The Establishment Of Warfarin PK/PD Model Based NONlinear Mixed Effects Model

ObjectiveTo establish warfarin population pharmacokinetics/pharmacodynamics (PK/PD) model and provide reference for clinical individual administration.Methods1. Establish RP-HPLC-UV methodology to determinate warfarin plasma concentrations.2. Determine the experimental conditions of the polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method to identify CYP2C91075A/C and VKORC1-1639G/A genotype.3. Collecting inpatient cases of warfarin therapy, taking different blood concentrations of2-4points, using RP-HPLC-UV to determine the plasma concentrations of warfarin, PCR-RFLP method was used to identify CYP2C9-1075and VKORC1-1639allele. Meanwhile record patient information.4. Study the influence among patient gender, age, weight, CYP2C9and VKORC1genotype, concomitant medications, liver and kidney function, disease status and warfarin clearance and anticoagulant effects by nonlinear mixed-effects model (NONMEM);then establish warfarin population pharmacokinetics/pharmacodynamics (PPK/PD) model.5. Evaluate the model fitting by plots; Evaluate internal validity and stability of the final model for warfarin by Bootstrap. Evaluate the predictive ability of the model by normalized predictive distribution error(NPDE) Results1. The established RP-HPLC-UA plasma concentrations of warfarin, diazepam as internal standard, retention time12.8mino The regression equationa of standard curve was Y=0.1444X+1.0264(r=0.9994), over the range of48.0～2500.0ng·mL-1.The detection limit was16ng·mL-1(S/N>3).Recoveries101%-103%. The intra-and inter-day RSD were less than2.10%.2.The results of genotyping CYP2C9and VKORC1of the73patients is CYP2C9*1/*1,63cases;*1/*3,10cases;VKORC1-1639AA,62cases,GA,10cases; GG, one case.3. Based on190concentrations、263INR monitoring values and other information of the73patients, the warfarin population pharmacokinetics/pharmacodynamics final model is Among them, the patient’s weight, CYP2C9and VKORC1genotype are major influence factors in warfarin pharmacokinetic/pharmacodynamic parameters.4.Model fitting results shows the goodness fit of the final model. Goodness of fit for Plasma concentrations and INR superior than the simplest model. The estimated value of final model was in the95%CI by Bootstrap method and related deviations were less than6.30%. NPDE showed homogeneity of variance and a normal distribution.Conclusion1. The established RP-HPLC-UV method is accurate, reproducible, and simple, suitable for the concentrations determination of warfarin in human plasma and study drug action. 2. Weight, CYP2C9and VKORC1genotyping are the main factors that affect warfarin PK/PD parameters.3. The established warfarin population PK/PD model is stable and effective and of good goodness and predictive ability, can provide a reference to develop individualized dosing regimens.