotional features, one had avoidant personality disorder, and two had no evident psychiatric diagnosis. The minimum period of follow-up was 17 months. The average period of follow-up was 28 months. Nine have been followed up for less than 2 years, six for 2 to 3 years, and four for more than 3 years. Thirteen (68%) of the 19 have had resolution of hypercortisolism during follow-up. Study Protocol The 24-hour urine excretion of free cortisol, 17-hydroxycorticosteroid, and creatinine was measured for 2 days while patients took no glucocorticoids and subsequently while they were receiving 0.5 mg of dexamethasone orally every 6 hours for 2 days starting at 6 AM (the low-dose dexamethasone suppression test), as described previously [25]. The intra-assay and interassay variabilities were, respectively, 4% to 11% and 6% to 18% for urine free cortisol, 5% to 14% and 6% to 21% for 17-hydroxycorticosteroid, and 1% and 2% for creatinine. Each subject also underwent two 8 AM CRH stimulation tests. Ovine CRH (Bachem, Torrance, Calif) was administered as an intravenous bolus injection at a dose of 1 micrograms/kg of body weight. Plasma samples were assayed for cortisol and adrenocorticotropin by Hazelton Laboratories, Vienna, Va, at four basal time points (-15, -10, -5, and 0 minutes) and then at 5, 15, 30, 45, and 60 minutes after CRH administration. The first CRH test was performed while the patient took no glucocorticoids. A second CRH stimulation test (the dexamethasone-CRH test) was performed 2 hours after the patient had completed a course of 0.5 mg of dexamethasone orally every 6 hours for eight doses, starting at noon 2 days before the dexamethasone-CRH test. Thus, the last dose of dexamethasone was administered at 6 AM, 2 hours before the scheduled start of the dexamethasone-CRH test. Plasma cortisol and adrenocorticotropin levels were obtained at the aforementioned times. The sensitivity for the adrenocorticotropin assay ranged from 0.9 to 2.2 pmol/L and for cortisol from 5.5 to 22 nmol/L. The intra-assay and interassay variabilities were 7% to 12% and 12% to 25% for adrenocorticotropin, and 6% and 13% for cortisol. Each cortisol sample was also measured in a second cortisol assay (Abbott TDX kit, Abbott Laboratories, North Chicago, Ill). Results were quite similar, and each cut point identified using the Hazelton assay had the same diagnostic accuracy for Abbott kit cortisol determinations. Therefore, only results from the Hazelton cortisol assay were reported. Each administration of dexamethasone was separated by at least 2 days to allow for recovery between tests. Data Analysis Data were analyzed on a Macintosh IIfx with Superanova and StatView II (Abacus Concepts Inc, Berkeley, Calif). Analysis of variance with repeated measures (with logarithmic data transformation, where appropriate) was performed to detect between-group and time-related differences. Post hoc tests were performed, where appro- priate, and interpreted with the Bonferroni adjustment for multiple comparisons. The receiver operating characteristics [26] of each test for the diagnosis of Cushing's syndrome were then determined using the RuleMaker program (Digital Medicine Inc, Hanover, NH). Sensitivity, specificity, positive and negative predictive values, and diagnostic accuracy [27] were calculated for each test variable at multiple criteria. For each test, the variables and criteria with the highest sensitivity, given 100% specificity for the diagnosis of Cushing's syndrome, were identified and used for between-test comparisons. For the low-dose dexamethasone suppression test, criteria were determined for urine free cortisol and 17-hydroxycorticosteroid levels obtained on day 2 (no dexamethasone) and day 4 (during dexamethasone), and for the percentage of suppression by dexamethasone on day 4 compared with day 2. The 17-hydroxycorticosteroid excretion was also corrected for creatinine excretion; but since results were not improved by this addition, these data are not presented. For the CRH stimulation tests, criteria were established for adrenocorticotropin and cortisol values at each individual time point and for the peak response. Criteria were also determined for the sum of all values during CRH tests, for all combinations of the sums of the 5-, 15-, 30-, 45-, and 60-minute