Identifying Drug-Drug Interactions: Power of the Population Approach
This talk will provide an overview of the emerging role of modeling and simulation (M&S) in drug development, and will present an example of how M&S could be used to identify potential drug-drug interactions.
M&S is assuming an increasingly central role in drug development. This enhanced role of M&S has largely been due to the recognition by the FDA  that the application of model-based approaches  have the potential to improve the quality of new drug applications (NDAs), as well as improve the efficiency (time and cost) of drug development. The models employed may be broadly classified as pharmacokinetic or pharmacodynamic models, based on the response of interest. Pharmacokinetic models describe the concentration-time profile of drugs and metabolites in tissues of interest, whereas pharmacodynamic models describe physiological responses to a drug. It is of interest to characterize both the central tendency of a response, as well as the variability in the response, as a means of ensuring that a drug is safe and efficacious for the average patient as well as for susceptible individuals. Nonlinear mixed effects (“population”) models are now the method of choice for characterizing pharmacokinetic and pharmacodynamic responses. These pharmacostatistical models are specified in terms of fixed effects and two or more levels of random effects. The fixed effects characterize the central tendency of the response, while the random effects characterize, interoccasion, interindividual, and within subject (residual error) variability.
A major advantage of population pharmacokinetic models is that the parameters of these models can be estimated from sparsely sampled concentration-time data. This allows drug exposure to be assessed in subjects without the burden of intensive blood sampling, and facilitates the acquisition of pharmacokinetic data from a larger number of subjects than would otherwise have been practical. A common application of these models is to identify subject specific factors (covariates such as age, body weight, or creatinine clearance) that affect the response in a systematic manner. An example is presented in which concomitant medication is used as a covariate, and the statistical power to detect a drug-drug interaction is assessed for alternative clinical trial designs.