Predicting drug interactions in vivo from experiments in vitro. Human studies with paclitaxel and ketoconazole.

This study was performed to evaluate whether concomitant treatment with ketoconazole could reduce the clearance of paclitaxel given to ovarian cancer patients. Paclitaxel, 175 mg/m2, was given as a 3-hour continuous intravenous infusion and repeated every 21 days. Initially, ketoconazole, 100 to 1600 mg, was given as a single oral dose 3 hours after paclitaxel. Later, ketoconazole, 200 mg, was given perorally 3 hours before paclitaxel. Plasma drug concentrations were measured by high-pressure liquid chromatography (HPLC), and cytochrome P450 3A (CYP3A) activity was measured with the erythromycin breath test (ERMBT). Ketoconazole did not alter plasma concentrations of paclitaxel or its principal metabolite, 6 alpha-hydroxypaclitaxel. Although there was marked inter- and intrapatient variability in ketoconazole pharmacokinetics, peak plasma concentrations in all but one course were below the 50% inhibitory concentration (IC50) point determined for inhibition of paclitaxel metabolism in vitro. Therefore, paclitaxel and ketoconazole can be coadministered safely without dose adjustments. There was no correlation between ERMBT measurements and serial plasma concentrations of paclitaxel. The erythromycin breath-test measurements did correlate with the corresponding ketoconazole plasma concentrations. The erythromycin breath test is a valuable tool for measuring instantaneous CYP3A activity in vivo. This clinical study confirms the results of prior studies with human-derived materials in vitro, reinforcing the notion that such studies are useful predictors of drug pharmacokinetics and interactions in vivo.