BACKGROUND

Little is known about the population pharmacokinetics (PPK) of tacrolimus (TAC) in pediatric patients with primary nephrotic syndrome (PNS). In this study, we aimed to compare the predictive performance between nonlinear and linear PK models and investigate the significant factors influencing TAC PK characteristics in pediatric PNS.

METHODS

Data were obtained from 71 pediatric patients with PNS, along with 525 TAC trough concentrations at steady-state. Patients' demographic, medical, and treatment details were collected. Genetic polymorphisms of CYP3A41G, CYP3A53, and ABCB1-C3435T were analyzed. The PPK models were developed using nonlinear mixed-effects model (NONMEM®) software. Two modeling strategies, linear compartmental and nonlinear Michaelis-Menten (MM) models, were evaluated and compared.

RESULTS

Body weight, age, daily dose of TAC, co-therapy drugs (including azole antifungal agents and diltiazem), and CYP3A53 genotype were the important factors in the final linear model (one-compartment model), whereas only body weight, co-therapy drugs, and CYP3A53 genotype were the important factors in the final nonlinear MM model. Apparent clearance and volume of distribution in the final linear model were 7.13 L/h and 142 L, respectively. The maximal dose rate (V) of the nonlinear MM model was 1.92 mg/day and the average concentration at steady-state at half-V (K) was 1.98 ng/mL. The nonlinear model described the data better than the linear model. Dosing regimens were proposed based on the nonlinear PK model.

CONCLUSION

Our findings demonstrated that the nonlinear MM model showed better predictive performance than the linear compartmental model, providing reliable support for optimizing TAC dosing and adjustment in children with PNS.