经皮肾镜碎石取石术中、术后严重出血的列线图模型建立

Establishment of a nomogram model for severe bleeding after percutaneous nephrolithotomy

目的探讨经皮肾镜取石(PCNL)术中、术后严重出血的危险因素并建立其风险预测的列线图模型。方法选取2018年1月至2019年6月该院泌尿外科诊治的363例上尿路结石行PCNL患者为研究对象,依据患者血红蛋白下降量分为出血组和对照组。采用logistic回归分析筛选PCNL术中、术后并发严重出血的独立危险因素,然后用纳入筛选的独立危险因素建立列线图预测模型,并对模型的预测性及准确度进行验证。结果纳入的363例患者中,出血组119例,对照组244例。性别、BMI、糖尿病、结石位置、鹿角形结石、多发结石、手术时间、手术分期是PCNL术中、术后严重出血的独立危险因素(P<0.05);受试者工作特征(ROC)曲线对该模型预测效能进行评价,结果显示模型联合预测效能的曲线下面积为0.826,95%CI:0.782~0.870;将拟合出的logistic多因素回归模型,通过列线图实现模型的可视化,使用区分度和符合度校验此列线图的效能,结果显示随着BMI、糖尿病、结石位置、鹿角形结石、手术时间的增加,评分逐渐增加,严重出血发生的风险也逐渐增加。依logistic回归模型建立的列线图模型区分度好,模型的符合度好(C-index=0.807)。结论构建的列线图预测模型对PCNL发生严重出血具有较好的预测效能。

Objective To explore the risk factors of severe bleeding after percutaneous nephrolithotomy(PCNL)and establish a risk prediction nomogram model.Methods The clinical data of 363 patients with upper urinary tract calculi who underwent PCNL from January 2018 to June 2019 in hospital were retrospectively analyzed.The patients were divided into the bleeding group and the control group according to hemoglobin reduction.Logistic regression analysis was used to screen the independent risk factors of severe bleeding after PCNL operation.Then,a nomogram prediction model was established with independent risk factors selected for screening,and the predictive accuracy and accuracy of the model were verified.Results Among the 363 patients,119 cases were in the bleeding group.The control group consisted of 244 patients.Gender,BMI,diabetes,stone location,staghorn calculi,multiple stones,operative time and staged operation were independent risk factors for severe postoperative bleeding in PCNL(P<0.05).Receiver operating characteristic(ROC)curve was used to evaluate the predictive effectiveness of the model,the results showed that area under the curve of the model′s combined prediction efficiency was 0.826,95%CI:0.782-0.870.The visualization of the model was realized by the nomogram,and the efficiency of the nomogram was verified by the division and coincidence.With the increase of BMI,diabetes,stone location,staghorn calculi and operation time,the score increased gradually,and the risk of serious bleeding gradually increased.The nomogram model established by logistic regression model was well distributed.The compliance of the model was good(C-index=0.807).Conclusion The nomogram prediction model has a good predictive effect on the occurrence of severe bleeding in PCNL.