Results of a subanalysis of data from the multinational risperidone trial (RIS-INT-2) are reported. Patients with chronic schizophrenia were treated with risperidone at 1 mg/day (n = 25), 4 mg/day (n = 27), 8 mg/day (n = 29), 12 mg/day (n = 31), or 16 mg/day (n = 29), or 10 mg/day of haloperidol for 8 weeks. According to the Positive and Negative Syndrome Scale (PANSS) total and subscale scores, improvements were noted in each treatment group from baseline to treatment endpoint. On each scale the magnitude of improvement was greater in the risperidone patients than in the haloperidol patients. The onset of action of risperidone was faster than haloperidol. By treatment week 2, over half of the patients receiving > or = 4 mg/day of risperidone were clinically improved (> or = 20% reduction in total PANSS scores). This rate of improvement was not seen until week 6 in the haloperidol patients. Severity of extrapyramidal symptoms (scores on the Extrapyramidal Symptom Scale) was significantly lower in patients receiving 1 or 4 mg/day of risperidone than in patients receiving higher risperidone doses and in haloperidol patients. The optimal dose of risperidone, in terms of both efficacy and safety, was 4 mg/day. These results confirm the findings of the controlled trials of risperidone conducted in North America and the multinational trial.