上尿路结石梗阻所致尿源性脓毒症的危险因素及其预测模型构建与验证  

Risk factors of urosepsis caused by upper urinary tract stone obstruction and construction and verification of its predictive model

作　　者：王悦 张宗绵 唐钟祥 蔡耿鑫 何晓君 庄峥志 林润培 WANG Yue;ZHANG Zongmian;TANG Zhongxiang;CAI Gengxin;HE Xiaojun;ZHUANG Zhengzhi;LIN Runpei(Department of Critical Care Medicine,Guangdong Provincial People′s Hospital Nanhai Hospital,Foshan 528251,China;不详)

机构地区：[1]广东省人民医院南海医院重症医学科,广东佛山528251 [2]佛山市南海区人民医院重症医学科 [3]广东省中西医结合医院重症医学科

出　　处：《山东医药》2022年第25期16-20,共5页Shandong Medical Journal

基　　金：广东省医学科研基金立项项目(B2021323);佛山市医学科研立项课题(20210201);佛山市“十四五”医学重点和培育专科建设(FSPY145037)。

摘　　要：目的分析上尿路结石梗阻所致尿源性脓毒症的危险因素,构建并验证其列线图预测模型。方法选择上尿路结石梗阻所致尿源性脓毒症患者122例(观察组),同期因上尿路结石梗阻所致尿路感染但未达到脓毒症患者60例(对照组),收集两组基线资料,采用多因素Logistic回归模型分析上尿路结石梗阻所致尿源性脓毒症的危险因素。采用随机分层抽样法将182例患者分为训练集和验证集,在训练集中通过rms程序包构建列线图预测模型,在验证集中采用校准曲线、一致性指数及受试者工作特性(ROC)曲线验证其预测效能。结果两组性别、年龄、体温、心率、平均动脉压、血WBC计数、血RBC计数、PLT计数、血钠、血钾及血清白蛋白(ALB)、C反应蛋白(CRP)、降钙素原(PCT)和机械通气、连续性肾脏替代治疗比较P均<0.05,两组尿WBC计数、结石数量、肾积水程度、慢性病、住院病死率比较P均>0.05。以是否发生尿源性脓毒症为因变量,以上述有统计学差异的指标为自变量,纳入多因素Logistic回归模型。结果发现,女性、高体温及低ALB、高CRP、高PCT是上尿路结石梗阻所致尿源性脓毒症的独立危险因素(P均<0.05)。基于以上危险因素,以训练集为基础构建了列线图预测模型,利用验证集数据绘制校准曲线,结果显示上尿路结石梗阻所致尿源性脓毒症的预测值与实际值基本一致;ROC曲线分析显示,上尿路结石梗阻所致尿源性脓毒症的预测模型在训练集中的曲线下面积为0.972(95%CI:0.933~1.000),在验证集中的曲线下面积为0.943(95%CI:0.901~0.986)。结论女性、高体温及低ALB、高CRP、高PCT是上尿路结石梗阻所致尿源性脓毒症的独立危险因素,据此构建的列线图预测模型,经验证能够准确预测上尿路结石梗阻所致尿源性脓毒症的发生风险。Objective To investigate the risk factors of urosepsis caused by upper urinary tract stone obstruction and to construct and verify the nomogram predictive model.Methods Totally 122 patients with urosepsis caused by obstruction of upper urinary tract stones(observation group)and 60 patients with urinary tract infection without sepsis caused by obstruction of upper urinary tract stones(control group)were enrolled,and their baseline data were collected retrospectively.Logistic regression analysis was performed to detect the risk factors of urosepsis caused by upper urinary tract stone obstruction.A total of 182 patients were divided into the training set and validation set according to randomized stratified sampling method.The predictive nomogram was constructed by rms package in the training set,and the prediction performance of nomogram was assessed by calibration curve,consistency index and receiver operating characteristic(ROC)curve in the validation set.Results Significant differences were found in the gender,age,temperature,heart rate,mean arterial pressure,serum WBC count,RBC count,PLT count,sodium,potassium,serum albumin(ALB),C-reaction protein(CRP),procalcitonin(PCT),using of mechanical ventilation,and continuous renal replacement therapy between the two groups(all P<0.05),while there was no difference in urinary WBC count,stone number,hydronephrosis degree,chronic disease or in-hospital mortality between these two groups(all P>0.05).A multivariate Logistic regression model was constructed with urosepsis as the dependent variable,and all above as the independent variables.The results showed that female gender,high temperature,low ALB,high CRP and PCT were independent risk factors for urosepsis caused by upper urinary tract stone obstruction(all P<0.05).Based on the above risk factors,a nomogram was constructed by data from training set,and the calibration curve was drawn using the data from validation set.It was shown that the predicted value of urosepsis caused by upper urinary tract stone obstruction was

关 键 词：尿源性脓毒症 上尿路结石梗阻 危险因素 列线图预测模型 

分 类 号：R631[医药卫生—外科学]