Diagnostic value of standard deviation score of log-transformed serum dehydroepiandrosterone sulfate in patients with hypothalamic-pituitary-adrenal axis insufficiency

Abstract

Serum dehydroepiandrosterone sulfate (DHEA-S) levels reflect the state of adrenocorticotropic hormone (ACTH) secretion. However, it is difficult to use serum DHEA-S to diagnose hypothalamic-pituitary-adrenal (HPA) axis insufficiency due to its non-normal and highly skewed distribution. In this study, we focused on HPA insufficiency caused by hypothalamic and/or pituitary dysfunction and evaluated the usefulness of the standard deviation score of log-transformed DHEA-S (ln DHEA-S SD score), which was calculated from the established age- and sex-specific reference values. We retrospectively reviewed the medical records of 94 patients suspected of having HPA insufficiency, in whom serum DHEA-S measurement and the rapid ACTH stimulation test were performed, and included 65 patients who met our criteria in this study. The ln DHEA-S SD scores were distributed more normally than measured DHEA-S levels and were significantly higher in patients with a peak cortisol level ≥18 μg/dL than in those below this value, suggesting that this score is a legitimate and strong indicator of adrenocortical function. The optimal cut-off value for impaired HPA function was –0.853, with a sensitivity of 70.3% and a specificity of 100%. Among the 37 patients whose peak cortisol levels were below 18 μg/dL, 11 patients with ln DHEA-S scores ≥–0.853 exhibited significantly higher basal ACTH and basal and peak cortisol levels than the 26 patients with scores <–0.853. Thus, this score plays a supportive role in evaluating HPA axis function, particularly in patients with borderline cortisol responses to ACTH.