文献速递衰弱、肌少症与跌倒老年慢病共病管理

尿酸与高密度脂蛋白胆固醇比值作为肌少症新型生物标志物:一项结合机器学习的全国性研究

Uric acid to high-density lipoprotein cholesterol ratio as a novel biomarker for sarcopenia: a national study with machine learning insights.
PubMed 文献 · J Gerontol A Biol Sci Med Sci · 2026-01-01

该研究基于中国健康与养老追踪调查9853名中老年人,按亚洲肌少症工作组2019共识定义肌少症,采用多因素logistic回归和限制性立方样条评估尿酸/高密度脂蛋白胆固醇比值(UHR)与肌少症的关系。TyG指数和eGFR分别介导16.44%和2.03%的关联。机器学习模型ROC-AUC为0.73-0.77,XGBoost表现最佳,SHAP分析显示UHR是仅次于年龄的第二重要预测因子。

要点

  • TyG指数和eGFR分别介导UHR与肌少症关联的16.44%和2.03%。
  • 机器学习模型中UHR是仅次于年龄的肌少症第二重要预测因子。

核心结论

  • 在中国中老年人群中,较高的UHR与较低的肌少症风险独立相关,部分通过代谢和肾脏途径介导,UHR可能作为肌少症风险分层的简易生物标志物。

研究设计与样本

  • 研究分析了中国健康与养老追踪调查(CHARLS)中的9853名成年人,肌少症按亚洲肌少症工作组2019共识标准定义。
  • 采用多因素logistic回归和限制性立方样条评估UHR与肌少症的关系,并进行中介分析探讨TyG指数和eGFR的作用。
  • 开发并评估了五种机器学习算法,训练集占70%、测试集占30%,使用Boruta和LASSO进行特征选择,SHAP评估模型可解释性。

主要发现

  • UHR与较好的握力和起立-坐下表现相关。
  • TyG指数和eGFR分别介导UHR与肌少症关联的16.44%和2.03%。
  • 机器学习分析中ROC-AUC范围为0.73-0.77,XGBoost表现最佳。
  • 基于XGBoost模型的SHAP分析显示UHR是仅次于年龄的肌少症第二重要预测因子。

关键数值与推荐

  • 肌少症定义采用亚洲肌少症工作组2019共识标准。
  • 中介分析中TyG指数和eGFR分别解释16.44%和2.03%的关联。
  • 机器学习模型ROC-AUC范围为0.73-0.77。

临床怎么用

  • UHR计算简便,或可用于中老年人群肌少症的风险分层,但需进一步外部验证。
  • UHR与肌少症关联部分由胰岛素抵抗和肾功能介导,提示代谢和肾脏状态在肌少症评估中的潜在价值。
  • 机器学习模型(如XGBoost)结合UHR和年龄等特征可能辅助肌少症风险预测。

局限性

  • 原文未报告研究设计类型为横断面或纵向,无法判断因果方向。
  • 原文未报告UHR的具体计算单位或分界值。
  • 原文未报告机器学习模型的外部验证结果。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: Sarcopenia is a major age-related health burden. Although the uric acid to high-density lipoprotein cholesterol ratio (UHR) has been linked to sarcopenia, its underlying pathways and clinical utility remain unclear. METHODS: We analyzed 9853 adults from the China Health and Retirement Longitudinal Study. Sarcopenia was defined by Asian Working Group for Sarcopenia 2019 Consensus criteria. We assessed the UHR-sarcopenia relationship using multivariable logistic regression and restricted cubic splines. Mediation analyses examined roles of insulin resistance (TyG index) and renal function (eGFR). Five machine learning algorithms were developed and evaluated in training (70%) and testing (30%) sets. Feature selection was performed using Boruta and LASSO, and model interpretability was assessed using SHapley Additive exPlanations (SHAP). RESULTS: Higher UHR was independently associated with lower odds of sarcopenia, with an adjusted odds ratio per-standard deviation increase of 0.75 (95% CI: 0.69-0.81; p < .001), and showed a nonlinear dose-response relationship. UHR correlated with better handgrip strength and chair-stand performance. The TyG index and eGFR mediated 16.44% and 2.03% of this association, respectively. In machine learning analyses, ROC-AUC values ranged from 0.73 to 0.77, with XGBoost showing the best performance. SHAP analysis based on the XGBoost model further identified UHR as the second most important predictor of sarcopenia, following age. CONCLUSIONS: Higher UHR was associated with lower sarcopenia risk in Chinese middle-aged and older adults, partly through metabolic and renal pathways. UHR may serve as a simple biomarker for sarcopenia risk stratification.

来源:J Gerontol A Biol Sci Med Sci,查看原文。版权归原作者所有。

引用

  1. Yang R, Cheng Z, Yang P, et al. Uric acid to high-density lipoprotein cholesterol ratio as a novel biomarker for sarcopenia: a national study with machine learning insights. J Gerontol A Biol Sci Med Sci. 2026. PMID: 42172590. doi:10.1093/gerona/glag132

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。