Purpose: The main purpose of the present study was to identify predictor variables with significant influence on seizure outcome after discontinuation of treatment in children with uncomplicated epilepsy and to analyze whether these variables, included in a prognostic model could identify children in whom 1-year treatment would be sufficient.
Methods: Before initiation of treatment in children aged 2-16 years with uncomplicated epilepsy, the duration of treatment was randomized to 1 year (group I) or 3 years (group II). At the end of the allotted period, treatment was discontinued in 161 children who had been seizure fre during the previous 6 months. The mean follow-up period after treatment was 5.8 years. Twenty-three predictor variables were analyzed by survival methods regarding their influences on the outcome.
Results: At the latest follow-up check, 60 children (37%) had relapsed. The following predictor variables were selected by multiple regression analysis and constituted a model with a simple scoring system: age at seizure onset; seizure type; generalized, irregular spike-wave activity on EEG after 1 year of treatment; and persistent 3-Hz spike-wave activity after 6 months of treatment in children with absence epilepsy. In group I, the remission rate was 73% in children with high prognostic scores, 10% in children with low scores, and 40% in those with intermediate scores (log-rank test, p = 0.0001).
Conclusions: After 1 year of treatment, our prognostic model identified children in whom treatment could be withdrawn at that time. Our model should be easily applicable in clinical practices and may be of clinical importance in determining the duration of treatment in children with uncomplicated epilepsy.