Associations between body composition phenotypes and physical function in community-dwelling older adults: an...

To identify body composition phenotypes and compare physical function across phenotypes among community-dwelling older adults, with a secondary exploratory analysis of limb muscle mass asymmetry.

This cross-sectional study included 600 community-dwelling adults aged 60-79 years in Bengbu, Anhui, China. Body composition was assessed using bioelectrical impedance analysis. Body fat percentage, trunk fat mass, and lower-limb muscle mass were standardized within sex and entered into latent profile analysis to identify body composition phenotypes. Upper- and lower-limb muscle mass asymmetry indices (AIs) were calculated and standardized for secondary exploratory analyses. Physical function was assessed using the 30-s chair stand, one-leg stance with eyes closed, grip strength, and 2-min high-knee stepping in place. ANCOVA, linear regression, logistic regression, subgroup analyses, and extreme-group analyses were performed, with FDR correction applied to regression-related analyses.

Three distinct body composition phenotypes were identified: low-adiposity phenotype (C1, 27%), balanced phenotype (C2, 55%), and high-adiposity-high-muscle phenotype (C3, 18%). After adjustment for age and sex, significant differences across phenotypes were observed in grip strength (P = 0.003), one-leg stance with eyes closed (P = 0.002), and 30-s chair stand performance (P = 0.003), but not in 2-min high-knee stepping in place. The high-adiposity-high-muscle phenotype had the highest grip strength, whereas the low-adiposity phenotype had the longest one-leg stance time and the best 30-s chair stand performance. Linear regression analyses showed no significant associations between upper- or lower-limb AI and continuous functional performance outcomes after FDR correction. In logistic regression analyses, upper-limb AI was nominally associated with sample-defined low one-leg stance performance (OR = 1.215, 95% CI: 1.017-1.452, P = 0.032), but this association did not remain significant after FDR correction. No significant association was observed for lower-limb AI.

Community-dwelling older adults showed heterogeneous body composition phenotypes with significant differences in several physical function domains. Limb muscle mass asymmetry was not robustly associated with functional performance after correction for multiple comparisons. These findings suggest that body composition phenotypes may provide useful information for understanding functional heterogeneity in older adults, whereas the role of limb muscle mass asymmetry requires further investigation.