Evaluation of body composition. Current issues

Abstract

In the selection of body composition field methods and prediction equations, exercise and health practitioners must consider their clients' demographics. Factors, such as age, gender, level of adiposity, physical activity and ethnicity influence the choice of method and equation. Also, it is important to evaluate the relative worth of prediction equations in terms of the criterion method used to derive reference measures of body composition for equation development. Given that hydrodensitometry, hydrometry and dual-energy x-ray absorptiometry are subject to measurement error and violation of basic assumptions underlying their use, none of these should be considered as a 'gold standard' method for in vivo body composition assessment. Reference methods, based on whole-body, 2-component body composition models, are limited, particularly for individuals whose fat-free body (FFB) density and hydration differ from values assumed for 2-component models. Use of field method prediction equations developed from 2-component model (Siri equation) reference measures of body composition will systematically underestimate relative body fatness of American Indian women, Black men and women, and Hispanic women because the average FFB density of these ethnic groups exceeds the assumed value (1.1 g/ml). Thus, some researchers have developed prediction equations based on multicomponent model estimates of body composition that take into account interindividual variability in the water, mineral, and protein content of the FFB. One multicomponent model approach adjusts body density (measured via hydrodensitometry) for total body water (measured by hydrometry) and/or total body mineral estimated from bone mineral (measured via dual-energy x-ray absorptiometry). Skinfold (SKF), bioelectrical impedance analysis (BIA), and near-infrared interactance (NIR) are 3 body composition methods used in clinical settings. Unfortunately, the overwhelming majority of field method prediction equations have been developed and cross-validated for White populations and are based on 2-component model reference measures. Because ethnicity may affect the composition of the FFB and regional fat distribution, race-specific prediction equations may need to be developed for some ethnic groups. To date, race-specific SKF (American Indian women, Black men, and Asian adults), BIA (American Indian women and Asian adults), and NIR (American Indian women and White women) equations have been developed. However, these equations need to be cross-validated on additional samples from these ethnic groups. In summary, research strongly suggests that multicomponent models need to be used in order to quantify differences in FFB composition due to ethnicity so that accurate SKF, BIA, and NIR prediction equations can be developed. Assessment of body composition in vivo may be enhanced by using advanced technologies such as dual-energy x-ray absorptiometry and hydrometry to refine hydrodensitometry. Practitioners should carefully select and use only those prediction equations that have been developed and cross-validated for specific ethnic groups. Additional research is needed to test the accuracy and applicability of previously published prediction equations for the American Indian, Asian, Black, and Hispanic populations.