Vigabatrin (VGB) has been shown in a number of clinical trials with varying designs to be effective and well-tolerated as both add-on therapy and monotherapy in epilepsy with partial seizures with or without secondary generalization as well as in infantile spasms. The present study is an open, long-term (1 year) extension of a randomized double-blind placebo-controlled multicentre Canadian trial of VGB in resistant partial adult epilepsy. The present study was designed to examine the safety and long-term efficacy of VGB. Completers of the preceding double-blind study had their dose of VGB titrated to 4 g/day over 3 weeks. Patients were evaluated every 2-4 weeks and at week 14 were allowed to continue only if they achieved a 50% seizure reduction compared with pre-VGB baseline. In addition to neurological and physical examinations, safety was assessed by a cognitive psychosocial test battery, visual and somatosensory evoked potentials and MRI scans. Ninety-seven of 100 eligible patients entered the study, 53 of whom completed the 52 weeks. Fifty-eight percent of the patients had a greater than 50% seizure reduction in seizures vs. pre-VGB baseline. Seizure reductions of 56% and 45%, respectively, were seen in the VGB and placebo groups from the preceding study. Fifty-four percent of patients were judged by the investigators to have experienced at least a moderate therapeutic effect. Discontinuations were 29% for lack of efficacy and 12% for adverse effects. There was a mean weight gain of 3.7 +/- 0.2 kg by end of study. Neurologica/psychiatric side effects were the most common reason for withdrawal including three behavioral reactions attributed to the drug which required temporary hospitalization. There were no abnormalities on laboratory or special tests and there was a tendency for improvement on most tests of cognitive function and mood. Vigabatrin, as an add-on agent, is well-tolerated and can be of long-term benefit in a substantial proportion of patients with intractable partial epilepsy.