post-CRH time points, and for peak excursion of the response from the baseline. Correlated 2x2 tables were constructed for comparison of test criteria, and Cochran Q statistics were interpreted with the Bonferroni adjustment for multiple comparisons. RESULTS Low-Dose Dexamethasone Suppression Test Urine 17-hydroxycorticosteroid, free cortisol, and creatinine measurements were available before and during dexamethasone administration from all 58 patients (Figs 1 and 2). Daily creatinine measurements varied by no more than 17%, and the range of individual values was between 8.9 and 25.2 mmol/d (1.0 to 2.8 g/24 h). Both urine 17-hydroxycorticosteroid and free cortisol excretion were significantly greater in the group with Cushing's syndrome, whether or not dexamethasone was administered (P<.0001). Although the dexamethasone-suppressed values of both urine 17-hydroxycorticosteroid and free cortisol excretion were significantly lower in patients with pseudo-Cushing's states (P<.001), there was a large overlap in responses between the two groups. ---------------------------------------------- Figure 1. Basal 24-hour urine collections in 58 patients with hypercortisolism. Patients were selected if they had a urine free cortisol level between 250 and 1000 nmol/d. Left, Urine free cortisol excretion. Right, 17-hydroxycorticosteroid excretion. Closed circle indicates Cushing's disease; triangle, ectopic adrenocorticotropin; square, primary adrenal disease; and open circle, pseudo-Cushing's states ---------------------------------------------- The low-dose dexamethasone test had 74% specificity, 69% sensitivity, and 71% diagnostic accuracy for the diagnosis of Cushing's syndrome when the standard criterion, 17-hydroxycorticosteroid excretion level greater than 11 micromole/d on the second day of dexamethasone administration, was used Figure 2and Table 2. To achieve 100% specificity, a criterion for Cushing's syndrome based on 17-hydroxycorticosteroid level greater than 14 micromole/d was required; this yielded 54% sensitivity and 69% diagnostic accuracy Table 2. ---------------------------------------------- Figure 2. Results of low-dose dexamethasone suppression test in 58 patients with hypercortisolism. Criteria shown have 100% specificity. Left, Urine free cortisol on fourth day of low-dose dexamethasone test. Right, 17-hydroxycorticosteroid excretion on fourth day of low-dose dexamethasone test. Definition of symbols as in Figure 1. Horizontal lines indicate criterion levels ---------------------------------------------- ---------------------------------------------- Table 2. Comparison of Criteria Chosen to Yield 100% Specificity for Diagnosis of Cushing's Syndrome in 58 Patients With Mild Hypercortisolism ---------------------------------------------- A diagnostic accuracy of 71% was achieved using a criterion for Cushing's syndrome based on urine free cortisol level greater than 100 nmol/d (36 micrograms/24 h) on the second day of dexamethasone administration Figure 2and Table 2. With this criterion, urine free cortisol had 100% specificity, but only 56% sensitivity (P=.48 vs 17-hydroxycorticosteroid). CRH Stimulation Test Without Dexamethasone Pretreatment Data were available from all patients for the CRH test without dexamethasone (Figs 3 and 4). Although both adrenocorticotropin and cortisol levels were significantly (P<.005) higher in patients with Cushing's syndrome at all time points, the overlap in individual values was too great to afford diagnostic utility at any single point. The best criterion with 100% specificity using adrenocorticotropin was a value greater than 35 pmol/L at any time point during the CRH test, which yielded 13% sensitivity and 41% diagnostic accuracy Table 2. The best sensitivity (64%) with 100% specificity was obtained using the criterion of a sum of post-CRH cortisol levels greater than 3450 nmol/L Figure 4, which was significantly better than the sensitivity based on the adrenocorticotropin criterion (P<.02). ---------------------------------------------- Figure 4. No caption available ---------------------------------------------- Attempts to find criteria with improved diagnostic accuracy by combining results from the low-dose dexamethasone suppression test with those from the CRH test without dexamethasone