Prevention of relapse is a major goal of maintenance treatment in patients with psychotic disorders. We performed a long-term comparison of a newer, atypical antipsychotic drug, risperidone, and an older, conventional neuroleptic drug, haloperidol, in terms of the rate of relapse in patients with schizophrenia and schizoaffective disorder.

In a double-blind, prospective study at 40 sites, we randomly assigned adult outpatients in stable condition with chronic schizophrenia or schizoaffective disorder to receive treatment with flexible doses of either risperidone or haloperidol for a minimum of one year.

Of the 397 patients who underwent randomization, data from 2 were excluded because they did not receive study medication; data from all 30 patients from one site were excluded by the sponsor, the Janssen Research Foundation, because of concern about the integrity of the data. The median duration of treatment was 364 days in the risperidone group and 238 days in the haloperidol group (P=0.02). Of the 177 patients assigned to risperidone and the 188 assigned to haloperidol who remained in the analysis, 44.1 percent and 52.7 percent, respectively, discontinued treatment for reasons other than relapse. The Kaplan–Meier estimate of the risk of relapse at the end of the study was 34 percent for the risperidone group and 60 percent for the haloperidol group (P<0.001); the risk ratio for relapse with haloperidol, from the Cox model, was 1.93 (95 percent confidence interval, 1.33 to 2.80; P<0.001). Early discontinuation of treatment for any reason was more frequent among haloperidol-treated patients (risk ratio, 1.52; 95 percent confidence interval, 1.18 to 1.96). Patients in the risperidone group had greater reductions in the mean severity of both psychotic symptoms and extrapyramidal side effects than those in the haloperidol group.

Adult outpatients with clinically stable schizophrenia or schizoaffective disorder have a lower risk of relapse if they are treated with risperidone than if they are treated with haloperidol.