建立一种痔术后尿潴留风险预测模型

 目的 研究痔术后尿潴留的危险因素，并构建预测尿潴留风险的列线图。方法 回顾性收集复旦大学附属中山医院青浦分院2016年1月至2020年12月159例痔术后未留置尿管的患者的临床资料。采用LASSO方法筛选出尿潴留的危险因素，通过Logistic回归分析绘制痔术后需要留置尿管的列线图模型，并评估此模型的准确性（C-index），通过内部验证评价模型。结果 该模型由年龄、性别、BMI、手术方式、外痔切除、术后补液量和术后镇痛等7个变量组成，具有较高的C-index（0.841，95%CI：0.774~0.908）和校准曲线，内部验证组仍有较高的C-index（0.793）。临床曲线分析得出阈值率在0.02~0.83时，该模型具有更好的临床使用价值。结论 痔术后尿潴留的风险预测模型具有较高的准确性和良好的稳定性，有助于临床医师评估痔术后是否留置尿管。

Establishment of an prediction model for the risks of urinary retention after hemorrhoids surgery

Objective To study the risk factors of urinary retention after hemorrhoid surgery, and to construct a column line chart to predict the risks of urinary retention.Methods From Jan 2016 to Dec 2020 in Qinpu Branch of Zhongshan Hospital, Fudan University, clinical data of 159 patients without indwelling urinary catheters after hemorrhoid surgery were retrospectively collected. The risk factors were screened by LASSO method.The column line graph model of the need for indwelling urinary catheter was drawn by Logistic regression analysis.The accuracy of this model (C-index) was assessed, followed by internal validation to evaluate this model.Results The prediction model consisted of 7 variables: age, sex, BMI, surgical procedure, external hemorrhoid removal, postoperative rehydation and postoperative analgesia. The model had a high C-index of 0.841 (95% CI: 0.774-0.908) and calibration curve, and still had a high C-index of 0.793 in the internal validation group. And the clinical curve analysis yielded a threshold rate at 0.02-0.83, suggesting a better clinical use of the model.Conclusion This model has quite accurate and provide a reference in whether to leave a urinary catheter in place after hemorrhoid surgery.