pretreatment did not improve diagnostic accuracy over that found with the low-dose dexamethasone test alone. CRH Test With Dexamethasone Pretreatment All patients completed the dexamethasone-CRH test (Figs 5 and 6). Compared with those who had pseudo-Cushing's states, the patients with Cushing's syndrome had significantly elevated cortisol (P<.001) and adrenocorticotropin (P<.01) values at all time points, both before and after CRH administration. The criterion of a basal plasma cortisol level greater than 38 nmol/L (1.4 micrograms/dL) during the dexamethasone-CRH test had 100% specificity, 90% sensitivity, and 91% diagnostic accuracy Table 2. After administration of CRH, those patients with Cushing's disease whose cortisol levels were less than 38 nmol/L had increases in cortisol level, whereas those with pseudo-Cushing's states generally did not. A single plasma cortisol level greater than 38 nmol/L, determined 15 minutes after administration of CRH, separated all cases of Cushing's syndrome from all cases of pseudo-Cushing's states and yielded 100% sensitivity, specificity, and diagnostic accuracy Figure 5. The 15-minute, post-CRH cortisol value had significantly greater diagnostic accuracy than the dexamethasone-CRH basal cortisol before CRH administration (P<.05). At time points more than 15 minutes after administration of CRH, cortisol values exceeded 38 nmol/L in some patients with pseudo-Cushing's states and overlapped with the responses seen in Cushing's syndrome. However, excellent discrimination was still found for the 30- and 45-minute cortisol samples (99% and 98% diagnostic accuracy, respectively). The highest cortisol level recorded for any patient in the pseudo-Cushing's group during the dexamethasone-CRH test was 96 nmol/L (3.5 micrograms/dL) at 60 minutes. ---------------------------------------------- Figure 5. Criteria for best-diagnostic accuracy for dexamethasone-corticotropin-releasing hormone (CRH) test; plasma cortisol levels obtained 15 minutes after administration of CRH. Definition of symbols as in Figure 1. Horizontal rule indicates criterion level ---------------------------------------------- For the adrenocorticotropin levels during the dexamethasone-CRH test, a plasma adrenocorticotropin level that exceeded 3.5 pmol/L at 30 minutes after CRH administration offered the highest diagnostic accuracy (83%) and had 74% sensitivity and 100% specificity Table 2. The plasma adrenocorticotropin value during the dexamethasone-CRH test was also examined when patients with adrenal causes of Cushing's syndrome were excluded (because their adrenocorticotropin values are expected to be very low). Under these conditions, a dexamethasone-CRH plasma adrenocorticotropin level greater than 3.83 pmol/L obtained 30 minutes after administration of CRH had 89% diagnostic accuracy, 83% sensitivity, and 100% specificity. Thus, regardless of whether patients with adrenal disease were excluded, the dexamethasone-CRH test, 15-minute cortisol criterion had significantly greater sensitivity, negative predictive value, and diagnostic accuracy when compared with the best dexamethasone-CRH adrenocorticotropin criterion (P<.01). Test Comparison Criteria with 100% specificity and maximal sensitivity were used to compare the tests Table 2. The cortisol value of the dexamethasone-CRH test obtained 15 minutes after CRH administration had significantly greater sensitivity, negative predictive value, and diagnostic accuracy when compared either to the low-dose dexamethasone test (P<.01) or to the CRH stimulation test performed without dexamethasone administration (P<.01). COMMENT The dexamethasone-CRH test, a new test for the differential diagnosis of hypercortisolism, distinguished patients with mild Cushing's syndrome from those with pseudo-Cushing's states with greater accuracy than was possible using either the low-dose dexamethasone test or the CRH test alone. The improvement in diagnostic accuracy of the combined dexamethasone-CRH test may be explicable by the pathophysiology of Cushing's syndrome. Patients with primary adrenal pathology and most patients with the syndrome of ectopic adrenocorticotropin generally show little feedback regulation of cortisol production by glucocorticoid and little response to CRH [19]. These patients are