Abstracts

Compared to young adulthood, the incidence of epilepsy is thought to increase substantially after age 60, and may be further elevated among males and African-Americans. The reason for this rise is uncertain. Although elevated risk may be due in part to stroke, dementia, and other risk factors common in older adults, it has been postulated that aging itself may be associated with increased risk of epilepsy. We therefore sought to determine the incidence of epilepsy in a relatively healthy, racially-diverse, elderly cohort. 1919 community-dwelling subjects (age 70+) followed in a prospective longitudinal aging study were screened for seizures. For all subjects who screened positive and a sample of subjects who screened negative (n=100), we determined which subjects had epilepsy with onset after age 45 by means of standardized telephone interviews and review of aging study records. For screen-positive subjects, we sought hospital medical records to verify self-reported age of onset. Age-specific and cumulative incidence of epilepsy were estimated using reconstructed cohort methods and the Kaplan-Meier procedure. We assessed sex and race as risk factors using Cox proportional hazards regression. 45 subjects screened positive for seizures, and 1874 screened negative. Of the 45 screen-positive subjects, 21 had epilepsy with onset after age 45, 21 had likely seizures or epilepsy with onset before age 45, 2 experienced single provoked seizures, and 1 subject had possible epilepsy. None of the 100 screen-negative subjects interviewed had epilepsy. Age-specific incidence of epilepsy was generally low but increased with age. Incidence (per 100,000 person-years with 95%CI) was 10.6 [2.2, 30.9] between 45-59 years, 25.8 [10.4, 38.1] between 60-74 years, and 101.1 [41.4, 160.8] between 75-89 years. No difference was observed between males and females, but African-American race was associated with greater cumulative incidence of epilepsy (p=0.048) and was a significant risk factor in the regression model (hazard ratio with 95%CI: 2.63 [1.07, 6.37]). The incidence of epilepsy was low in this relatively healthy elderly cohort. This study was unique in that we evaluated race as a covariate and observed elevated risk among African-Americans. Future analyses are planned to estimate the incidence among elderly individuals who do not have competing risk factors for epilepsy such as stroke and dementia, and to determine whether our race-specific findings are modified or confounded by cardiovascular risk factors. (Supported by NIH grant P01 AG03949 and an unrestricted grant from GlaxoSmithKline.)