CROSS-VALIDATION OF BODY COMPOSITION EQUATIONS FOR CHILDREN USING DUAL-ENERGY X-RAY ABSORPTIOMETRY

Abstract

926 This study assessed the predictive accuracy of ten existing skinfold (SKF) and bioelectrical impedance (BIA) equations in estimating the body composition of 120 hispanic and white children, aged 10 to 18 years. The cross-validation (XV) criteria recommended by Lohman (1992) and Houtkooper et al. (1992) were used to assess the predictive accuracy of the equations, using dual-energy x-ray absorptiometry as the reference method. Only one equation, the Slaughter et al. (1988) triceps plus subscapular SKF equation for boys met the XV criteria (r = .980, SEE = 2.49%BF). All other SKF and BIA equations significantly and consistently under-or overestimated reference measures of body composition (SEEs ranged from 2.28 to 3.16kg and 3.61 to 4.27%BF). Therefore, a SKF equation for girls and a BIA equation for children were developed using hierarchical regression. The sample was randomly divided into validation and XV groups for these analyses. There were no significant differences between groups for any potential predictors or the reference measure. Cross-validation of the new SKF and BIA equations was acceptable (SKF: r = .942, SEE = 3.21%BF; BIA: r = .976, SEE = 2.06kg); therefore, the samples were pooled to yield more stable regression coefficients: SKF: %BF = 0.583(sum4SKF) − 0.0017 (sum4SKF)2 + 1.856 (R = .993, SEE = 3.37%BF), where sum4SFK = medial calf + midaxillary + subscapular + triceps SKFs. BIA: FFM = 0.832(HT2/R) + 0.0478(BW) + 0.150(Xc) + 0.324(age) − 12.772 (R = .983, SEE = 1.77kg), where R = resistance and Xc = reactance. Supported by a grant from OMRON Healthcorp Inc.