Application Of Quantitative Pharmacological Modeling To Improve Drug Therapy In Older Adults

Abstract

Consideration of the elderly as a special population in pharmacological treatment is often overlooked. The consequence of this is an increased likelihood of therapeutic failure or adverse drug effects in elderly patients. Particularly given the huge disease burden and challenges with clinical trial recruitment in this population, the development and utilization of approaches that allow better understanding of pharmacological differences in older patients is important. This dissertation aimed to use model-based approaches to examine and characterize age-related differences in the drug disposition and tolerability of topiramate and gabapentin, two neurology drugs commonly used in elderly patients. In alignment with the support for model-based drug development in the elderly population, a modeling and simulation tool that helps facilitate the workflow of pharmacometrics tasks was also developed, as part of this dissertation. Based on data pooled from three randomized, crossover studies in healthy subjects, we characterized the relationship between topiramate plasma concentration and cognitive impairment, as measured by the Symbol Digit Modality Test (SDMT) using population pharmacokinetic-pharmacodynamic modeling. In this analysis, both age and number of prior test administered were important determinants of the baseline SDMT score, with an estimated decrease of 1.13% in baseline SDMT score per year increase in age, and 7% improvement in test scores after two prior testing. Differences in sensitivity to cognitive effect of topiramate based on age, could not be discerned. In the population pharmacokinetic analysis of gabapentin, data comprising of a wide range of renal function and ages from patients (aged 18 years and above) was pooled and the effects of factors such as renal function, age, dose, total daily dose etc. on pharmacokinetic parameters examined, with the objective of providing more information for dosing recommendations, especially in elderly. This data was sufficiently described by a one-compartment pharmacokinetic model, with saturable absorption of gabapentin characterized through a nonlinear function of dose on the extent of absorption. Similar to earlier reports, gabapentin clearance was dependent on renal function in this analysis, and there was no additional effect of age on clearance after this has been accounted for in the model. In addition, no pharmacokinetic differences between nursing home and community dwelling elderly were found, and no improvement in the model fit was observed with inclusion of age as a covariate on other parameters. Hence, from this analysis, no adjustment of dose by age in adults appears to be warranted. Lastly, we developed the R package, Phxnlme, as an auxiliary tool to address existing limitations in the population pharmacokinetic/pharmacodynamics software, Phoenix NLME. With the development and sharing of this R package on a public repository, additional ease is now provided for checking of model assumptions and exploring relationships in the output through graphical visualization. Furthermore, users will also be able to multi-task modeling activities using Phoenix NLME through the use of this R package. This is a functionality that was not readily available with the use of Phoenix NLME as a desktop software. Through this R package release, we hope to facilitate use of model-based drug development, particularly in the elderly population. In line with this, video demonstrations for use of the R package was made and a manual for its use is provided.