基于真实世界数据住院患者发生尿路感染的相关因素研究  

Related factors of urinary tract infections in inpatients based on real world data

作　　者：边春红 潘玥[2] 谭雅楠[2] 张丽敏[1] 王荣祺[2] 张国军[1,3,4] Bian Chunhong;Pan Yue;Tan Yanan;Zhang Limin;Wang Rongqi;Zhang Guojun(Laboratory Diagnosis Center,Beijing Tiantan Hospital Affiliated to Capital Medical University,Beijing 100070,China;Clinical Laboratory,Beijing Haidian Hospital,Beijing 100089,China;Beijing Engineering Research Center of Immunological Reagents Clinical Research,Beijing 100070,China;Key Laboratory for Quality Control of In Vitro Diagnostics,National Medical Products Administration,Beijing 100070,China)

机构地区：[1]首都医科大学附属北京天坛医院实验诊断中心,北京100070 [2]北京市海淀医院检验科,北京100089 [3]北京市免疫试剂临床工程技术研究中心,北京100070 [4]国家药监局体外诊断试剂质量控制重点实验室,北京100070

出　　处：《中华预防医学杂志》2022年第11期1636-1641,共6页Chinese Journal of Preventive Medicine

基　　金：北京市医院管理局临床医学发展专项资金(扬帆计划)(ZYLX202108);北京市医管中心人才培养计划“登峰”项目(DFL20220505);北京市高层次公共卫生技术人才培养计划(2022-2-013);2021年海淀区科研培育计划(HP2021-30-10102)。

摘　　要：探讨临床住院患者尿路感染的相关因素。回顾性分析2019年10月至2021年5月期间在北京市海淀医院住院并进行尿液细菌培养的1875例患者病例资料,根据尿路感染诊断标准分为感染组和非感染组,分析感染组病原菌的种类分布情况,并对病例资料和实验室指标进行单因素分析,选取具有统计学意义的变量进行二元logistic回归分析尿路感染的风险因素并建立预测模型,对纳入模型的各参数绘制受试者工作特征曲线(ROC)并计算曲线下面积(AUC),评价各参数单独使用和联合使用对尿路感染的诊断和预测效能。结果显示,非感染组1162例,感染组713例,在培养的病原菌中,革兰阴性菌构成比57.2%(408/713),革兰阳性菌构成比35.9%(256/713),真菌构成比6.9%(49/713)。多因素分析结果显示:年龄、置管天数(>7 d)、脑卒中和骨科手术是尿路感染发病的风险因素,抗菌药物应用是保护性因素。年龄、置管天数(>7 d)、脑卒中、骨科手术、抗菌药物应用、尿白细胞酯酶、尿亚硝酸盐及红细胞体积分布宽度变异系数纳入尿路感染的预测模型,其诊断和预测AUC面积为0.835(95%CI 0.816~0.855),敏感度为70.7%、特异度为82.8%。综上,上述8项参数联合应用可能对住院患者能较好的辅助诊断和预测尿路感染。To analyze the risk factors for urinary tract infection(UTI)among inpatients.The case data of 1875 inpatients receiving urinary bacterial culture in Beijing Haidian Hospital from October 2019 to May 2021 were analyzed retrospectively.According to the etiological diagnostic criteria of UTI,they were divided into infection group and non-infection group.The species and distribution of pathogens in the infection group were analyzed,and the case data and laboratory indexes were subjected to univariate analysis.The variables with statistical significance were selected for binary logistic regression to analyze the risk factors of urinary tract infection and establish a prediction model.The receiver operating characteristic(ROC)curve was drawn for each parameter included in the model,and the area under the curve(AUC)was calculated.The diagnostic and predictive efficacy of each parameter alone and their combination for UTI were evaluated.So,a total of 1162 patients with non-infection group and 713 patients with UTI were detected.Among the cultured pathogens,the constituent ratio of Gram-negative bacteria,Gram-positive bacteria and fungi was 57.2%(408/713),35.9%(256/713)and 6.9%(49/713)respectively.Multivariate analysis showed that,Age,duration of urinary catheterization>7 d,stroke and orthopedic surgery were the risk factors of UTI among inpatients.The use of antibiotics is a protective factor for urinary tract infections.The prediction model of UTI was established by the risk factors,age,duration of urinary catheterization>7 d,stroke,orthopedic surgery,urinary leukocyte esterase,urinary nitrite and Coefficient of variability of red blood cell volume distribution width(RDW-CV).The AUC of the combination of the eight parameters in diagnosing and predicting UTI was 0.835(95%CI:0.816-0.855),with the sensitivity of 70.7%and the specificity of 82.8%.In conclusion,the combination of the eight parameters can better assist in the diagnosis and prediction of UTI,and provide an experimental basis for clinicians to judge UTI.

关 键 词：泌尿系统感染 病原菌 风险 LOGISTIC回归模型 

分 类 号：R691.3[医药卫生—泌尿科学]