Double-blind substitution of vigabatrin and valproate in carbamazepine-resistant partial epilepsy. 012 Study group.

Patients from 12 countries reporting two or more partial seizures per month despite treatment with optimal doses of CBZ were randomised to additional vigabatrin (VGB, 2-4 g daily) or sodium valproate (VPA, 1-2 g daily) using a double-blind, double-dummy design. The study included a 6 month retrospective baseline on unchanged CBZ dosage, a month's prospective baseline, a short titration phase, and an assessment period lasting 3 months on duotherapy. CBZ was withdrawn over a further 2 months in responders ( > or = 50% monthly seizure reduction compared with baseline), who continued on alternative monotherapy for 3 or more months. If seizure control deteriorated, CBZ was reinstated and these patients were also followed up for 3 months. A total of 215 patients (108 VGB, 107 VPA) reporting a mean of seven partial seizures per month fulfilled the criteria for the intention-to-treat analysis. 53 and 51% of patients in the VGB and VPA group respectively achieved a monthly reduction in seizure numbers > or = 50%, respectively. 27 and 31% maintained alternative monotherapy. Overall, 17% (7% monotherapy, 10% duotherapy) of the VGB treated patients and 19% (8% monotherapy, 11% duotherapy) of the VPA group remained seizure-free during the final 3 month treatment period. VGB and VPA, which increase neuronal inhibition mediated by gamma aminobutyric acid, can be added to or substituted for CBZ when this Na+ channel blocker fails to control partial seizures. This lends credence to the hypothesis in support of a mechanistic approach to the management of epilepsy.