Seizure recurrence after a first unprovoked seizure in childhood: a prospective study

Abstract

Purpose: To study the risk of recurrence after a first unprovoked seizure in childhood.

Methods: All consecutive patients aged less than 14 years with one or more unprovoked seizures who were attended between January 1, 1987, and June 1, 1996, were included in a prospective study. Clinical features of patients attended after a first seizure and those attended after two or more seizures were compared. Recurrence risk in both groups was estimated by Kaplan-Meier curves. Univariate and multivariate analyses of the potential predictors of recurrence risk were performed for the group of patients attended after a first seizure using the Cox proportional hazards model.

Results: Included in the study were 217 children. Kaplan-Meier estimate of recurrence risk was 64% at 5 years, when only patients being attended after a first epileptic seizure were included, compared with 74% when all patients were included. Significant differences in several clinical features were found between patients attended after a first seizure and those attended after two or more seizures. Univariate and multivariate analyses showed that in the overall cohort of patients attended after a first seizure, a symptomatic etiology increased the risk of recurrence, whereas a patient age of 3 to 10 years decreased this risk. In particular, the recurrence risk was 96% at 2 years for symptomatic seizures, compared with 46% for idiopathic/cryptogenic seizures. In the group of patients with idiopathic/cryptogenic seizures, an abnormal electroencephalogram and the occurrence of seizures during sleep increased the recurrence risk, whereas a patient aged 3 to 10 years reduced it. In the group of patients with symptomatic etiology, univariate analysis revealed that there was a lower recurrence risk for patients aged 3 to 10 years. This last finding was not maintained, however, in multivariate analysis.

Conclusions: The recurrence risk depends on the inclusion criteria for enrolling patients. Several factors enable us to predict the recurrence risk after a first unprovoked seizure; the most important of these factors is the etiology of the seizures.