Authors :
Presenting Author: Marissa Kellogg, MD, MPH, FAES – Portland VA Healthcare System & Oregon Health Sciences University (OHSU)
Megan Amuan, MA – Salt Lake City VAMC; Zulfi Haneef, MD – Houston VA HCS; Stephan Eisenschenk, MD – Gainesville VAMC; Glenn Graham, MD, PhD – VA Central Office; Neurology; Linda Williams, MD – Indianapolis VAMC; Mary Jo Pugh, PhD, MPH, RN – Salt Lake City VA
Rationale:
Cerebrovascular events/accidents (CVA; e.g., ischemic/hemorrhagic stroke) are a common etiology of epilepsy in adults. The literature indicates epilepsy can also be a risk factor for CVA, suggesting a bidirectional relationship between epilepsy and CVA. This study uses data from the Long-term Impact of Military-Relevant Brain Injury Consortium (LIMBIC-CENC: a study designed to examine long-term outcomes of mild TBI) to explore the relationship between new-onset (incident) epilepsy and subsequent CVA in post-9/11 Veterans with no prior CVA.
Methods:
This retrospective cohort study used administrative data to identify participants with care documented in US Department of Defense (three years of care) and Department of Veterans Affairs (10/1/99 to 9/30/19), with follow up data through 9/30/20. Inclusion criteria included: incident epilepsy, as defined by a validated algorithm using diagnostic codes and pharmacy records. CVA was defined using an algorithm adopted from Stewart et al (JAMA Neuro 2022). Exclusion criteria included: CVA before epilepsy, seizure history not meeting epilepsy criteria, and TBI for which severity could not be classified. Fine-Gray hazard models (death as competing event) were used to identify characteristics of veterans most likely to have CVA after epilepsy diagnosis.
Results:
Among 2,530,847 veterans, 44,823 met epilepsy criteria; after exclusions, 35,871 were included. Participants were predominantly young (median age 31 yrs [IQR 25-40]) at epilepsy index date, and male (77%, n=27,634). 4,183 developed CVA endpoints during the study period. 2,558 died.
Over half of CVAs occurred within one year of epilepsy onset (median 12.6 mos [IQR 0.5-48.6]). Veterans with CVA were significantly older (median age 37 yrs [IQR 27-47]) than veterans without CVA and death (median age 30 yrs [IQR 25-38]) (p< 0.0001 ). Compared to veterans without a history of TBI, diagnoses of moderate/severe TBI (hazard ratio [HR]1.49; 95% CI, 1.34-1.65) and penetrating TBI (HR 1.97; 95% CI, 1.73-2.23) but not mild TBI (HR 0.95; 95% CI, 0.88-1.03) were associated with CVA in adjusted models. Other diagnoses significantly associated with increased risk of CVA were anoxic brain injury