Physical Exercise as Treatment for PTSD: A Systematic Review and Meta-Analysis

Post-traumatic stress disorder (PTSD) is a cluster of physical and psychiatric symptoms following military or civilian trauma. The effect of exercise on PTSD symptoms has previously been investigated in several studies. However, it has not been fully determined what type of exercise most impacts PTSD symptoms. The aim of the present study was to systematically review the effects of different types of exercise on PTSD symptom severity and symptoms of coexisting conditions in adults.

Electronic searches were conducted in the databases PubMed, APA PsycInfo, and SportDiscus, from database inception up until February 1, 2021. Inclusion criteria were randomized controlled trials published in English, participants having a PTSD diagnosis or clinically relevant symptoms, and participants randomly allocated to either a non-exercising control group or an exercise group. Data concerning the number of participants, age, exercise type and duration, PTSD symptom severity (primary outcome), and symptoms of coexisting conditions (secondary outcomes) were extracted. The subgroup analysis included high or low training dose, military trauma versus non-military trauma, the type of intervention (yoga versus other exercise), active or passive control condition, group training versus individual exercise, and study quality. The study quality and risk of bias were assessed using grading of recommendation assessment, development and evaluation (GRADE) guidelines. A meta-analysis was performed with a mixed-effects model and restricted maximum likelihood as model estimator, and effect size was calculated as the standardized difference in mean and 95% CI.

Eleven studies were included in the present review. Results showed a main random effect of exercise intervention (0.46; 95% CI: 0.18 to 0.74) and a borderline significant interaction between more voluminous (>20 hours in total) and less voluminous (≤20 hours in total) exercise interventions (P = .07). No significant findings from the subgroup analysis were reported. The secondary outcome analysis showed a small but significant effect of exercise on depressive symptoms (0.20, 95% CI: 0.01 to 0.38), and a larger effect on sleep (0.51, 95% CI: 0.29 to 0.73). For substance use (alcohol and drugs combined) and quality of life, we found significant effects of 0.52 (95% CI: 0.06 to 0.98) and 0.51 (95% CI: 0.34 to 0.69), respectively. No significant effect was found for anxiety (0.18, 95% CI: -0.15 to 0.51), and no sign of publication bias was found.

Exercise can be an effective addition to PTSD treatment, and greater amounts of exercise may provide more benefits. However, as there were no differences found between exercise type, possibly due to the inclusion of a low number of studies using different methodologies, further research should aim to investigate the optimal type, dose, and duration of activity that are most beneficial to persons with PTSD.