• suicide;
  • epidemiology;
  • psychiatric disorders;
  • antiepileptic drugs;
  • geriatrics


To examine the association between antiepileptic drug (AED) receipt and suicide-related behavior (SRB) in older veterans.


Retrospective database analysis.


Veterans Health Administration (VHA) inpatient and outpatient care.


Veterans aged 65 and older in 2004 to 2006.


SRB was identified using International Classification of Diseases, Ninth Revision, Clinical Modification codes, and new AED monotherapy was identified using the VHA product variable in pharmacy data. Comorbid conditions and medications were also identified as potential confounders using previously validated algorithms. Cox proportional hazards models controlling for the propensity to receive AEDs examined the association between any AED exposure, specific AEDs, and time to SRB.


Within the eligible sample of 2.15 million individuals, 332 cases of SRB were found. Overall, 98% of participants were male, and 67% were non-Hispanic white. Affective disorders and severe psychiatric conditions were strongly associated with SRB and were included in the propensity score. AED exposure displayed a significant association with SRB (odds ratio = 4.10, 95% confidence interval (CI) = 3.85–6.63) after adjusting for propensity to receive AEDs. Stratified analyses found similar results for those with (hazard ratio (HR) = 4.00, 95% CI = 2.9–5.5) and without (HR = 4.57, 95% CI = 1.15–18.20) mental health comorbidities. Gabapentin, phenytoin, lamotrigine, levetiracetam, topiramate, and valproate were significantly associated with SRB.


Exposure to five common AEDs was associated with SRB in older VHA beneficiaries. Given the strong associations between psychiatric comorbidity and SRB, clinicians treating elderly adults should weigh this potential adverse effect into their consideration for treatment of those receiving AEDs. Particular attention should be given to depression and suicidality screening in people prescribed AEDs.