The primary objective of this 9-week open-label extension trial was to assess the effects of risperidone monotherapy in patients with acute bipolar I disorder who completed treatment in two preceding 3-week double-blind trials. Patients with DSM-IV bipolar I disorder, experiencing an acute manic episode, received a flexible dose of risperidone (1-6 mg/day) or placebo in two independent double-blind, randomized, 3-week monotherapy trials. Completers who required ongoing treatment were eligible to enter this open-label 9-week extension trial during which all patients received risperidone. The primary efficacy measure was the mean change in the Young Mania Rating Scale (YMRS) total score. Secondary efficacy measures included the Clinical Global Impressions-Severity Scale, Montgomery-Asberg Depression Rating Scale, Positive and Negative Syndrome Scale and Global Assessment Scale. Safety assessments included adverse event reports, laboratory tests, and the Extrapyramidal Symptom Rating Scale (ESRS). Of the 283 patients who entered the extension study, 160 had previously received risperidone (RIS/RIS) in the acute treatment trial and 123 had received placebo (PLA/RIS). This study was completed by 71% of these patients. The mean+/-SE modal dose of risperidone was 4.6+/-1.5 mg/day. Patients in both the RIS/RIS and PLA/RIS groups improved significantly at the endpoint of the 9-week open-label study compared to their open-label baseline scores (-5.2+/-0.69, P<0.001 and -9.12+/-1.44, P<0.001, respectively) on the YMRS. Furthermore, changes from double-blind baseline to open-label endpoint were -29.4+/-1.0 in the RIS/RIS group and -23.9+/-1.4 in the PLA/RIS group. Significant improvements from both double-blind and open-label baseline were seen at week 1 of the open-label trial (P<0.001) and at each subsequent timepoint. A similar pattern was observed on the secondary measures of efficacy. Most frequent adverse events were extrapyramidal disorder (18%) and somnolence (12%). Most adverse events were mild or moderate in severity. The mean score for the Parkinsonism subscale of the ESRS was 1.1 at open-label baseline, and decreased by 0.1 at endpoint. Mean increase in body weight from open-label baseline was 0.6 kg in patients treated with placebo in the preceding double-blind trial and 1.2 kg in patients previously treated with risperidone. Risperidone treatment was well tolerated and resulted in further improvement during the 9-week extension, beyond the 3 weeks of acute treatment. Patients switched from placebo to risperidone improved markedly. Risperidone treatment did not induce depression.