Age-related variability in body composition methods for assessment of percent fat and fat-free mass in men aged 20–74 years

Open Access
Publisher Website
Google Scholar
Abstract
Introduction: assessing body composition is important because of the association between excess body fat or obesity and the increased risk for coronary artery disease, diabetes, hyperlipidaemia and hypertension. Methodological differences among indirect body composition techniques are a source of variability that can affect estimations of fat-free mass (FFM) or percent fat, especially for different age groups. Purpose: this study examined the effect of age on the estimation of body composition by three methods based on the two-component model. One hundred and fifty-three men were placed into 5-year age groups from 20–24 to 70–74 years. Body composition was assessed by hydrodensitometry (underwater weighing; UWW), 40K spectroscopy (K40) and anthropometry. Results: day-to-day reliability was excellent for each method. Correlation coefficients between methods, independent of age, ranged from r=0.71 (K40 and UWW) to r=0.83 (UWW and anthropometry) for % fat and from r=0.76 (K40 and anthropometry) to r=0.90 (UWW and anthropometry) for FFM. Correlations between techniques weakened and showed greater variability as age increased. UWW produced the highest % fat and lowest FFM for the different age groups, while K40 and anthropometry yielded similar body composition values. All three methods detected age-related differences in % fat; however, only UWW determined a significant age effect in FFM. The greatest discrepancies between the methods occurred for the oldest age groups (60–74 years), but noticeable differences began occurring after the age of 40. Conclusions: each method produced reliable data. Care should be taken when choosing any method based on the two-component model, especially when different age groups are being examined, since estimating % fat and FFM is dependent on the age-related chemical composition of an individual and the limitations of each method.

This publication has 0 references indexed in Scilit: