Prediction of percent body fat in adult males using dual energy x-ray absorptiometry, skinfolds, and hydrostatic weighing

The purpose of this study was to compare the prediction of percent body fat (%FAT) by dual energy x-ray absorptiometry (DXA), skinfolds (SF), and hydrostatic weighing (HW) in adult males. Subjects were 35 adult male Caucasians (mean +/- SD; age: 39.1 +/- 14.0 yr, height: 180.6 +/- 5.3 cm, weight: 81.0 +/- 11.1 kg). %FAT, determined by HW with residual volume determined via O2 dilution, served as the criterion. DXA %FAT was determined by the Norland XR-26 (XR-26) bone densitometer and by the SF equations of Jackson and Pollock (JP) (1978), and Lohman (LOH) (1981). Criterion referenced validation included analyzing mean (+/- SD) %FAT values using a one-way ANOVA for significance, comparison of mean differences (MD), correlations (r), standard error of estimates (SEE), and total errors (TE). Significant differences were found between means of each method. The r (0.91) and SEE (3.0 %FAT) for DXA compare favorably with the established SF methods of JP and LOH for predicting %FAT; however, DXA demonstrated the largest MD (3.9 %FAT) and TE (5.2 %FAT). Regression analysis yields HW = 0.79* DXA + 0.56. The results do not support earlier research that found no significant difference between HW and DXA %FAT in males. The study suggests the density of the fat-free body (DFFB) is not constant, and that the variation in bone mineral content affects the DFFB, which contributes to the differences between DXA and HW %FAT. We recommend further research to identify inconsistencies between manufacturers of DXA equipment in prediction of %FAT in males.