Twenty hospitalized schizophrenic patients on haloperidol (doses 6 to 80 mg/day; median, 30 mg/day) underwent 4 days of placebo washout before being treated for 6 weeks with remoxipride, a new benzamide derivative with selective D2-dopamine receptor blocking properties. All patients completed the clinical trial period with week 6 doses ranging from 75 to 500 mg/day (median, 225 mg/day). Comparison of final scores with end of placebo washout showed improvement in schizophrenic symptoms in 10 patients and a reduction in the mean score for Clinical Global Impression of severity of illness (14.1%) and Brief Psychiatric Rating Scale total score (23.0%). Remoxipride caused less parkinsonism than the prior neuroleptic therapy and appeared to have little masking effect on tardive dyskinesia. Only slight evidence of serum neuroleptic activity was shown by radio-receptor assay measurements using [3H]spiperone binding and calf caudates, and the drug's effect on prolactin elevation was short-lasting (less than 10 hours). The mean elimination half-life of remoxipride was 5.9 hours. These results add to the consistent impression that D2 receptor blockade predicts clinical antipsychotic effects.