Using Body Composition Metrics to Predict Exposure Between Japanese and Caucasian Populations
G. Morrish(1), C.Kirkpatrick(1), N. Byrne(2), H. Kinoshita(3) & B. Green(1)
(1) School of Pharmacy, University of Queensland, Brisbane, Australia; (2) Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Australia; (3) Department of Clinical Pharmacology, Chugai Pharmaceutical Co. Ltd, Tokyo, Japan
Objectives: The Japanese Regulatory Authority require bridging studies for NCE’s to ensure exposure in this population is similar to Caucasians. We believe exposure in different races might be highly predictable if body composition is considered, and have a mechanistic basis to purport the use of lean bodyweight (LBW) for this purpose . The aims of this study were to evaluate the predictive performance of our LBW model  in an external Caucasian and Japanese population, and to then evaluate how well LBW predicts clearance (CL) across Japanese and Caucasian populations.
Methods: Predicted LBW using the model of Janmahasatian et al  was compared to "true" LBW (measured by DEXA) in 189 Caucasian and 139 Japanese females. The predictive performance of Janmahasatian’s LBW was evaluated using the mean error (ME) and root mean square error (RMSE), as measures of bias and precision. A population pharmacokinetic analysis of "Drug A" (hepatically metabolized anti-cancer agent) after twice daily oral administration to 24 Caucasian and 20 Japanese females was performed using NONMEM VI. Data from intensive plasma sampling on days 1 and 14 were available totaling 498 concentrations. Body composition metrics such as total body weight (WT), body surface area (BSA) and LBW were investigated as covariates for CL.
Results: Evaluation of LBW yielded a ME 0.81 kg and RMSE of 3.86 kg in the Caucasian population and ME of -0.24 kg and RMSE 2.16 kg in the Japanese population. Preliminary pharmacokinetic analysis indicates a 1 compartment first-order elimination model with a transit compartment absorption system best describes the pharmacokinetics of Drug A. Between subject and between occasion variability were included on all parameters. Incorporation of WT or LBW as covariates on CL resulted in an improved model fit (p<0.05), with LBW being preferred over WT.
Conclusions: The semi-mechanistic LBW model  shows good predictive performance when evaluated prospectively in female Caucasian and Japanese populations. Preliminary pharmacokinetic results show LBW is the preferred covariate for CL when data from Caucasians and Japanese are fitted simultaneously, suggesting the incorporation of LBW as a covariate for CL in analysis of Caucasian pharmacokinetic data may allow for better prediction of drug exposure in the Japanese compared to body weight alone.
 Han, P., et al., Dosing in obesity: a simple solution to a big problem; Clin Pharmacol Ther. 2007; 82(5): 505-8
 Janmahasatian S., et al., Quantification of Lean Bodyweight; Clin Pharmacokinet. 2005; 44(10): 1051-1065.