In the Canadian multicenter, double-blind clinical trial of risperidone, 135 hospitalized chronic schizophrenic patients were randomly assigned to one of six parallel treatment groups for 8 weeks: risperidone, 2, 6, 10, or 16 mg/day; haloperidol, 20 mg/day; or placebo. Risperidone (6 to 16 mg)-treated patients showed significantly (p < 0.05) lower dyskinetic scores than those receiving placebo, whereas in haloperidol- and placebo-treated patients, no significant differences for dyskinetic symptoms were noted. To explore the antidyskinetic effect of risperidone, a post hoc analysis was performed on two selected patient samples: (1) patients meeting Research Diagnosis Criteria (RDC) for tardive dyskinesia (TD) at baseline or during double-blind treatment (N = 49) and (2) patients with RDC TD and with a Clinical Global Impression (CGI) Severity of dyskinesia score > or = 5 (at least moderately severe) (N = 48). The composition of the two subsamples was found to be almost identical because all but one of the patients with RDC TD (N = 49) were members of the group with at least moderately severe TD (N = 48). Analysis of four parameters (Extrapyramidal Symptom Rating Scale-dyskinesia total score, CGI severity of dyskinesia, buccolinguomasticatory [BLM] factor score, and extremities [choreoathetoid factor] score confirmed the antidyskinetic effect that was noted in the intent-to-treat analysis, which included all patients, whether they had RDC TD or not. Results indicated that risperidone at 6 mg/day had the most beneficial effect on TD, especially on the BLM syndrome, without inducing significant parkinsonism while treating psychotic symptoms. This antidyskinetic effect was greater than with either placebo or haloperidol.