摘要目的 建立一个指导中国汉族人群华法林用量的预测方法,并对其预测准确性进行验证.方法 选取2012年5月至2014年12月在北京医院住院或门诊就诊的正在服用华法林抗凝的中国汉族患者512例,服药指征包括人工瓣膜置换术、心房颤动及肺栓塞.直接测序分析患者维生素K环氧化物还原酶复合物1(VKORC1)和细胞色素P450 2C9(CYP2C9)基因多态性(野生型、*3型和*13型),记录患者人口统计学信息、华法林日服用量及合并使用胺碘酮情况;利用随机数字表法进行随机分组,从研究人群随机抽取384例作为建模组,剩余128例作为验证组.通过多元线性回归对建模组患者的数据进行分析,建立华法林稳定剂量预测方法;应用该方法对验证组患者的华法林稳定剂量进行预测,将结果与实际剂量进行比对,评价该方法的预测准确率.结果 患者的年龄、体重、身高、VKORC1、CYP2C9基因型以及是否合并应用胺碘酮与华法林的稳定剂量相关,利用这几项变量建立的预测方法对于华法林稳定剂量的预测效果较好(R2 =0.682,P ＜0.01),VKORC1和CYP2C9基因多态性共解释了50％以上的剂量个体差异;验证组中62.5％ (80/128)的患者预测剂量为理想预测剂量,7.81％ (10/128)的患者预测剂量过低,29.69％ (38/128)的患者预测剂量过高.结论 本预测方法可用于指导中国汉族患者华法林稳定用药量的预测.

abstractsObjective To develop and evaluate a warfarin-dosing algorithm method which can be used to guide the adjustment of warfarin maintenance dose in Chinese Han population.Methods A total of 512 patients with steady warfarin taking were recruited from Beijing Hospital during May 2012 to December 2014.Indications for warfarin prescribing included prosthetic heart valve,atrial fibrillation and pulmonary embolism.Genomic DNAs were extracted from blood samples and used for the genetic polymorphism analysis of VKORC1 and CYP2C9 (include *3 and * 13 alleles).Warfarin dose,demographic variabilities and amiodarone compliance were recorded during regular visit.These patients were randomly divided into groups using the method of random number table,384 patients were randomly selected as derivation group,the remaining 128 cases as the validation group.Using data from derivation group,a warfarin-dosing algorithm was established based on the genetic information,demographic characteristics and concomitant compliance by a multiple linear regression analysis parameter.Then the accuracy of newly developed algorithm method was further evaluated by comparing the predicting dose with the actual dose in the validation group.Results The stable dose of warfarin was tightly associated with factors like age,height,weight,VKORC1-1639G ＞ A,CYP2C9 *3,CYP2C9 * 13 and amiodarone usage.Newly developed algorithm method exhibited better prediction effect (R2 =0.682,P ＜ 0.01) as compared with that of previously reported algorithm methods.The weights of VKORC1 and CYP2C9 for predicting of warfarin dosage were estimated to more than 50％.Using this method,62.5％ of patients in the validation group could be well recognized,in which the predicting dose of warfarin was within 20％ of the actual dose,and only 7.81％ patients showed underestimated prediction warfarin dose while 29.69％ patients showed overestimated values.Conclusion Newly developed algorithm method can be used for the guidance of warfarin maintenance dose adjustment in Chinese Han population.