From allostasis to addiction: Cortisol dysregulation, daily stress and relations with 9‐year alcohol and substance use disorder severity in community adults

Abstract Background and aims Diurnal cortisol dysregulation and daily stressor severity may jointly forecast long‐term alcohol use disorder (AUD) and substance use disorder (SUD) severity, yet most evidence is cross‐sectional and relies on single cortisol indices. This study examined whether three cortisol indices nonlinearly interacted with daily stressor severity to predict 9‐year AUD and SUD severity. Design Longitudinal observational study with a 9‐year follow‐up. Analyses were not pre‐registered. Setting Community‐based sample in the United States of America. Participants 1617 community‐dwelling adults aged 33–83 years at baseline. Measurements Ecological momentary assessments of salivary cortisol and daily stressor severity were collected at baseline. AUD and SUD severity were self‐reported at baseline and 9‐year follow‐up. Cortisol indices included diurnal cortisol slope, cortisol awakening response and area under the curve (AUC). Findings Parametric hurdle Poisson models yielded null main effects and interactions (all P > 0.050). Generalized additive models with negative binomial family revealed three statistically significant nonlinear interactions: diurnal cortisol slope × stress predicting AUD severity [effective degrees of freedom (edf) = 2.09, χ 2 = 12.56, P = 0.005], cortisol awakening response × stress predicting SUD severity (edf = 2.88, χ 2 = 11.60, P = 0.019) and AUC × stress predicting AUD severity (edf = 4.08, χ 2 = 16.34, P = 0.006). Adjusting for age, sex, education, income and race nullified all three interactions; however, AUC × stress interaction → AUD remained statistically significant after controlling for anxiety and depression. Conclusions Nonlinear interactions between cortisol indices and daily stressor severity appear to predict 9‐year alcohol use disorder and substance use disorder severity, with cortisol area under the curve emerging as the most robust indicator. Findings are predictive rather than causal. Cognitive‐behaviorally based prevention programs targeting stress and cortisol regulation may mitigate substance use risk.