Development and Validation of An Integrated Traditional Chinese and Western Medicine Clinical Prediction Model for Adverse Cardiovascular Events in Patients with Acute Myocardial Infarction

Author：LIU Pengyu1, SHANG Juju2

Unit：1.Beijing Health Service Management and Guidance Center, Beijing 101160, China; 2.Beijing Hospital of Traditional Chinese Medicine, Capital Medical University, Beijing 100010, China

Quote：引用：刘鹏宇，尚菊菊.急性心肌梗死患者不良心血管事件中西医结合临床预测模型的建立与验证[J].中医药导报,2026,32(2):87-95,100.

DOI：10.13862/j.cn43-1446/r.2026.02.014

PDF： Download PDF

Abstract：

Objective: To develop and validate a clinical prediction model for major adverse cardiovascular events (MACE) in patients with acute myocardial infarction (AMI). Methods: A retrospective analysis was conducted on 398 AMI patients hospitalized in the Department of Cardiology, Beijing Hospital of Traditional Chinese Medicine, Capital Medical University, from January 1, 2017, to January 1, 2021. These patients were followed up and divided into a MACE group (n=174) and a non-MACE group (n=224) based on the occurrence of MACE (including cardiac death, heart failure, non-fatal stroke, non-fatal myocardial infarction, and rehospitalization for ischemic heart disease) during follow-up. The Least Absolute Shrinkage and Selection Operator (LASSO) algorithm was used to screen for predictive factors of MACE in AMI patients. Variables were further refined using near-zero variance (NZV) analysis, collinearity screening, and recursive feature elimination (RFE). The selected variables were incorporated into a Cox regression model to construct the prediction model, and a nomogram was developed for clinical use. Internal validation of the model was performed using the bootstrap resampling method. The area under the curve (AUC) and concordance index (C-index) were used to assess the model's discriminative ability. Calibration curves were plotted to evaluate its calibration. Decision curve analysis (DCA) was applied to assess the clinical utility of the prediction model. Results: LASSO regression identified seven predictive factors: Killip class, age, stent implantation, creatinine (Cr), albumin (Alb), C-reactive protein (CRP), and holistic pulse diagnosis. The internally validated C-index of the prediction model based on these factors was 0.697. The AUC values for predicting MACE at six months, one year, and three years in the study population were 0.682, 0.644, and 0.698, respectively. The calibration curves showed good agreement between the predicted and observed outcomes. Decision curve analysis indicated that the nomogram had favorable clinical utility. Conclusion: The prediction model constructed in this study demonstrated a reasonable ability to estimate the risk of MACE across different time points, which may aid clinical risk stratification and provide a reference for enhancing the scientific rationale of medical decision-making.

Key words：acute myocardial infarction; major adverse cardiovascular events; predictive factors; clinical prediction model

摘要：目的：在急性心肌梗死（AMI）患者中建立并验证预测主要不良心血管事件（MACE）的临床预测模型。方法：回顾性收集2017年1月1日至2021年1月1日期间在首都医科大学附属北京中医医院心内科住院的398例AMI患者作为研究对象，进行随访，根据患者在随访期间是否发生MACE事件（心源性死亡、心力衰竭、非致死性卒中、非致死性心肌梗死、缺血性心脏病再入院）将其分为MACE组（174例）和非MACE组（224例）。采用最小绝对收缩和选择算子算法（LASSO）筛选AMI患者MACE事件的预测因素。使用近零方差查询（NZV）、共线性筛查和自变量递归特征消除（RFE）对变量进一步筛选，将筛选后的指标纳入Cox回归构建预测模型并绘制列线图以供临床使用。使用bootstrap重抽样方法对模型进行内部验证。采用曲线下面积（AUC）、一致性指数（C-index）评估模型区分度。采用校准曲线评估模型校准度。采用决策曲线分析（DCA）评估预测模型的临床有效性。结果：LASSO回归共筛选了7个预测因素：Killip分级、年龄、支架植入、肌酐（Cr）、白蛋白（Alb）、C反应蛋白（CRP）和整体脉象。基于上述预测因素建立的预测模型在内部验证中C-index值为0.697，研究人群在半年、1年和3年MACE事件的AUC分别为0.682、0.644、0.698。校准曲线结果显示预测曲线与校准曲线重合度较好。决策曲线表明列线图在指导临床实践中具有良好的效果。结论：本研究构建的预测模型在不同时期具有较好的MACE事件风险预测能力，能辅助临床开展风险分层，进而为提高医疗决策的科学性与合理性提供参考。

关键词：急性心肌梗死；不良心血管事件；预测因素；临床预测模型

Release time：2026-03-05

click-through rate：42