A population pharmacokinetic model with time-dependent covariates measured with errors |
| |
Authors: | Li Lang Lin Xihong Brown Morton B Gupta Suneel Lee Kyung-Hoon |
| |
Affiliation: | Division of Biostatistics, Indiana University, Indianapolis, Indiana 46254, USA. lali@iupui.edu |
| |
Abstract: | We propose a population pharmacokinetic (PK) model with time-dependent covariates measured with errors. This model is used to model S-oxybutynin's kinetics following an oral administration of Ditropan, and allows the distribution rate to depend on time-dependent covariates blood pressure and heart rate, which are measured with errors. We propose two two-step estimation methods: the second-order two-step method with numerical solutions of differential equations (2orderND), and the second-order two-step method with closed form approximate solutions of differential equations (2orderAD). The proposed methods are computationally easy and require fitting a linear mixed model at the first step and a nonlinear mixed model at the second step. We apply the proposed methods to the analysis of the Ditropan data, and evaluate their performance using a simulation study. Our results show that the 2orderND method performs well, while the 2orderAD method can yield PK parameter estimators that are subject to considerable biases. |
| |
Keywords: | Differential equations Laplace approximation Measurement error Nonlinear mixed models Pharmacokinetics Two-compartment model |
本文献已被 PubMed 等数据库收录! |
|