An investigation into the validity of percentage body fat estimations by a commercially available bioelectrical impedance analyser

Abstract

The present investigation examined the validity of bioelectrical impedance analysis (BIA) and selected skinfolds (Sum 4) when estimating percent body fat (% fat) in young, physically active adults (MEAN ± SEM = 18.3 ± 1.2 % fat) by comparing the estimates with values obtained from densitometry (D). Thirty-five Caucasian volunteers (21 males, 14 females; MEAN ± SEM = 22.9 ± 0.4 yr) served as subjects. The statistical analysis involved calculation of the bias and 95% limits of agreement. The results indicated that the Bodystat 1500 BIA system agreed better with D (bias and 95% limits = 0.7 ± 7.4 % fat) than Sum 4 (bias and 95% limits = 2.2 ± 8.5 % fat). However, the error observed for both predictive methods was too large to recommend use in assessing % fat in a young, physically active population, unless only a general estimation of % fat is required for work such as epidemiological studies.