心力衰竭易损期心源性死亡中西医风险预警模型的建立与验证  

Establishment and validation of a traditional Chinese and Western medicine risk prediction model for cardiac death in the vulnerable phase of heart failure

作　　者：李润民 戴国华[1] 高武霖 管慧 任丽丽 王兴蒙 曲惠文 LI Runmin;DAI Guohua;GAO Wulin;GUAN Hui;REN Lili;WANG Xingmeng;QU Huiwen(Shandong University of Traditional Chinese Medicine,Jinan 25000,China;Affiliated Hospital of Shandong University of Traditional Chinese Medicine,Jinan 250000,China)

机构地区：[1]山东中医药大学,济南250000 [2]山东中医药大学附属医院,济南250000

出　　处：《中华中医药杂志》2024年第7期3637-3645,共9页China Journal of Traditional Chinese Medicine and Pharmacy

基　　金：国家重点研发计划(No.2019YFC1710401);国家自然科学基金项目(No.82174172);国家自然科学基金青年科学基金项目(No.82204942);山东省自然科学基金青年项目(No.ZR2022QH123);中国博士后科学基金面上资助项目(No.2022M721998)。

摘　　要：目的:分析心力衰竭易损期患者发生心源性死亡的中西医预警指标,建立并验证预警模型。方法:将山东省7家中医院符合研究标准的2627例心力衰竭患者纳入队列研究。运用最小绝对收缩和选择算子(Lasso)回归结合Cox回归模型筛选心力衰竭易损期心源性死亡预警指标,绘制预警模型列线图。通过受试者工作特征曲线(ROC)下面积(AUC)、校准图、决策曲线分析(DCA)验证模型效能。结果:将3个保护性指标(利尿类药物、益气类注射液、降血脂类药物)和9个损害性指标[血浆氨基末端脑利钠肽前体(NT-ProBNP)]、吸烟史、纽约心脏病学会(NYHA)心功能分级、2型糖尿病、脑血管病、脉细数无力或结代、阳气亏虚血瘀证、气阴两虚血瘀证、面色口唇紫暗作为预警指标构建预警模型。该模型在训练集与验证集的AUC值分别为0.776与0.772,预警准确性良好。训练集与验证集校准曲线较为一致,预警稳定性良好。训练集与验证集在17%~56%和18%~44%的风险范围内显示出较好的临床有效性。结论:基于中西医预警指标构建心力衰竭易损期心源性死亡预警模型的准确性、稳定性与临床有效性较好,可为心力衰竭易损期高危人群识别与中西医结合防治提供参考。Objective:To analyze the early warning indicators of traditional Chinese and Western medicine for cardiac death in patients with vulnerable phase of heart failure(HF),and to establish and verify the risk prediction model.Methods:A total of 2627 patients with HF who met the research criteria in 7 traditional Chinese medicine hospitals in Shandong Province were enrolled in this cohort study.The least absolute selection and shrinkage operator(Lasso)regression combined with Cox regression was used to identify the early warning indicators,and a nomogram was drawn for the risk prediction model of cardiac mortality in the vulnerable phase of HF.The area under the receiver operating characteristic(ROC)curve(AUC),calibration plot and decision curve analysis(DCA)were used to verify the performance of the model.Results:Three protective indicators(diuretic drugs,benefiting-qi injection,lipid-lowering drugs)and nine damaging indicators(NT-proBNP,smoking history,NYHA heart function classification,type 2 diabetes mellitus,cerebrovascular disease,weakness or knotted intermittent of pulse,yang-qi deficiency and blood stasis syndrome,qi-yin deficiency and blood stasis syndrome,dark complexion and lips)were used as early warning indicators to construct the risk prediction model.The AUC values of the model in the training set and the validation set were 0.776 and 0.772,respectively,suggested the predictions were accurate.The calibration plot of the training set and the validation set were consistent,suggested the predictions were stable.The training set and the validation set showed good clinical effectiveness in the risk range of 17%-56%and 18%-44%.Conclusion:The risk prediction model of cardiac death in the vulnerable phase of HF based on the early warning indicators of traditional Chinese and Western medicine is accurate,stable,and good clinically valid.That could serve as a guide for identifying at-risk populations at the vulnerable phase of HF and for preventing and curing with Western medicine and traditional Chinese medicine

关 键 词：心力衰竭易损期 中西医预警指标 预警模型 列线图 最小绝对收缩和选择算子回归 COX回归 心源性死亡 

分 类 号：R541.6[医药卫生—心血管疾病]