Limited sampling models for oral midazolam: midazolam plasma concentrations, not the ratio of 1-hydroxymidazolam to midazolam plasma concentrations, accurately predicts AUC as a biomarker of CYP3A activity.

Oral midazolam is used as a phenotyping probe for cytochrome P450 (CYP) 3A activity and requires multiple plasma samples to measure drug exposure. Limited sampling is a useful strategy for optimizing sampling and reducing costs and labor. We studied limited sampling models using multiple linear regressions to predict the area under the concentration versus time curve (AUC) of midazolam using either midazolam plasma concentrations or the ratio of 1-hydroxymidazolam (1-OH MDZ) to midazolam plasma concentrations. CYP3A baseline activity data for oral midazolam from previous studies were used (45 healthy adults for models using midazolam plasma concentrations and 41 healthy adults for models using the ratios of 1-OH MDZ to midazolam plasma concentrations). Limited sampling models were derived, validated, and evaluated for precision and bias. Two equations using the time points at 0.5 and 6 hours and 0.5, 2, and 6 hours were acceptable and predictive of midazolam AUC using midazolam plasma concentrations. No 1-OH MDZ to midazolam plasma concentration ratios accurately predicted midazolam AUC.