Optimal sampling strategies for bayesian estimation of docetaxel (Taxotere) clearance

Docetaxel activity has been documented in many solid tumors, including metastatic breast cancer and non-small cell lung cancer. However, as clinical studies in other tumor types are now being conducted, the validation of an optimal sampling strategy would allow the performance of pharmacokinetics/pharmacodynamics studies with minimum inconvenience for the patient. Six optimal sampling strategies with one to six sampling times were computed, based on the D-optimality theory, using population pharmacokinetic parameters estimated from a large pharmacokinetic database of 547 patients treated in previous Phase II studies. Validation of these sampling strategies was performed on a set of 35 patients, from two Phase I studies, who received docetaxel as a 1-h infusion at doses ranging from 50 to 100 mg/m2. Validation consisted of comparing clearance assessed by maximum likelihood estimation obtained using complete plasma concentration-time data (considered as the reference) and clearance determined by Bayesian estimation with an optimal design. For all of the optimal sampling strategies tested, clearance was found to be well estimated when at least two samples were taken. Bayesian estimation with two measured levels (at the end of infusion and at 6 h after the start of infusion) can be selected, because it allows adequate estimation of clearance with a nonsignificant bias of +1.37% and a precision of 12.3%.