Beyond a simple marker of stress: A multiverse analysis of biopsychosocial associations with hair cortisol and cortisone

BACKGROUND: Hair concentrations of cortisol and cortisone are increasingly used as potential biomarkers of chronic stress. However, their validity as indicators of psychosocial stress remains debated, as prior research shows variable associations with stress-related psychological and behavioral factors. This study systematically investigates a broad set of biopsychosocial correlates of hair glucocorticoids in young adults, leveraging a multiverse framework to evaluate the robustness of associations often reported as variable in prior work. METHODS: Using data from the Zurich Project on the Social Development from Childhood to Adulthood, a large community-based age cohort, we employed a multiverse analysis to explore relationships between more than 60 biological, psychological, and contextual variables and three hair-based outcomes (cortisol, cortisone, and a combined score of both) at ages 20 (n = 1,002) and 24 (n = 887). Specification curve analyses were performed on the full sample and separately by sex, systematically varying model choices (i.e., inclusion/exclusion of covariates). RESULTS: Hair characteristics (i.e., color, washing intensity, chemical treatment), sampling season, body mass index, and, among women, hormonal contraception, showed strong associations with glucocorticoids and were included as covariates. Sweating intensity was associated with glucocorticoids primarily in females, but not in males. Hair concentrations of drugs (i.e., Δ9-tetrahydrocannabinol, cocaine, 3,4-methylenedioxymethamphetamine [MDMA, "Ecstasy"], paracetamol) and the hormone testosterone were consistently linked to steroid outcomes. Several psychosocial factors were associated with hair glucocorticoids: migration background, perceived stress, symptoms of anxiety/depression, sports activity, and social exclusion. These factors exhibited stronger correlations with cortisol, while media use, aggression, and effective coping were more closely associated with cortisone. The combined steroid score mainly reflected associations detected for cortisol, but with weaker effect sizes. Associations varied descriptively by age and sex, with more associations detected in males at age 20 and in females at age 24. CONCLUSION: In cohort studies, corticosteroids in hair can reflect specific facets of psychosocial stress when critical confounders and sex are considered.