Percent body fat measured by BIA and DEXA in obese, African-American adolescent girls

OBJECTIVE: To compare the estimation of body fat between bioelectrical impedance analysis (BIA) and dual energy X-ray absorptiometry (DEXA) in overweight, African-American female adolescents. SUBJECTS: In total, 54 African-American adolescent female subjects were recruited for Study 1. Each adolescent's body mass index was greater than the 85th percentile and their average body fat was 45% according to DEXA. A total of 26 African-American adolescent female subjects were available for Study 2, and had an average body fat of 26% according to DEXA. MEASUREMENTS: Percent body fat was measured by DEXA and BIA. Seven different BIA equations were tested. Both sets of data were analyzed using Bland-Altman regression analyses, utilizing percent body fat measured by DEXA as the criterion. RESULTS: The Kushner equation provided estimates that were unaffected by body fat in both studies. Estimates were unbiased when applied to the exclusively overweight sample and biased when utilized with the separate sample of normal weight and obese girls. The remaining equations were biased, provided inconsistent estimates across body weight, or were biased and provided inconsistent estimates. Ethnicity-specific and ethnicity-combined equations performed similarly in the obese sample, but became more disparate when applied to a sample encompassing a wider body weight range. The limits of agreement between all BIA equations and the DEXA estimates ranged from 6 to 9%. CONCLUSION: The study suggests that the Kushner BIA equation is appropriate for use with African-American female adolescents across the weight spectrum, while the majority of BIA equations underestimated percent body fat as body fat increased. International Journal of Obesity (2005).