Methods and Applications of Statistics in Clinical Trials, Volume 2: Planning, Analysis, and Inferential Methods

Chapter 33

Population Pharmacokinetic and Pharmacodynamic Methods

Jürgen B. Bulitta and Nicholas H.G. Hoi ford

33.1 Introduction

Population pharmacokinetics (PK) and pharmacodynamics (PD) originated in the 1970s and have been adopted extensively in drug development and clinical pharmacology [1]. Population PKPD can incorporate all available and relevant data from a clinical study in a single data analytical step [2,3]. Therefore, population PKPD methods provide a sound basis for data analysis and for interpretation of results. Population PKPD models allow one to study the effect of multiple covariates like renal function, body size, body composition, and age on PK and PD simultaneously. Population PKPD can be distinguished from hypothesis testing procedures such as analysis of variance (ANOVA) and analysis of covariance (ANCOVA) in that population PKPD can account for the full time course of PK, PD, and disease progression. This difference becomes especially clear when attempting to make predictions. ANOVA-based methods cannot be extended to predict unstudied circumstances.

As the time course of concentrations and the time course of effects are usually non-linear, almost all PKPD models are nonlinear. ANOVA and ANCOVA can only account for nonlinearity by transformation of the independent or dependent variable(s).

Because a substantial amount of between-subject variability (BSV) often occurs, it is very important to account for BSV when describing PK and PD parameters. If the goal ...