Drugs metabolized by cytochrome CYP3A isoenzymes have wide interindividual variability and normally distributed plasma clearance distributions. This makes precise dosing difficult to achieve clinically, which may compromise safe therapy. We hypothesized that with potent inhibition of CYP3A, we could clinically render patients "poor metabolizer" phenotype status, and thus reduce interindividual pharmacokinetic variability of midazolam, a well-known CYP3A substrate. Intravenous bolus midazolam at doses of 2.5 mg and 1 mg were administered to 28 and 29 patients with cancer with and without co-administration of 200 mg of oral ketoconazole twice per day respectively for 3 days, starting 1 day before midazolam. Pharmacokinetic analyses of midazolam on both groups were derived using noncompartmental methods and compared. The mean clearance (CL) of midazolam was reduced 6 times by ketoconazole. Midazolam CL were normally distributed in both groups, and ranged from 1.7 to 51.9 and 1.4 to 8.2 L/hour in the control and ketoconazole groups, respectively, corresponding to a 7-fold reduction in dispersion between the 2 groups. Area-under-the-curve variability was reduced by >100%. A limited sampling model consisting of time points at 15 and 300 minutes was validated as a phenotype for CYP3A activity to facilitate the use of midazolam as a probe drug for CYP3A activity. Potent inhibition of CYP3A by ketoconazole reduced midazolam CL and area-under-the-curve variability, allowing for more precise achievement of therapeutic target drug exposure. Prospective evaluation of this approach, together with dose adjustment based on limited sampling, seems warranted.