Associations of Late-Onset Epilepsy With Myocardial Infarction and Nonstroke Vascular Death

Cerebrovascular disease is associated with increased risk of late-onset epilepsy (LOE). Because vascular disease is often systemic, coexisting in multiple vascular beds, LOE may be a marker of increased systemic vascular risk outside the brain. We sought to determine whether stroke-free middle-aged and older adults with incident myocardial infarction (MI) subsequently have increased risk of LOE and whether those with incident LOE subsequently have increased risks of incident MI and nonstroke vascular death.

The Northern Manhattan Study (NOMAS) is a population-based cohort study of participants aged 40 years and older enrolled from 1993 to 2008, with follow-up through May 2023 for the present analysis. Participants free of a history of stroke, MI, or epilepsy at enrollment were followed prospectively for up to 30 years (mean 14 years). We used Cox proportional hazards regression with censoring at incident stroke to assess the associations of (1) incident MI with subsequent incident LOE; (2) incident LOE with subsequent incident MI; and (3) incident LOE with nonstroke cardiovascular death, adjusting for demographics, health behaviors, and comorbid diagnoses.

The mean (SD) age of 3,174 participants at enrollment was 69.1 (10.4) years, and 63.5% were women. We identified 296 participants (9.3%) who developed incident MI, 120 (3.8%) who developed incident LOE, and 794 (25.0%) who died of nonstroke vascular causes. Incident LOE occurred at a rate of 7.02 cases per 1,000 person-years (PYs) after incident MI compared with 2.49 per 1,000 PYs without MI (adjusted hazard ratio [aHR] 2.12; 95% CI 1.06-4.25; p = 0.035). Incident MI occurred at a rate of 17.68 cases per 1,000 PYs after incident LOE compared with 6.46 per 1,000 PYs without LOE (aHR 1.99; 95% CI 0.98-4.05; p = 0.059). Nonstroke vascular death occurred at a rate of 99.24 deaths per 1,000 PYs after incident LOE compared with 16.29 per 1,000 PYs without LOE (aHR 2.82; 96% CI 2.09-3.80; p < 0.001). Sensitivity analyses yielded similar results.

The bidirectional associations we observed suggest that LOE might be a marker of increased systemic vascular risk. This warrants further study in additional populations.