Pharmacokinetically Based Estimation of Patient Compliance with Oral Anticancer Chemotherapies
Background and objectives
More and more anticancer chemotherapies are now available as oral formulations. This relatively new route of administration in oncology leads to problems with patient education and non-compliance. The aim of this study was to explore the performances of the ‘inverse problem’, namely, estimation of compliance from pharmacokinetics. For this purpose, we developed and evaluated a method to estimate patient compliance with an oral chemotherapy in silico (i) from an a priori population pharmacokinetic model; (ii) with limited optimal pharmacokinetic information collected on day 1; and (iii) from a single pharmacokinetic sample collected after multiple doses.
Population pharmacokinetic models, including estimation of all fixed and random effects estimated on a prior dataset, and sparse samples taken after the first dose, were combined to provide the individual POSTHOC Bayesian pharmacokinetic parameter estimates. Sampling times on day 1 were chosen according to a D-optimal design. Individual pharmacokinetic profiles were simulated according to various dose-taking scenarios.
To characterize compliance over the n previous dosing times (supposedly known without error), 2n different compliance scenarios of doses taken/not taken were considered. The observed concentration value was compared with concentrations predicted from the model and each compliance scenario. To discriminate between different compliance profiles, we used the Euclidean distance between the observed pharmacokinetic values and the predicted values simulated without residual errors.
This approach was evaluated in silico and applied to imatinib and capecitabine, the pharmacokinetics of which are described in the literature, and which have quite different pharmacokinetic characteristics (imatinib has an elimination half-life of 17 hours, and α-fluoro-β-alanine [FBAL], the metabolite of capecitabine, has an elimination half-life of 3 hours). 1000 parameter sets were drawn according to population distributions, and concentration values were simulated at several timepoints under various compliance patterns to compare with the predicted ones. In addition, several simulation scenarios were run in order to explore the impact of the quality of the error model, interoccasion variability (IOV), error in the number of pills taken, and the performance of the compliance estimation method.
The best compliance estimate was obtained with pharmacokinetic samples taken 5 hours after the last dose. Performance of the method varied between simulation scenarios. In both the imatinib and capecitabine basic simulations, patient compliance was correctly estimated on the two last scheduled doses (with better results for imatinib). The magnitude of the error model also had a great impact on the quality of the compliance estimate.
We highlight the effect of three parameters on the quality of compliance estimates based on limited pharmacokinetic information: the plasma elimination half-life, interdose interval and magnitude of the error model. Nevertheless, the pharmacokinetic method is not informative enough and should be used with electronic monitoring, which provides additional information on compliance. Our method will be used in a future phase IV clinical trial where the relationships between compliance, efficacy and tolerability will be assessed.
- 4.Grober SE, Carpenter RC, Glassman M, et al. A comparison of patients’ perceptions of oral cancer treatments and intravenous cancer treatments: what the health care team needs to know [abstract no. 3000]. Proc Am Soc Clin Oncol 2003; 22: 746Google Scholar
- 5.Catania C, Didier F, Sbanotto A, et al. Fully oral chemotherapy regimens: patient’s or physician’s preference [abstract no. 3122]. Proc Am Soc Clin Oncol 2003; 22: 777Google Scholar
- 24.Vrijens B, Tousset E, Rode R, et al. Successful projection of the time course of drug concentration in plasma during a 1-year period from electronically compiled dosing-time data used as input to individually parameterized pharmacokinetic models. J Clin Pharmacol 2005 Apr; 45(4): 461–7PubMedCrossRefGoogle Scholar
- 26.Duffull S, Denman M, Eccleston J, et al. WinPOPT user guide (version 1.1). Dunedin: University of Otago, 2006Google Scholar
- 28.Beal SL, Sheiner LB. NONMEM user’s guides. San Francisco (CA): NONMEM Project Group, University of California, 1992Google Scholar
- 29.Insightful Corporation. S-Plus 6 user’s guide. Seattle (WA): Insightful Corporation, 2002Google Scholar