Four methods of assessing body composition were compared in 55 black and 35 white, Division 1, American football players. Percent body fat (%BF) was estimated with hydrostatic weighing at residual volume, corrected for race; seven-site skinfolds (7 SF), corrected for race; bioelectrical impedance analysis (BIA); and near-infrared spectrophotometry (NIR). Percent body fat with HW in blacks (mean = 14.7%) and whites (19.7%) did not differ (P > .05) from %>BF with 7 SF (blacks, 14.7%; whites, 19.0%). In relation to HW, BIA significantly (P < .05) overpredicted (blacks: 20.1%, SEE = 3.2%; whites; 22.3%, SEE = 4.3%) and NiR underpredicted %BF (blacks; 12.6%, SEE = 3.9%; whites; 17.7%, SEE = 3.6%). The contribution of BIA variables (resistance, phase angle, conductance) and NIR optical density to predict %BF was trivial compared to body mass index. It appears that race may not substantially influence %BF prediction by NIR and BIA. It was concluded that when considering the cost and expertise required with NIR and BIA, SF measurements appear to be a superior alternative for rapid and accurate body composition assessment of athletes, independent of race.
The authors are with the Human Performance Laboratory, Sports Medicine Bldg., Fast Carolina University, Greenville, NC 27858-4353. O'Brien is with the Biostatistics Program at ECU.