经皮肾镜术后感染性发热危险因素及风险预测模型

目的探讨经皮肾镜术(PCNL)后感染性发热的危险因素,建立风险预测模型。方法选择2016年1月-2019年12月于滕州市中心人民医院行PCNL术172例患者为研究对象,根据患者术后是否发生感染性发热,分为感染组31例和未感染组141例。回顾性分析患者年龄,性别、体质量指数(BMI)、有无糖尿病史、高血压史、术前肾积水、结石直径、结石部位、术前尿培养结果、手术时间、术中灌注量等临床资料,单因素及多因素Logistic回归分析PCNL术后感染性发热的危险因素,并建立风险预测模型,Hosmer-Lemeshow检验评估模型拟合度,受试者工作特征(ROC)曲线检测Logistic回归模型的预测价值。结果患者有糖尿病史(OR=3.853,P=0.006)、有术前肾积水(OR=3.185,P=0.036)、结石直径≥2 cm(OR=2.709,P=0.026)及术前尿培养阳性(OR=3.846,P=0.006)均为PCNL术后感染性发热的独立危险因素;风险预测模型的预测概率P=1/1+e~((-4.328+1.349×(糖尿病史)+1.158×(术前肾积水)+0.997×(结石直径)+1.347×(术前尿培养))], Hosmer-Lemeshow χ²=3.381,P=0.848;该模型预测术后感染性发热的曲线下面积(AUC)为0.823,95%CI为0.744~0.902,具有良好的预测效能。结论多因素Logistic回归模型可以较好的预测PCNL术后感染性发热的发生,临床可重点关注有糖尿病史、术前肾积水、结石直径≥2 cm和术前尿培养阳性的患者。

Risk factors and risk prediction model of postoperative infectious fever in percutaneous nephrolithotomy patients

OBJECTIVE To explore the risk factors for postoperative infectious fever in patients undergoing percutaneous nephrolithotomy(PCNL)and establish the risk prediction model.METHODS A total of 172 patients who underwent PCNL in the Central People′s Hospital of Tengzhou from Jan 2016 to Dec 2019 were recruited as the study objects and were divided into the infection group with 31 cases and the non-infection group with 141 cases according to the status of postoperative infectious fever.The clinical data of the patients,including age,gender,body mass index(BMI),history of diabetes mellitus,history of hypertension,preoperative hydronephrosis,diameter of stone,sites of stone,result of preoperative urine culture,operation duration and intraoperative perfusion volume,were retrospectively analyzed.Univariate analysis and multivariate logistic regression analysis were performed for the risk factors for the postoperative infectious fever in the PCNL patients,the risk prediction model was established,the fitting degree of the model was evaluated by Hosmer-Lemeshow test,and the predictive value of the logistic regression model was tested by means of receiver operating characteristic(ROC)curve.RESULTS History of diabetes mellitus(OR=3.853,P=0.006),preoperative hydronephrosis(OR=3.185,P=0.036),diameter of stone no less than 2 cm(OR=2.709,P=0.026)and positive urine culture before surgery(OR=3.846,P=0.006)were the independent risk factors for the postoperative infectious fever in the PCNL patients.The prediction probability of the established risk prediction model was as the following:P=1/1+e((-4.328+1.349×(diabetes history+1.158×(preoperative hydronephrosis)+0.997×(stone diameter)+1.347×(preoperative urine culture))],Hosmer-Lemeshowχ²=3.381,P=0.848;the area under curve(AUC)of the model was 0.823 in prediction of the postoperative infectious fever,with 95%CI 0.744-0.902,and it had favorable predictive efficiency.CONCLUSION The multivariate logistic regression model can predict the postoperative infectious fever in the PCNL patients.It is necessary for the hospital to focus on the patients with history of diabetes mellitus,preoperative hydronephrosis,diameter of stone no less than 2 cm and positive urine culture before surgery.