Concomitant Medication Use Can Confound Interpretation of the Combined Dexamethasone-Corticotropin Releasing Hormone Test in Cushing's Syndrome

Context: The ability of combined dexamethasone-corticotropin releasing hormone (Dex-CRH) testing to distinguish pseudo-Cushing's syndrome (PCS) from Cushing's syndrome is controversial. One factor potentially impairing diagnostic efficacy is the concomitant use of commonly prescribed medications that may alter dexamethasone metabolism. Objective: Our objective was to assess the diagnostic accuracy of the Dex-CRH test and evaluate the potential impact of concomitant drugs. Design: The study was a retrospective one. Participants: Participants included 101 patients [60 Cushing's disease (CD); 41 PCS] who underwent 112 Dex-CRH tests. Patients were divided into two groups, depending on use of medications potentially interfering with dexamethasone metabolism: 58 tests were classified as No Meds (32 CD; 26 PCS) and 54 as Meds (34 CD; 20 PCS). The latter group was further subdivided into patients taking one medication vs. those taking multiple medications. Main Outcome Measures: Diagnostic accuracy of different serum cortisol and ACTH thresholds at baseline and 15 min after CRH injection was assessed. Results: The specificity of a baseline post-low-dose-dexamethasone-suppressed test cortisol lower than 1.4 mu g/dl (38 nmol/liter) was significantly higher in the No Meds vs. the Meds group (P = 0.014). Sensitivity and specificity using a post-CRH cortisol cutoff of 1.4 mu g/dl (38 nmol/liter) were 93.1% (95% confidence interval = 88.4-97.8) and 92.3% (95% confidence interval = 87-97.6) in the No Meds group. The specificity of a cortisol lower than 1.4 mu g/dl (38 nmol/l) at 15 min after CRH was significantly higher in patients taking only one medication vs. those on multidrug treatment (P < 0.05). Conclusions: Medications commonly prescribed in hypercortisolemic patients undergoing Dex-CRH testing may contribute to the variable diagnostic accuracy of this test. Prospective studies to address this issue are needed. (J Clin Endocrinol Metab 94: 4851-4859, 2009)