OC07 - Morning cortisol and central adrenal insufficiency: new thresholds from low- dose ACTH test and second-generation assay

Background Central adrenal insufficiency (CAI) often presents with nonspecific symptoms, and its diagnosis relies on complex dynamic tests. Current guidelines recommend early morning serum cortisol as the initial screening tool: values <30 µg/L are diagnostic of CAI, values >150 µg/L exclude it, while intermediate values (30–150 µg/L) require dynamic testing. However, these thresholds are based on limited evidence and based on old low specific assay. Purpose To identify morning cortisol cut-offs capable of confirming CAI with high specificity (SP ≥95%) or excluding it with high sensitivity (SE ≥95%), in order to reduce reliance on dynamic tests without compromising diagnostic accuracy. Methods We retrospectively analysed 435 patients with hypothalamic-pituitary disorders who underwent the low-dose ACTH stimulation test to assess the hypothalamic-pituitary-adrenal axis. Based on the peak cortisol response, patients were classified as having CAI (≤180 µg/L; n=229) or normal adrenal function (>180 µg/L; n=206). Additional analyses were performed using stricter thresholds (≤136 µg/L and ≤127 µg/L), identifying 74 and 53 patients with CAI, respectively. ROC curves were used to determine optimal basal cortisol cut-offs. Results Using the 180 µg/L peak cortisol threshold, ROC analysis (AUC 0.776) identified a morning cortisol value of ≤102 µg/L as the optimal cut-off, providing the best balance between SE (66.9%) and SP (82.8%). A value ≤78 µg/L achieved SP 95.9% (SE 38.6%), while >143.2 µg/L excluded CAI with SE 95.2% (SP 30.8%). With more stringent thresholds (AUC 0.836 for ≤136 µg/L and 0.860 for ≤127 µg/L), the optimal morning cortisol cut-off was ≤87 µg/L in both cases. Cortisol levels predictive of CAI were 65 µg/L (SE 48.7%, SP 95.8%) and 60.9 µg/L (SE 54.7%, SP 96.3%), respectively. In both models, values >141 µg/L excluded CAI with SE ≥95% (SP ∼21%). Conclusions Basal morning cortisol remains a valuable tool in the diagnostic work- up of CAI. Values ≤60.9 µg/L and >141 µg/L reliably confirmed or excluded CAI with SP or SE ≥95%, respectively. Based on these findings from a large cohort and second-generation immunoassays, we propose an updated diagnostic algorithm that refines cortisol thresholds to streamline the diagnostic pathway and reduce the need for dynamic testing.