| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
BRIEF REPORT |
Departments of Endocrinology (B.G., P.R., A.T.) and Nuclear Medicine (J.-B.C.), University Hospital of Bordeaux, 33600 Pessac, France; Department of Endocrinology (O.C., M.M.), University Hospital of Grenoble, 38000 Grenoble, France; and Department of Endocrinology (C.C.), University Hospital of Lille, 59037 Lille, France
Address all correspondence and requests for reprints to: Antoine Tabarin, Endocrinology, Hopital Haut Leveque, Avenue de Magellan, 33600 Pessac, France. E-mail: antoine.tabarin{at}chu-bordeaux.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: Our objective was to reevaluate the performance of the DST-CRH test to differentiate CD from PCS and compare it with that of midnight plasma cortisol measurement.
Setting: The study took place at three specialized tertiary care university hospitals.
Design: Fourteen patients with PCS and 17 patients with CD matched for 24-h urinary free cortisol were retrospectively studied.
Main Outcome Measure: Diagnosis or exclusion of CD was the main outcome measure.
Results: A 55 nmol/liter cortisol concentration after dexamethasone (DST) yielded 94% sensitivity, 86% specificity, and 90% diagnostic accuracy. Using the historical 38 nmol/liter threshold for plasma cortisol 15 min after CRH administration, the DST-CRH test achieved 100% sensitivity, 50% specificity, and 77% diagnostic accuracy. Increasing the threshold to 110 nmol/liter improved the specificity and diagnostic accuracy to 86 and 93.5%, respectively. However, diagnostic accuracy was not significantly different from that of the DST. A midnight plasma cortisol concentration of more than 256 nmol/liter was consistent with the diagnosis of CD with 100% sensitivity, specificity, and diagnostic accuracy.
Conclusion: The diagnostic performance of the DST-CRH test for the differential diagnosis between PCS and mild CD was lower than previously reported. Although the specificity of the test is improved using a revised cortisol threshold, its diagnostic accuracy is not better than that of the standard DST. Our study supports the preferential use of the DST and midnight plasma cortisol measurement as first-line diagnostic tests in equivocal cases.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Patients recruited in three endocrine units of university hospitals between 1997 and 2005 were retrospectively studied. The criteria for PCS were: 1) presence of clinical and biochemical features compatible with CS, 2) lack of specific clinical features (proximal myopathy, purple striae, easy bruising, or skin atrophy), 3) associated condition known to activate the hypothalamic-pituitary-adrenal axis, 4) lack of progression of clinical and biochemical abnormalities during follow-up and/or normalization of biochemical abnormalities after cure of the associated pathological condition, and 5) normal pituitary magnetic resonance imaging including contrast enhancement.
Fourteen patients (five males and nine females; age, 39.2 ± 4.2 yr) with PCS are reported. All had been evaluated by endocrinologists before referral to the specialized endocrine units and were then followed during 28.5 ± 2.5 (18–48) months. One was alcoholic, five had binge eating disorders, and eight had endogenous depression. Nine patients displayed elevated 24-h UFC [170 ± 17 (111–267) µg/24 h; normal, <100). The five remaining patients with normal UFC, 33–87 µg/24 h, displayed nonsuppressed plasma cortisol after the 1-mg overnight DST [347 ± 63 (230–546) nmol/liter]. The 0800-h ACTH plasma concentrations were in the normal range in all patients (Table 1
).
|
Methods
All investigations were performed during hospitalization.
Combined DST-CRH test
The protocol described by Yanovski et al. (1) was followed except that 100 µg CRH instead of 1 µg/kg were administered at 0800 h. Human-sequence CRH (Ferring, Kiel, Germany) was used in most patients. Ovine-sequence CRH (Biogenesis, Poole, UK) was used in 6 patients with CD.
Midnight plasma cortisol sampling
All patients had at least one night of habituation to the hospital environment, and an indwelling venous catheter was placed at least 6 h before blood sampling.
Assays
Plasma cortisol concentration was determined by a solid-phase RIA (Coat-a-Count; Diagnostic Products Corp., Los Angeles, CA), chemiluminescence with automated analyzer (ADVIA; Bayer, Puteaux, France) and chemiluminescence with automated analyzer Advantage (Nichols, San Clemente, CA) in Bordeaux, Grenoble, and Lille, respectively. The lowest detectable level was 27 nmol/liter in all centers. The intraassay and interassay coefficients of variation at low cortisol concentrations (64–107 nmol/liter) ranged from 3.0–8.6% and 5.0–17.0%, respectively. ACTH was assayed by solid-phase immunoradiometric assay (Nichols) in Bordeaux and Grenoble and by chemiluminescence with automated analyzer Immulite 2000 (Diagnostic Products) in Lille.
