输尿管镜钬激光碎石术后尿路感染风险评分工具的构建及应用

Construction and application of a risk scoring tool for urinary tract infection after ureteroscopic holmium laser lithotripsy

摘要:

目的 构建、验证输尿管镜钬激光碎石术后尿路感染(UTI)风险评分工具。 方法 便利抽样选取2018年1月—2021年12月在宜昌市某三级甲等综合医院泌尿外科住院的414例输尿管镜钬激光碎石术后患者作为研究对象, 按照随机数字表法将所有患者按7∶3的比例随机分为建模组(288例)和验证组(126例)。采用单因素和多因素logistic回归分析输尿管镜钬激光碎石术后患者发生UTI的危险因素, 采用建模组数据构建输尿管镜钬激光碎石术后UTI风险预测模型并绘制列线图, 使用验证组数据对模型进行验证。采用受试者工作特征(ROC)曲线和Hosmer-Lemeshow检验验证模型的拟合优度及预测效果。在风险预测模型的基础上构建风险评分工具, 并在临床中应用。 结果 414例患者中有68例发生术后UTI, 发生率为16.43%;单因素分析结果显示, 年龄、吸烟史、糖尿病史、慢性病史、肾积水史、术前UTI史、术前血清蛋白水平、氮质血症史、肝功能异常、肾功能异常、术中灌注时间、结石大小、留置导尿管日数、抗菌药物使用日数、是否联合使用抗菌药物是输尿管镜钬激光碎石术后患者发生UTI的影响因素。多因素logistic回归分析显示, 年龄、糖尿病史、肾功能异常、术中灌注时间、是否联合使用抗菌药物是输尿管镜钬激光碎石术后发生UTI的独立影响因素。预测模型ROC曲线下面积(AUC)为0.735, 约登指数最大值为0.448, 对应的灵敏度为73.3%, 特异度为71.5%, 可得临界值为5.5分。以6分为界, 依据评分工具将患者分为低危组(0~5分)和高危组(6~13分)。将输尿管镜钬激光碎石术后UTI风险评分工具应用到临床, 模型的准确率为86.8%, 术后UTI的风险评分工具临床上应用良好。 结论 风险评分工具可为护理人员提供指导, 个性化预测输尿管镜钬激光碎石术后患者发生UTI的风险, 协助护理人员提前干预。

Abstract:

Objective To construct and validate a risk scoring tool for urinary tract infection (UTI) after ureteroscopic holmium laser lithotripsy(UHLL). Methods A total of 414 post-UHLL patients in the department of urology in a tertiary first-class hospital in Yichang City from January 2018 to December 2021 were selected as study objects through convenient sampling. According to a random number table, 288 patients were randomly assigned to the modeling group, and 126 patients to the validation group, in a 7 ∶3 ratio. Risk factors for UTI in patients after UHLL were identified via univariate and multivariate logistic regression analysis. Based on the data from the mode-ling group, a risk prediction model for post-UHLL UTI was constructed and a nomogram was depicted accordingly, the model was validated through data in validation group. The goodness of fit and predictive performance of the model were assesed using receiver operating characteristic (ROC) curve and Hosmer-Lemeshow test. A risk scoring tool based on the risk prediction model was constructed and applied in clinical practice. Results Among the 414 patients, 68 (16.43%) experienced postoperative UTI. Univariate analysis showed that age, history of smoking, diabetes, chronic diseases, hydronephrosis, and preoperative UTI, as well as preoperative serum protein level, history of azotemia, abnormal liver function, abnormal renal function, intraoperative perfusion time, stone size, urinary catheter retention days, duration of antimicrobial use, and combined use of antimicrobial were influencing factors for the occurrence of UTI in patients after UHLL. Multivariate logistic regression analysis revealed that age, history of diabetes, abnormal renal function, intraoperative perfusion time and combined use of antimicrobial agents were independent influencing factors for post-UHLL UTI. The area under ROC curve of the prediction model was 0.735, the maximum Yoden index was 0.448, the corresponding sensitivity and specificity were 73.3% and 71.5%, respectively and the critical value was determined as 5.5 points. With a cut-off value of 6 points, patients were divided into a low-risk group (0-5 points) and a high-risk group (6-13 points). Applying the post-UHLL UTI risk scoring tool in clinical practice yielded an accuracy of 86.8%, demonstrating favorable clinical utility. Conclusion The risk scoring tool provides guidance for nursing staff by predicting personalized risk of post-UHLL UTI and assisting them in implementing early interventions.