discriminated from patients with pseudo-Cushing's states by virtue of their lack of cortisol suppression by low-dose dexamethasone. In contrast to patients with cortisol-producing adrenal disorders or patients with ectopic adrenocorticotropin-producing tumors, most individuals with Cushing's disease show some suppression of adrenocorticotropin and cortisol by dexamethasone and stimulation of adrenocorticotropin and cortisol by CRH. However, cortisol production decreases greatly in some patients with Cushing's disease during low-dose dexamethasone administration, possibly because of diminished dexamethasone clearance [28,29]. Such patients would be misclassified as having pseudo-Cushing's states if evaluated only by suppression of cortisol production after low-dose dexamethasone administration. Thus, the dexamethasone-CRH test basal plasma cortisol measurement greater than 38 nmol/L correctly identified only 91% of the patients in this series. By the addition of CRH to stimulate greater adrenocorticotropin and cortisol secretion (either by pituitary or adrenal [30] mechanisms) in patients with Cushing's disease who are CRH-sensitive, those patients with unusual sensitivity to dexamethasone suppression achieved adrenocorticotropin and cortisol levels above those of patients with pseudo-Cushing's states. Thus, the dexamethasone-CRH test, 15-minute cortisol test correctly identified more of the patients with Cushing's syndrome than the dexamethasone-CRH test basal plasma cortisol, or the low-dose dexamethasone suppression test, either in our series or in others [9,10,11,12,13,14,15,16,17,18]. Administration of CRH following dexamethasone suppression allowed the criterion for the diagnosis of Cushing's syndrome to be 100% specific with a much higher sensitivity than was possible with the dexamethasone test alone (100% vs 71%, respectively). The dexamethasone-CRH test should be reserved for those patients who the clinician suspects have Cushing's syndrome who are found to fail to suppress 8 AM plasma cortisol after 1 mg of dexamethasone has been administered at 11 PM, or who have an elevated 24-hour excretion of free cortisol. After 2 days of low-dose dexamethasone suppression, a single plasma cortisol level obtained 15 minutes after CRH stimulation may suffice to discriminate patients with mild Cushing's syndrome from those with pseudo-Cushing's states. Individuals with an elevated dexamethasone-CRH plasma cortisol level could then be selected for further evaluation of Cushing's syndrome. Plasma cortisol offers several practical advantages over the classic response measure used in the low-dose dexamethasone suppression test, urine 17-hydroxycorticosteroid excretion. The problems of patient adherence are diminished when 24-hour urine collections are not required. Furthermore, the radioimmunoassay for plasma cortisol is easier to perform, more reproducible, more readily available, and less expensive than urine 17-hydroxycorticosteroid measurement. These factors, combined with its excellent diagnostic accuracy, make the dexamethasone-CRH test a better diagnostic test for the presence or absence of Cushing's syndrome. The dexamethasone-CRH test also compares favorably with the insulin tolerance test as a method to distinguish those with and without Cushing's syndrome. Patients with Cushing's syndrome often show little elevation in plasma cortisol after insulin-induced hypoglycemia, while those with pseudo-Cushing's states are said to elevate plasma cortisol normally. However, the insulin tolerance test fails to identify as many as 18% of patients with Cushing's syndrome (diagnostic accuracy <76%) [9,10,31,32], because they may show normal responsiveness to hypoglycemia. Thus, the dexamethasone-CRH test appears to be superior to previously described tests for the differential diagnosis of hypercortisolism. Because individuals with mild Cushing's syndrome may not develop overt signs of Cushing's syndrome for many years, it is conceivable that some of the patients found to have a pseudo-Cushing's state by our evaluation and by the dexamethasone-CRH test might subsequently develop apparent Cushing's syndrome. To assess theoretically the potential impact of progression to Cushing's syndrome in some of these patients, we reexamined the five patients diagnosed