Background: Steroid hormone dysregulations have frequently been implicated in posttraumatic stress disorder (PTSD) pathogenesis. However, the translation into naturalistic clinical settings as markers of symptomatology and treatment success remains complex. Particularly, there is little longitudinal data on steroid secretion over the course of interventions. Objective: This study examined the potential of long-term steroid hormone secretion assessed in hair as diagnostic and intervention-related biomarkers among medicated, multimorbid inpatients with PTSD. Method: As part of a secondary analysis of a randomised controlled trial, 54 female inpatients with a primary diagnosis of PTSD received standardised treatment and provided hair samples at pre-treatment, post-treatment, and 3-month follow-up. Cortisol, cortisone, and dehydroepiandrosterone (DHEA) were determined, alongside clinical assessments. Results: Cross-sectional results showed a negative association of pre-treatment DHEA with anxiety symptoms and a trend-level association with lifetime trauma exposure. While inpatients improved in PTSD symptomatology during treatment, neither pre-treatment steroids, nor treatment-induced steroid changes predicted PTSD symptoms at post-treatment or 3-month follow-up. Conclusion: The study highlights the challenges of establishing biomarkers in naturalistic clinical populations. While the association of attenuated DHEA with anxiety symptoms warrants further exploration, our data points towards the potential necessity of patient sub-sample selection to understand, and in the long run clinically target, the endocrine mechanisms in PTSD.