皖南地区汉族人群华法林稳定剂量模型预测准确性的评价与建立

目的:比较并评价4种华法林稳定剂量预测模型的预测准确性并建立华法林剂量预测模型。方法:收集483名服用华法林患者的临床资料，检测患者CYP2C9*3和VKORC1基因型，以预测百分比和平均绝对误差分析4种华法林稳定剂量预测模型的准确性；将纳入影响华法林稳定剂量的相关因素进行多元线性回归分析，得到相应的剂量预测模型。结果:CYP2C9*3/*3型患者华法林稳定剂量为(0.83±0.19) mg·d^-1,显著低于*1/*1型和*1/*3型患者(P<0.05),VKORC1 GG型患者华法林稳定剂量为(4.17±1.49) mg·d^-1,显著高于AA型和GA型患者(P<0.05);4种华法林剂量预测模型的预测剂量与实际剂量差异均显著相关(P<0.01),IWPC模型的相关性最好，r=0.519;MAE最低的是TAN模型，为(0.70±0.53) mg·d^-1,指南模型的MAE最高，为(0.86±0.60) mg·d^-1;华法林稳定剂量模型建立结果D(mg·d^-1)=...

Evaluation and establishment of prediction accuracy of warfarin stable dose model in Han population in southern Anhui

OBJECTIVE To compare and evaluate the prediction accuracy of four warfarin stable dose prediction models, and to establish the warfarin dose prediction model.METHODS The clinical data of 483 patients taking warfarin were collected, CYP2C9*3 and VKORC1 genotypes were detected, and the accuracy of four warfarin stable dose prediction models was analyzed by prediction percentage and mean absolute error; multiple linear regression analysis was carried out on the related factors affecting the stable dose of warfarin, and the corresponding dose prediction model was obtained.RESULTS The stable dose of warfarin in the patients with CYP2C9*3/*3 was (0.83±0.19) mg·d^-1, which was significantly lower than that in the patients with *1/*1 and *1/*3(P＜0.05), and that in the patients with VKORC1GG was (4.17±1.49) mg·d^-1, which was significantly higher than that in the patients with AA and GA (P＜0.05); the predicted doses of the four warfarin dose prediction models were significantly correlated with the actual doses (P＜0.01), and the correlation of IWPC model was the best (r=0.519); the lowest MAE was (0.70±0.53) mg·d^-1 in TAN model and the highest MAE was (0.86±0.60) mg·d^-1 in the guideline model; establishment result of warfarin stable dose model was as follows: D(mg·d^-1)=2.826-0.018×Age+0.544×BSA-0.322×CYP2C9+0.482×VKORC1-0.336 ×amiodarone-0.801×atrial fibrillation.CONCLUSION CYP2C9 and VKORC1 gene polymorphisms have certain influence on the stable dose of warfarin. The established warfarin dose prediction model can accurately predict the stable dose, optimize the dosing schedule, and provide guidance for clinical individualized medication.