Population pharmacokinetic-pharmacodynamic modeling and model-based prediction of docetaxel-induced neutropenia in Japanese patients with non-small cell lung cancer.

Abstract

Docetaxel is used to treat many cancers, and neutropenia is the dose-limiting factor for its clinical use. A population pharmacokinetic-pharmacodynamic (PK-PD) model was introduced to predict the development of docetaxel-induced neutropenia in Japanese patients with non-small cell lung cancer (NSCLC). Forty-seven advanced or recurrent Japanese patients with NSCLC were enrolled. Patients received 50 or 60mg/m2 docetaxel as monotherapy, and blood samples for a PK analysis were collected up to 24h after its infusion. Laboratory tests including absolute neutrophil count data and demographic information were used in population PK-PD modeling. The model was built by NONMEM 7.2 with a first-order conditional estimation using an interaction method. Based on the final model, a Monte Carlo simulation was performed to assess the impact of covariates on and the predictability of neutropenia. A three-compartment model was employed to describe PK data, and the PK model adequately described the docetaxel concentrations observed. Serum albumin (ALB) was detected as a covariate of clearance (CL): CL (L/h)=32.5×(ALB/3.6)0.965×(WGHT/70)3/4. In population PK-PD modeling, a modified semi-mechanistic myelosuppression model was applied, and characterization of the time course of neutrophil counts was adequate. The covariate selection indicated that α1-acid glycoprotein (AAG) was a predictor of neutropenia. The model-based simulation also showed that ALB and AAG negatively correlated with the development of neutropenia and that the time course of neutrophil counts was predictable. The developed model may facilitate the prediction and care of docetaxel-induced neutropenia.