Performance of DHEAS as a Screening Test for Autonomous Cortisol Secretion in Adrenal Incidentalomas: A Prospective Study.

Abstract

Autonomous cortisol secretion (ACS) affects up to 30% of patients with adrenal incidentalomas (AIs). The current guidelines for ACS diagnosis are not decisive. A lower dehydroepiandrosterone sulfate (DHEAS) level is a potential biomarker, but the evidence is conflicting. This prospective study aimed to evaluate and validate the ACS screening and diagnostic accuracy of DHEAS. Recruited patients with AI were screened for adrenal medullary and cortisol hypersecretion. The diagnosis of ACS was based on a serum cortisol level ≥ 50 nmol/L following a 1-mg dexamethasone suppression test (DST) and a low-dose DST. Age- and sex-specific DHEAS ratios were also calculated. In the development cohort (45 ACS and 242 non-ACS patients), the areas under the receiver operator characteristic curves (AUCs) of DHEAS and the DHEAS ratio were 0.869 (95% CI 0.824-0.906) and 0.799 (95% CI 0.748-0.844), respectively. The optimal DHEAS cutoff for diagnosing ACS was 60 μg/dL, with a sensitivity of 75.6% (95% CI 60.5-87.1) and a specificity of 81.4% (95% CI 76.4-86.5). The midnight serum cortisol level had moderate diagnostic accuracy [AUC 0.875 (95% CI 0.831-0.911)]. Suppressed adrenocorticotropic hormone (≤2.2 pmol/L) had a lower sensitivity (55.6%), and the 24-hour urinary free cortisol lacked sensitivity and specificity [AUC 0.633 (95% CI 0.603-0.721)]. In the validation cohort (14 ACS and 45 non-ACS patients), the sensitivity and specificity of the optimized DHEAS cutoff were 71.4% (95% CI 41.9-91.6) and 82.2% (95% CI 68.0-92.0), respectively. A single basal measurement of DHEAS is valuable for identifying ACS. Because of its stability and ease of use, the DHEAS level could be used as an ACS screening test.