Anthropometric correlates of fat mass and fat-free mass in preschool children: a cross-sectional bioelectrical impedance study
Litsenziya: cc-by
Annotatsiya
Objectives. To describe sex-related differences in anthropometry and body composition and to examine the unadjusted associations between commonly used anthropometric indicators and bioelectrical impedance analysis (BIA)-derived fat mass (FM) and fat-free mass (FFM) in preschool children from Uzbekistan. Materials and methods. This cross-sectional study included 750 children aged 3.0-6.0 years (370 boys and 380 girls) from four territories of Jizzakh Region. Height, weight, body mass index (BMI), waist circumference, and triceps skinfold thickness were measured according to a standardized protocol. FM and FFM were estimated by direct segmental multi-frequency BIA using an InBody 270S analyzer. Sex differences and Pearson correlations were analyzed without multivariable adjustment; 95% confidence intervals (CIs) are reported for the main estimates. Results. Girls had higher mean FM than boys (3.6 ± 1.2 vs. 3.2 ± 1.1 kg; mean difference, 0.40 kg; 95% CI, 0.24-0.56; Hedges g = 0.35; p < 0.001), whereas boys had higher mean FFM than girls (14.7 ± 2.1 vs. 13.8 ± 2.0 kg; mean difference, 0.90 kg; 95% CI, 0.61-1.19; Hedges g = 0.44; p < 0.001). In unadjusted analyses, FM correlated most strongly with triceps skinfold thickness (r = 0.74; 95% CI, 0.71-0.77), BMI (r = 0.69; 95% CI, 0.65-0.73), and waist circumference (r = 0.66; 95% CI, 0.62-0.70). FFM correlated most strongly with weight (r = 0.76; 95% CI, 0.73-0.79) and height (r = 0.72; 95% CI, 0.68-0.75). BMI shared approximately 48% of the observed variance in FM in the unadjusted bivariate analysis. Conclusions. Simple anthropometric measures reflect different aspects of body composition and may support population-level assessment. These unadjusted associations with BIA-derived outcomes do not establish measurement agreement, individual diagnostic accuracy, or causal relationships; adjusted and externally validated models are needed before clinical prediction can be considered.
Hali tarjima qilinmagan