OBJECTIVE: To conduct the first randomized, controlled trial assessing the prophylactic efficacy of gabapentin in bipolar disorder. METHOD: We conducted a 1-year, double-blind, randomized, comparative, placebo-controlled, parallel-group, multicenter study. As this was a pure prophylactic trial, only euthymic bipolar I and II patients (DSM-IV) were randomly assigned in a 1:1 ratio to gabapentin (N = 13) or placebo (N = 12) added to the current treatment (lithium, valproate, carbamazepine, or any combination but not antipsychotics or antidepressants). Subjects participated in the study for 12 months. The primary efficacy parameter was the Clinical Global Impressions scale for Bipolar Illness, Modified (CGI-BP-M), which was assessed at all visits. Other assessments were the Young Mania Rating Scale (YMRS), Hamilton Rating Scale for Depression (HAM-D), Hamilton Rating Scale for Anxiety (HAM-A), Pittsburgh Sleep Quality Index (PSQI), and the systematic collection of reported adverse events. Data were collected from May 1999 to February 2004. RESULTS: The change from baseline to month 12 in mean CGI-BP-M scores between groups was statistically significant (p = .0046). Mean score change from baseline to endpoint in the gabapentin group was -2.1, and the mean score change in the placebo group was -0.6. No emerging manic or depressive symptoms were seen in either group as measured with the YMRS, HAM-D, HAM-A, and PSQI. In the PSQI-6 subscale (use of sleeping medication), the mean score change at month 12 in the gabapentin group was 0.9, and the mean score change in the placebo group was 0.05 (p = .0267). Overall, gabapentin was well tolerated. CONCLUSION: This small, randomized clinical trial comparing the prophylactic efficacy of adjunctive gabapentin to placebo suggests that, despite lack of acute efficacy, treatment with gabapentin might provide some benefit on the long-term outcome of bipolar disorder.