高风险腹主动脉瘤的预测模型  

Prediction model for high-risk abdominal aortic aneurysm

作　　者：霍桂军 郑进 曾宇琪 姚志超 周大勇[1] 黄剑[1] Huo Guijun;Zheng Jin;Zeng Yuqi;Yao Zhichao;Zhou Dayong;Huang Jian(Department of Vascular Surgery,Gusu School of Nanjing Medical University/Affiliated Suzhou Hospital of Nanjing Medical University/Suzhou Municipal Hospital(Headquaters),Suzhou 215002,Jiangsu,China)

机构地区：[1]南京医科大学姑苏学院/南京医科大学附属苏州医院/苏州市立医院(本部)血管外科,江苏苏州215002

出　　处：《血管与腔内血管外科杂志》2024年第9期1051-1055,1072,共6页Journal of Vascular and Endovascular Surgery

基　　金：苏州市“科教兴卫”青年科技项目(KJXW2021031)。

摘　　要：目的探讨高风险腹主动脉瘤(AAA)的预测模型。方法收集2010年11月至2024年1月于南京医科大学姑苏学院/南京医科大学附属苏州医院/苏州市立医院(本部)收治的136例AAA患者的临床资料,按是否进行手术治疗将其分为高风险组(n=69)和低风险组(n=67)。系统收集患者入院后的临床特征,通过套索回归(LASSO)筛选特征变量,在多因素逻辑回归分析的基础上构建高风险AAA预测模型,通过曲线下面积(AUC)、HosmerLemeshow校准曲线和决策曲线分析(DCA)评估所构建模型的预测效能和临床应用价值。结果单因素分析结果显示,AAA直径、中性粒细胞/淋巴细胞比值(NLR)、甘油三酯(TG)、吸烟、高血压、二甲双胍服药史、抗血小板服药史均是高风险AAA的影响因素(P﹤0.05)。LASSO分析将19个影响因素构建系数分布图,结果显示,最佳变量参数分别为AAA直径、NLR、高血压、二甲双胍服药史,据此构建高风险AAA预测模型,其AUC为0.893(95%CI:0.838~0.947);经Hosmer-Lemeshow检验显示模型拟合良好(P﹥0.05);DCA结果显示,在10%~95%的阈值范围内,高风险AAA预测模型净获益值最高为0.51。结论由AAA直径、高血压、NLR、二甲双胍服药史构建的高风险AAA预测模型能够有效预测AAA患者高风险事件的发生率,可为AAA患者提供更加个体化的风险评估及治疗方案,值得临床推广应用。Objective To investigate the prediction model of high risk abdominal aortic aneurysm(AAA).Method Clinical data of 136 patients with AAA admitted to Gusu School of Nanjing Medical University/Affiliated Suzhou Hospital of Nanjing Medical University/Suzhou Municipal Hospital(Headquaters)from November 2010 to January 2024 were collected,they were divided into high risk group(n=69)and low risk group(n=67)according to whether they underwent surgical treatment.Clinical information of patients after admission was systematically collected,the characteristic variables were screened by least absolute shrinkage and selection operator(LASSO),and the high-risk AAA prediction model was constructed on the basis of multi-factor logistic regression analysis.Finally,area under curve(AUC),Hosmer-Lemeshow calibration curve and decision curve analysis(DCA)were used to evaluate the predictive efficacy and clinical application value of the constructed model.Result The results of univariate analysis showed that AAA diameter,neutrophil-to-lymphocyte ratio(NLR),triglyceride(TG),smoking,hypertension,history of metformin use,and history of antiplatelet use were all influencing factors for high-risk AAA(P<0.05).LASSO regression analysis constructed coefficient distribution maps for 19 influencing factors,the results showed that the optimal variable parameters were AAA diameter,NLR,hypertension,and metformin medication history.Based on this,the high-risk AAA prediction model was constructed,the AUC was 0.893(95%CI:0.838~0.947);the Hosmer-Lemeshow test showed that the model fit well(P>0.05);the DCA results showed that within the threshold range of 10%to 95%,the highest net benefit value of the high-risk AAA prediction model was 0.51.Conclusion The high-risk AAA prediction model constructed by AAA diameter,hypertension,NLR and metformin medication history can effectively predict the probability of high-risk events in AAA patients,which will provide a more individualized risk assessment and treatment plan for AAA patients,it is worthy of clinical

关 键 词：腹主动脉瘤 风险评估 预测模型 二甲双胍 

分 类 号：R543[医药卫生—心血管疾病]