Incidence of poststroke seizures: A meta-analysis.

BACKGROUND: Late-onset poststroke seizure is highly associated with epilepsy but the incidence of it is variable due to different definition and follow-up time. We conducted a meta-analysis to determine the incidence of poststroke seizure and analyze the discrepancies among published studies.

METHODS: We searched the literature for relevant articles published in English between January 1, 1990, and December 31, 2014, by using the following search terms: "seizures after stroke," "poststroke seizures," "epilepsy after stroke," "poststroke epilepsy," "incidence," and "follow-up." Reference lists of the relevant articles were reviewed to identify eligible studies not captured by these terms.

RESULTS: Seven relevant cohort studies were identified and analyzed. Incidence density was defined as the number of episodes per 100 person-years. Using the definition of late-onset poststroke seizure as seizure occurring 14 days after a stroke, the incidence density was 1.12 (0.95-1.32) per 100 person-years. For poststroke seizures occurring 7 days after the stroke, the incidence density increased to 3.22 (2.94-3.52) per 100 person-years.

CONCLUSION: Our results are relevant to the epidemiology of late-onset poststroke seizure, with 14 days being the ideal cutoff time point. The pooled incidence density of late-onset poststroke seizure was 1.12 per 100 person-years in 4 cohort studies.