Baseline morning cortisol level as a predictor of pituitary-adrenal reserve: a comparison across three assays.

CONTEXT: The short ACTH stimulation test (250 μg) is the dynamic test most frequently used to assess adrenal function. It is possible that a single basal cortisol could be used to predict the dynamic response, but research has been hampered by the use of different assays and thresholds. OBJECTIVE: To propose a morning baseline cortisol criterion of three of the most commonly used modern cortisol immunoassays - Advia Centaur (Siemens), Architect (Abbott) and the Roche Modular System (Roche) - that could predict adrenal sufficiency. DESIGN: Observational, retrospective cross-sectional study at two centres. PATIENTS AND MEASUREMENTS: Retrospective analysis of the results of 1019 Short Synacthen tests (SSTs) with the Advia Centaur, 449 SSTs with the Architect and 2050 SSTs with the Roche Modular System assay. Serum cortisol levels were measured prior to injection of 250 μg Synacthen and after 30 min. Overall, we were able to collate data from a total of 3518 SSTs in 3571 patients. RESULTS: Using receiver-operator curve analysis, baseline cortisol levels for predicting passing the SST with 100% specificity were 358 nmol/l for Siemens, 336 nmol/l for Abbott and 506 nmol/l for Roche. Utilizing these criteria, 589, 158 and 578 SSTs, respectively, for Siemens, Abbott and Roche immunoassays could have been avoided. CONCLUSIONS: We have defined assay-specific morning cortisol levels that are able to predict the integrity of the hypothalamo-pituitary-adrenal axis. We propose that this represents a valid tool for the initial assessment of adrenal function and has the potential to obviate the need for dynamic testing in a significant number of patients.