Comparison of body composition assessment among lean black and white male collegiate athletes.

Variations in the density of the fat-free mass (DFFM) across ethnic groups is a critical factor that invalidates the use of body fat equations. It has also been suggested that resistance trained athletes may have higher body densities (BDs) than untrained subjects. Thus, the validity of using anthropometric (ANT) equations, which have mainly been derived on white nonathletic groups, has been questioned for athletic white and black men. This study compared BD and percent body fat (%BF) between 34 white (20 +/- 1 yr, 184 +/- 11 cm, 84 +/- 12 kg, 25 +/- 3 BMI) and 30 black (20 +/- 1 yr, 182 +/- 9 cm, 84 +/- 12 kg, 25 +/- 2 BMI) male collegiate athletes and determined the accuracy of 5 ANT equations in estimating %BF. Subjects were underwater weighed (UWW), and skinfold measurements were obtained from the chest, mid-axillary, abdomen, suprailiac, subscapula, triceps, and thigh. BD was obtained from UWW and estimated from the five skinfold equations. From UWW, significant (P < or = 0.05) differences were found for BD (1.075 +/- 0.007 vs 1.0817 +/- 0.009), but not for %BF (10.49 +/- 2.8 vs 11.59 +/- 3.4) for white and black subjects, respectively. Differences were noted for subcutaneous skinfold sites (abdominal (vertical), suprailiac, and thigh), sum of three and seven skinfolds, and proportion of subscapular subcutaneous fat. One out of five and five out of five ANT equations (Siri conversions) yielded significantly lower estimates compared with UWW %BF for the white and black athletes, respectively. Use of the Schutte equation for the black athletes resulted in overpredictions of %BF for five out of five equations. In addition, the Schutte equation offered slightly greater accuracy than did the Siri equation for estimating %BF in black athletes. These data confirm earlier concerns that ANT equations derived on general populations may not be as accurate for athletic populations and also suggest that correction equations are necessary for converting BD into %BF for populations differing with respect to race or training status.