Do Seizures in Older Adults Predict Stroke Risk?

Abstract & Commentary

Sources: Cleary P, et al. Late-onset seizures as a predictor of subsequent stroke. Lancet. 2004;363:1184-86; Sudlow, CM. Epilepsy and stroke (commentary). Lancet. 2004;363: 1175-1176.

Using data from the UK General Practice Research Database (GPRD), Cleary and associates evaluated the hypothesis that late-onset (60-year-old subjects and older) seizures increase the risk of subsequent stroke. The 4709 subjects (with no prior history of cerebrovascular disease, other brain injury, brain tumor, alcohol, or drug abuse, or dementia) were compared to 4709 age- and gender-matched seizure-free individuals from the same database. There was a highly significant (P < 0.0001) difference in subsequent stroke-free survival in seizure patients vs controls. By Cox’s model, the relative risk of stroke was 2.89 (95% confidence interval, 2.45-3.41).

In an accompanying editorial, Sudlow suggests that "it seems reasonable for general practitioners, general physicians, geriatricians, and neurologists . . . to assess their patients’ vascular risk factors, and to consider treatment to prevent stroke (and other vascular disease)."


There are limitations to the present study, which go to the heart of the association of new-onset seizures and subsequent stroke risk. The main question is whether seizures are a true marker for stroke or an epiphenomenon. Seizures in older adults are more likely to be symptomatic than in younger individuals, with idiopathic and cryptogenic epilepsy being more common in the latter. We suspect that the GPRD does not capture important data to solidify the claim that seizures are an independent predictor of stroke risk. First, while a prior history of cerebrovascular disease was an exclusion criterion for the study group, there is no indication whether this exclusion was rigorously defined by neuroimaging in addition to clinical history and examination. Second, it is unlikely that the evaluation of the patients’ new-onset seizures included diffusion-weighted imaging (to exclude ischemic stroke) or echo gradient MRI sequences (to exclude subtle focal cortical hemorrhage). Without these data, one is left to wonder whether many late-onset seizures are due to existing cerebrovascular disease. If so the conclusion of the study is not particularly novel: previous stroke predicts subsequent stroke.

Even if seizures in older individuals are not an independent marker of cerebrovascular disease, we agree with the investigators that these seizures should be a red flag to consider treatment to modify stroke risk (be it primary or secondary prevention). — Andy Dean

Dr. Dean, Assistant Professor of Neurology and Neuroscience; Director of the Epilepsy Monitoring Unit, Department of Neurology, New York Presbyterian Hospital Cornell Campus, is Assistant Editor of Neurology Alert.