Population pharmacokinetic/pharmacodynamic modeling of warfarin by nonlinear mixed effects model / 药学学报

The study aimed to establish a population pharmacokinetic/pharmacodynamic (PPK/PD) model of warfarin. PCR-RFLP technique was used to genotype the CYP2C9 and VKORC1 polymorphisms of 73 patients. RP-HPLC-UV method was used to determine the 190 plasma concentrations of warfarin. Application of NONMEM, the clinical information and 263 international normalized ratio (INR) monitoring data were used to investigate the effect of genetic, physiological, pathological factors, other medication on clearance and anticoagulant response. The final model of warfarin PPK/PD was described as follows: CL = θCL · (WT/60)θWT · θCYP · eηCL (if CYP2C9*1/*1, θCYP = 1; if *1/*3, θCYP = 0.708); EC50 = θEC50 · θVKOR · eηEC50 (if VKORC1- 1639AA, θVKOR = 1; if GA, θVKOR = 2.01; V = θV; K(E0) = θK(E0); Emax = θEmax; E0 = θE0 · eηE0. Among them, the body weight (WT), CYP2C9 and VKORC1 genotype had conspicuous effect on warfarin PK/PD parameters. The goodness diagnosis, Bootstrap, NPDE verification showed that the final model was stable, effective and predictable. It may provide a reference for opitimizing the dose regimen of warfarin.

Population pharmacokinetic modeling of flurbiprofen / 药学学报

The paper is to report the establishment of a population pharmacokinetic model for flurbiprofen (FP), an active metabolite of flurbiprofen axetil (FA). 246 FP serum concentration and clinical data were perspectively collected from 23 general anaesthesia patients receiving FA intravenously before operation in Dentofacial Surgery and Otorhinolaryngology Department of the First Affiliated Hospital of Fujian Medical University. Population pharmacokinetic data analysis was performed using NONMEM software. The measure of Bootstrap was applied for internal validation, while Visual Predictive check was adopted for external validation. The data of FP correspond with two-compartment model. The body weight (WT) had conspicuous effect on clearance and volume of central compartment, while sex, age and daily dose of administration had no marked effect on pharmacokinetic parameter of FP. The basic model was described as follows: CL (L x h(-1)) = 1.28x EXP(ETA(1)), V1 (L) = 5.03x EXP(ETA(2)), Q (L x h(-1)) = 8.5 x EXP(ETA(3)), V2 (L) = 4.39 x EXP(ETA(4)). The final model was described as follows: CL (L x h(-1)) = 1.32 x (WT/60) x EXP(ETA(1)), V1 (L) = 5.23 x (WT/60) x EXP(ETA(2)), Q (L x h(-1)) = 8.45 x EXP(ETA(3)), V2 (L) = 4.37 x EXP(ETA(4)). The population typical value of CL, V1, Q and V2 were: 1.32 L x h(-1), 5.23 L, 8.45 L x h(-1) and 4.37 L, respectively. Bootstrap and visual predictive check show that the final model of FP is stable, effective and predictable. A novel population pharmacokinetic model is developed to estimate the individual pharmacokinetic parameter for patients intravenous injecting FA in terms of patients' characteristics and dosing history, and to design a prior dosage regimen.