Internal and external validation with sparse, adaptive-design data for evaluating the predictive performance of a population pharmacokinetic model of tacrolimus
Johan E. Wallin(1,2), Martin Bergstrand(1), Mats O. Karlsson(1), Henryk Wilczek(3), Christine E. Staatz(1,4)
(1) Department of Pharmaceutical Biosciences, Uppsala University, Sweden (2). PK/PD/TS, EliLilly, United Kingdom (3) Division of Transplantation Surgery, Karolinska Institute, Sweden (4) School of Pharmacy, University of Queensland, Australia
Introduction: Tacrolimus is a potent immunosuppressant agent used to prevent and treat rejection in paediatric liver transplantation. Tacrolimus has a narrow therapeutic window and displays considerable between and within-subject pharmacokinetic (PK) variability. The PK of tacrolimus change markedly in the immediate post-transplant period. We have previously developed a population PK model of tacrolimus with the intent of capturing this process(1). Commonly used simulation based diagnostics are unsuitable when using adaptive design data, but visual evaluation of the predictive performance can be performed with prediction corrected VPC (pcVPC), where observed and simulated observations are normalized based on the population prediction (2). This model has been used to suggest a revised initial dosing schedule and forms the basis for a dose adaptation tool.
Objectives: To evaluate the predictive performance of the model in comparison to two previously published models by Sam et al (3) and Staatz et al (4), by simulation based diagnostics as well as by prediction of data collected from an independent group of paediatric liver patients.
Methods: pcVPC:s were constructed using all available data and the three models. PK data from the first two weeks following liver transplantation was collected retrospectively from the medical records of 12 paediatric patients. Population and individual predicted drug concentrations were compared to measured concentrations. To evaluate the models’ potential for Bayesian forecasting in dose adaptation, individual predicted drug concentrations based on one or three prior measurements were evaluated. Predictive performance was compared by calculation of MPE and RMSE.
Results: The graphical diagnostics (pcVPCs) indicated a strong over-prediction of typical plasma concentrations for the Sam model, while the Staatz model exhibited a pronounced under-prediction during the early post operative time. The proposed model demonstrated overall satisfactory predictive performance with only a marginal and quickly transient initial over prediction. Accuracy and precision in external validation was significantly better for the proposed model compared to prior models, indicating the possibility of using the model for dose schedule development and Bayesian forecasting.
Conclusions: The proposed PK model predicted the validation data reasonably well, and was superior to the previously published models in the early post-transplantation phase.
 Wallin J et al. Population pharmacokinetics of tacrolimus in paediatric liver transplant recipients: a model to describe early post-transplantation apparent clearance, In thesis: Wallin, Johan. Dose Adaptation Based on Pharmacometric Models. 2009. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-100509
 Bergstrand M et al. Prediction Corrected Visual Predictive Checks. ACoP (2009) Abstr F7.
 Sam WJ et al. Population pharmacokinetics of tacrolimus in Asian paediatric liver transplant patients. Br J Clin Pharmacol 2000; 50 (6): 531.
 Staatz CE et al. Population pharmacokinetics of tacrolimus in children who receive cut-down or full liver transplants. Transplantation 2001; 72 (6): 1056.