大型肾结石行微创经皮肾镜取石术后发生尿源性脓毒症的预测模型构建

Construction of a predictive model for urinary sepsis after minimally invasive percutaneous nephrolithotomy for large kidney stones

目的分析大型肾结石患者行微创经皮肾镜取石术(mPCNL)后尿源性脓毒症的发生率和危险因素,并构建列线图预测模型用于指导临床诊断尿源性脓毒症。方法回顾性选取2017年2月至2022年2月在本院诊断的188例大型肾结石(结石长径≥25 mm)患者为研究对象,均采用单通道钬激光mPCNL进行治疗。术后7 d,根据欧洲泌尿外科学会指南关于尿源性脓毒症的诊断标准将患者分为脓毒症组(28例)和无脓毒症组(160例)。分别采用单因素和多因素logistic回归分析筛选尿源性脓毒症的危险因素,并建立列线图模型。结果单因素分析结果显示,脓毒症组的术前尿白细胞计数、尿亚硝酸盐阳性率、相邻肾盏内结石阳性率、结石最大横截面积、肾积水直径、尿细菌培养阳性率和手术时间均较无脓毒症组明显增加,差异均有统计学意义(均P<0.001)。多因素logistic回归分析结果显示,尿白细胞计数、尿亚硝酸盐阳性、相邻肾盏内结石阳性、结石最大横截面积和肾积水直径均是mPCNL术后发生尿源性脓毒症的独立危险因素(均P<0.001)。R软件建立列线图模型,总分220分。受试者工作特征(ROC)曲线分析结果显示,列线图预测尿源性脓毒症的曲线下面积(AUC)为0.876(95%CI:0.810~0.923)。绘制校准曲线显示模型评估与实际结果有较好的一致性。结论大型肾结石患者行mPCNL术后有一定的尿源性脓毒症发生风险,其中术前尿白细胞计数、尿亚硝酸盐阳性、相邻肾盏内结石阳性、结石最大横截面积和肾积水直径是影响其发生的独立危险因素。建立操作简单、可视化强的列线图模型对指导临床预测尿源性脓毒症有较好的应用价值,值得临床推广。

Objective To analyze the incidence and risk factors of urinary sepsis after minimally invasive percutaneous nephrolithotomy(mPCNL)in large renal calculus patients,and to construct a nomogram prediction model for clinical diagnosis of urinary sepsis.Methods A retrospective analysis was performed on 188 patients with large kidney stones(longest diameter≥25 mm)diagnosed in our hospital from February 2017 to February 2022.All patients were treated with single-channel holmium laser mPCNL.Seven days after surgery,patients were divided into sepsis group(28 cases)and no sepsis group(160 cases)according to the diagnostic criteria of the European Society of Urology guidelines on urogenic sepsis.The risk factors of urogenic sepsis were screened by univariate and multivariate logistic regression analysis,and the nomogram model was established.Results Single-actor comparison showed that the preoperative urinary white blood cell count,urinary nitrite positive rate,positive rate of stones in adjacent calyces,maximum cross-sectional area of stones,diameter of hydronephrosis,positive rate of urinary bacterial culture and operation time after mPCNL in sepsis group were significantly increased compared with those in non-sepsis group,with statistical significance(all P<0.001).Multivariate logistic regression analysis showed that urine white blood cell count,urine nitrite positive,stone positive in adjacent calyces,stone maximum cross-sectional area and hydronephrosis diameter were independent risk factors for urogenic sepsis(all P<0.001).R software built a nomogram model,with a total score of 220.Receiver operating characteristics(ROC)curve results showed that the area under the curve(AUC)predicted by the nomogram for urogenic sepsis was 0.876(95%CI:0.810~0.923).The calibration curve was drawn to show that the model evaluation was in good agreement with the actual results.Conclusions Large renal calculi after mPCNL have a certain risk of urogenic sepsis.Preoperative urinary leukocyte count,urinary nitrite positive,adjacent renal calyceal calculi positive,maximum cross-sectional area of calculi and diameter of hydronephrosis are the independent risk factors.Establishment of a simple and visual nomogram model has good value in guiding clinical prediction of urogenic sepsis,which is worthy of popularization.