| Objective:To establish a dose prediction model for warfarin pharmacogenomics by analyzing the clinical data,CYP2C9,VKORC1 gene test results and the dosage of warfarin in a third-class hospital in Jilin Province.The model was used to predict the dosage of patients,and the difference between the actual dosage and the predicted dosage was compared to evaluate the accuracy of the model.Methods:A total of 190 patients of Han nationality in Jilin Province who were hospitalized in Jilin People’s Hospital from June 2016 to July 2018 and took warfarin for a long time(>1 year)were selected as subjects.General information(including age,sex,height,weight,etc.),living habits(including smoking,drinking,etc.),clinical data(combined use of drugs,daily average dose of warfarin in patients),CYP2C9 and VKORC1 gene test results were collected.The database was established by EpiData3.1 software,and SPSS 22 statistical software was used for data analysis.The measurement data were described by means of standard deviation(mean±SD),and the counting data were described by rate and composition(%).The variables were measured by t test and the chi square test was used for the enumeration variables.The patients were divided into modeling group(143 cases)and validation group(47 cases).The actual dosage of the modeling group was dependent variable when the treatment window was INR2.0-3.0.The multiple linear regression model was established to predict the stable dose of warfarin.Patient dose.Paired samples t test was used to compare the actual dose and the predicted dose of warfarin in the validation group.Calculate the correlation coefficient in the group to test the consistency between predicted dose and actual dose.Take P<0.05 as the boundary value with statistical difference.Results:1.Warfarin daily stable dose model was correlated with age,BSA,CYP2C9,VKORC1 and other factors(P<0.05),but not with gender,smoking,drinking,and combination of drugs(P>0.05).2.CYP2C9 and VKORC-1639 alleles were polymorphic in 190 patients of Han nationality in Jilin province.The frequencies of CYP2C9*1 and*3 alleles were96.32%and 3.68%respectively.The frequencies of VKORC1-1693A and G alleles were 91.05%and 8.95%respectively,and the genotype distribution frequency obeyed Hardy-Weinberg’s law of genetic balance.It shows that the sample population has group representation.3.Based on the information of 143 patients in the modelling group,after stepwise multivariate linear regression analysis and excluding unobvious factors,the daily stable dose model of warfarin with the target INR value of 2-3.0 in Jilin area was established:D(mg/day)=[1.369-0.003×(Age,in year)+0.241(BSA,m~2)-0.254×CYP2C9*3(*1*1=0,*1/*3=1)+0.066*VKORC1-1639A>G(AA=0,AG=1)]~2.This model can explain 69.5%individual differences in daily dose of Han nationality in Jilin province(P<0.001).4.The average prediction error between the dose predicted by the model established in this study and the actual dose predicted by the validation group was0.41.The average prediction error between the dose predicted by the IWPC model and the actual dose predicted by the validation group was 0.47.Therefore,the prediction of warfarin dose by this model is better than IWPC model.Conclusion:1.Among 190 patients taking warfarin anticoagulant therapy in Jilin area,CYP2C9 and VKORC-1639 gene loci were polymorphic,and the genotype distribution frequency was representative of the population,which accorded with Hardy-Weinberg equilibrium.2.The steady dose model of warfarin day obtained from this study:D(mg/day)=D(mg/day)=[1.369-0.003×(Age,in year)+0.241(BSA,m~2)-0.254×CYP2C9*3(*1*1=0,*1/*3=1)+0.066*VKORC1-1639A>G(AA=0,AG=1)]~23.The model can explain 69.6%of the individual differences in warfarin doses,CYP2C9 and VKORC1 are one of the reasons for the stable dose of warfarin.4.The dose model can guide the dosage of warfarin daily dose in Jilin area. |
PDF Full Text Request