RESUMO
PURPOSE: To investigate the accuracy of cutoff values of the morning serum cortisol (MSC) using the cortisol stimulus test (CST) insulin tolerance test (ITT) and 250 mcg short Synacthen test (SST) as the reference standard tests, to better define its clinical role as a tool in the diagnostic investigation of adrenal insufficiency (AI) AI. METHODS: An observational study was conducted with a retrospective analysis of MSC in adult patients who had been submitted to a CST to investigate AI between January 2014 and December 2020. The normal cortisol response (NR) to stimulation was defined based on the cortisol assay. RESULTS: 371 patients underwent CST for suspected AI, 121/371 patients (32.6%) were diagnosed with AI. ROC curve analysis showed an area under the curve (AUC) for MSC of 0.75 (95% CI 0.69 - 0.80). The best MSC cutoff values to confirm AI were < 3.65, < 2.35 and < 1.5 mcg/dL with specificity of 98%, 99%, and 100%, respectively. MSC > 12.35, > 14.2 and > 14.5 mcg/dL had sensitivity of 98%, 99%, and 100%, respectively, being the best cutoff values to exclude AI. Almost 25% of patients undergoing CST for possible AI had MSC values between < 3.65 mcg/dL (6.7% of patients) and > 12.35 mcg/dL (17.5% of patients), making the formal CST testing unnecessary if we consider these cutoff values. CONCLUSION: With the most modern cortisol assays, MSC could be used as a diagnostic tool, with high accuracy to confirm or exclude AI, avoiding unnecessary CST; thus, reducing expenses and safety risks during AI investigation.