Abstracts

To evaluate peer-reviewed, published RCTs of AEDs in childhood epilepsy.
Medline (Pubmed) searches (1966-2002) for RCTs in children comparing the efficacy of 2 or more AEDs, exploring therapy for untreatable syndromes or primary proof of efficacy trials with negative results. Primary proof of efficacy trials with an AED more effective than placebo were excluded.
Only 27 RCTs met our criteria. Without reference to the methodological quality of each trial, we list the conclusions of these trials. In newly treated children, carbamazepine, oxcarbazepine, phenytoin, valproic acid, clobazam and phenobarbital have equal efficacy. Phenobarbital is equivalent to mephobarbital and may have more side effects than carbamazepine, phenytoin and valproic acid. Carbamazepine[rsquo]s effect on memory may be greater than phenytoin or valproic acid. Gabapentin is ineffective for childhood absence and Benign Rolandic epilepsy. Stiripental has some efficacy in combination with other AEDs for severe myoclonic epilepsy of infancy. For West syndrome, low dose ACTH is as effective as high dose: both are better than oral prednisolone. ACTH is better than vigabatrin except in tuberous sclerosis where vigabatrin is superior to ACTH and oral hydrocortisone. High dose vigabatrin is more effective than low dose. For acute seizures, IM midazolam may act faster than IV diazepam and rectal lorazepam may be more effective than diazepam.
Because many children have epilepsy, these results are discouraging. The choice of AED treatment for most children is without a clear evidence base. Collaboration for RCTs should be a priority for childhood epilepsy specialists.