Statistical analysis
Results are presented as mean ± SEM (range). The prevalence of symptoms was compared using 
2 test. Biochemical comparisons were done by t tests and ANOVA for repeated time-points analysis followed by Fisher post hoc test. Significance was accepted for P < 0.05. Specificity, sensitivity, diagnostic accuracy, and receiver-operating characteristic curves were calculated according to standard statistical methods for the various results (3). Comparisons between diagnostic accuracies were performed using the paired exact test.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Central weight gain, hypertension, and hirsutism were equally prevalent in the two groups of patients. Ecchymoses, purple striae, and proximal myopathy were observed only in 64, 41, and 41% of patients with CD, respectively (P < 0.05 for all).
DST
Post-DST cortisol concentrations were significantly higher in patients with CD than in patients with PCS [243.6 ± 39.6 (48–643) vs. 63.9 ± 23.1 (27–452) nmol/liter, respectively; P < 0.0004] (Fig. 1A). Plasma cortisol was more than 50 nmol/liter in 16 of 17 patients with CD and in three patients with PCS (54, 183, and 328 nmol/liter) (Table 1
and Fig. 1B). Thus, the 50 nmol/liter threshold yielded 94% sensitivity [94% confidence interval (CI) 71.2–99], 79% specificity (95% CI 49.2–95.1), and 90% diagnostic accuracy for the diagnosis of CD. A 55 nmol/liter threshold yielded the best performances (sensitivity 94%, CI 71.2–99; specificity 86%, CI 57.2–97.8; and diagnostic accuracy 90%).
|
After CRH stimulation, plasma cortisol and ACTH increased significantly more in patients with CD than in patients with PCS at all time points (Fig. 1A). Plasma cortisol 15 min after CRH injection was superior to the 38 nmol/liter historical threshold in all patients with CD (1) but also in half of patients with PCS (40–516 nmol/liter). No patient with CD had plasma cortisol concentration 15 min after CRH that was less than 110 nmol/liter, whereas only two patients with PCS had cortisol concentration above this threshold (Fig. 1C). Thus, a 110 nmol/liter threshold for plasma cortisol 15 min after CRH increased the specificity of the test from 50 (CI 23.1–76.9) to 86% (CI 57.2–97.8) and the diagnostic accuracy from 77 to 93%, respectively, keeping its 100% sensitivity.
CRH induced a significant increase (more than four times the intraassay coefficient of variation) in plasma cortisol in 14 of 17 patients with CD but also in eight patients with PCS with highest cortisol values observed between 30 and 60 min after its administration. Thus, the accuracy of the test was not improved if the relative cortisol response to CRH or the absolute concentration at a given time point was considered. We analyzed the diagnostic performance of ACTH measurement. The best diagnostic performance was obtained with a plasma ACTH concentration 15 min after CRH greater than 3.5 pmol/liter (sensitivity 100%, specificity 85%, and diagnostic accuracy 93%). Receiver-operating characteristic analysis of various thresholds confirmed that the highest diagnostic accuracy was obtained with a 110 nmol/liter threshold for plasma cortisol concentration 15 min after CRH. However, the diagnostic accuracy of the combined test did not significantly differ from that of the DST.
Subanalyses of CD patients with ovine CRH stimulation did not change the diagnostic performances of the combined test.
Midnight plasma cortisol
The midnight plasma cortisol concentration was significantly different between the two groups of patients [429.9 ± 21.8 (267–556) vs. 119.8 ± 18.1 (43–256) nmol/liter, CD vs. PCS, respectively; P < 0.0001]. Midnight plasma cortisol correctly identified all patients because it was less than 257 nmol/liter in all patients with PCS and above this threshold in all patients with CD (Fig. 1D).
