A previous study reported a 2- and 3-timepoint limited sampling strategy (LSS) model accurately predicted oral midazolam area under the concentration time curve (AUC), and thus cytochrome P450 (CYP) 3A activity.
Objective: This study evaluated whether the LSS models predict midazolam AUC during CYP3A baseline, inhibition and induction/activation.
Materials and methods: Plasma midazolam concentrations from 106 healthy adults from 6 published studies were obtained where oral midazolam was co-administered alone or with ketoconazole, double-strength grapefruit juice, Ginkgo biloba extract, pleconaril, or rifampin. Observed and predicted midazolam AUCs were determined. Bias and precision of the LSS models were determined.
Results: Contrasting results were observed for the 2- and 3-timepoint LSS models in accurately predicting midazolam AUC during baseline CYP3A conditions. With the exception of 1 study (single dose, double-strength grapefruit juice), the 2- and 3-timepoint LSS models did not accurately predict midazolam AUC during conditions of CYP3A inhibition and induction/activation.
Conclusion: The previously reported 2- and 3-timepoint oral midazolam LSS models are not applicable to the evaluated conditions of CYP3A baseline, inhibition, and induction/ activation.