Abstract
The safe and effective use of medicines requires that prescribers use the right drug for the right patient at the right dose. The choice of dosing regimen requires a quantitative understanding of the magnitude of the drug effect for any given dose, the time course over which desired and adverse effects are expected to occur, and how variability between patients will impact dose requirements. The methods used by prescribers to achieve individualised dosing in practice are poorly understood and it is likely that a trial and error process prevails for many drugs. A model-based approach to drug dosing provides a means of predicting drug response in individual patients and, therefore, constitutes a useful tool for designing safe and effective dosing regimens in clinical practice.
A model-based approach was explored for the dosing of warfarin, simvastatin and allopurinol. Each drug presented a different challenge in the clinical setting with regards to safe and effective dosing and this necessitated the use of different pharmacometric methodologies in each case.
For warfarin, a Bayesian forecasting method for dose individualisation was developed. The method required the identification and evaluation of a suitable prior model and the development of a novel optimal International Normalised Ratio (INR) sampling design for Bayesian parameter control. The performance of this design was evaluated using simulation-estimation techniques. It was predicted that this method will result in a substantial improvement in INR control and time-in-the-therapeutic range compared to currently dosing methods.
For simvastatin, a simulation based study using a published pharmacokinetic-pharmacodynamic model was conducted. This analysis addressed a clinical research question related to the practice of dosing simvastatin at bedtime. The model predicted that circadian low-density lipoprotein (LDL) production had a negligible impact on simvastatin effect and that dosing in the morning should be considered for patients who may be less compliant with bedtime dosing.
For allopurinol, a novel population parent-metabolite model was developed. A covariate analysis found that renal function, fat free mass and diuretic use determined the differences in allopurinol and oxypurinol exposure between patients. This pharmacokinetic model provided the basis for an integrated PKPD model, intended to describe the effect of allopurinol on serum urate. While a suitable PD model could not be developed with the data available, optimal design methodologies were used to evaluate future study designs and alternative models for the allopurinol-urate effect.
A PKPD model-based approach to inform rational drug dosing was successfully demonstrated for warfarin, simvastatin, and allopurinol. By quantifying the magnitude and time course of drug effects and by elucidating the patient characteristics which determine drug response between individuals, the model-based approach to drug dosing provides a useful tool to aid the safe and effective use of medicines in clinical practice.