期刊文献+

Lasso-Logistic回归模型拟合临床因素、NF-κB/NLRP3信号通路预测心肌梗死后缺血性心肌病价值

Lasso-Logistic regressionmodel fitting clinical factors and NF-κB/NLRP3 predictive value of ischemic cardiomyopathy after myocardial infarction
下载PDF
导出
摘要 目的基于Lasso-Logistic回归分析心肌梗死后缺血性心肌病(ICM)影响因素,探讨临床因素、核因子-κB(NF-κB)/核苷酸结合寡聚结构域样受体家族3(NLRP3)信号通路及Lasso-Logistic回归模型对心肌梗死后ICM的预测价值,为本病防治提供参考。方法选取2020年9月—2023年9月秦皇岛市第一医院收治的342例心肌梗死患者为研究对象进行前瞻性研究,按照7∶3比例分为建模组239例、验证组103例,依据经皮冠状动脉介入术(PCI)术后6个月内是否发生ICM分为ICM亚组、非ICM亚组。采用Lasso筛选心肌梗死后ICM发生相关变量,以有统计学意义变量构建临床因素模型,以NF-κB/NLRP3信号通路构建NF-κB/NLRP3信号通路模型,以临床因素、NF-κB/NLRP3联合建立混合模型(Lasso-Logistic回归模型)。对比不同预测模型对心肌梗死后ICM的预测价值。结果建模组ICM发生率为27.97%,验证组ICM发生率为26.47%;Lasso筛选出5个预测变量为NF-kB mRNA、NLRP3 mRNA、Gensini评分、LVEF、饮酒,Logistic回归分析显示,Gensini评分、NLRP3 mRNA、NF-κB mRNA、饮酒是心肌梗死后ICM影响因素(P<0.05);混合模型预测心肌梗死后ICM的AUC、敏感度、特异度分别为0.921、80.30%、88.82%,临床因素模型分别为0.886、78.79%、85.29%,NF-κB/NLRP3信号通路模型分别为0.873、74.24%、87.06%,混合模型的AUC高于临床因素模型、NF-κB/NLRP3信号通路模型(P<0.05)。结论Gensini评分、NLRP3 mRNA、NF-κB mRNA、饮酒是心肌梗死后ICM危险因素,联合上述影响因素建立Lasso-Logistic回归模型,该模型对心肌梗死后ICM具有一定预测效能,有助于临床早期筛查高危人群,并予以相应干预措施,以降低ICM发生风险。 Objective To analyze the influential factors of ischemic cardiomyopathy(ICM)after myocardial infarction based on Lasso-Logistic regression,and to investigate the predictive value of clinical factors,nuclear factorκB(NF-κB)/nucleotide-bound oligomuctor-like receptor family 3(NLRP3)and Lasso-Logistic regression model for ICM after myocardial infarction.Methods 342 patients with myocardial infarction admitted to our hospital from September 2020 to September 2023 were selected as the study objects,and were divided into the modeling group(239 cases)and the verification group(103 cases)according to the ratio of 7∶3.The patients were divided into the ICM subgroup and the non-ICM subgroup according to the occurrence of ICM within 6 months after percutaneous coronary intervention(PCI).Lasso was used to screen the variables related to the occurrence of ICM after myocardial infarction,the clinical factor model was constructed with statistically significant variables,the NF-κB/NLRP3 model was constructed with NF-κB/NLRP3,and the mixed model(Lasso-Logistic regression model)was established with the combination of clinical factors and NF-κB/NLRP3.The prediction value of different prediction models for ICM after myocardial infarction was compared.Results The incidence of ICM in the modeling group was 27.97%,and the incidence of ICM in the validation group was 26.47%.Lasso screened out 5 predictive variables as NF-κB mRNA,NLRP3 mRNA,Gensini score,LVEF,and alcohol consumption,and logistic regression analysis showed that Gensini score,NLRP3 mRNA,NF-κB mRNA,and alcohol consumption were the influencing factors of ICM after myocardial infarction(P<0.05).The AUC,sensitivity,and specificity of the mixed model for predicting ICM after myocardial infarction were 0.921,80.30%,and 88.82%,respectively,while those of the clinical factor model were 0.886,78.79%,and 85.29%,and those of the NF-κB/NLRP3 model were 0.873,74.24%,and 87.06%,respectively.The AUC of the mixed model was higher than that of the clinical factor model and NF-κB/NLRP3 model(P<0.05).Conclusion Gensini score,NLRP3 mRNA,NF-κB mRNA,and alcohol consumption are risk factors for ICM after myocardial infarction.A Lasso-Logistic regression model was established using these factors,which has certain predictive power for ICM after myocardial infarction.This model can help clinicians screen high-risk populations early and provide appropriate interventions to reduce the risk of ICM.
作者 杜然 滕腾 赵云凤 方钱超 蔡丽丽 DU Ran;TENG Teng;ZHAO Yunfeng;FANG Qianchao;CAI Lili(Department of CCU(2),First Hospital of Qinhuangdao,Qinghuangdao Hebei 066000,China)
出处 《中国急救复苏与灾害医学杂志》 2024年第6期705-709,747,共6页 China Journal of Emergency Resuscitation and Disaster Medicine
基金 河北省科学技术研究发展计划项目(项目:202004A028)。
关键词 心肌梗死 缺血性心肌病 Lasso回归 LOGISTIC回归分析 核因子-ΚB 核苷酸结合寡聚结构域样受体家族3 预测 Myocardial infarction Ischemic cardiomyopathy Lasso regression Logistic regression analysis Nuclear factor-κB Nucleotide-bound oligomeric domain-like receptor family 3 Prediction
  • 相关文献

参考文献6

二级参考文献35

共引文献169

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部