摘要
目的构建重症监护病房(intensive care unit,ICU)患者多重耐药菌(multidrug-resistant organisms,MDRO)感染的风险列线图预测模型,为早期识别、干预、降低MDRO传播风险提供参考依据。方法采用巢式病例对照研究分析2018—2022年上海市某三甲医院ICU的住院患者,发生MDRO感染的患者为病例组,未发生MDRO感染的患者为对照组,将12207例住院患者按照7∶3的比例随机分为训练集(8544例)和验证集(3663例),使用竞争风险分析法建立列线图预测模型并通过一致性指数(C指数)、受试者操作特征曲线下面积(area under curve,AUC)进行验证。结果2018—2022年ICU患者的MDRO医院感染发生率为3.65%(446/12207)。MDRO患者的主要感染部位为呼吸机相关性肺炎191例(42.83%)、下呼吸道(跟导管无关)141例(31.61%)、导尿管相关尿路感染63例(14.13%);主要检出耐药菌分别为耐碳青霉烯类肠杆菌科细菌(carbapenem-resistant enterobacteriaceae,CRE)234例(52.47%)、耐碳青霉烯鲍曼不动杆菌(carbapenem-resistant Acinetobacter baumannii,CRAB)196例(43.95%)、耐甲氧西林金黄色葡萄球菌(methicillin-resistant Staphylococ cusaureus,MRSA)41例(9.19%)。训练集和验证集的基本特征差异无统计学意义(P>0.05),列线图风险预测模型显示,手术风险比(hazard ratio,HR)为2.01、入住ICU时间(10~<20 d:HR=1.78;20~<30 d:HR=3.01;≥30 d:HR=3.46)、中心静脉导管留置时间(1~<7 d:HR=2.73;7~<14 d:HR=4.81;14~<21 d:HR=4.86;≥21 d:HR=5.13)、气管插管或切开时间(1~<7 d:HR=3.47;7~<14 d:HR=4.36;14~<21 d:HR=3.79;≥21 d:HR=2.92)是发生MDRO感染的独立危险因素(P<0.05)。验证结果显示,训练集和测试集预测入ICU 7 d时MDRO感染风险的AUC值分别为75.93(72.58~79.22)、73.31(68.06~78.63),均高于单一影响因素的AUC值。结论基于竞争风险分析建立ICU患者MDRO医院感染风险的列线图预测模型具有良好的区分度与预测价值,能为临床识别高危患者,采取更具针对性的干预措施提供参考。
Objective To develop a risk nomogram for predicting infections caused by multidrug-resistant organisms(MDRO)in intensive care unit(ICU)patients,providing a reference for early identification,intervention,and reduction of MDRO transmission risk.Methods A nested case-control study was used to analyze ICU inpatients from a tertiary hospital in Shanghai from 2018 to 2022.Patients with MDRO infections were selected as the case group,while those without MDRO infections were selected as the control group.They were randomly divided into a training set(n=8544 cases)and a validation set(n=3663 cases)in a 7:3 ratio.Competing risk analysis was employed to establish the nomogram prediction model,validated using the concordance index(C-index)and the area under the receiver operating characteristic curve(AUC).Results The incidence of MDRO hospital infections among ICU patients from 2018 to 2022 was 3.65%(446/12207).The main infection sites of MDRO patients are ventilator-associated pneumonia in 191cases(42.83%),lower respiratory tract(unrelated to catheterization)in 141 cases(31.61%),and urinary tract infections related to catheterization in 63 cases(14.13%).The main detected drug-resistant bacteria are carbapenem-resistant Enterobacteriaceae(CRE)in 234 cases(52.47%),carbapenem-resistant Acinetobacter baumannii(CRAB)in 196 cases(43.95%),and methicillin-resistant Staphylococcus aureus(MRSA)in 41 cases(9.19%).No statistically significant differences were found in the baseline characteristics between the training set and validation set(P>0.05)The nomogram identified independent risk factors for MDRO infections,including surgical risk ratio(hazard ratio,HR=2.O1),duration of ICU stay(10-<20 days:HR=1.78;20-<30 days:HR=3.01;≥30 days:HR=3.46),duration of central venous catheterization days(1-<7 days:HR=2.73;7-<14 days:HR=4.81;14-<21 days:HR=4.86;≥21 days:HR=5.13),and duration of intubation or tracheostomy(1-<7 days:HR=3.47;7-<14 days:HR=4.36;14-<21 days:HR=3.79;≥21 days:HR=2.92)(P<0.05).Validation results showed that the AUC values for predicting MDRO infection risk at 7 days in the ICU were 75.93(72.58~79.22)for the training set and 73.31(68.06~78.63)for the validation set,both exceeding the AUC values of individual risk factors.Conclusions The nomogram developed based on competing risk analysis demonstrates good discrimination and predictive value for MDRO hospital infections in ICU patients,serving as a valuable tool for clinicians to identify high-risk patients and implement targeted interventions.
作者
江佳艳
钱香玲
王欣
田琛霞
尹贤哲
杨惠英
刘银梅
JIANG Jiayan;QIAN Xiangling;WANG Xin;TIAN Chenxia;YIN Xianzhe;YANG Huiying;LIU Yinmei(Department of Nosocomial Infection and Disease Control of Shanghai Tenth People's Hospital,Shanghai200072 China)
出处
《中国预防医学杂志》
2025年第6期715-721,共7页
Chinese Preventive Medicine
基金
上海申康医院发展中心临床研究数据共享和模拟RCT项目(SHDC2024CRI035)。
关键词
多重耐药菌
竞争风险模型
列线图
预测模型
Multidrug-resistant organisms
Competing risk model
Nomogram
Prediction model
