Validity of U.S. Department of Defense Circumference Equation in Detecting Body Composition Changes.

Abstract

Body composition assessment is a regular part of the U.S. Army's Weight Control Program. Soldiers failing initial screening with a height and weight table are further evaluated with a circumference based method that predicts body fat percentage (BF%). However, research investigating the ability of circumference based predictive equations to detect body composition changes is limited. PURPOSE: The purpose of this study was to examine the validity of the U.S. Department of Defense (DOD) BF% prediction equation against air-displacement plethysmography (ADP) in detecting body composition changes. METHODS: Body composition of 23 Caucasian male, Army ROTC cadets (21.00 ± 2.17 yrs), was assessed using ADP and the current DOD circumference equation for males. Cadets were prospectively followed and assessed at baseline and 4 months later. Paired t-tests were used to compare mean BF% from baseline to follow-up for ADP and the DOD method and to compare change scores from ADP and the DOD method. A Bland-Altman plot and regression 872 analysis was used to examine the relationship between BF% changes measured by ADP and DOD methods with ADP used as the criterion measure. Alpha was set ≤.05 for all analyses. RESULTS: Mean BF% from ADP at baseline (17.02 ± 5.78%) and follow-up (15.50 ± 5.32%) were significantly different, t = 2.86, p <.01, while the DOD method indicated no significant difference in mean BF% from baseline (15.85 ± 5.16%) to follow-up (15.96 ± 4.26%). The change score in mean BF% for ADP (- 1.51 ± 2.53%) was significantly greater than the DOD method (.11 ± 2.41%; t=2.94, p <.01). The Bland-Altman plot did not indicate any strong systematic trend although there was a general underestimation of BF% change by the DOD method. However, regression analysis showed a lack of agreement in BF% change between the DOD method and ADP as both the slope (.45, SE =.21) and intercept (-1.56, SE =.49) of the resulting fitted regression line significantly differed from the line of identity (p =.01 and p <.01, respectively) with only a moderate correlation between the two methods (r =.43). CONCLUSION: Results suggest the DOD circumference equation cannot adequately detect body composition changes in comparison to ADP among an ethnically homogenous sample of male Army ROTC cadets.