Consumption of typical quantities of grapefruit juice (GFJ) increases the oral bioavailability of several CYP3A4 substrates without affecting their elimination, consistent with selective inhibition of intestinal but not hepatic CYP3A4. However, increases in the AUCs of CYP3A4 substrates recently associated with the consumption of large amounts of GFJ were similar to those observed with potent inhibitors of hepatic CYP3A4. The current study compared the effects of consuming large quantities and more typical amounts of GFJ on the activity of hepatic and intestinal cytochrome P450 3A4 in vivo, employing the erythromycin breath test (EBT) and oral midazolam pharmacokinetics. This was a two-phase, randomized, placebo-controlled crossover study, with each phase conducted with a separate panel of subjects. In Phase I, 8 male volunteers were randomized to the order of receiving one glass (240 mL) of water (placebo) or double-strength (DS) GFJ tid for 2 days and then 90, 60, and 30 minutes prior to administration of probe drugs on the 3rd day. In Phase II, 16 male volunteers were randomized to the order of receiving one glass of (1) single-strength (SS) GFJ, (2) DS GFJ, and (3) water (placebo). All treatments were administered in a fasted state. There was at least a 7-day washout period between treatments. Probe drugs, administered 30 minutes or 1 hour following each treatment in Phase I or II, respectively, consisted of oral midazolam (2 mg) coadministered with IV [14G N-methyl] erythromycin (0.03 mg). The EBT was performed 20 minutes following erythromycin administration. Blood was collected during the 24 hours following probe drug administration for the analysis of midazolam pharmacokinetics. In Phase I, consumption of one glass of DS GFJ tid for 3 days increased the Cmax of midazolam 3-fold, the AUC 6-fold, and the t1/2 2-fold and decreased the amount of exhaled 14CO2 in all 8 subjects, with a mean decrease in EBT of 18%. In Phase II, consumption of one glass of DS GFJ significantly increased the AUC and Cmax of midazolam approximately 2-fold without a significant effect on the t1/2 of midazolam or the EBT. The effects of consuming one glass of SS GFJ on midazolam pharmacokinetics and the EBT were not significantly different from those of one glass of DS GFJ. It was concluded that consumption of one glass of DS GFJ tid for 3 days significantly increased the AUC, Cmax, and t1/2 of midazolam and reduced EBT values, reflecting inhibition of both hepatic and intestinal CYP3A4. In contrast, consumption of one glass of SS or DS GFJ increased midazolam AUC and Cmax, with little effect on the midazolam t1/2 and EBT values, reflecting preferential inhibition of intestinal CYP3A4. Alterations of midazolam AUC and Cmax induced by nine glasses of DS GFJ were significantly greater than those produced by one glass of SS or DS GFJ. These data suggest that GFJ inhibits intestinal and hepatic CYP3A4 in an exposure-dependent fashion and that patients taking medications that are CYP3A4 substrates are at risk for developing drug-related adverse events if they consume large amounts of grapefruit juice.