The efficacy and safety of felbamate monotherapy were evaluated in 52 patients with refractory partial seizures with or without secondary generalization in a double-blind, randomized, placebo-controlled trial. Each patient completed a routine evaluation for epilepsy surgery and was randomized to receive either felbamate, titrated to a maximum daily dose of 3600 mg over 2 days, or placebo during the 10-day, inpatient, treatment phase. An intent-to-treat analysis was performed on the data of all 52 patients who received study medication, while a separate efficacy analysis also was performed on the data of 43 evaluable patients, which excluded protocol violators. The endpoint of the trial was completing 10 days of treatment or the occurrence of a fourth seizure. The primary efficacy variable was the average daily seizure frequency during the treatment phase for each patient. For the intent-to-treat analysis based on all 52 patients who received study medications, the mean rank of the daily seizure frequency for patients treated with felbamate was 21.6 compared to 29.6 for patients treated with placebo (P = 0.065). In the analysis based on the 43 evaluable patients, the mean rank of the daily seizure frequency for felbamate-treated patients was 17.0 compared to 25.4 for placebo-treated patients. This difference was statistically significant (P = 0.032) in favor of felbamate. Seizure frequency was decreased by 89.5% compared to baseline in nine patients who completed 10 days of felbamate therapy. This study permitted the rapid determination of the anticonvulsant activity of felbamate and demonstrated that felbamate is effective as monotherapy for the treatment of partial seizures.