老年肺部耐碳青霉烯肠杆菌感染风险预测模型的构建和验证  

Establishment and validation of model for the risk of Carbapenem-Resistant Enterobacteriaceae infection in elderly patients with pulmonary infection

作　　者：赵诗琦 王书 张颖 王莎莎 李静 唐文娟[3] 李静[4] 张学敏 李芳 ZHAO Shiqi;WANG Shu;ZHANG Ying;WANG Shasha;LI Jing;TANG Wenjuan;LI Jing;ZHANG Xuemin;LI Fang(Cadre Ward,the Third Affiliated Hospital of Anhui Medical University,Hefei,Anhui 230000,China;不详)

机构地区：[1]安徽医科大学第三附属医院干部保健科,安徽合肥230000 [2]合肥市滨湖医院干部保健科,安徽合肥230000 [3]安庆市立医院全科医学科,安徽安庆246000 [4]中国科学技术大学第一附属医院老年医学科,安徽合肥230000 [5]安徽医科大学附属阜阳医院老年医学科,安徽阜阳236000 [6]安徽医学高等专科学校临床医学院,安徽合肥230000

出　　处：《中华全科医学》2023年第11期1860-1864,1945,共6页Chinese Journal of General Practice

基　　金：合肥市卫生健康委2022年度重点应用医学研究项目(Hwk2022zd003);安徽省高等学校科学研究项目(自然科学类)重点项目(2022AH052334)。

摘　　要：目的分析及筛选影响老年肺部耐碳青霉烯肠杆菌(CRE)感染的高危因素,构建并验证风险Nomogram预测模型。方法回顾性对安徽地区共894例的老年(≥60岁)肺部感染痰培养为肠杆菌科细菌的住院患者的临床资料进行收集,抽取2个医院作为验证组(250例),其余作为建模组(644例),使用单因素Lasso回归和多变量logistic回归分析对建模组数据进行分析,确定CRE感染的独立危险因素。构建Nomogram预测模型,并对模型进行多中心验证。结果气管插管≥72 h、长期卧床、抗生素使用种类≥2种、脑血管疾病、营养不良是老年肺部耐碳青霉烯肠杆菌感染的独立危险因素。基于上述5项独立危险因素,建立预测风险Nomogram模型,建模组和验证组2组AUC值分别为0.862、0.858;且Hosmer-Lemeshow检验可见建模组P=0.761,验证组P=0.339;Decision曲线分析显示,模型在较大的阈值内获益性均良好。结论构建老年肺部CRE感染风险预测模型具有一定的临床意义,能够针对可控因素进行早期个体化干预。Objective To analyze and screen the high-risk factors for CRE infection in elderly people with pulmonary infection,and to construction and validate a risk nomogram prediction model.Methods Clinical data of a total of 894 elderly(≥60 years old)hospitalized patients with pulmonary infection cultured by Enterobacteriaceae bacteria in Anhui Province were retrospectively collected.Two hospitals were selected as verification group(250 cases),and the rest were used as training group(644 cases).Univariate Lasso regression and multivariate logistic regression analysis were used to analyze the modeling group data to identify independent risk factors for CRE infection.Construct a Nomogram prediction model and verify the model in multiple centers.Results The independent risk factors of CRE infection in the elderly were endotracheal intubation≥72 h,long-term bed rest,≥2 types of antibiotics,cerebrovascular disease and malnutrition.Based on the above five independent risk factors,a Nomogram model for predicting risk was established,and the AUC values of the training group and verification group were 0.862 and 0.858.The Hosmer-Lemeshow test shows that the training group P=0.761 and the verification group P=0.339.Decision curve analysis shows that the model has good benefit within large thresholds.Conclusion The construction of a risk prediction model for the risk of CRE infection in the lungs of elderly people have some clinical significance,allowing early individualized intervention according to controllable factors.

关 键 词：耐碳青霉烯类肠杆菌 肺部感染 危险因素 预测模型 多中心 

分 类 号：R563.1[医药卫生—呼吸系统]