山东第二医科大学学报

目的 构建并验证前列腺增生(benign prostatic hyperplasia, BPH)患者经尿道前列腺手术后尿潴留的风险预测列线图模型。方法 回顾性分析山东第二医科大学附属医院2015年7月～2024年3月收治的2893例行经尿道前列腺手术(包括等离子切除术、等离子剜除术及钬激光剜除术) BPH患者的临床资料。根据术后是否发生尿潴留分为尿潴留组(n=58)和非尿潴留组(n=2835)。应用LASSO回归初步筛选变量，采用Logistic回归分析确定独立危险因素，构建预测列线图模型。采用Bootstrap法(重复抽样1000次)进行内部验证，ROC曲线分析预测效能。结果 术后尿潴留发生率为2.01%(58/2893)。多因素Logistic回归分析显示：术前尿潴留史(OR=10.496,95%CI:5.234～21.987)、年龄≥70岁(OR=1.832,95%CI:1.235～2.724)、前列腺体积≥80 mL(OR=2.156,95%CI:1.456～3.187)、术前国际前列腺症状评分术前(IPSS)评分≥20分(OR=1.756,95%CI:1.234～2.498)、膀胱逼尿肌功能不全(OR=2.867,95%CI:1.987～4.134)、残余尿量≥100 mL(OR=2.123,95%CI:1.456～3.098)均为独立危险因素(P<0.05),差异有统计学意义。列线图模型经Bootstrap内部验证的C-index为0.875(95%CI:0.834～0.916),校准曲线显示预测概率与实际概率吻合良好。模型预测术后尿潴留的灵敏度为86.21%,特异度为85.47%,AUC为0.912(95%CI:0.876～0.948)。结论 基于临床危险因素构建的预测列线图模型具有良好的预测效能和临床实用性，可为术前风险评估提供客观依据。随着手术技术的进步，术后尿潴留的发生率显著降低，该模型在早期识别高危患者方面具有重要价值。

Objective To develop and validate a nomogram model for predicting postoperative urinary retention risk in patients undergoing transurethral prostate surgery for benign prostatic hyperplasia(BPH).Methods A retrospective analysis was conducted on clinical data from 2,893 BPH patients who underwent transurethral prostate surgery(including plasmakinetic resection, plasmakinetic enucleation, and holmium laser enucleation) at the Affiliated Hospital of Shandong Second Medical University between July 2015 and March 2024.Patients were divided into urinary retention group(n=58) and non-retention group(n=2835).Variables were initially screened using LASSO regression, independent risk factors were identified through Logistic regression analysis, and a predictive nomogram model was constructed.Internal validation was performed using Bootstrap method(1,000 resamples),and predictive performance was analyzed using ROC curves.Results The postoperative urinary retention rate was 2.01%(58/2893).Multivariate Logistic regression analysis identified the following independent risk factors(P<0.05):Preoperative urinary retention history(OR=10.496,95%CI:5.234～21.987),age ≥70 years(OR=1.832,95%CI:1.235～2.724),prostate volume ≥80 mL(OR=2.156,95%CI:1.456～3.187),preoperative IPSS score ≥20(OR=1.756,95%CI:1.234-2.498),detrusor dysfunction(OR=2.867,95%CI:1.987～4.134),and residual urine volume ≥100 mL(OR=2.123,95%CI:1.456～3.098).The nomogram model showed a C-index of 0.875(95%CI:0.834～0.916) after Bootstrap internal validation, with good calibration between predicted and actual probabilities.The model demonstrated sensitivity of 86.21%,specificity of 85.47%,and AUC of 0.912(95%CI:0.876～0.948).Conclusion The predictive nomogram model based on clinical risk factors shows good predictive performance and clinical utility, providing an objective basis for preoperative risk assessment.With the advancement of surgical techniques, the incidence of postoperative urinary retention has significantly decreased, making this model valuable for early identification of high-risk patients.