Sex- and age-specific risk factors for possible sarcopenic obesity in community-dwelling older adults: a cross-sectional study
- 1Geriatric Medicine Department, The Second People’s Hospital of Nanning, Guangxi, China
- 2Department of Respiratory and Critical Care Medicine, Guangxi Jiangbin Hospital, Nanning, Guangxi China
Abstract
Background: Possible Sarcopenic obesity (PSO) carries higher risks than either condition alone, but whether its risk factors differ by sex and age remains unclear. We investigated the prevalence, risk factors, and sex/age differences of PSO using traditional regression, Bayesian analysis, and machine learning.
Methods: This cross-sectional study included 427 community-dwelling adults aged 60–98 years. PSO was defined as concurrent possible sarcopenia (AWGS 2019: low handgrip strength + slow gait speed) and central obesity (waist circumference ≥ 90 cm in men, ≥ 85 cm in women). Multivariable logistic regression, sex- and age-stratified analyses, dose–response analysis, mediation analysis, and E-value sensitivity analysis were performed. Bayesian hierarchical models and XGBoost with SHAP were used for validation.
Results: PSO prevalence was 8.7% (37/427). Overall, age (OR = 1.12, 95%CI:1.05–1.19) was a risk factor and hemoglobin (OR = 0.97, 95%CI:0.94–0.99) was protective. Sex-stratified analysis revealed marked differences: in men, BMI was a strong risk factor (OR = 1.54, 95%CI:1.19–2.01); in women, low hemoglobin (OR = 0.96, 95%CI:0.92–0.99) and high white blood cell count (OR = 1.43, 95%CI:1.04–1.98) were significant. Age-stratified analysis showed that in those < 75 years, glucose (OR = 1.26) and white blood cell count (OR = 1.23) were important; in those ≥ 75 years, BMI showed a strong effect (OR = 1.91, 95%CI:1.33–2.76). Bayesian analysis produced consistent findings with narrow credible intervals. The XGBoost model achieved good discrimination (AUC = 0.86), with SHAP confirming age, BMI, and hemoglobin as top predictors. Dose–response analysis showed a linear relationship between hemoglobin and PSO (p for nonlinearity = 0.743). E-value analysis confirmed robustness to unmeasured confounding (E-values ≥ 1.47).
Conclusions: Risk factors for PSO differ substantially by sex and age. These findings support targeted prevention strategies: weight control in men, nutritional and inflammatory monitoring in women, metabolic management in younger-old adults, and weight control in the oldest-old.
Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s12877-026-08088-0.
Related
- Effects of the EatWelLog® Mobile Application for Managing Sarcopenic Obesity Among Community-Dwelling Older Adults: A Pilot Randomized Controlled Trial
- Factors associated with health-promoting lifestyles among community-dwelling older adults in Zigong, China: a cross-sectional study
- Association between duration of smoking cessation and depressive symptoms among community-dwelling older adults: a cross-sectional study
- Quality of life among community-dwelling older adults receiving primary care in the Podkarpackie region (Poland): a cross-sectional study
- Sex-specific Association Between High-frequency Hearing Thresholds and Sarcopenia in Community-dwelling Older Adults
BibTeX
@article{chen2026agespecific,
title = {{Sex- and age-specific risk factors for possible sarcopenic obesity in community-dwelling older adults: a cross-sectional study}},
author = {Cong Chen and Yuhai Dang},
journal = {BMC Geriatrics},
year = {2026},
month = aug,
doi = {10.1186/s12877-026-08088-0},
url = {https://doi.org/10.1186/s12877-026-08088-0},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Sex- and age-specific risk factors for possible sarcopenic obesity in community-dwelling older adults: a cross-sectional study
AU - Cong Chen
AU - Yuhai Dang
PY - 2026
DA - 2026/08/06
JO - BMC Geriatrics
DO - 10.1186/s12877-026-08088-0
UR - https://doi.org/10.1186/s12877-026-08088-0
AB - Background: Possible Sarcopenic obesity (PSO) carries higher risks than either condition alone, but whether its risk factors differ by sex and age remains unclear. We investigated the prevalence, risk factors, and sex/age differences of PSO using traditional regression, Bayesian analysis, and machine learning. Methods: This cross-sectional study included 427 community-dwelling adults aged 60–98 years. PSO was defined as concurrent possible sarcopenia (AWGS 2019: low handgrip strength + slow gait speed) and central obesity (waist circumference ≥ 90 cm in men, ≥ 85 cm in women). Multivariable logistic regression, sex- and age-stratified analyses, dose–response analysis, mediation analysis, and E-value sensitivity analysis were performed. Bayesian hierarchical models and XGBoost with SHAP were used for validation. Results: PSO prevalence was 8.7% (37/427). Overall, age (OR = 1.12, 95%CI:1.05–1.19) was a risk factor and hemoglobin (OR = 0.97, 95%CI:0.94–0.99) was protective. Sex-stratified analysis revealed marked differences: in men, BMI was a strong risk factor (OR = 1.54, 95%CI:1.19–2.01); in women, low hemoglobin (OR = 0.96, 95%CI:0.92–0.99) and high white blood cell count (OR = 1.43, 95%CI:1.04–1.98) were significant. Age-stratified analysis showed that in those < 75 years, glucose (OR = 1.26) and white blood cell count (OR = 1.23) were important; in those ≥ 75 years, BMI showed a strong effect (OR = 1.91, 95%CI:1.33–2.76). Bayesian analysis produced consistent findings with narrow credible intervals. The XGBoost model achieved good discrimination (AUC = 0.86), with SHAP confirming age, BMI, and hemoglobin as top predictors. Dose–response analysis showed a linear relationship between hemoglobin and PSO (p for nonlinearity = 0.743). E-value analysis confirmed robustness to unmeasured confounding (E-values ≥ 1.47). Conclusions: Risk factors for PSO differ substantially by sex and age. These findings support targeted prevention strategies: weight control in men, nutritional and inflammatory monitoring in women, metabolic management in younger-old adults, and weight control in the oldest-old. Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s12877-026-08088-0.
SN - 1471-2318
ER -