A Physiologically Based Pharmacokinetic Approach to Recommend an Individual Dose of Tacrolimus in Adult Heart Transplant Recipients

Abstract

Tacrolimus is the principal immunosuppressive drug which is administered after heart transplantation. Managing tacrolimus therapy is challenging due to a narrow therapeutic index and wide pharmacokinetic (PK) variability. We aimed to establish a physiologically based pharmacokinetic (PBPK) model of tacrolimus in adult heart transplant recipients to optimize dose regimens in clinical practice. A 15-compartment full-PBPK model (Simbiology^® Simulator, version 5.8.2) was developed using clinical observations from 115 heart transplant recipients. This study detected 20 genotypes associated with tacrolimus metabolism. CYP3A5*3 (rs776746), CYP3A4*18B (rs2242480), and IL-10 G-1082A (rs1800896) were identified as significant genetic covariates in tacrolimus pharmacokinetics. The PBPK model was evaluated using goodness-of-fit (GOF) and external evaluation. The predicted peak blood concentration (C_max) and area under the drug concentration-time curve (AUC) were all within a two-fold value of the observations (fold error of 0.68-1.22 for C_max and 0.72-1.16 for AUC). The patients with the CYP3A5*3/*3 genotype had a 1.60-fold increase in predicted AUC compared to the patients with the CYP3A5*1 allele, and the ratio of the AUC with voriconazole to alone was 5.80 when using the PBPK model. Based on the simulation results, the tacrolimus dosing regimen after heart transplantation was optimized. This is the first PBPK model used to predict the PK of tacrolimus in adult heart transplant recipients, and it can serve as a starting point for research on immunosuppressive drug therapy in heart transplant patients.

Keywords: genetic polymorphism; heart transplantation; physiologically based pharmacokinetics model; tacrolimus; therapeutic drug monitoring.

Figure 1

Tacrolimus clearance of CYP3A5*3 and CYP3A4*18B genotypes in heart transplant patients (p < 0.05). The box represents the Bayesian estimate of tacrolimus clearance in the population model. The edges of the box represent the 25th and 75th percentiles; the red lines inside the box represent the median; the dotted lines represent the 2.5th and 97.5th percentiles; the plus signs show outliers.

Figure 2

Simulated blood concentration–time profiles after a 5 mg oral dose of tacrolimus in healthy CYP3A5 expressers. The thick line represents the mean simulated concentration; the orange shadow represents the 5th and 95th percentiles of simulations; the solid dots represent the mean observed data. TAC, tacrolimus.

Figure 3

Goodness-of-fit plots of the PBPK model. (A) The plot of the observations versus population prediction. (B) The plot of the observations versus individual prediction. (C) The plot of residuals versus time. (D) The plot of residuals versus population prediction.

Figure 4

Predicted blood concentration–time profiles of tacrolimus after oral dose of 2 mg (A,B) and 3 mg (C,D) in healthy CYP3A5 expressers and non-expressers. The thick line represents the mean predicted concentration; the orange shadow represents the 5th and 95th percentiles of the prediction; the solid dots represent the mean observed data. TAC, tacrolimus.

Figure 5

Predicted blood concentration–time profiles of tacrolimus in heart transplant patients. The thick line represents the mean predicted data; the orange shadow represents the 5th and 95th percentiles of the prediction; the blue solid dots represent observed data from heart transplant patients. TAC, tacrolimus.

Figure 6

Performance of the tacrolimus PBPK model. Linear analysis for external validation of the predicted and observed tacrolimus concentration. “*” represents “×”.

Figure 7

Predicted blood concentration–time profiles of tacrolimus after a single oral dose of 2.5 mg tacrolimus (A), predicted blood concentration–time profiles of tacrolimus after multidose of tacrolimus (2.5 mg, q12h for 6 days) (B). TAC, tacrolimus; EM: extensive metabolizers (CYP3A5*1*1 or *1*3 and CYP3A4*18B*18B or CYP3A4*1*18B); PM: poor metabolizers (CYP3A5*3*3 and CYP3A4*1*1). The thick line represents the mean predicted data; the yellow and blue shadow represents the 5th and 95th percentiles of the prediction.