Predictors of Current DSM-5 PTSD Diagnosis and Symptom Severity Among Deployed Veterans: Significance of Predisposition, Stress Exposure, and Genetics.

Background: Previously we reported a genetic risk score significantly improved PTSD prediction among a trauma-exposed civilian population. In the current study, we sought to assess this prediction among a trauma-exposed military population.

Methods: We examined current PTSD diagnosis and PTSD symptom severity among a random sample of 1042 community-based US military veterans. Main effects and interaction effects were assessed for PTSD genetic risk by trauma exposure using cross-product terms for PTSD x trauma exposures, including combat, lifetime trauma, and adverse childhood exposures. The PTSD risk variants studied were within genetic loci previously associated with PTSD, including , , , and genetic variants, which were used to calculate a total PTSD genetic risk score (range=0-8, mean=3.6, SD=1.4).

Results: Based on DSM-5 PTSD criteria, 7.1% of veterans (95% CI=5.6-8.8) met criteria for current PTSD. The PTSD genetic risk count was significantly higher among PTSD cases vs non-cases (3.92 vs 3.55, p=0.027). Since the PTSD genetic risk score was not significant in the PTSD diagnosis model, we assessed this association using PTSD symptom severity. Because these symptom data were skewed (mean=9.54, SD=12.71, range=0-76), we used negative binomial regression to assess this outcome. This symptom model included a PTSD genetic risk score, demographic factors, trauma exposures, current insomnia, current depression, concussion history, and attention-deficit disorder, expressed as incident rate ratios (IRR), which is an estimate of one-unit increase in PTSD severity, given other variables are held constant. Variables in the final model included age and sex (both p<0.001), PTSD genetic risk (IRR=1.02, p=0.028), warzone tours (IRR=0.94, p=0.003), childhood abuse (IRR=1.50, p<0.0001), current depression (IRR=1.89, p<0.0001), current insomnia (IRR=2.58, p<0.0001), low social support (IRR=1.19, p<0.0001), attention-deficit disorder (IRR=1.51, p<0.0001), agreeable personality (IRR=0.77, p<0.0001), and concussion (IRR=1.38, p<0.0001). Significant interactions were detected for combat and lifetime trauma exposure by PTSD genetic risk (both p<0.0001), suggesting that the impact of trauma exposures on PTSD severity was lower when the PTSD genetic risk was higher.

Conclusion: Both warzone and non-warzone factors predicted current PTSD symptoms among veterans, including a PTSD genetic risk score. Interaction effects were detected for combat exposure and lifetime trauma by genetic risk score for PTSD symptoms, suggesting that PTSD symptom manifestation was more dependent on PTSD risk variants than the level of trauma or combat exposure. This suggests that controlling for other factors, the absence of genetic risk variants may confer PTSD resilience. Further research is planned.