经皮肾镜碎石取石术后发生全身炎症反应综合征的列线图模型建立

Establishment and evaluation of a nomogram model of systemic inflammatory response syndrome after percutaneous nephrolithotomy

目的经皮肾镜碎石取石术(PCNL)后全身炎症反应综合征(SIRS)的主要原因尚不清楚。文中旨在分析PCNL后发生SIRS的危险因素,建立列线图模型。方法回顾性分析2017年12月至2018年12月桂林医学院附属医院泌尿外科213例因上尿路结石行PCNL患者的临床资料。依据SIRS发生情况将患者分为SIRS组(SIRS患者)、对照组(无SIRS患者)。采用logistic回归分析PCNL后SIRS的风险因素,基于logistic回归模型建立列线图模型。结果SIRS组54例,对照组159例。性别(OR=2.547,95%CI:1.229~5.275)、糖尿病(OR=5.027,95%CI:1.442~17.525)、结石表面积(OR=2.657,95%CI:1.206~5.853)、术后即刻NLR(OR=3.793,95%CI:1.749~8.02)、手术时间(OR=2.985,95%CI:1.305~6.826)、输血(OR=12.50,95%CI:1.954~80.056)是PCNL术后SIRS发生的危险因素(P<0.05)。基于上述logistic多因素回归模型结果,通过列线图实现模型的可视化显示,随着术后即刻NLR比值、手术时间、结石表面积等的增加,评分逐渐增加,SIRS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=0.791)。结论根据性别、糖尿病病史、结石表面积、术后即刻NLR等危险因素,构建的列线图模型具有较好的预测效能,对临床具有指导意义。

Objective The main cause of systemic inflammatory response syndrome(SIRS)after percutaneous nephrolithotomy(PCNL)was still unclear. The purpose of this study was to investigate the risk factors associated with SIRS after PCNL and establish the nomogram model. Methods A retrospective analysis of 213 cases of PCNL patients due to upper urinary calculi admitted to urology department in affiliated hospital of guilin medical college from December 2017 to December 2018 was performed. According to the occurrence of SIRS,patients were divided into SIRS group(SIRS patients)and control group(patients without SIRS). Logistic regression was used to analyze the risk factors of SIRS after PCNL,and a nomogram model was established based on logistic regression model. Results There were 54 cases in the SIRS group and 159 in the control group. Gender(OR=2.547,95%cl:1.229-5.275),diabetes (OR=5.027,95%cl:1.442-17.525),calculi surface area (OR= 2.657,95%cl:1.206-5.853),NLR immediately after surgery(OR=3.793,95%cl:1.749-8.02),operation time(OR=2.985, 95%cl:1.305-6.826),and blood transfusion(OR=12.50,95%cl:12.50). 1.954-80.056)were the risk factors of SIRS after PCNL(P <0.05). Based on the results of the logistic multi-factor regression model mentioned above,visualized display of the model was achieved by using column and diagram. As the NLR ratio,operation time and stone surface area increased immediately after the operation,the score gradually increased,and the risk of SIRS gradually increased. The nomogram model established according to logistic regression model has good differentiation and model consistency(c-index =0.791). Conclusion According to the risk factors,such as gender,diabetes history,stone surface area,immediate postoperative NLR,the constructed nomogram model has good predictive efficacy,which is of guiding significance for clinical practice.