某三甲医院老年住院患者下尿路症状的关联因素分析和风险预测模型的构建

Factors associated with lower urinary tract symptoms and related prediction model in elderly inpatients in a tertiary hospital

  • 摘要: 背景 随着人口老龄化,下尿路症状(lower urinary tract symptoms,LUTS)已成为世界范围内的一个重大公共卫生问题,因此,深入探究其有效防治策略变得尤为迫切。目的 对老年LUTS的关联因素进行分析并构建风险评估模型。方法 本研究为回顾性队列研究,选取2023 年1 月至2024 年1 月在解放军总医院第一医学中心住院的60 岁以上老年人作为研究对象,根据是否有LUTS分为LUTS组和无LUTS组,并按性别分层进行。采用多因素Logistic回归分析老年LUTS的独立关联因素,绘制森林图,并绘制受试者工作特征(receiver operating characteristic,ROC)曲线、临床校准曲线、决策分析曲线(decision curve analysis,DCA)对模型进行评估。结果 本研究共纳入4 122 例老年住院患者,其中男性2 394 例,女性1 728例。594 例(14.48%)有LUTS,其中男性438 例(18.30%),女性159 例(9.20%)。多因素Logistic 回归模型分析显示,年龄(OR=1.051,95% CI:1.022 ~ 1.084)、便秘(OR=1.583,95% CI:1.250 ~ 2.002)、糖尿病(OR=2.591,95% CI:1.949 ~ 3.443)、手术史(OR=2.334,95% CI:1.781 ~ 3.043)、前列腺增生(OR=2.920,95% CI:2.262 ~ 3.781)、高血压(OR=2.771,95% CI:2.082 ~ 3.689)、脑血管疾病(OR=2.620,95% CI:1.991 ~ 3.438)、冠心病(OR=1.451,95% CI:1.162 ~ 1.823)是老年男性LUTS的独立相关因素;年龄(OR=1.042,95% CI:1.011 ~ 1.073)、手术史(OR=1.681,95% CI:1.320 ~ 2.139)、高血压(OR=2.850,95% CI:2.121 ~ 3.833)、糖尿病(OR=2.571,95% CI:1.958 ~ 3.373)、动脉粥样硬化(OR=3.220, 95% CI:2.480 ~ 4.181)、脑血管疾病(OR=3.051,95% CI:2.288 ~ 4.062)、帕金森病(OR=2.980,95% CI:2.252 ~ 3.947)、钙通道阻滞剂(OR=1.561,95% CI:1.233 ~ 1.981)是老年女性LUTS的独立相关因素。根据Logistic 回归建立的预测模型ROC曲线下面积分别为0.856(95% CI:0.840 ~ 0.872)和0.839(95% CI:0.819 ~ 0.860),校准曲线显示模型有较高的校准度(X2=5.689,P=0.772;X2=6.375,P=0.703),DCA曲线显示在一定区间内该模型净获益具有临床实用价值。结论 本研究明确了老年男性、女性住院患者下尿路症状的独立关联因素,并构建了性别分层的Logistic 风险评估模型。经多维度模型验证,该模型具备良好的区分能力、校准度与临床净获益,可为老年住院患者下尿路症状的早期筛查、个体化预防及临床决策提供可靠的量化参考。

     

    Abstract: Background With the aging population, lower urinary tract symptoms (LUTS) have emerged as a major global public health issue, therefore, conducting an in-depth exploration of its effective prevention and treatment strategies has become particularly urgent. Objective To analyze the relevant factors associated with LUTS in the elderly and construct a risk assessment model. Methods A retrospective cohort study was performed in elderly patients aged ≥60 years hospitalized at the First Medical Center of PLA General Hospital from January 2023 to January 2024. Participants were divided into LUTS group and non-LUTS group based on the presence or absence of LUTS. Given the significant gender differences in LUTS etiology and manifestations, all analyses in this study were stratified by sex. Clinical data were collected for both groups. Multivariate logistic regression was used to identify independent factors associated with elderly LUTS. Forest plots were constructed, and model performance was evaluated using receiver operating characteristic (ROC) curves, clinical calibration curves, and decision analysis curves (DCA). Results This study included a total of 4 122 hospitalized elderly patients, comprising 2 394 males and 1 728 females. Among them, 594 cases (14.48%) had LUTS, including 438 males (18.30%) and 159 females (9.20%). Multivariate logistic regression analysis showed that age (OR=1.051, 95% CI: 1.022 - 1.084), constipation (OR=1.583, 95% CI: 1.250 - 2.002), diabetes mellitus (OR=2.591, 95% CI: 1.949 - 3.443), history of surgery (OR=2.334, 95% CI: 1.781 - 3.043), benign prostatic hyperplasia (OR=2.920, 95% CI: 2.262 - 3.781), hypertension (OR=2.771, 95% CI: 2.082 - 3.689), cerebrovascular disease (OR=2.620, 95% CI: 1.991 - 3.438), and coronary heart disease (OR=1.451, 95% CI: 1.162 - 1.823) were independent risk factors for LUTS in elderly men. Age (OR=1.030, 95% CI: 1.010 - 1.050), surgical history (OR=1.960, 95% CI: 1.490 - 2.580), diabetes (OR=3.380, 95% CI: 2.600 - 4.420), cerebrovascular disease (OR=4.810, 95% CI: 3.330 - 6.980), Parkinson's disease (OR=4.350, 95% CI: 3.240 - 5.910), hypertension (OR=3.830, 95% CI: 2.820 - 5.240), atherosclerosis (OR=4.670, 95% CI: 3.590 - 6.100), and calcium channel blockers (OR=1.820, 95% CI: 1.390 - 2.390) were independent risk factors for LUTS in elderly women. The areas under the ROC curves (AUC) were 0.856 (95% CI: 0.840 - 0.872) and 0.839 (95% CI: 0.819 - 0.860), respectively. Calibration curves demonstrated high model calibration (X2=5.689, P=0.772; X2=6.375, P=0.703), while decision curve analysis indicated clinically meaningful net benefit within specific intervals. Conclusion The present study identifies the independent associated factors of LUTS in hospitalized elderly male and female patients, and develops gender‑stratified Logistic risk assessment models. Through multidimensional model validation, the models demonstrate good discrimination, calibration, and clinical net benefit, and provide a reliable quantitative reference for early screening, individualized prevention, and clinical decision‑making regarding LUTS in hospitalized elderly patients.

     

/

返回文章
返回