In-vitro studies were performed using human liver microsomes and c-DNA-expressed human P450 isoforms to identify the cytochrome P450 isoenzyme(s) involved in the back oxidation and N-dealkylation of reduced haloperidol. Back oxidation and N-dealkylation of reduced haloperidol were assessed by measuring the formation of haloperidol and 4-(4-chlorophenyI)-4- hydroxypiperidine (CPHP), respectively. The Haloperidol and CPHP formation rates as a function of substrate concentration, measured in three livers, followed monophasic enzyme kinetics. For haloperidol formation Km values ranged from 51-59 μM, and Vmax values from 190-334 pmol mg-1 min-1; for CPHP formation Km values were 44-49 μM, and Vmax values 74-110 pmol mg-1 min-1. haloperidol and CPHP formation rates in the nine liver preparations were significantly correlated with dextromethorphan N-demethylase activity (a marker of CYP3A4 activity), but not with the CYP2D6, CYP1A2 and CYP2C9 activity. Ketoconazole and troleandomycin, inhibitors of CYP3A4, inhibited competitively both haloperidol and CPHP formation, with a Ki value lower than 0.2 μM for ketoconazole and lower than 0.3 μM for troleandomycin. Sulphaphenazole (CYP2C9), furafylline (CYP1A2) and quinidine and paroxetine (CYP2D6) gave only little inhibition (IC50>60 μM). CPHP and haloperidol formation were, moreover, enhanced by a-naphthoflavone, an effect known for CYP3A4 mediated reactions. Anti-CYP3A4 antibodies strongly inhibited haloperidol and CPHP formation, whereas CYP2D6 antibodies did not. Among the recombinant human CYP isoforms tested, CYP3 A4 exhibited the highest activity with respect to haloperidol and CPHP formation rates, with no detectable effect of CYP1A2, CYP2D6 and CYP2C9. These results strongly suggest that back oxidation and N-dealkylation of reduced haloperidol in human liver microsomal preparations are mediated by CYP3A4.