急性心肌梗死患者住院期间死亡风险模型的构建  

Establishment of a prediction model for death during hospitalization in patients with acute myocardial infarction

作　　者：欧阳建 厉锦巧 徐淑英 王斌 陈剑平 Jian Ouyang;Jinqiao Li;Shuying Xu;Bin Wang;Jianping Chen(Department of Emergency Medicine,the Affiliated Dongyang Hospital of Wenzhou Medical University,Dongyang 322100,China)

机构地区：[1]温州医科大学附属东阳医院急诊科,浙江东阳322100

出　　处：《中华危重症医学杂志（电子版）》2023年第2期116-122,共7页Chinese Journal of Critical Care Medicine:Electronic Edition

摘　　要：目的识别住院期间高死亡风险的急性心肌梗死患者,尽早进行干预,减少病死率。方法收集2013年6月1日至2022年5月31日期间在温州医科大学附属东阳医院住院的1 251例急性心肌梗死患者的临床资料,分析该类患者住院期间死亡的独立危险因素,绘制列线图。采用受试者工作特征(ROC)曲线评估模型的判别能力,同时用校准图评估模型的校准度和决策曲线分析法(DCA)评估预测模型的临床有效性,并应用bootstrap法进行内验证。结果逐步回归分析显示,年龄[比值比(OR)= 1.060,95%置信区间(CI)(1.035,1.087),P < 0.001]及入院后24 h内肌酐[OR = 1.003,95%CI(1.001,1.005),P = 0.012]、乳酸[OR = 1.190,95%CI(1.064,1.332),P = 0.002]、肌酸激酶同工酶[OR = 1.003,95%CI(1.001,1.004),P < 0.001]、B型脑钠肽前体[OR = 1.006,95%CI(1.004,1.008),P < 0.001]及白细胞计数[OR = 1.055,95%CI(1.006,1.105),P = 0.024]的最大值为急性心肌梗死患者住院期间死亡的独立危险因素。建模后ROC曲线下面积为0.851(P < 0.001),校准图的P值为0.896,Brier scaled为0.055,calibration slope为1.000,R2为0.328,DCA曲线在两条极端曲线之上。内验证曲线提示校正曲线与理想曲线之间重合度高。结论入院后24 h内肌酐、乳酸、肌酸激酶同工酶、B型脑钠肽前体、白细胞计数的最大值及发病的年龄为急性心肌梗死患者住院期间死亡的独立危险因素,由上述指标建立的模型在预测急性心肌梗死患者住院期间死亡风险中有较好的意义。Objective To identify the acute myocardial infarction patients with bad outcomes during hospitalization and to help clinical physicians to take intervention measures to reduce the mortality.Methods Clinical data were collected from 1251 patients with acute myocardial infarction who were hospitalized in the Affiliated Dongyang Hospital of Wenzhou Medical University between 1st Jan 2013 and 31th May 2022.Independent risk factors of death during hospitalization were screened.Based on the screened variables,a nomogram prediction model was established.Then the model was evaluated for its prediction power by a receiver operating characteristic(ROC)curve,calibration accuracy by a GiViTI calibration curve and clinical effectiveness by decision curve analysis(DCA).Finally,the established prediction model was validated by a bootstrap assay.Results Stepwise regression analysis showed that age[odds ratio(OR)=1.060,95%confidence interval(CI)(1.035,1.087),P<0.001]and the maximum values of creatinine[OR=1.003,95%CI(1.001,1.005),P=0.012],lactic acid[OR=1.190,95%CI(1.064,1.332),P=0.002],creatine kinase isoenzymes[OR=1.003,95%CI(1.001,1.004),P<0.001],B-type natriuretic peptide precursor[OR=1.006,95%CI(1.004,1.008),P<0.001]and white blood cell[OR=1.055,95%CI(1.006,1.105),P=0.024]within 24 hours after hospitalization were significantly associated with death in hospital among patients with acute myocardial infarction.The area under the ROC curve of the prediction model was 0.851(P<0.001),with P value of 0.896,calibration slope of 1.000,R2 of 0.328 and Brier scaled value of 0.055 in the calibration curve and with the DCA curve above the two extreme curves.In validation using bootstrap,the bias-corrected curve was closed to the idea curve.Conclusions Age and the maximum values of creatinine,lactic acid,creatine kinase isoenzymes,B-type natriuretic peptide precursor and white blood cell within 24 hours after hospitalization are independent risk factors of death in hospital among patients with acute myocardial infarction.The established

关 键 词：心肌梗死 危险因素 预测模型 预后 BOOTSTRAP法 列线图 逐步回归分析 

分 类 号：R542.22[医药卫生—心血管疾病]