with a pseudo-Cushing's state who had the highest levels of 17-hydroxycorticosteroid excretion during the low-dose dexamethasone test. These five patients had values greater than 11 micromole/d and, thus, would be classified as having Cushing's syndrome by the classic criterion for this test. Two of these individuals have been followed for more than 4 years without development of the signs of Cushing's syndrome and probably do not have the disorder. The remaining three patients could conceivably represent incorrect diagnoses. According to this theoretical analysis, the dexamethasone-CRH test, 15-minute cortisol level greater than 38 nmol/L would have 100% specificity, 95% sensitivity, and 95% diagnostic accuracy; the low-dose dexamethasone test would have 87% specificity, 71% sensitivity, and 76% diagnostic accuracy for the diagnosis of Cushing's syndrome. Thus, even in the circumstance of eventual progression to Cushing's disease in some of the patients diagnosed with pseudo-Cushing's states, the dexamethasone-CRH test would remain superior to the low-dose dexamethasone suppression test. The patients with pseudo-Cushing's states in this study did not include individuals with severe melancholic depression, anorexia nervosa, cortisol resistance syndrome, renal failure, or recent surgery. Therefore, we cannot comment on whether the criteria presented will be useful for differentiating such patients from those with Cushing's syndrome. Patients who have appetite disturbances and weight loss (as observed in those with vegetative signs of depression or anorexia nervosa) typically show resolution of hypercortisolism when they regain weight [8]. Because of this distinction, they are not often confused with patients who may have Cushing's syndrome. However, several of our patients, both with pseudo-Cushing's states and Cushing's syndrome, met Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition [24] criteria for affective disorder. Thus, we believe that the dexamethasone-CRH test may prove especially useful in the evaluation of patients with affective dysfunction and equivocal physical or biochemical findings. Like all tests for the differential diagnosis of hypercortisolism, the utility of the dexamethasone-CRH test has been examined only in patients who were actively producing excessive cortisol. Thus, its value in classifying those who are not hypercortisolemic at the time of testing, but who are believed to have Cushing's syndrome with periodic hormonogenesis [33], remains unknown. Repeated 24-hour urine collections must be performed in such individuals to document hypercortisolism. Other known limitations of the uses of dexamethasone pertain. For example, results from the dexamethasone-CRH test should be interpreted with caution in patients receiving medications that affect the metabolism or clearance of dexamethasone or cortisol-binding globulin. Another limitation of this test is that it has been studied only in adults. It is likely that a weight-adjusted dose of dexamethasone (eg, 5 to 7.5 micrograms/kg per dose [17]) will be necessary for the test to be applicable to children. Previous limited studies have proposed the use of plasma cortisol measurements after dexamethasone administration for the differential diagnosis of hypercortisolism [9,10,34,35,36]. As in the present study, most have shown some overlap in the responses of those with and without Cushing's syndrome. Although others have performed CRH tests in clinically normal controls and in patients with Cushing's syndrome or depression following a variety of dexamethasone administration regimens to demonstrate the increased resistance to glucocorticoid suppression of patients with Cushing's syndrome [37,38,39,40,41], none has used dexamethasone doses capable of inhibiting the normal corticotroph's ability to secrete adrenocorticotropin and compared the results in patients with Cushing's syndrome and in patients with pseudo-Cushing's states. Thus, the dexamethasone-CRH test described in this article is a novel test for the differential diagnosis of hypercortisolism. In our series, the dexamethasone-CRH test identified all proven patients with Cushing's syndrome without appearing to misclassify any patient with a pseudo-Cushing's state. Properly performed, the dexamethasone-CRH test should prove useful in the outpatient evaluation of patients with suspected Cushing's syndrome who have a mildly elevated urine free cortisol and equivocal physical findings. REFERENCES 1. Plotz CM, Knowlton A, Ragan C. The natural history of Cushing's syndrome. Am J Med. 1921;13:597-614. 