Association between the ratio of uric acid to high-density lipoprotein cholesterol (UHR) and the abnormal risk of sarcopenia: Evidence from two large population-based surveys and interpretable machine learning-driven sarcopenia screening.

Authors

Yunnong Luo, Aisheng Wang, Minfei Wu, Yang Wang

Published in

Therapeutic advances in endocrinology and metabolism. Volume 19. Pages 20420188261489474. Epub Sep 11, 2026.

Abstract

Sarcopenia is a geriatric syndrome linked to nutritional intake, chronic inflammation and metabolic dysregulation. While uric acid (UA) and high-density lipoprotein cholesterol (HDL-C) are recognized biomarkers, their specific relationship with sarcopenia remains debated. The uric acid-to-HDL-C ratio (UHR) has emerged as a novel, integrated biomarker. This study investigates the UHR-sarcopenia association using representative US and South Korean populations and develops an interpretable machine learning screening framework.
To investigate the association between UHR and sarcopenia in two nationally representative populations and to develop and externally validate an interpretable machine-learning model for sarcopenia screening.
A population-based cross-sectional study using NHANES 2011-2018 as the primary cohort and KNHANES 2024 as the external validation cohort.
This cross-sectional study included 7,314 participants from NHANES (2011-2018) and 3,274 from KNHANES (2024). The UHR-sarcopenia relationship was evaluated using multivariable logistic regression, smooth curve fitting, and subgroup analyses. To develop a screening model for prevalent sarcopenia, feature selection was performed using LASSO regression and the Boruta algorithm. Machine learning models were trained on the NHANES cohort, interpreted using SHAP values, and rigorously validated externally using the KNHANES dataset.
Higher UHR was significantly associated with greater odds of prevalent sarcopenia. In fully adjusted models, participants in the highest UHR quartile had higher odds of sarcopenia than those in the lowest quartile in NHANES (OR 2.014, 95% CI 1.552-2.625; P<0.001) and KNHANES (OR 1.658, 95% CI 1.183-2.342; P=0.004). Among seven machine-learning models, LightGBM demonstrated the most balanced performance, achieving an AUC of 0.810 in NHANES and maintaining good discrimination during external validation in KNHANES (AUC 0.823).
Elevated UHR is significantly associated with sarcopenia. The UHR-integrated LightGBM model demonstrates robust discriminative capacity, serving as a practical and interpretable tool for sarcopenia screening in clinical practice.

PMID:
42732258
Bibliographic data and abstract were imported from PubMed on 13 Sep 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.