Sung Jin Kim, Young Goo Lee, Ohseong Kwon, Sahyun Pak, Sung Gon Park, Sung Tae Cho
{"title":"韩国绝经后妇女握力与尿失禁的关联:一项机器学习研究。","authors":"Sung Jin Kim, Young Goo Lee, Ohseong Kwon, Sahyun Pak, Sung Gon Park, Sung Tae Cho","doi":"10.1007/s00192-025-06310-y","DOIUrl":null,"url":null,"abstract":"<p><strong>Introduction and hypothesis: </strong>Urinary incontinence (UI) is a common condition in older women and is associated with frailty and functional decline. Handgrip strength is a simple, noninvasive marker of muscle function that may help identify women at risk for UI. This study aimed to examine the association between handgrip strength and UI in postmenopausal Korean women and evaluate its predictive utility using machine learning models.</p><p><strong>Methods: </strong>We analyzed data from 4213 postmenopausal women aged ≥45 years who participated in the Korean Longitudinal Study of Aging (2006 -2022). UI was defined as reporting urine leakage on ≥5 days during the past month. Handgrip strength was measured using a standardized digital dynamometer protocol. Five machine learning models-logistic regression, random forest (RF), support vector machine, extreme gradient boosting (XGB), and deep neural network-were used to predict UI. Model performance was evaluated using fivefold cross-validation. Logistic regression and generalized additive models were used to validate findings.</p><p><strong>Results: </strong>Among postmenopausal women, the prevalence of UI was 26.0%. Compared to the continent group, individuals with UI were older (73.5 ± 6.7 vs. 68.3 ± 6.1 years, p < 0.001), exhibited lower handgrip strength (20.4 ± 4.8 vs. 22.5 ± 4.9 kg, p < 0.001), and had a higher burden of comorbidities and functional impairments. In multivariable logistic regression models adjusting for age, BMI, and other demographic, clinical, lifestyle, and frailty-related symptom variables, lower handgrip strength remained independently associated with UI (adjusted OR 0.971, 95% CI 0.951-0.992 per 1 kg decrease). RF and XGB models demonstrated robust predictive performance, with AUCs of 0.887 and 0.860, respectively. A handgrip strength threshold of 15.75 kg was identified as a clinically relevant cutoff below which UI risk increased.</p><p><strong>Conclusion: </strong>Handgrip strength is independently associated with UI in postmenopausal women. As a practical and noninvasive measure, it may be useful in identifying individuals at increased risk of incontinence and guiding early interventions.</p>","PeriodicalId":14355,"journal":{"name":"International Urogynecology Journal","volume":" ","pages":""},"PeriodicalIF":1.8000,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Association of Handgrip Strength With Urinary Incontinence in Postmenopausal Korean Women: A Machine Learning Study.\",\"authors\":\"Sung Jin Kim, Young Goo Lee, Ohseong Kwon, Sahyun Pak, Sung Gon Park, Sung Tae Cho\",\"doi\":\"10.1007/s00192-025-06310-y\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Introduction and hypothesis: </strong>Urinary incontinence (UI) is a common condition in older women and is associated with frailty and functional decline. Handgrip strength is a simple, noninvasive marker of muscle function that may help identify women at risk for UI. This study aimed to examine the association between handgrip strength and UI in postmenopausal Korean women and evaluate its predictive utility using machine learning models.</p><p><strong>Methods: </strong>We analyzed data from 4213 postmenopausal women aged ≥45 years who participated in the Korean Longitudinal Study of Aging (2006 -2022). UI was defined as reporting urine leakage on ≥5 days during the past month. Handgrip strength was measured using a standardized digital dynamometer protocol. Five machine learning models-logistic regression, random forest (RF), support vector machine, extreme gradient boosting (XGB), and deep neural network-were used to predict UI. Model performance was evaluated using fivefold cross-validation. Logistic regression and generalized additive models were used to validate findings.</p><p><strong>Results: </strong>Among postmenopausal women, the prevalence of UI was 26.0%. Compared to the continent group, individuals with UI were older (73.5 ± 6.7 vs. 68.3 ± 6.1 years, p < 0.001), exhibited lower handgrip strength (20.4 ± 4.8 vs. 22.5 ± 4.9 kg, p < 0.001), and had a higher burden of comorbidities and functional impairments. In multivariable logistic regression models adjusting for age, BMI, and other demographic, clinical, lifestyle, and frailty-related symptom variables, lower handgrip strength remained independently associated with UI (adjusted OR 0.971, 95% CI 0.951-0.992 per 1 kg decrease). RF and XGB models demonstrated robust predictive performance, with AUCs of 0.887 and 0.860, respectively. A handgrip strength threshold of 15.75 kg was identified as a clinically relevant cutoff below which UI risk increased.