Abstracts

Rationale:Epilepsy and Traumatic brain injury (TBI) may have a bi-directional relationship. Whereas prior studies have demonstrated higher mortality among veterans with epilepsy (VWE), research examining the impact of TBI timing in relation to epilepsy onset on mortality is rare. The purpose of this study is to examine the relationship between mortality with epilepsy, and the roles of TBI timing and uncontrolled epilepsy in this association.

Methods: The study cohort was derived from the Veterans Health Administration (VHA) (FY 2002-2018) and Department of Defense (DoD) (FY 2002-2019) for Veterans who used at least 2 years of VHA and DoD care between 2002 and 2014. We used ICD9/10 codes and concomitant use of antiseizure medications (ASMs) to identify epilepsy; the epilepsy index date was the first concomitant ASM or seizure diagnosis date. TBI was identified using the VHA comprehensive TBI evaluation or/and ICD9/10 codes; the TBI index date was based on the first TBI documentation. The cohort consisted of 6 groups: 1) no TBI nor epilepsy [reference group], 2) TBI only, 3) epilepsy only, 4) TBI before epilepsy, 5) TBI 180 days after the epilepsy diagnosis, 6) TBI within a proximal period. Because historical TBI may be diagnosed during epilepsy evaluation, we identified 180 days after epilepsy index date as “proximal period.” We identified possible uncontrolled epilepsy each year for veterans having > three ASMs or with urgent/acute (emergency/hospital) care for seizures. We used Kaplan–Meier estimates and log-rank tests to compare cumulative mortality rates among groups and Cox proportional hazard models to compute hazard ratios (HR, 95% CI), adjusting for covariates (demographics and comorbidities).

Results:Among 1,055,873 veterans 27,438 (2.60%) were VWE. The mortality rate was 6.43% for VWE and 1.51% for others. Table 1 shows adjusted Hazard Ratios (HR) and 95% confidence intervals for mortality. Model 1 indicates that compared to the reference group, VWE had significantly higher HR. TBI within the proximal period had the highest HR followed by TBI after epilepsy, epilepsy only, TBI before epilepsy, and TBI only. Models 3 and 4 demonstrated that possible uncontrolled epilepsy based on acute/urgent seizure care was associated with significant higher mortality, while receipt of >3 ASM was not.

Conclusions: Consistent with our prior work, we found a high mortality rate among VWE. In addition, we showed that VWE who had TBI in proximity (180 days) to epilepsy diagnosis had a significantly higher mortality rate than those with earlier TBI. Providers often focus on post-traumatic epilepsy, but these findings raise concern about VWE who had a TBI in proximity or following epilepsy diagnosis. The increased mortality risk associated with hospitalization and ER visits for seizures highlights the importance of managing intractable epilepsy to mitigate mortality risk in VWE.: Funding: CDMRP Epilepsy Research Program W81XWH-18-1-0247.