摘要
目的分析急性心力衰竭(AHF)患者住院期间心血管不良事件发生的主要影响因素,构建心血管不良事件风险预测模型并进行验证。方法回顾性选取2021年2月—2022年2月宁夏医科大学总医院心脑血管病医院收治的515例AHF患者,根据患者住院期间是否发生心血管不良事件将其分为发生组与未发生组。采用R4.2.1统计软件对两组各项指标进行单因素Logistic分析并进行Lasso回归初步筛选,再经多因素一般Logistic回归分析确定预测变量,构建AHF患者住院期间心血管不良事件预测模型,并通过绘制列线图、受试者工作特征(ROC)曲线及校准曲线对模型的性能进行评价,最后采用Bootstrap法对模型进行内部验证。结果515例AHF患者中有62例发生心血管不良事件。心血管不良事件发生组与未发生组的冠心病、扩心病、脑卒中、睡眠差、呼吸频率、合并心房颤动(以下简称房颤)比较,差异均有统计学意义(P<0.05);先后经单因素Logistic分析及Lasso回归初步筛选变量后,最终通过多因素一般Logistic回归分析确定AHF患者住院期间发生心血管不良事件的主要影响因素为脑卒中[O^R=7.99(95%CI:3.29,19.59)]、白细胞计数[O^R=1.25(95%CI:1.00,1.52)]、血尿素氮[O^R=1.13(95%CI:1.04,1.23)]、D-二聚体[O^R=1.13(95%CI:1.03,1.24)]、左心室短轴缩短率[O^R=0.89(95%CI:0.81,0.96)]、合并房颤[O^R=2.40(95%CI:1.07,5.31)]。基于以上6个影响因素构建风险预测列线图模型,结果显示,模型的ROC曲线下面积(AUC)为0.82(95%CI:0.76,0.88),校准曲线贴近理想曲线。经Bootstrap法内部验证所得AUC值为0.81(95%CI:0.75,0.87)。结论AHF患者住院期间发生心血管不良事件的主要影响因素是脑卒中、白细胞计数、血尿素氮、D-二聚体、左心室短轴缩短率、合并房颤,以此构建的心血管不良事件风险预测模型具有良好的效能。
Objective To analyze the main influencing factors of cardiovascular adverse events during hospitalization in patients with acute heart failure(AHF),construct a risk prediction model for cardiovascular adverse events,and validate its performance.Methods We retrospectively selected 515 cases of AHF patients admitted to the Cardiovascular Hospital of General Hospital of Ningxia Medical University from February 2021 to February 2022.Patients were divided into the occurrence group and non-occurrence group based on whether cardiovascular adverse events occurred during hospitalization.R 4.2.1 statistical software was used for univariate Logistic analysis and Lasso regression preliminary screening of various indicators in both groups.Subsequently,multifactorial Logistic regression analysis was performed to determine the predictive variables,construct a risk prediction model for cardiovascular adverse events during AHF patients'hospitalization.Model performance was evaluated through line charts,receiver operating characteristic(ROC)curves,and calibration curves.Bootstrap method was employed for internal validation of the model.Results Among the 515 AHF patients,62 cases experienced cardiovascular adverse events.There were statistically significant differences in aspects such as coronary heart disease,dilated cardiomyopathy,stroke,poor sleep,respiratory rate,and atrial fibrillation(AF)between the cardiovascular adverse event occurrence group and non-occurrence group(P<0.05).After univariate Logistic analysis and Lasso regression preliminary screening,the main influencing factors for cardiovascular adverse events during AHF patients'hospitalization were determined through multifactorial Logistic regression analysis.These factors included stroke[O^R=7.99(95%CI:3.29,19.59)],white blood cell count[O^R=1.25(95%CI:1.00,1.52)],blood urea nitrogen[O^R=1.13(95%CI:1.04,1.23)],D-dimer[O^R=1.13(95%CI:1.03,1.24)],left ventricular fractional shortening[O^R=0.89(95%CI:0.81,0.96)],and concomitant AF[O^R=2.40(95%CI:1.07,5.31)].A risk prediction line chart model was constructed based on these six influencing factors,with an area under the ROC curve(AUC)of 0.82(95%CI:0.76,0.88),and the calibration curve closely approximating the ideal curve.The AUC value obtained through Bootstrap internal validation was 0.81(95%CI:0.75,0.87).Conclusion The main influencing factors for cardiovascular adverse events during hospitalization in AHF patients are stroke,white blood cell count,blood urea nitrogen,D-dimer,left ventricular fractional shortening,and concomitant AF.The risk prediction model constructed based on these factors demonstrates good efficacy.
作者
张世昌
马萍
马萌雪
徐清斌
董丹丹
马晓彦
Zhang Shi-chang;Ma Ping;Ma Meng-xue;Xu Qing-bin;Dong Dan-dan;Ma Xiao-yan(Department of Cardiology,Cardiovascular and Cerebrovascular Disease Hospital of General Hospital of Ningxia Medical University,Yinchuan,Ningxia 750002,China;Intensive Care Unit Yinchuan,General Hospital of Ningxia Medical University,Yinchuan,Ningxia 750002,China)
出处
《中国现代医学杂志》
CAS
2024年第1期7-15,共9页
China Journal of Modern Medicine
基金
宁夏医科大学校级科研项目(No:XM2022028)。
关键词
急性心力衰竭
心血管不良事件
危险因素
预测模型
heart failure,acute
adverse events,cardiovascular
risk factors
prediction model