</p><p><strong>Conclusion: </strong>Handgrip strength is independently associated with UI in postmenopausal women. As a practical and noninvasive measure, it may be useful in identifying individuals at increased risk of incontinence and guiding early interventions.</p>\",\"PeriodicalId\":14355,\"journal\":{\"name\":\"International Urogynecology Journal\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":1.8000,\"publicationDate\":\"2025-09-25\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Urogynecology Journal\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1007/s00192-025-06310-y\",\"RegionNum\":3,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"OBSTETRICS & GYNECOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Urogynecology Journal","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1007/s00192-025-06310-y","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"OBSTETRICS & GYNECOLOGY","Score":null,"Total":0}
引用次数: 0
摘要
介绍和假设:尿失禁(UI)是老年妇女的常见病,与身体虚弱和功能下降有关。握力是一种简单、无创的肌肉功能指标,可以帮助识别女性是否有尿失禁的风险。本研究旨在研究绝经后韩国女性握力与UI之间的关系,并使用机器学习模型评估其预测效用。方法:我们分析了参加韩国老龄化纵向研究(2006 -2022)的4213名年龄≥45岁的绝经后妇女的数据。尿漏定义为在过去一个月内报告尿漏≥5天。使用标准化的数字测功机协议测量了手握力。使用逻辑回归、随机森林(RF)、支持向量机、极端梯度增强(XGB)和深度神经网络五种机器学习模型来预测用户界面。采用五重交叉验证对模型性能进行评估。使用逻辑回归和广义加性模型来验证研究结果。结果:绝经后妇女尿失禁患病率为26.0%。与大陆组相比,UI患者年龄更大(73.5±6.7岁vs. 68.3±6.1岁,p < 0.001),握力更低(20.4±4.8 kg vs. 22.5±4.9 kg, p < 0.001),合并症和功能障碍负担更高。在调整了年龄、BMI和其他人口统计学、临床、生活方式和虚弱相关症状变量的多变量logistic回归模型中,较低的握力仍然与UI独立相关(调整OR 0.971, 95% CI 0.951-0.992 /每减少1 kg)。RF和XGB模型具有较好的预测性能,auc分别为0.887和0.860。15.75 kg的握力阈值被确定为临床相关临界值,低于该临界值,尿失禁风险增加。结论:握力与绝经后妇女尿失禁独立相关。作为一种实用且无创的措施,它可能有助于识别失禁风险增加的个体并指导早期干预。
Association of Handgrip Strength With Urinary Incontinence in Postmenopausal Korean Women: A Machine Learning Study.
Introduction and hypothesis: Urinary incontinence (UI) is a common condition in older women and is associated with frailty and functional decline. Handgrip strength is a simple, noninvasive marker of muscle function that may help identify women at risk for UI. This study aimed to examine the association between handgrip strength and UI in postmenopausal Korean women and evaluate its predictive utility using machine learning models.
Methods: We analyzed data from 4213 postmenopausal women aged ≥45 years who participated in the Korean Longitudinal Study of Aging (2006 -2022). UI was defined as reporting urine leakage on ≥5 days during the past month. Handgrip strength was measured using a standardized digital dynamometer protocol. Five machine learning models-logistic regression, random forest (RF), support vector machine, extreme gradient boosting (XGB), and deep neural network-were used to predict UI. Model performance was evaluated using fivefold cross-validation. Logistic regression and generalized additive models were used to validate findings.
Results: Among postmenopausal women, the prevalence of UI was 26.0%. Compared to the continent group, individuals with UI were older (73.5 ± 6.7 vs. 68.3 ± 6.1 years, p < 0.001), exhibited lower handgrip strength (20.4 ± 4.8 vs. 22.5 ± 4.9 kg, p < 0.001), and had a higher burden of comorbidities and functional impairments. In multivariable logistic regression models adjusting for age, BMI, and other demographic, clinical, lifestyle, and frailty-related symptom variables, lower handgrip strength remained independently associated with UI (adjusted OR 0.971, 95% CI 0.951-0.992 per 1 kg decrease). RF and XGB models demonstrated robust predictive performance, with AUCs of 0.887 and 0.860, respectively. A handgrip strength threshold of 15.75 kg was identified as a clinically relevant cutoff below which UI risk increased.
Conclusion: Handgrip strength is independently associated with UI in postmenopausal women. As a practical and noninvasive measure, it may be useful in identifying individuals at increased risk of incontinence and guiding early interventions.
期刊介绍:
The International Urogynecology Journal is the official journal of the International Urogynecological Association (IUGA).The International Urogynecology Journal has evolved in response to a perceived need amongst the clinicians, scientists, and researchers active in the field of urogynecology and pelvic floor disorders. Gynecologists, urologists, physiotherapists, nurses and basic scientists require regular means of communication within this field of pelvic floor dysfunction to express new ideas and research, and to review clinical practice in the diagnosis and treatment of women with disorders of the pelvic floor. This Journal has adopted the peer review process for all original contributions and will maintain high standards with regard to the research published therein. The clinical approach to urogynecology and pelvic floor disorders will be emphasized with each issue containing clinically relevant material that will be immediately applicable for clinical medicine. This publication covers all aspects of the field in an interdisciplinary fashion