We studied the effects of the CYP2C8 inhibitor trimethoprim and CYP2C8 genotype on the pharmacokinetics of the antidiabetic pioglitazone. In a randomized crossover study, 16 healthy volunteers with the CYP2C8(*)1/(*)1 (n = 8), (*)1/(*)3 (n = 5), or (*)3/(*)3 (n = 3) genotype ingested 160 mg of trimethoprim or placebo twice daily for 6 days. On day 3, they ingested 15 mg of pioglitazone. The effects of trimethoprim on pioglitazone were characterized in vitro. Trimethoprim raised the area under the plasma pioglitazone concentration-time curve (AUC(0-infinity)) by 42% (p < 0.001) and decreased the formation rates of pioglitazone metabolites M-IV and M-III (p < 0.001). During the placebo phase, the weight-adjusted AUC(0-infinity) of pioglitazone was 34% smaller in the CYP2C8(*)3/(*)3 group and 26% smaller in the CYP2C8(*)1/(*)3 group than in the CYP2C8(*)1/(*)1 group (p < 0.05). Trimethoprim inhibited M-IV formation in vitro (inhibition constant 38.2 muM), predicting the in vivo interaction. In conclusion, drug interactions and pharmacogenetics affecting the CYP2C8 enzyme may change the safety of pioglitazone.