Scott B Going

PMID: 14636813;Abstract:

Background. Pathways, a randomized trial, evaluated the effectiveness of a school-based obesity prevention program on body composition changes in American Indian children. Several body composition methods were compared in intervention and control schools for assessing body composition changes. Methods. Body composition methods, including skinfolds, bioelectric impedance analysis (BIA), body mass index (BMI), and using a combination of body composition methods were selected to assess 3-year changes in PBF in 705 children within 21 intervention schools and 663 children within 20 control schools. The study equation using skinfolds, BIA, and BMI was developed on a previous sample of American Indian children using deuterium oxide dilution as the criterion method. Results. Body fat changes among methods for the intervention sample ranged from 5.4% (BMI method) to 7.1% (combination of methods) and for the control sample, from 5.8% (BMI method) to 7.3% (combination of methods). The study equation estimates were significantly higher than the other methods and the BMI equation estimates were significantly lower than the other methods except by BIA. The BIA equation showed a significantly larger standard deviation of the difference over the 3-year intervention than each of the other methods indicating less reliability for detecting body composition changes. Conclusions. Within the Pathways large scale intervention trial with American Indian children, we found comparable yet significantly different mean PBF changes among methods. However, BIA was not as reliable as skinfolds and the combination of BIA, skinfolds, and body weight in assessing PBF changes. © 2003 American Health Foundation and Elsevier Science (USA). All rights reserved.

PMID: 6530218;Abstract:

The estimation of body fatness by the densitometric method assumes a constant density of the fat-free body. Water constitutes the largest single component of the fat-free body and because of its relatively low density (.9934 gm/cc at 37°C) exerts the greatest influence on the density of the fat-free body. The purpose of this study was to test the assumption that the water content of the fat-free body (% water-FFB) is constant during growth and development. The sample consisted of 292 Black and White males and females (ages 8-30 years) and was classified via maturational assessment as prepubertal, (N = 54), pubertal (N = 50), postpubertal (N = 107) and adult (N = 81). Body water was measured by deuterium oxide dilution and body density was measured by the underwater weighing method. Body water as a percentage of body weight (% water-BW) was higher (p .01) for males (X̄ = 61.6%) than females (X̄ = 55.8%) with the greatest difference observed in the pubertal, postpubertal and adult levels. This finding likely reflected the increased relative fatness of females at puberty and is in agreement with skinfold thickness measurements. The % water-FFB progressively decreased from prepubescence to adulthood by 2.8% overall at a rate of 0.38% per year for both sexes with the lowest values found in the adult samples. Although the % water-FFB of the prepubertal and pubertal groups was similar (p > .05), both groups were significantly (p .01) higher than the postpubertal and adult groups, and the postpubertal groups was higher (p .05) than the adult group. Moreover, the % water-FFB was lower (p .05) for males (X̄ = 73.5%) than females (X̄ = 74.2%). No difference in % water-FFB was found between the Black and White samples. These findings suggest that the water content of the FFB is not constant during growth and development and that adult equations for estimating fat from density and other indirect methods may not be appropriate for children, overestimating fatness by at least 4%.

Abstract:

Estimation of energy requirements relies on adequate values of basal metabolic rate (BMR). Prediction equations recommended for international use have been shown to overestimate BMR in populations living in the tropics. We have previously shown the inadequacy of these equations in samples of Brazilians living in tropical and temperate regions of the country. We sought to investigate whether BMR could adequately be estimated by prediction equations in a sample of Brazilians living in a different setting: the Sonoran desert of the Southwestern USA. BMR was measured under standard conditions in 33 subjects (14 men). Mean bias (estimated-measured) varied from 404.4 to 708.6 kJday^-1 in women and 566.8 to 1122.8 kJday^-1 in men, representing 8.5-15 and 8.9-17.6% overestimation, respectively, using the Schofield equations. Bland and Altman analyses showed large, relevant limits of agreement. The results using the recommended equations for the American population (IOM, 2005) were only 2% different from the Schofield equations. The Harris and Benedict equations yielded higher overestimations (15.0 and 16.8% for women and men, respectively) and the Henry and Rees equations also overestimated BMR (8.5 and 8.9%) even though they were developed for populations from the tropics, although to a lesser degree. It is concluded that the equations currently recommended for international use are not appropriate for Brazilians living in the Southwestern USA.