INTERACTION BETWEEN STRUCTURAL, STATISTICAL, AND COVARIATE MODELS IN POPULATION PHARMACOKINETIC ANALYSIS

The influence of the choice of pharmacokinetic model on subsequent det ermination of covariate relationships in population pharmacokinetic an alysis was studied using both simulated and real data sets. Simulation s and data analysis were both performed with the program NONMEM. Data were simulated using a two-compartment model, but at late sample times , so that preferential selection of the two-compartment model should h ave been impossible. A simple categorical covariate acting on clearanc e was included. Initially, on the basis of a difference in the objecti ve function values, the two-compartment model was selected over the on e-compartment model. Only when the complexity of the one-compartment m odel was increased in terms of the covariate and statistical models wa s the difference in objective function values of the two structural mo dels negligible. For two real data sets, with which the two-compartmen t model was not selected preferentially, more complex covariate relati onships were supported with the one-compartment model than with the tw o-compartment model. Thus, the choice of structural model can be affec ted as much by the covariate model as can the choice of covariate mode l be affected by the structural model; the two choices me interestingl y intertwined. A suggestion on how to proceed when building population pharmacokinetic models is given.