Lack of in vivo correlation between indinavir and saquinavir exposure and cytochrome P450 3A phenotype as assessed with oral midazolam as a phenotype probe.

Study Objective: To investigate a potential correlation between exposure to oral midazolam, a commonly used cytochrome P450 (CYP) 3A probe, and saquinavir and indinavir exposure.

Design: Open-label, prospective, pharmacokinetic study.

Setting: Outpatient research center.

Subjects: Thirty-six healthy volunteers aged 22-50 years.

Intervention: Subjects received a single oral dose of midazolam 8 mg; 4 hours later, blood was drawn to determine their serum midazolam concentrations. Midazolam phenotyping was followed by successive administration of the protease inhibitors indinavir and saquinavir, with blood sampling and pharmacokinetic analyses performed at steady state.

Measurements And Main Results: Pharmacokinetic parameters of each protease inhibitor were evaluated to assess for a potential relationship with 4-hour concentrations of midazolam. No correlations between phenotype results for midazolam and any pharmacokinetic parameter for indinavir or saquinavir were identified (r(2)=0.00002-0.073). When the results were analyzed based on race, significant correlations were identified in five African-American subjects, including correlations between 4-hour midazolam levels and apparent oral clearance of saquinavir (r(2)=0.734, p=0.064), area under the plasma concentration-time curve from 0-8 hours (r(2)=0.914, p=0.011), minimum concentration (r(2)=0.857, p=0.024), and maximum concentration (r(2)=0.969, p=0.002). These findings for African-American subjects were not seen with indinavir. No correlation was found between indinavir and saquinavir pharmacokinetic parameters (r(2)=0.017-0.261).

Conclusion: Oral midazolam was not a useful probe for predicting saquinavir or indinavir exposure at steady state. Reasons for the lack of correlation likely included differences between midazolam and protease inhibitor P-glycoprotein specificity, differences in the relative contribution of CYP3A5-mediated metabolism, and/or variation in intestinal and hepatic CYP3A specificity. The strong correlation between midazolam phenotype and pharmacokinetic parameters for saquinavir in African-American subjects indicated a racial difference in one or more of these confounding variables.