Prevalence and predictors of seizure clusters: A prospective observational study of adult patients with epilepsy

Objective: The purpose of this prospective observational study was to describe the prevalence and adverse outcomes associated with seizure clusters (defined as >= 2 seizures in a 6-hour period) in a large sample of adult patients with a range of epilepsy severities and to identify clinical characteristics predictive of clustering. Methods: Patients maintained a seizure diary and were contacted monthly to verify compliance and data accuracy. Logistic regression models were utilized to test associations between individual patient demographic/clinical characteristics and seizure clustering. Fisher's exact test was utilized to test associations between rescue medication use and adverse seizure-related outcomes. Results: A total of 300 patientswere followed prospectively for one year; 247 patients qualified for final analysis. Six-hour seizure clusters occurred in 45.8% of patients with active epilepsy at enrollment, including 62.7% of thosewith prior day-clusters and 30.0% of thosewithout prior day-clusters. The odds of clusteringweremarkedly greater among patientswho reported a higher seizure frequency (>4 seizures per year vs. 1-4 seizures per year) (adjusted odds ratio (OR): 8.9; 95% confidence interval (CI): 3.2-24.6; p < 0.0001) and among patientswith prior day-clusters (adjusted OR: 11.0; 95% CI: 1.2-104.2; p=0.036). Rescue medication use was associated with significantly fewer injuries and emergency department visits, but rescue medication was underutilized. Conclusions: Seizure clusters are common, occurring in nearly half of adult patientswith active epilepsy followed prospectively over one year, and are more frequent in those with higher seizure frequencies and prior day-clusters. Although underutilized, rescue medication was associated with fewer injuries and emergency department visit. (c) 2018 Elsevier Inc. All rights reserved.