New-onset Vascular Events in Older People with Epilepsy: A Cohort Study of the Canadian Longitudinal Study on Aging (CLSA)
Abstract number :
1.397
Submission category :
16. Epidemiology
Year :
2023
Submission ID :
218
Source :
www.aesnet.org
Presentation date :
12/2/2023 12:00:00 AM
Published date :
Authors :
Presenting Author: Jimmy Li, MD – Centre hospitalier univesitaire de Sherbrooke (CHUS)
Nathan Shlobin, BA – Northwestern University Feinberg School of Medicine; Roland Thijs, MD PhD – Stichting Epilepsie Instellingen Nederland (SEIN); Charles Deacon, MD – Centre hospitalier universitaire de Sherbrooke (CHUS); Mark Keezer, MDCM PhD – Centre hospitalier universitaire de Montréal (CHUM)
Rationale: There is some evidence that epilepsy represents a risk factor for cerebral and cardiac ischemic events, although the mechanisms behind this association are poorly understood. We aimed to estimate the odds of new-onset vascular events (VEs) over three years in older people with versus without epilepsy. We also aimed to investigate the potential mediation effects of enzyme-inducing antiseizure medication (EIASM) use and of the Framingham cardiovascular risk score on the association between epilepsy and new-onset VEs.
Methods: This study focused on a cohort of 30,097 Canadians aged 45-85 years from the Canadian Longitudinal Study on Aging. A new-onset VE was defined as a self-reported stroke, TIA, or myocardial infarction occurring between baseline and the three-year follow-up. We fitted two weighted quasibinomial logistic models for new-onset VEs as a function of lifetime history of epilepsy, one (basic) adjusting for age, sex, and province, and the other (complete) also adjusting for history of stroke and myocardial infarction. With the covariates from the complete model, we then performed causal mediation analyses using a difference of coefficients approach and the Monte Carlo method. The proportion of the effect of epilepsy on new-onset VEs mediated by EIASM use and by the Framingham risk score was calculated. A directed acyclic graph (DAG, Figure 1) was used to determine which variables required adjustment in the logistic models to eliminate confounding and which variables to test for mediation. Missing data were handled using multiple imputations (200 iterations).
Results: Within our study cohort, 514 had a lifetime history of epilepsy, and 716 had new-onset VEs (194 strokes, 279 TIAs, and 299 myocardial infarctions). Mean age was 62.96 years, and 50.90% of individuals were of female sex. Table 1 provides a detailed description of the cohort, with and without epilepsy. The basic model generated ORs of 2.46 (1.71-3.55) for epilepsy, 1.06 (1.05-1.07) for age, 0.46 (0.40-0.54) for female sex, and 1.40 (1.05-1.87) for Manitoba versus Alberta (other provinces had non-significant ORs). The complete model generated ORs of 2.18 (1.52-3.14) for epilepsy, 1.05 (1.05-1.06) for age, 0.52 (0.45-0.60) for female sex, 1.46 (1.10-1.94) for Manitoba versus Alberta (other provinces had non-significant ORs), 4.22 (3.26-5.48) for history of stroke, and 2.75 (2.26-3.35) for history of myocardial infarction. EIASM use and the Framingham risk score mediated 16% (0.4%-38%) and 3.8% (1.3%-7.8%) of the effect of epilepsy on new-onset VEs, respectively.
Conclusions: Over three years, new-onset VEs were more common in PWE. Although this phenomenon could partially be attributed to EIASM use and vascular risk factors, other mechanisms probably underly the effect of lifetime epilepsy on new-onset VEs. Indeed, our findings provide evidence that there may be a direct effect of epileptic seizures on vascular atherosclerosis.
Funding: The authors received no grant from any funding agency in the public, commercial, or not-for-profit sectors.
Epidemiology