Stopping medication in children with epilepsy: predictors of outcome.

Anticonvulsant therapy was discontinued in 68 children with epilepsy who had had no seizures for four years, in an attempt to find the risk of relapse and the factors predictive of recurrence. In this population the probability of remaining free of seizures for four years after discontinuation of medication was 69 per cent. Children were more likely to have recurrent seizures if they were mentally retarded, if their seizures had begun before two years of age, if they had had many generalized seizures before control, or if they had had a definitely abnormal electroencephalogram before medication was discontinued. Multivariate analysis showed that the best predictors of outcome were the electroencephalogram taken at cessation of medication and the number of seizures before control. We conclude that in the normal child who has not had many seizures and whose electroencephalogram is normal or mildly abnormal, the risks of discontinuing medication after four seizure-free years are acceptable.