Exercise improves stress perception and sleep quality and reduces repetitive negative thinking in patients with various mental disorders. However, it is unclear whether changes in these processes mediate treatment effects on psychopathology in a transdiagnostic sample.

Physically inactive adult outpatients with depressive disorders, agoraphobia, panic disorder, post-traumatic stress disorder, and/or nonorganic primary insomnia were randomly allocated to ImPuls – a 6-month transdiagnostic group exercise intervention – plus treatment-as-usual (n = 198), or to a treatment-as-usual alone control group (n = 201) at 10 study sites between March 2021 and May 2022. The primary outcome was global symptom severity; perceived stress, repetitive negative thinking, and sleep quality were included as mediators. All variables were assessed at baseline, 6 months, and 12 months using validated rating scales. As a secondary analysis of an RCT, intention-to-treat analyses were performed using structural equation modeling to test whether changes in stress perception, repetitive negative thinking, and sleep quality mediate treatment effects on changes in global symptom severity in two path models (from baseline to 6 and 12 months, respectively).

Treatment effects on global symptom severity were fully mediated by changes in perceived stress (6 months: β = −0.99, p = .024; 12 months: β = −1.28, p = .014) and repetitive negative thinking (6 months: β = −1.34, p = .004; 12 months: β = −0.94, p = .024).

Our results suggest that changes in perceived stress and repetitive negative thinking may be key transdiagnostic mechanisms underlying the treatment effect of exercise on global symptom severity.