We studied the time course for the reversal of rifampin's effect on the pharmacokinetics of oral midazolam (a cytochrome P450 (CYP) 3A4 substrate) and digoxin (a P-glycoprotein (P-gp) substrate). Rifampin increased midazolam metabolism, greatly reducing the area under the concentration-time curve (AUC(0-∞)). The midazolam AUC(0-∞) returned to baseline with a half-life of ~8 days. Rifampin's effect on the AUC(0-3 h) of digoxin was biphasic: the AUC(0-3 h) increased with concomitant dosing of the two drugs but decreased when digoxin was administered after rifampin. Digoxin was found to be a weak substrate of organic anion-transporting polypeptide (OATP) 1B3 in transfected cells. Although the drug was transported into isolated hepatocytes, it is not likely that this transport was through OATP1B3 because the transport was not inhibited by rifampin. However, rifampin did inhibit the P-gp-mediated transport of digoxin with a half-maximal inhibitory concentration (IC(50)) below anticipated gut lumen concentrations, suggesting that rifampin inhibits digoxin efflux from the enterocyte to the intestinal lumen. Pharmacokinetic modeling suggested that the effects on digoxin are consistent with a combination of inhibitory and inductive effects on gut P-gp. These results suggest modifications to drug-drug interaction (DDI) trial designs.