BACKGROUND Gabapentin is widely approved as add-on therapy for epilepsy treatment for partial seizures with and without secondary generalization. To investigate the efficacy of gabapentin administered as monotherapy in patients with newly diagnosed partial epilepsy, a randomized double-blind trial was performed. METHODS Eligible patients were randomized to receive one of three masked doses of gabapentin (300, 900, or 1,800 mg/day) or open-label carbamazepine (600 mg/day) and kept daily seizure diaries throughout the study. After titration, patients entered a 24-week evaluation phase. Patients were required to exit the study if they experienced an exit event, defined as a total of three simple or complex partial seizures, one generalized tonic-clonic (GTC) seizure, or status epilepticus. Patients could be withdrawn for lack of efficacy, adverse events, or noncompliance. Kaplan-Meier statistics were used to estimate the probability that patients would continue in the study without having an exit event. RESULTS Time to exit event was longer for patients on 900 mg/day (n = 72) or 1,800 mg/day (n = 74) of gabapentin than for patients receiving 300 mg/day (n = 72; p = 0.0395 and 0.0175, respectively). The most clinically relevant measure of retention on treatment (exit event plus adverse event withdrawal rate) was similar for carbamazepine (n = 74) and 1,800 mg/day gabapentin (54% versus 57%) but was lower (better) for 900 mg/day gabapentin (44%). No unexpected new adverse events emerged with gabapentin monotherapy. CONCLUSIONS Gabapentin at 900 or 1,800 mg/day is effective and safe as monotherapy for patients with newly diagnosed partial epilepsy.