Abstract des Autors

The purpose of this investigation was to study whether meaningful differences occurred when percent body fat (%BF) values were estimated for white prepubertal children using six different regression equations found in the literature. Skinfold (SF) thickness and bioelectrical impedance analysis (BIA) measurements were taken on 107 boys and 105 girls. All children were pubertal stage 1 according to Tanner, except two 11 year old girls, who were in stage 2. Analyses of variance showed that there was a significant main effect of the regression equations on %BF values for boys (F=50.6; p<0.001) and girls (F=55.4; p<0.001). The post hoc comparisons indicated significant differences between almost all equations for boys and girls. Mean %BF calculated using different SF and BIA equations ranged from 11.0+/-3.9 to 19.7+/-4.9% for boys and from 14.7+/-4.3 to 25.1+/-5.1% for girls. In addition, the calculated sums of SF were highly correlated (p<0.001) with the Slaughter et al., and Boileau et al., age-specific SF regression equations in boys (r=0.93-0.99) and girls (r=0.92-0.97). The correlation coefficients with other SF prediction equations were also significant but somewhat lower (boys: r=0.44-0.57; girls: r=0.63-0.73). The mean %BF calculated using two BIA equations significantly correlated with all sums of SF (boys: r=0.51; girls: r=0.59-0.64). Our results demonstrate the need for caution when using these field techniques to estimate %BF of prepubertal children. According to the results of this investigation, it is suggested that the most appropriate regression equations for use with this group of children are the Boileau et al., SF and the Houtkooper et al., BIA equations. In addition, the calculated sums of SF, especially the sum of two SF (triceps and calf) could be used to monitor changes in body composition in 9-11 year old children. Verf.-Referat