Abstracts

Rationale: Patients with epilepsy often experience seizures in clusters, also known as acute repetitive seizures. In this retrospective study, we examined the prevalence of physician-confirmed clustering and identified potential predictors. Methods: As part of the Columbia/Yale AED Database Project, we reviewed medical records of 2435 adult outpatients with epilepsy seen over a 12-year period. We included anyone with "seizure clusters" checked as yes after chart review, which included reviewing patient and physician reports. There was no specific definition of seizure clusters. To investigate potential factors associated with seizure clustering, we examined 82 different variables including patient demographics, epilepsy details, medical and psychiatric history, total number and types of AEDs taken, and various epilepsy risk factors. We used logistic regression to test the correlation between each individual variable and occurrence of seizure clustering. Results: Overall, 8.3% (203/2435) of all patients had at least one seizure cluster reported in the chart. Of the 203 patients who had seizure clusters, 54.2% (n = 110) had one or more rescue medication(s) prescribed. Being on benzodiazepine rescue medication(s) was significantly associated with reported clustering [p < 0.001; OR, 1.93 (1.45-2.57)]. Patients who had seizure cluster(s) also had shorter periods of seizure freedom (14.6 ± 22.5 months) compared to that of those who did not report seizure clusters (27.4 ± 42.2 months, p < 0.001). In univariate analysis, factors that have significant and positive associations with reported seizure clustering were: bilateral seizures [p = 0.007; OR, 2.09 (1.22-3.57)], multifocal epilepsy [p = 0.001; OR, 3.03 (1.57-5.86)], history of birth complications [p = 0.011; OR, 3.00 (1.29-6.95)], history of CNS infection [p < 0.001; OR, 3.05 (1.78-5.24)], simple partial seizures [p < 0.001; OR, 1.78 (1.28-2.47)], complex partial seizures [p < 0.001; OR, 2.55 (1.79-3.63)], history of status epilepticus [p < 0.001; OR, 2.98 (1.82-4.88)], secondary generalized seizures [p < 0.011; OR, 1.53 (1.02-2.12)], and being on multiple AEDs [p < 0.001; OR, 2.33 (1.64-3.33]. In multivariate analysis, duration of seizure freedom [p < 0.001; OR, 0.99 (0.98-1.00)], complex partial seizures [p < 0.001; OR, 2.27 (1.58-3.28)], status epilepticus [p = 0.005; OR, 2.10 (1.26-3.51)], and polytherapy [p < 0.001; OR, 1.74 (1.22-2.48)] remain significantly associated with seizure clustering. Conclusions: Our study identified potential predictors of reported seizure clustering. There appears to be a significant association between clusters and a history of status epilepticus, polytherapy (likely to be an indirect measure of seizure severity and intractability), complex partial seizures, and duration of seizure freedom. Surprisingly, almost half of patients with clusters did not have a rescue plan documented in the charts. We believe that increased awareness and discussion of seizure clusters, their predictors, and more practical rescue medications will lead to improved care of these patients.