老年人工心脏瓣膜置换患者术后心脏骤停风险的Nomogram预测模型建立  

Nomogram prediction model of cardiac arrest after artificial heart valve replacement in the elderly patients

作　　者：叶菊 陆真 Ye Ju;Lu Zhen(Cardiac and Great Vessel Surgical Monitoring,Jiangsu Provincial People′s Hospital,the First Affiliated Hospital of Nanjing Medical University,Nanjing 210000,China)

机构地区：[1]南京医科大学第一附属医院江苏省人民医院心脏大血管外科监护室,江苏南京210000

出　　处：《中国急救医学》2023年第4期312-318,共7页Chinese Journal of Critical Care Medicine

摘　　要：目的分析老年人工心脏瓣膜置换术后发生心脏骤停(CA)的危险因素,并构建Nomogram风险预测模型。方法选取我院2019年5月至2022年4月收治的986例行人工心脏瓣膜置换术老年患者设为训练集,另选取2022年4月至2023年3月收治的175例行人工心脏瓣膜置换术老年患者设为验证集,训练集用于Nomogram模型的构建、评价和内部验证,验证集用于检验模型外部的重复性验证。训练集患者根据是否发生CA将其分为CA组和非CA组。分析所选患者的临床资料,采用单因素、Lasso和Logistic回归分析筛选人工心脏瓣膜置换术后老年患者发生CA的危险因素,根据独立危险因素构建Nomogram模型并进行拟合优度检验。结果训练集中986例老年人工心脏瓣膜置换术后患者发生CA有50例,其发生率为5.07%;体重指数(BMI)>28 kg/m 2(OR=2.952,95%CI 1.373~6.347,P=0.006)、术后胸痛(OR=3.437,95%CI 1.604~7.362,P=0.001)、左室射血分数(LVEF)≤40%(OR=0.348,95%CI 0.163~0.745,P=0.007)、主动脉断流时间>120 min(OR=3.505,95%CI 1.618~7.594,P=0.032)、术后血流动力学紊乱(OR=3.064,95%CI 1.418~6.622,P=0.004)、术后低氧血症(OR=3.117,95%CI 1.436~6.768,P=0.004)和脑利钠肽(BNP)>5000 pg/mL(OR=2.821,95%CI 1.333~5.969,P=0.007)是老年患者行人工心脏瓣膜置换术后发生CA的独立危险因素。Nomogram模型显示,BMI>28 kg/m 2为86分、术后胸痛为99分、LVEF≤40%为85分、主动脉断流时间>120 min为100分、术后血流动力学紊乱为89分、术后低氧血症为91分、BNP>5000 pg/mL为83分。列线图模型一致性指数(C-index)为0.841,H-L拟合优度检验为χ^(2)=11.5248,P=0.082,校正曲线趋近于理想曲线,阈值概率在5%~82%范围内时净获益值较高,内部验证的AUC为0.812,同时,外部验证ROC曲线AUC为0.808。结论构建的Nomogram模型能够较为可靠地评估老年患者行人工心脏瓣膜置换术后发生CA的风险。Objective To analyze the risk factors of cardiac arrest(CA)after artificial heart valve replacement in the aged,and to construct a Nomogram risk prediction model.Methods 986 elderly patients receiving artificial heart valve replacement in our hospital from May 2019 to April 2022 were selected as the validation set,and 175 elderly patients receiving artificial heart valve replacement from April 2022 to March 2023 were selected as the validation set.The training set was used to construct,evaluate and verify the Nomogram.Validation set was used to verify repeated validation outside the model.They were classified into CA group and non-CA group according to whether CA occurred.The clinical data of the selected patients were analyzed,and univariate,Lasso and Logistic regression analysis were used to screen the risk factors of CA in the elderly patients after artificial heart valve replacement.A Nomogram model was constructed based on the independent risk factors and goodness-of-fit test was performed.Results CA occurred in 50 of 986 elderly patients after artificial heart valve replacement,with an incidence of 5.07%.BMI>28 kg/m 2(OR=2.952,95%CI 1.373-6.347,P=0.006),postoperative chest pain(OR=3.437,95%CI 1.604-7.362,P=0.001),LVEF≤40%(OR=0.348,95%CI 0.163-0.745,P=0.007),aortic rupture time>120 min(OR=3.505,95%CI 1.618-7.594,P=0.032),postoperative hemodynamic disturbance(OR=3.064,95%CI 1.418-6.622,P=0.004),postoperative hypoxemia(OR=3.117,95%CI 1.436-6.768,P=0.004)and BNP>5000 pg/mL(OR=2.821,95%CI 1.333-5.969,P=0.007)were independent risk factors for CA after artificial heart valve replacement in the elderly patients.Nomogram model showed that BMI>28 kg/m 2 was 86 points,postoperative chest pain was 99 points,LVEF≤40%was 85 points,aortic rupture time>120 min was 100 points,postoperative hemodynamic disturbance was 89 points,postoperative hypoxemia was 91 points,and BNP>5000 pg/mL was 83 points.The C-index of the line graph was 0.841,the H-L goodness-of-fit test wasχ^(2)=11.5248,P=0.082,the correction curve was close

关 键 词：人工心脏瓣膜置换术 心脏骤停 老年 危险因素 列线图 预测 

分 类 号：R654.2[医药卫生—外科学]