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
All patients with CD displayed elevated plasma cortisol 15 min after CRH administration. It rose to 489 nmol/liter in the CD patient with dexamethasone-suppressed plasma cortisol. Thus, and as previously suggested (1, 2), one specific utility of the DST-CRH test is to identify the small proportion of patients with CD who are missed by the DST test because an absent cortisol response excludes CD. However, half of patients with PCS also had cortisol concentrations above the historical 38 nmol/liter threshold resulting in a poor specificity (50%) of the DST-CRH test. Extensive analysis of data suggested that the best diagnostic performance was obtained with a 110 nmol/liter threshold for plasma cortisol concentration 15 min after CRH. The need to increase the historical threshold to maximize specificity has been recently advocated in two studies (2, 6). Significant differences with the original protocol in the study of Martin et al. (6) hampers direct comparisons with the results of the National Institutes of Health study (1). Pecori Giraldi et al. (2) also advocated the use of a 110 nmol/liter threshold to improve the specificity of the test. However, and similarly to our findings, the DST-CRH test did not allow complete separation of the two groups of patients. Indeed, plasma cortisol increased significantly after the DST-CRH test in some patients with PCS, indicating that normal pituitary corticotrophs may retain a degree of responsiveness to pharmacological doses of exogenous CRH while suppressed with dexamethasone. Plasma cortisol reached 452 and 516 nmol/liter in a 20-yr-old binge-eating woman and a 40-yr-old depressive nightshift worker man refusing psychiatric treatment. During follow-up (25 and 48 months, respectively), patients showed episodic normalization of 24-h UFC, permanent preservation of cortisol circadian rhythm, and lack of response to the desmopressin test (2, 7, 8). Thus, although speculative, the diagnosis of PCS is highly probable in these two patients. Such cases typically illustrate the difficulty of establishing the diagnosis in patients with mild ACTH-dependent hypercortisolism, lack of specific clinical features, normal pituitary magnetic resonance imaging, and the absence of a foolproof test.
Midnight plasma cortisol measurement has between 90 and 100% sensitivity and 96 and 100% specificity to differentiate PCS from the CS (9, 10, 11). Our data confirm its usefulness because a concentration of more than 256 nmol/liter allowed absolute discrimination between the two groups of patients. In accordance with the results of other studies, the narrowness of the gap between cortisol concentrations of the two groups of patients (256–267 nmol/liter) intuitively suggest that the absolute diagnostic accuracy that we found would not be confirmed with an increased number of cases.
In conclusion, our study confirms the difficulty of the differential diagnosis between mild CD and PCS. In our series, the specificity of the DST-CRH was lower than previously reported, and its diagnostic performance was not superior to that of the low-dose DST and was inferior to that of midnight plasma cortisol measurement. Because the combined DST-CRH test is expensive and cumbersome, our results support the preferential use of the low-dose DST and midnight cortisol measurement as first-line tests in equivocal cases.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online July 17, 2007
Abbreviations: CD, Cushing’s disease; CI, conficence interval; DST, dexamethasone suppression test; PCS, pseudo-Cushing syndrome; UFC, urinary free cortisol.
Received December 20, 2006.
Accepted July 11, 2007.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Boscaro and G. Arnaldi Approach to the Patient with Possible Cushing's Syndrome J. Clin. Endocrinol. Metab., September 1, 2009; 94(9): 3121 - 3131. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Arnaldi, G. Tirabassi, R. Papa, G. Furlani, L. Trementino, M. Cardinaletti, E. Faloia, and M. Boscaro Human corticotropin releasing hormone test performance in the differential diagnosis between Cushing's disease and pseudo-Cushing state is enhanced by combined ACTH and cortisol analysis Eur. J. Endocrinol., June 1, 2009; 160(6): 891 - 898. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Reimondo, S. Bovio, B. Allasino, S. De Francia, B. Zaggia, I. Micossi, A. Termine, F. De Martino, P. Paccotti, F. Di Carlo, et al. The combined low-dose dexamethasone suppression corticotropin-releasing hormone test as a tool to rule out Cushing's syndrome Eur. J. Endocrinol., November 1, 2008; 159(5): 569 - 576. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. L. Batista, N. Courcoutsakis, J. Riar, M. F. Keil, and C. A. Stratakis Severe Obesity Confounds the Interpretation of Low-Dose Dexamethasone Test Combined with the Administration of Ovine Corticotrophin-Releasing Hormone in Childhood Cushing Syndrome J. Clin. Endocrinol. Metab., November 1, 2008; 93(11): 4323 - 4330. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. K. Nieman, B. M. K. Biller, J. W. Findling, J. Newell-Price, M. O. Savage, P. M. Stewart, and V. M. Montori The Diagnosis of Cushing's Syndrome: An Endocrine Society Clinical Practice Guideline J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1526 - 1540. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, CD; 
, PCS. *, P < 0.05. A, Plasma cortisol (nanomoles per liter) during the combined DST-CRH test. Dashed arrow, DST; bold arrow, CRH test. B, Individual concentrations of plasma cortisol (nanomoles per liter) after DST. The dotted line represents the threshold of 55 nmol/liter. C,Individual concentrations of plasma cortisol (nanomoles per liter) 15 min after CRH stimulation. The solid line represents the threshold cutoff of 38 nmol/liter. The dotted line represents the threshold cutoff of 110 nmol/liter. D, Individual concentrations of midnight plasma cortisol (nanomoles per liter). The solid line represents the 256 nmol/liter threshold.