2. Lamberts SW, Klijn JG, de Jong FH, et al. Hormone secretion in alcohol-induced pseudo-Cushing's syndrome. JAMA. 1979;242:1640-1643. 3. Wand GS, Dobs AS. Alterations in the hypothalamic-pituitary-adrenal axis in actively drinking alcoholics. J Clin Endocrinol Metab. 1991;72:1290-1295. [Medline Link] 4. Reincke M, Allolio B, Wurth G, Winkelman W. The hypothalamic-pituitary-adrenal axis in critically ill patients: response to CRH and dexamethasone. Endocr Soc Abstr. 1991;73:250. Abstract. 5. Wallace EZ, Rosman P, Toshav N, et al. Pituitary-adrenocortical function in chronic renal failure: studies of episodic secretion of cortisol and dexamethasone suppressibility. J Clin Endocrinol Metab. 1980;50:46-51. [Medline Link] 6. Gold PW, Gwirtsman HE, Avgerinos PC, et al. Abnormal hypothalamic-pituitary-adrenal function in anorexia nervosa. N Engl J Med. 1986;314:1335-1342. [Medline Link] 7. Smith SR, Bledsoe T, Chhetri MKL. Cortisol metabolism and the pituitary-adrenal axis in adults with protein-calorie malnutrition. J Clin Endocrinol Metab. 1975;40:43-52. [Medline Link] 8. Laue L, Gold PW, Richmond A, Chrousos G. The hypothalamic-pituitary-adrenal axis in anorexia nervosa and bulimia nervosa: pathophysiologic implications. Adv Pediatr. 1991;38:287-316. [Medline Link] 9. Crapo L. Cushing's syndrome: a review of diagnostic tests. Metabolism. 1979;9:955-977. 10. Besser GM, Edwards CRW. Cushing's syndrome. Clin Endocrinol Metab. 1972;1:451-490. 11. Werner S, Thoren M, Gustafsson J, Bronnegard M. Glucocorticoid receptor abnormalities in fibroblasts from patients with idiopathic resistance to dexamethasone diagnosed when evaluated for adrenocortical disorders. J Clin Endocrinol Metab. 1992;75:1005-1009. [Medline Link] 12. Schteingart DE, Gregarman RI, Conn JW. A comparison of the characteristics of increased adrenocortical function in obesity and in Cushing's syndrome. Metabolism. 1963;12:484-497. 13. Liddle GW. Tests of pituitary-adrenal suppressibility in the diagnosis of Cushing's syndrome. J Clin Endocrinol Metab. 1960;20:1539-1561. 14. Eddy RL, Jones AL, Gilliland PF, et al. Cushing's syndrome: a prospective study of diagnostic methods. Am J Med. 1973;55:621-630. [Medline Link] 15. Schteingart DE. Cushing's syndrome. Endocrinol Metab Clin North Am. 1989;18:311-338. [Medline Link] 16. Burke CW, Beardwell CG. Cushing's syndrome: an evaluation of the clinical usefulness of urinary free cortisol and other urinary steroid measurements. Q J Med. 1973;42:175-204. [Medline Link] 17. Streeten DHP, Stevenson CT, Dalakos TG, et al. The diagnosis of hypercortisolism: biochemical criteria differentiating patients from lean and obese normal subjects and from females on oral contraceptives. J Clin Endocrinol Metab. 1969;29:1191-1211. [Medline Link] 18. Hankin ME, Theile HM, Steinbeck AW. An evaluation of laboratory tests for the detection and differential diagnosis of Cushing's syndrome. Clin Endocrinol (Oxf). 1977;6:185-196. [Medline Link] 19. Nieman L, Chrousos G, Oldfield E, et al. The ovine corticotropin-releasing hormone stimulation test and the dexamethasone suppression test in the differential diagnosis of Cushing syndrome. Ann Intern Med. 1986;105:862-867. [Medline Link] 20. Gold PW, Loriaux DL, Roy A, et al. Responses to corticotropin-releasing hormone in the hypercortisolism of depression and Cushing's disease. N Engl J Med. 1986;314:1329-1335. [Medline Link] [PsycINFO Link] 21. Rupprecht R, Lesch KP, Muller U, et al. Blunted adrenocorticotropin but normal beta-endorphin release after human corticotropin-releasing hormone administration in depression. J Clin Endocrinol Metab. 1989;69:600-603. [Medline Link] 22. Grossman AB, Howlett TA, Perry L, et al. CRF in the differential diagnosis of Cushing's syndrome: a comparison with the dexamethasone suppression test. Clin Endocrinol (Oxf). 1988;29:167-178. 23. Gold EM. The Cushing syndromes: changing views of diagnosis and treatment. Ann Intern Med. 1979;90:829-844. [Medline Link] 24. American Psychiatric Association, Committee on Nomenclature and Statistics. Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition. Washington, DC: American Psychiatric Association; 1987. 25. Flack M, Oldfield E, Cutler G, et al. Use of urine free cortisol as an end point in the high-dose dexamethasone suppression test for the differential diagnosis of the Cushing syndrome. Ann Intern Med. 1992;116:211-217. [Medline Link] 26. Swets JA. Measuring the accuracy of diagnostic systems. Science. 1988;240:1285-1293. [Medline Link] 27. McNeil BJ, Keeler E, Adelstein SJ. Primer on certain elements of medical decision making. N Engl J Med. 1975;293:211-215. [Medline Link] 28. Kapcala LP, Hamilton SM, Meikle AW. Cushing's disease with 'normal suppression' due to decreased dexamethasone clearance. Arch Intern Med. 1984;144:636-637. [Medline Link] 29. Caro JF, Meikle AW, Check JH, Cohen SN. 'Normal suppression' to dexamethasone in Cushing's disease: an expression of decreased metabolic clearance for dexamethasone. J Clin Endocrinol Metab. 1978;47:667-670. [Medline Link] 30. Andreis PC, Neri G, Mazzocchi G, Musajo F, Nussdorfer F. Direct secretagogue effect of corticotropin-releasing factor on the rat adrenal cortex: the involvement of the zona medullaris. Endocrinology. 1992;131:69-72. [Medline Link] 31. Butler PWP, Besser GM. Pituitary-adrenal function in severe depressive illness. Lancet. 1968;1:1234-1236. [Medline Link] 32. Krieger DT, Luria M. Plasma ACTH and cortisol responses to TRF, vasopressin, or hypoglycemia in Cushing's disease and Nelson's syndrome. J Clin Endocrinol Metab. 1977;44:361-368. [Medline Link] 33. Brown RD, Van Loon GR, Orth DN, Liddle GW. Cushing's disease with periodic hormonogenesis: one explanation for paradoxical response to dexamethasone. J Clin Endocrinol Metab. 1973;36:445-451. [Medline Link] 34. Kennedy L, Atkinson A, Johnston H, Sheridan B, Hadden D. Serum cortisol concentrations during low dose dexamethasone suppression test to screen for Cushing's syndrome. BMJ. 1984;289:1188-1191. 35. Gambardella S, Tamburrano G, Giaccari A, et al. Intravenous dexamethasone and subsequent ACTH test in comparison with dexamethasone oral test in the diagnosis of Cushing's syndrome: a report of 20 cases. J Endocrinol Invest. 1989;12:163-170. [Medline Link] 36. Ashcraft M, Van Herle A, Vener S, Geffner D. Serum cortisol levels in Cushing's syndrome after low- and high-dose dexamethasone suppression. Ann Intern Med. 1982;97:21-26. [Medline Link] 37. Heuser IJE, Holsboer F. The CRH-stimulation test in affective and anxiety disorder: comparison to physical stress. Clin Neuropharmacol. 1992;15(suppl No. 1):159A-160A. 38. Holsboer-Trachsler E, Stohler R, Hatzinger M. Repeated administration of the combined dexamethasone-human corticotropin releasing hormone stimulation test during treatment of depression. Psychiatry Res. 1991;38:163-171. [Medline Link] 39. Watson AC, Rosenfield RL, Fang VS. Recovery from glucocorticoid inhibition of the responses to corticotropin-releasing hormone. Clin Endocrinol (Oxf). 1988;28:471-477. [Medline Link] 40. Lytras N, Grossman A, Tomlin PS, et al. Corticotrophin releasing factor: responses in normal subjects and patients with disorders of the hypothalamus and pituitary. Clin Endocrinol (Oxf). 1984;20:71-84. [Medline Link] 41. Muller OA, Hartwimmer J, Hauer A, et al. Corticotropin-releasing factor (CRH): stimulation in normal controls and in patients with Cushing's syndrome. Psychoneuroendocrinology. 1986;11:49-60. [Medline Link] ---------------------------------------------- ---------------------------------------------- Figure 3. Results of corticotropin-releasing hormone stimulation test, without low-dose dexamethasone pretreatment. Top, Plasma adrenocorticotropin values. Bottom, Plasma cortisol values. Means and SEMs are given. Definition of symbols as in Figure 1 ---------------------------------------------- ---------------------------------------------- Figure 6. Results of dexamethasone-corticotropin-releasing hormone test, following low-dose (2 mg/d) dexamethasone pretreatment. Top, Plasma adrenocorticotropin values. The data for pseudo-Cushing's states and for primary adrenal disease are superimposed. Bottom, Plasma cortisol values. Means and SEMs are given. Definition of symbols as in Figure 1 ---------------------------------