ESPE Abstracts

The steroid constellation in affective disorders suggests a stress-related pathogenesis or response and may pave the way for novel therapeutic strategies. In adolescents, previous studies have identified hypothalamic-pituitary-adrenal (HPA) axis dysregulation with hypercortisolemia. However, evidence so far is limited to a subset of the steroid spectrum, predominantly focusing on glucocorticoids. In order to gain a more holistic understanding, we analyzed urine metabolites of adrenal, gonadal and neuroendocrine steroids using gas chromatography-mass spectrometry (GC-MS). In this observational cross-sectional trial, the urinary excretion rates (µg/24h) of 39 steroid metabolites were measured in 75 adolescent psychiatric patients (64 females, 15.3 ± 1.3 years) with at least mild depressive symptoms (Beck Depression Inventory II total score > 13) and 75 controls (63 females, 15.6 ± 1.3 years) matched for age, sex, and pubertal status. Rank-based analyses of variance, adjusted for age, sex, and body mass index, revealed a distinct upregulation of the HPA-axis in patients. This was characterized by significantly increased excretion rates (µg/24h) for the sum of corticosterone metabolites (patients: media n = 608.4, interquartile range (IQR) 342.4-1208.2; controls: media n = 321.1, IQR 243.9-443.8)) and for specific progesterone and glucocorticoid metabolites. Moreover, increased excretion rates were observed for the sum of dehydroepiandrosterone (DHEA) metabolites (patients: media n = 1253.8, IQR 569.8-2796.2; controls: media n = 519.5, IQR 254.0-1028.7), 11-deoxygenated, and 11-oxygenated adrenal androgen metabolites (overall androgen metabolites: patients: media n = 6721.0, IQR 4185.6-9395.8; controls: media n = 3680.4, IQR 2510.8-5419.0). In contrast, estradiol excretion rates were lower in adolescents with depressive symptoms compared to controls (patients: media n = 4.0, IQR 2.9-5.8; controls: media n = 5.8, IQR 4.3-7.7). Although neuroactive steroids such as allopregnanolone, pregnanolone, and 3α-androstanediol showed nominally increased excretion rates, these differences were non-significant. Furthermore, enzyme activities were estimated by ratios (precursor/product) of steroid metabolites. These ratios were then utilized to discriminate between patients and controls. Multivariate unsupervised machine-based learning with random-forest classifiers yielded satisfactory differentiating performance utilizing 20 different ratios (area under the curve AUC = 0.829, 95%-CI 0.777-0.904). The ratio of tetrahydro-11-deoxycorticosterone (TH-DOC) to corticosterone metabolite excretion rates (as an indicator for 11β-hydroxylase activity) emerged as one of the most effective single measures for discrimination (AUC = 0.800, 95%-CI 0.702-0.882). In conclusion, the urinary steroid metabolome indicates substantial alterations in HPA-axis functioning in adolescents with depressive symptoms. Future studies should examine the potential of the urinary steroid metabolome as a tool to identify persons at risk for recurrent disease or to guide personalized therapeutic approaches.