Xuan Zijun, Yu Zhikang, Tan Guobin, Ding Ning, He Huibin, Yu Shichao, Liu Guoqing, Zhu Xiping, Zhu Bo, Liu Zhe
Department of Urology, Dongguan Kanghua Hospital, Dongguan, China.
Department of Urology, Maoming People's Hospital, Maoming, China.
Transl Androl Urol. 2022 Feb;11(2):228-237. doi: 10.21037/tau-22-34.
The occurrence of systemic inflammatory response syndrome (SIRS) is an early alert for sepsis after flexible ureteroscopy (fURS). Once sepsis occurs, it often leads to severe or fatal consequences. We aimed to identify SIRS patients preoperatively by developing and validating a feasible prognostic nomogram model based on retrospective cohort analysis.
A total of 311 patients who underwent fURS in Dongguan Kanghua Hospital (Dongguan, China) between 2016 and 2020 were included and randomly divided into a primary cohort (n=219) and validation cohort (n=92). Single factor regression analysis was used to identify the primary cohort's meaningful characters between SIRS and non-SIRS groups. Factors of the primary cohort were then identified by least absolute shrinkage and selection operator (LASSO) regression analysis, and a nomogram was built to execute the subsequent analysis using these factors. Finally, we analyzed and drew the calibration curve, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA) curve to validate the prognostic value of the nomogram in calibration and discrimination.
Review of the single regression analysis of characters in the primary cohort showed gender, stone burden, diabetes, neutrophil (N), lymphocyte (L), neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), lymphocyte-to-monocytes ratio (LMR), urine-WBC, nitrite (Nit), urine culture, and surgery time as significant factors between the SIRS and non-SIRS groups (P<0.05). The LASSO regression analysis suggested NLR, PLR, and urine culture were substantial factors in predicting SIRS postoperatively, lambda.min and lambda.1se (standard error, SE) were 0.01491 and 0.0796. A nomogram built with the three factors showed good calibration and discrimination, with the Brier values 0.064 and 0.034 and the area under curve (AUC) values 0.897 (95% CI: 0.837-0.957) and 0.976 (95% CI: 0.947-1.000) in the primary and validation cohort, respectively. DCA demonstrated the nomogram was clinically useful, and the predict probability of SIRS's occurrence was very close to the actual rate as the risk threshold increased by higher than 60% in clinical impact curve analysis.
NLR, PLR, and urine culture were significantly related to the occurrence of SIRS's after fURS. The nomogram with these three factors showed excellent calibration, discrimination, and clinical usefulness.
全身炎症反应综合征（SIRS）的发生是输尿管软镜检查（fURS）后脓毒症的早期预警信号。一旦发生脓毒症，往往会导致严重或致命后果。我们旨在通过基于回顾性队列分析开发并验证一个可行的预后列线图模型，在术前识别SIRS患者。
纳入2016年至2020年期间在东莞康华医院（中国东莞）接受fURS的311例患者，并随机分为原始队列（n = 219）和验证队列（n = 92）。采用单因素回归分析确定原始队列中SIRS组和非SIRS组之间有意义的特征。然后通过最小绝对收缩和选择算子（LASSO）回归分析确定原始队列的因素，并使用这些因素构建列线图以进行后续分析。最后，我们分析并绘制校准曲线、受试者工作特征（ROC）曲线和决策曲线分析（DCA）曲线，以验证列线图在校准和鉴别方面的预后价值。
对原始队列特征的单因素回归分析显示，性别、结石负荷、糖尿病、中性粒细胞（N）、淋巴细胞（L）、中性粒细胞与淋巴细胞比值（NLR）、血小板与淋巴细胞比值（PLR）、淋巴细胞与单核细胞比值（LMR）、尿白细胞、亚硝酸盐（Nit）、尿培养和手术时间是SIRS组和非SIRS组之间的显著因素（P<0.05）。LASSO回归分析表明，NLR、PLR和尿培养是预测术后SIRS的重要因素，lambda.min和lambda.1se（标准误，SE）分别为0.01491和0.0796。用这三个因素构建的列线图显示出良好的校准和鉴别能力，原始队列和验证队列中的Brier值分别为0.064和0.034，曲线下面积（AUC）值分别为0.897（95%CI：0.837 - 0.957）和0.976（95%CI：0.947 - 1.000）。DCA表明列线图具有临床实用性，在临床影响曲线分析中，随着风险阈值增加高于60%，SIRS发生的预测概率与实际发生率非常接近。
NLR、PLR和尿培养与fURS后SIRS的发生显著相关。包含这三个因素的列线图显示出优异的校准、鉴别和临床实用性。
