基于药物遗传学的华法林剂量预测模型在甘肃地区心脏瓣膜置换术317例中的验证

Validation of Warfarin Dose Prediction Models Based on Pharmacogenetics in 317 Cases of Patients Undergoing Heart Valve Replacement in Gansu Province

目的探讨6种华法林剂量预测模型对甘肃地区心脏瓣膜置换术后患者华法林稳定剂量预测的准确性。方法选取317例行心脏瓣膜置换术(HVR)后使用华法林的患者,进行相关基因型检测,收集华法林给药剂量及患者临床信息。比较华法林实际稳定剂量与6种模型(Zhong等人、Chen等人、Wei等人、李等人、IWPC和Gage模型)的预测剂量,用预测剂量落入理想剂量(实际稳定剂量±20%)患者的百分比和平均绝对误差(MAE)来评估各模型的预测准确性。根据华法林稳定剂量范围将患者分为低、中间和高剂量组,评估各模型在不同剂量组的预测表现。结果在比较的6种模型中,Chen等人、李等人和Gage模型在稳定剂量整体预测准确性以及中间剂量范围和高剂量范围均显示出较高的预测准确性和较低的MAE。李等人的模型在低剂量组亦有较好表现。结论李等人模型预测准确性最高,更适用于甘肃地区HVR患者华法林剂量预测。

Objective To investigate the accuracy of six warfarin dose prediction models on the stable dose prediction of warfarin in patients after heart valve replacement(HVR)in Gansu Province.Methods A total of 317 patients who used warfarin after HVR and underwent related genotype tests were collected and recorded.To compare the actual stable dose of warfarin with the dose predicted by six models(models of Zhong et al.,Chen et al.,Wei et al.,Li et al.,IWPC and Gage)respectively,the predictive accuracy of all selected algorithms was assessed by the percentage of patients whose predicted dose was in the idea dose range(actual stable dose±20%)and the mean absolute error(MAE).Patients were stratified according to the dose range of warfarin so that the performance of six models would be evaluated further.Results Among the six algorithms compared,the algorithms of Chen et al.,Li et al.,and Gage showed higher accuracy and lower MAE in patients with stable dose prediction as well as the medium and high dose range.Conclusion Among the six models,Chen et al.,Li et al.,and Gage model were more suitable for warfarin dose prediction for HVR patients in Gansu Province,among which the model of Li et al.predicted warfarin dose most accurately.