Cytochrome P450 3A4 and P-glycoprotein mediate the interaction between an oral erythromycin breath test and rifampin

Background: The intravenous C-14-erythromycin breath test (ERMBTIV) does not measure aggregate liver and intestinal cytochrome P450 (CYP) 3A4 activity. Accordingly, we evaluated an oral stable-labeled (C-13) formulation of the test (ERMBToral) as an alternative CYP3A4 phenotyping probe. Methods: After an overnight fast, 14 young healthy volunteers (5 women and 9 men) received the ERMBTIV (0.07 mumol, 3 muCi), followed by the ERMBToral (500 mg). The next morning, the CYP3A4 inhibitor troleandomycin (500 mg) was given, and both ERMBTs were repeated. After at least 24 hours, the CYP3A4 and P-glycoprotein inducer rifampin (600 mg; INN, rifampicin) was given daily for 7 days, and both ERMBTs were repeated 24 hours after the last dose of rifampin. Plasma samples were collected for 10 hours with each administration of the ERMBToral, and erythromycin levels were measured by liquid chromatography-mass spectrometry. Finally, the effect of troleandomycin on erythromycin transport was examined in Caco-2 cell monolayers. Results: Compared with baseline values, the median ERMBTIV and ERMBToral results and erythromycin apparent oral clearance (CL/F) all significantly decreased, by at least 70%, with troleandomycin treatment (P = .001 for each comparison). With rifampin treatment, the median ERMBTIV result and CL/F increased 2-fold (P less than or equal to .01), but the median ERMBToral result was unchanged (P = .30). There were no rank-order correlations between the ERMBTIV and ERMBToral results or between either ERMBT result and CL/F within each treatment group (P greater than or equal to .07). In addition, troleandomycin had no effect on erythromycin transport in Caco-2 cells (P greater than or equal to .20). Conclusions: The ERMBToral was influenced by processes in addition to intestinal and hepatic CYP3A4 activity and therefore did not provide a straightforward measure of aggregate CYP3A4 phenotype. The erythromycin-rifampin interaction cannot be attributed to CYP3A4 induction alone and probably also reflected intestinal P-glycoprotein induction.