1. We have assessed the interaction of the antimalarial halofantrine with cytochrome P450 (CYP) enzymes in vitro, with the use of microsomes from human liver and recombinant cell lines. 2. Rac-halofantrine was a potent inhibitor (IC50 = 1.06 microM, Ki = 4.3 microM) of the 1-hydroxylation of bufuralol, a marker for CYP2D6 activity. Of a group of structurally related antimalarials tested, only quinidine (IC50 = 0.04 microM) was more potent. 3. Microsomes prepared from recombinant CYP2D6 and CYP3A4 cell lines were shown to catalyse halofantrine N-debutylation. 4. The metabolism of halofantrine to its N-desbutyl metabolite by human liver microsomes showed no correlation with CYP2D6 genotypic or phenotypic status and there was no consistent inhibition by quinidine. 5. The rate of halofantrine metabolism showed a significant correlation with both CYP3A4 protein levels (r = 0.88, P = 0.01) and the rate of felodipine metabolism (r = 0.86, P = 0.013), a marker substrate for CYP3A4 activity. Inhibition studies showed that ketoconazole is a potent inhibitor of halofantrine metabolism (IC50 = 1.57 microM). 6. In conclusion, we have demonstrated that halofantrine is a potent inhibitor of CYP2D6 in vitro and can also be metabolised by the enzyme. However, in human liver microsomes it appears to be metabolised largely by CYP3A4.