Evaluation of body composition methods for accuracy

The objective of this study was to evaluate the accuracy of a variety of portable methods and instruments used to estimate body composition or percentage body fat (%BF) in a systematic, comprehensive manner on a wide range of subjects. The %BF was estimated using four skinfold protocols, three girth measurement protocols, two bioelectric impedance analysis (BIA) instruments, and one near-infrared instrument on 121 subjects. The subjects ranged in age from 21 to 51 years; weight ranged from 105 to 226 pounds and %BF from 8.3% to 38.3%; and the group was 29% male. The %BF estimates were compared to the values obtained from the generally accepted reference standard, underwater weighing (UWW). The correlation coefficients (r) between the test methods and UWW ranged from 0.48 to 0.72. Regression analysis resulted in a range of slopes from 0.48 to 0.93, y-intercept range from 3.8 to 13.1, and standard error of the estimate range from 3.8 to 7.5. All of the methods tended to overestimate lower and underestimate higher %BF. With two exceptions, all methods appeared more accurate for males than females. In general, the near-infrared appeared least and BLA appeared most accurate. In conclusion, in our opinion, most of the methods were not sufficiently accurate to use on a wide range of individuals. Although they may be sufficiently accurate on the narrow sample of subjects on which they were developed, they may not be as accurate for the general population.