Page 110 - 《南京医科大学学报（自然科学版）》2026年第3期

南京医科大学学报（自然科学版） 第46卷第3期
·418 · Journal of Nanjing Medical University（Natural Sciences） 2026年3月

·临床研究·

脓毒症相关性心肌损伤患者临床转归分析及预测列线图构建

李加涌，朱 轶，罗春阳，陈旭锋 *
南京医科大学第一附属医院急诊医学科，江苏 南京 210029

［摘 要］ 目的：探讨脓毒症相关性心肌损伤（sepsis⁃associated myocardial injury，SAMI）的流行病学现状及其对预后的影响，
并通过构建列线图以期早期识别SAMI高危群体。方法：采用回顾性研究，收集2023年7月—2024年12月于南京医科大学第
一附属医院急诊医学科住院的脓毒症患者临床资料，统计SAMI发病率，绘制28 d Kaplan⁃Meier生存曲线比较SAMI对脓毒症
预后影响，通过最小绝对收缩和选择算子（least absolute shrinkage and selection operator，LASSO）回归以及Boruta算法分别对临
床变量进行筛选，并采用多因素 Logistic 回归分析构建 SAMI 早期预测模型。结果：共纳入 353 例脓毒症患者，其中 195 例
（55.2%）患者在病程中发生SAMI。SAMI组患者28 d死亡风险显著高于无SAMI患者（HR=2.342，P < 0.001）。通过LASSO回归
和Boruta算法行变量筛选并取交集，最终纳入年龄、冠心病史、肌酐、尿素氮、D⁃二聚体和降钙素原共6个变量构建预测模型并
绘制列线图，预测模型具有较好的区分度，Bootstrap重复抽样1 000次的受试者工作特征（receiver operating characteristic，ROC）
曲线下面积为0.770（95%CI：0.767~0.773，P < 0.001），校准曲线拟合良好，决策曲线分析示在阈值概率0~0.95区间内，预测模
型有较好的净收益。结论：SAMI是脓毒症患者常见并发症，并导致不良预后，基于临床变量构建的列线图具有较好的临床应
用前景。
［关键词］ 脓毒症相关性心肌损伤；LASSO回归；Boruta算法；列线图
［中图分类号］ R631 ［文献标志码］ A ［文章编号］ 1007⁃4368（2026）03⁃418⁃08
doi：10.7655/NYDXBNSN251304

Analysis of clinical outcomes and construction of predictive nomogram in patients with
sepsis⁃associated myocardial injury

LI Jiayong，ZHU Yi，LUO Chunyang，CHEN Xufeng *
Department of Emergency Medicine，the First Affiliated Hospital of Nanjing Medical University，Nanjing 210029，
China

［Abstract］ Objective：To explore the epidemiological status of sepsis ⁃ associated myocardial injury（SAMI）and its impact on
prognosis，and to construct a nomogram for early identification of high ⁃ risk groups of SAMI. Methods：A retrospective study was
conducted to collect clinical data of sepsis patients hospitalized in the Department of Emergency Medicine，the First Affiliated Hospital
of Nanjing Medical University from July 2023 to December 2024. The incidence of SAMI was analyzed，and 28⁃day Kaplan⁃Meier
survival curves were drawn to compare the impact of SAMI on the prognosis of sepsis. Clinical variables were screened by least
absolute shrinkage and selection operator（LASSO）regression and Boruta algorithm，respectively. Multivariate logistic regression
analysis was used to construct the early prediction model of SAMI. Results：A total of 353 patients with sepsis were included，of whom
195（55.2%）developed SAMI during the course of the disease. The 28⁃day mortality risk was significantly higher in patients with
SAMI than in patients without SAMI（HR=2.342，P < 0.001）. By using LASSO regression and Boruta algorithm，variables were
screened and intersections were taken. Finally，6 variables including age，history of coronary heart disease，creatinine，urea nitrogen，
D⁃dimer and procalcitonin were constructed and nomogram was drawn. The area under receiver operating characteristic curve of the
internal validation using the bootstrap method（resampling=1 000）was 0.770（95%CI：0.767-0.773，P < 0.001）. The calibration curve
fitted well，and the decision curve analysis showed that the prediction model had a good net benefit in the range of threshold probability
0-0.95. Conclusion：SAMI is a common complication of sepsis and leads to poor prognosis. Nomogram based on clinical variables has

［基金项目］ 江苏省科教能力提升工程（ZDXK202213）
通信作者（Corresponding author），E⁃mail：cxfyx@njmu.edu.cn（ORCID：0000⁃0003⁃3697⁃6446 ）
∗