胆总管结石ERCP术后并发胆道感染的高危因素及风险模型建立  

High-risk factors and risk model for post-ERCP biliary infection in patients with common bile duct stones

作　　者：张婷 秦文昊 周静 ZHANG Ting;QIN Wenhao;ZHOU Jing(Department of Gastroenterology(Endoscopy),the Third Affiliated Hospital of Naval Military Medical University,Shanghai 201805,China;Department of Gastroenterology,the Third Affiliated Hospital of Naval Military Medical University,Shanghai 201805,China;Department of Radiological Interventional,the Third Affiliated Hospital of Naval Military Medical University,Shanghai 201805,China)

机构地区：[1]海军军医大学第三附属医院消化内科(内镜科),上海201805 [2]海军军医大学第三附属医院消化内科,上海201805 [3]海军军医大学第三附属医院放射介入科,上海201805

出　　处：《东南大学学报(医学版)》2025年第2期236-242,共7页Journal of Southeast University(Medical Science Edition)

基　　金：上海市自然科学基金面上项目(23ZR1478200)。

摘　　要：目的:分析胆总管结石经内镜逆行性胰胆管造影术(ERCP)后并发胆道感染高危因素,建立风险模型并进行评价。方法:回顾性收集2023年1月至2024年10月在本院行ERCP的胆总管结石患者的临床资料,纳入术后1周并发胆道感染患者52例为胆道感染组,同期随机选择术后1周未并发胆道感染患者156例为对照组。比较两组基线资料、手术资料、既往病史及实验室检测结果,通过最小绝对收缩和选择算子(LASSO)回归优化筛选变量,二元Logistic回归模型确定独立预测变量,据此建立列线图风险模型,受试者工作特征(ROC)曲线评价模型预测性能,Hosmer-Lemeshow检验和校准曲线评价模型校准度,决策曲线分析(DCA)曲线、临床影响曲线(CIC)评价模型临床效用。结果:胆道感染组糖尿病、结石直径≥10 mm、结石部位壶腹段、胆道梗阻位置为高位、有十二指肠乳头憩室、Oddi括约肌功能障碍、手术时间≥30 min占比及术前总胆红素水平高于对照组,术前白蛋白(ALB)、血红蛋白水平低于对照组(P<0.05)。LASSO回归模型筛选出4个非零系数潜在预测变量(糖尿病、胆道梗阻位置、Oddi括约肌功能障碍、术前ALB),二元Logistic回归分析确定糖尿病、胆道梗阻位置、术前ALB为独立预测变量(P<0.05)。ROC曲线显示,列线图风险模型预测胆总管结石患者ERCP术后并发胆道感染的曲线下面积(AUC)为0.793,灵敏度为0.846,特异度为0.660,准确度为0.707。Hosmer-Lemeshow检验与校准曲线显示,模型校准度较高。DCA曲线显示,在0.07~0.71阈值概率范围内,列线图风险模型具有较高临床净获益。CIC分析显示,当阈值概率超过0.6时,模型判定为术后并发胆道感染高风险患者与实际发生术后并发胆道感染患者高度匹配。结论:胆总管结石ERCP术后并发胆道感染的预测因素包括糖尿病、胆道梗阻位置、术前ALB,且据此构建的风险模型预测性能良好。Objective:To analyze the high-risk factors for biliary infection after endoscopic retrograde cholangiopancreatography(ERCP)in patients with common bile duct stones,and to establish and evaluate a risk prediction model.Methods:Clinical data of patients with common bile duct stones who underwent ERCP from January 2023 to October 2024 were retrospectively analyzed.Fifty-two patients who developed biliary infection within one week post-operation were assigned to the biliary infection group,and 156 patients without biliary infection were randomly selected as the control group.Baseline characteristics,surgical data,medical history,and laboratory data were collected and compared between the twe groups.The least absolute shrinkage and selection operator(LASSO)regression was used for optimized variable selection,binary Logistic regression was performed to identify independent predictors,and a nomogram risk model was established accordingly.The model s predictive performance was evaluated using receiver operating characteristic(ROC)curve,calibration was assessed using Hosmer-Lemeshow test and calibration curve,and clinical utility was evaluated using decision curve analysis(DCA)and clinical impact curve(CIC).Results:The biliary infection group showed higher proportions of diabetes,stone diameter≥10 mm,stone location at ampulla,high-position biliary obstruction,duodenal papillary diverticulum,Oddi sphincter dysfunction,operation time≥30 min,and higher preoperative total bilirubin levels,while preoperative albumin(ALB)and hemoglobin levels were lower compared to the control group(P<0.05).LASSO regression identified four non-zero coefficient potential predictors(diabetes,biliary obstruction position,Oddi sphincter dysfunction,preoperative ALB),and binary Logistic regression confirmed diabetes,biliary obstruction position,and preoperative ALB as independent predictors(P<0.05).ROC curve analysis showed the nomogram risk model had an area under the curve(AUC)of 0.793,sensitivity of 0.846,specificity of 0.660,and accuracy of

关 键 词：胆总管结石 经内镜逆行性胰胆管造影术 胆道感染 风险模型 

分 类 号：R657.42[医药卫生—外科学]