双镜联合治疗复杂性肾结石后尿路感染发生的列线图预警模型构建与验证

Construction of a nomogram warning model for the occurrence of urinary tract infection after dual-lens combined treatment of complicated renal calculi

目的通过分析双镜联合治疗复杂性肾结石后尿路感染的影响因素来构建并验证列线图模型对尿路感染的预测价值。方法选取本院2020年3月至2022年1月收治的115例复杂性肾结石患者,根据术后是否出现尿路感染分为感染组(32例)与非感染组(83例)。比较两组患者的性别、年龄、体质量指数、糖尿病、高血压、冠心病、既往ESWL史、术前尿路感染、肾功能障碍、结石大小、手术时间、鹿角形结石、术前血糖水平、术中灌注量、术中通道类型、术后留置导尿管时间等资料,并进行尿路感染的单因素、多因素logistic回归分析,将筛选出的影响因素应用R软件拟合列线图模型,并绘制受试者工作特征(ROC)曲线评估列线图模型对尿路感染的预测价值。结果感染组与非感染组的性别、年龄、高血压、冠心病比较,差异均无统计学意义(均P>0.05)。感染组的体质量指数>24 kg/m^(2)、糖尿病、既往ESWL史、术前尿路感染、肾功能障碍、结石>3.5 cm、手术时间>100 min、鹿角形结石、术前血糖>6.1 mmol/L、术中灌注量>30 L、术中多通道、术后留置导尿管时间>7 d的占比均明显高于非感染组(均P<0.05)。经单因素及多因素logistic回归分析结果显示,术前尿路感染、肾功能障碍、结石大小、手术时间、鹿角形结石、术前血糖、术中灌注量是双镜联合治疗复杂性肾结石后尿路感染的影响因素(均P<0.05)。根据logistic回归分析结果绘制双镜联合治疗复杂性肾结石术后发生尿路感染的列线图模型,该模型预测尿路感染的ROC曲线下面积为0.889(95%CI:0.823~0.969)。Bootstrap法内部验证的平均绝对误差为0.020,预测曲线与标准曲线基本拟合,提示模型预测有较高的准确度。H-L拟合优度检验结果提示列线图模型预测概率与实际尿路感染率有较高一致性(χ^(2)=1.968,P=0.121)。结论术前尿路感染、肾功能障碍、结石大小、手术时间、鹿角形结石、术前血糖、术中灌注量是双镜联合治疗复杂性肾结石后尿路感染的影响因素。基于上述影响因素建立的列线图模型可有效预测患者的术后尿路感染风险。

Objective To construct and verify the predictive value of a nomogram model for urinary tract infection by analyzing the influencing factors of urinary tract infection after dual-lens combined treatment of complicated kidney stones.Methods From March 2020 to January 2022,115 patients withcomplicated kidney stones who were admitted to our hospital were selected and divided into infection group(32 cases)and non-infection group(83 cases)according to whether urinary tract infection occurred after operation.The gender,age,body mass index,diabetes mellitus,hypertension,coronary heart disease,previous extracorporeal shock wave lithotripsy(ESWL)history,preoperative urinary tract infection,renal dysfunction,stone size,operation time,staghorn stones,preoperative blood glucose level,and operation time were compared between the two groups.The univariate and multivariate logistic regression of urinary tract infection was carried out,and the selected influencing factors were fitted to the nomogram model by R software,and the receiver operating characteristic(ROC)curve was drawn to evaluate the predictive value of the nomogram model for urinary tract infection.Results There were no significant difference in gender,age,hypertension and coronary heart disease between the infection group and the non-infection group(all P>0.05).In the infection group,body mass index>24 kg/m2,diabetes mellitus,previous history of ESWL,preoperative urinary tract infection,renal dysfunction,stones>3.5 cm,operation time>100 min,staghorn stones,preoperative blood glucose>6.1 mmol/L,the percentages of intraoperative perfusion volume>30 L,intraoperative multi-channel and postoperative indwelling catheter>7 days were significantly higher than those in the non-infection group(all P<0.05).Univariate and multivariate logistic regression analysis showed that preoperative urinary tract infection,renal dysfunction,stone size,operation time,staghorn stones,preoperative blood sugar,and intraoperative perfusion volume were the most important factors in the urinary tract after double-endoscopy combined treatment of complicated kidney stones.Influencing factors of infection(all P<0.05).According to the logistic regression results,a nomogram model was drawn for the combined treatment of urinary tract infection after complicated renal calculi.The area under ROC curve of this model was 0.889(95%CI:0.823-0.969).The average absolute error of the internal verification of the Bootstrap method was 0.020,and the prediction curve basically fitted the standard curve,indicating that the model prediction had a high accuracy.The results of the H-L goodness-of-fit test indicated that the predicted probability of the nomogram model had high consistency with the actual rate of urinary tract infection(χ^(2)=1.968,P=0.121).Conclusions Preoperative urinary tract infection,renal dysfunction,stone size,operation time,staghorn stones,preoperative blood glucose,and intraoperative perfusion volume are the influencing factors of urinary tract infection after dual-endoscopy combined treatment of complicated kidney stones.The nomogram model established based on the above influencing factors can effectively predict the risk of postoperative urinary tract infection in patients.