Aim: To develop a population pharmacokinetic (PopPK) model of tacrolimus in healthy Chinese volunteers and liver transplant recipients for investigating the difference between the populations, and for potential individualized medication. Methods: A set of 1100 sparse trough concentration data points from 112 orthotopic liver transplant recipients, as well as 851 dense data points from 40 healthy volunteers receiving a single dose of tacrolimus (2 mg, po) were collected. PopPK model of tacrolimus was constructed using the program NONMEM. Related covariates such as age, hepatic and renal functions that were potentially associated with tacrolimus disposition were evaluated. The final model was validated using bootstrapping and a visual predictive check. Results: A two-compartment model of tacrolimus could best describe the data from the two populations. The final model including two covariates, population (liver transplant recipients or volunteers) and serum ALT (alanine aminotransferase) level, was verified and adequately described the pharmacokinetic characteristics of tacrolimus. The estimates of V2/F, Q/F and V3/F were 22.7 L, 76.3 L/h and 916 L, respectively. The estimated CL/F in the volunteers and liver transplant recipients was 32.8 and 18.4 L/h, respectively. Serum ALT level was inversely related to CL/F, whereas age did not influence CL/F. Thus, the elderly (≥65 years) and adult (<65 years) groups in the liver transplant recipients showed no significant difference in the clearance of tacrolimus. Conclusion: Compared with using the sparse data only, the integrating modeling technique combining sparse data from the patients and dense data from the healthy volunteers improved the PopPK analysis of tacrolimus.

We aimed to explore the new biomarkers influencing tacrolimus in vivo behavior in Chinese liver transplant recipients. A total of 418 drug concentration samples of 41 liver transplant patients were collected for modeling. A population pharmacokinetic model was developed using the nonlinear mixed-effects modeling approach. The potential covariates, such as postoperative day (POD), age, body weight, hepatic and renal function, and recipient genetic polymorphisms (ABCB1, CYP3A4, CYP3A5, NR1I2) were evaluated using forward-inclusion and backward-elimination methods. A 1-compartment model was used describing the in vivo behavior of tacrolimus in liver transplant patients. The estimates of CL/F and V/F were 8.88 L/h and 495.82 L, respectively. Two covariates, POD and NR1I2 rs2276707 genotypes, were incorporated into the final population pharmacokinetic model, and they could significantly impact the CL/F: CL/F (L/h) = 8.88 × (POD/16) 0.18 × e 0.91 × NR1I2 × e ηCL . The model evaluation and validation indicated a stable and precise performance of the final model. The functional annotation using ENCODE data indicated that rs2276707 was located on the higher peak of the H3K4Me1 and H3K4Me3 histone marker. To our knowledge, this is the first report indicating NR1I2 rs2276707 genotypes is another biomarker impacting tacrolimus clearance in liver transplant recipients. The NR1I2 gene polymorphism may affect the in vivo behavior of tacrolimus by regulating gene expression.