重症急性胰腺炎并发多重耐药菌感染列线图预测模型的构建与评价  

Construction and evaluation of a nomogram prediction model for multi-drug resistant organism infection in patients with severe acute pancreatitis

作　　者：陈燕春 邱洁净 田蓉 肖婉露 罗倩 罗微 廖小静 王平[1] CHEN Yanchun;QIU Jiejing;TIAN Rong;XIAO Wanlu;LUO Qian;LUO Wei;LIAO Xiaojing;WANG Ping(Department of Critical Care Medicine,the Fifth People's Hospital Affiliated to Chengdu University of Traditional Chinese Medicine,Chengdu 611130,China)

机构地区：[1]成都中医药大学附属第五人民医院重症医学科,611130

出　　处：《浙江医学》2025年第5期489-494,500,共7页Zhejiang Medical Journal

基　　金：成都市卫生健康委员会医学科研课题(2020031)。

摘　　要：目的分析重症急性胰腺炎(SAP)并发多重耐药菌(MDRO)感染的影响因素,并构建预测SAP患者发生MDRO感染的列线图模型。方法回顾性分析成都市第五人民医院2017年1月至2023年4月收治的473例SAP患者的临床资料,根据是否发生MDRO感染分为MDRO组(47例)和非MDRO组(426例)。比较两组临床资料,将差异有统计学意义的变量纳入到二元logistic回归,分析SAP患者发生MDRO感染的独立影响因素;根据独立影响因素构建列线图风险预测模型;采用ROC曲线、校准曲线评价预测模型的效能。结果473例SAP患者发生感染203例(42.9%),其中MDRO感染47例(9.9%)。MDRO组和非MDRO组患者年龄、是否有肾脏疾病史、病因、急性生理学与慢性健康状况评价Ⅱ评分、格拉斯哥昏迷评分(GCS)、入院24 h补液量、入院24 h液体平衡量、是否行连续性肾脏替代治疗、中心静脉导管留置时间、机械通气时间等比较,差异均有统计学意义(均P<0.05)。二元logistic回归分析结果显示年龄、肾脏疾病史、GCS、中心静脉导管留置时间、机械通气时间等5个变量是SAP患者并发MDRO感染的独立影响因素。由此构建的列线图预测SAP患者发生MDRO感染的AUC为0.821(95%CI:0.758~0.882),校正C指数为0.806,Hosemer-Lemeshow检验χ^(2)=6.239,P=0.512。结论根据临床资料构建的SAP并发MDRO感染的列线图风险预测模型预测效能较好,可为临床工作人员早期识别MDRO感染发生的高危人群提供参考。Objective To analyze the influencing factors for multidrug-resistant organism(MDRO)infection in severe acute pancreatitis(SAP)so as to construct a nomogram prediction model of MDRO infection.Methods A retrospective analysis was conducted on the clinical data of 473 SAP patients admitted to Chengdu Fifth People's Hospital between January 2017 and April 2023.The patients were divided into MDRO group(47 cases)and non-MDRO group(426 cases)based on the occurrence of MDRO infection.Variables with statistically significant differences after comparing clinical data of the two groups were included in binary logistic regression analysis to identify independent influencing factors for MDRO infection in SAP patients.A nomogram risk prediction model was constructed based on the independent influencing factors,and evaluated using ROC and calibration curves.Results Among the 473 patient,203(42.9%)developed infection,while 47(9.9%)were infected with MDRO.Univariate analysis revealed that age,history of chronic kidney disease,etiology,Acute Physiology and Chronic Health EvaluationⅡ,Glasgow Coma Scale(GCS),24-hour fluid resuscitation volume,24-hour fluid balance volume,continuous renal replacement therapy,duration of central venous catheterization,and duration of mechanical ventilation were statistically significant between the two groups(all P<0.05).Binary logistic regression analysis identified age,history of kidney disease,GCS,duration of central venous catheterization and duration of mechanical ventilation were independent influencing factors for MDRO infection in SAP patients.The AUC of the nomogram constructed based on these independent influencing factors for predicting MDRO infection of SAP patients was 0.821(95%CI:0.758-0.882),the calibration C-index was 0.806,and the Hosemer-Lemeshow test wasχ^(2)=6.239,P=0.512.Conclusion The nomogram model for predicting MDRO infection of SAP patients shows good performance and can serve as a reference tool for early identification of high-risk individuals for MDRO infection.

关 键 词：重症急性胰腺炎 多重耐药菌 预测模型 列线图 

分 类 号：R47[医药卫生—护理学]