Predicting Body Fat to the Accuracy of Dexa Using Standard Anthropometric Measures

The weight changes experienced during a season in football, as well as the need to increase body mass for several positions, make accuracy of measurement of body composition a necessity. The accuracy and reproducibility of dual X-ray absorptiometry (DEXA) measurements for body fat percentage have contributed to it quickly becoming the gold standard of body composition testing. Unfortunately, the high cost, accessibility, necessary training, and time associated with complete measurements make DEXA a difficult tool for the use in the applied exercise field and in sports. For this reason, it would be ideal to develop a standard body composition determination that is able to predict to the same level of accuracy of DEXA for the practical use in football populations. PURPOSE: The purpose of this study was to develop an equation to determine body composition in football players which provides the necessary reliability and accuracy associated with DEXA using standard anthropometric tools easily accessible to coaches and research personnel. METHODS: One hundred fifty-seven (N = 157) Division IA collegiate football athletes (ages 17-23, height 185.60 ± 6.5 cm, weight 103.1 ± 20.4 kg, DEXA%Fat 19.5 ± 9.1) were subjected to body composition testing using DEXA (GE), 7-site skinfold measurements (mm) and standard circumference measurements (cm) which included ankle, calf, thigh, hip, waist, umbilical, chest, wrist, forearm, arm, neck. Backward Multiple Linear Regression was used to determine the relationship between DEXA %Fat and each of the skinfold thickness and circumference measures. RESULTS: All skinfold thickness measures were shown to be highly correlated with DEXA%Fat (r^sup 2^>0.8). Likewise, waist, hip, and umbilical circumferences were also found to be highly correlated with DEXA%Fat (r^sup 2^>0.8). Collinearity diagnostics revealed that the abdominal skinfold thickness (SFAB), umbilical circumference (UMB), and hip circumference (HIP) provided the best model for prediction with all other sites being highly intercorrelated. The prediction equation developed based on the relationship found between DEXA%Fat and the above mentioned independent variables is as follows: DEXA%Fat = (SFAB x 0.371) + (UMB x 0.193) + ( HIP x 0.134) - 23.052 (r^sup 2^ = 0.932, SEE = 2.3). CONCLUSIONS: Abdominal skinfold thickness, hip circumference, and umbilical circumference are highly correlated with, and were shown able to be accurate predictors of DEXA%Fat in Division IA collegia