Both haloperidol and risperidone have been widely used in the treatment of schizophrenia. Because of wider therapeutic spectrum of risperidone, switching from haloperidol to risperidone is recommended in patients who do not sufficiently respond to haloperidol. The present study investigated the correlation between the steady-state plasma concentrations of haloperidol and risperidone together with the effects of CYP2D6 status on the steady-state kinetics of both drugs. Subjects were 22 schizophrenic inpatients. Eleven patients first received risperidone 6 mg/day and then haloperidol 12 mg/day, while the remaining 11 patients received these two treatments in the opposite sequence. The steady-state plasma concentrations of risperidone, 9-hydroxyrisperidone, haloperidol, and reduced haloperidol were measured after the subjects had been on the treatment for at least 2 weeks, and CYP2D6 genotypes were identified in all subjects. Neither the correlation between the steady-state plasma concentrations of haloperidol and those of risperidone (r = 0.061, ns) nor the active moiety (sum of concentration of risperidone and 9-hydroxyrisperidone) of risperidone (r = 0.141, ns) was significant. The mean (+/- SD) plasma concentration of risperidone in patients with mutated allele(s)for CYP2D6 was significantly higher than those without mutated allele (1.5 +/- 0.7 vs. 8.5 +/- 11.0, p < 0.05), while such a tendency for haloperidol was not observed. The present study suggests that the steady-state plasma concentration of risperidone is not predicted from that of haloperidol in the same individual, probably because of the much greater involvement of CYP2D6 in the metabolism of risperidone than in that of haloperidol.