Posttraumatic stress disorder, suicide, and unintended overdose death in later life: A national cohort study of veterans aged 50 and older.

BACKGROUND/

OBJECTIVES

Although studies have shown posttraumatic stress disorder (PTSD) associated with risk of suicide, the relationship in later life, especially for overdose death, remains unclear. Thus, the aim of the current study was to determine associations between PTSD, suicide, and unintended overdose death in mid- to late-life.

METHODS

A nationwide cohort study integrating Department of Veterans Affairs' (VA) data, Centers for Medicare & Medicaid Services data, and national cause-specific mortality data. Participants were US veterans aged ≥50 years with PTSD diagnoses at baseline (2012-2013) and were propensity-matched 1:1 with patients without PTSD based on sociodemographics, Charlson Comorbidity Index, and neuropsychiatric disorders (N = 951,018). Information on suicide attempts and unintended death by overdose through December 31, 2017 was provided by the VA's National Suicide Prevention Applications Network (non-fatal attempts) and Mortality Data Repository (death).

RESULTS

Veterans with PTSD (N = 475,509) had increased risk of suicide attempt (Hazard Ratio [HR], 1.59; 95% CI, 1.54-1.65; p < 0.001), non-fatal attempt (HR, 1.74; 95% CI, 1.67-1.81; p < 0.001), drug overdose death overall (HR, 1.32; 95% CI, 1.22-1.42; p < 0.001), and suicide overdose death (HR, 1.44; 95% CI, 1.15-1.80; p = 0.002), even after adjusting for sociodemographics, Charlson comorbidity index, and neuropsychiatric disorders. We found increased risk for overdose death by narcotics (HR, 1.30; 95% CI, 1.15-1.46; p < 0.001), antiepileptic/sedative-hypnotics (HR, 1.29; 95% CI, 1.02-1.62; p = 0.032), and for other/unspecified drugs (HR, 1.35; 95% CI, 1.20-1.51; p < 0.001), the last category indicative of polydrug. Results remained robust when examined for unintentional, suicide, and undetermined intent for cause-specific death by other/unspecified drugs.

CONCLUSIONS

PTSD persists throughout mid- to late-life with considerable increased risk for non-fatal suicide attempts and suicide overdose death. These findings suggest the importance of drug-monitoring in preventing late-life suicide.