Nine extensive metabolizers (EMs) and eight poor metabolizers (PMs) of sparteine took a single oral dose of 100 mg of desipramine HCI before and while taking paroxetine 20 mg per day. Before paroxetine, the median of the total desipramine clearance was 7 times higher in EMs than in PMs (102 and 15 l.h-1 respectively). This confirms that desipramine is extensively metabolized via the sparteine/debrisoquine oxidation polymorphism i.e. by CYP2D6. During paroxetine, the median clearances were 22 l.h-1 and 18 l.h-1 in EMs and PMs respectively. The 5-fold decrease in clearance in EMs when desipramine was co-administered with paroxetine confirms that paroxetine is a potent inhibitor of CYP2D6. The lack of effect on clearance in PMs shows that paroxetine is a selective inhibitor of CYP2D6, which is absent from the livers of PMs. Before paroxetine, the median of desipramine clearance via 2-hydroxylation was 40-times higher in EMs than in PMs (56 and 1.4 l.h-1 respectively), but during paroxetine, it was only 2-times higher (6 and 2.9 l.h-1 respectively). The increase in this clearance in PMs suggests that paroxetine is an inducer of the alternative, unidentified P450(s) which catalyze(s) the formation of 2-OH-desipramine in this phenotype. Before paroxetine, the median amounts of 2-OH-desipramine glucuronide recovered in urine were 69% and 68% of the total recovery of 2-OH-desipramine in urine in EMs and PMs respectively. During paroxetine, the corresponding values were 77% and 84%.(ABSTRACT TRUNCATED AT 250 WORDS)