ABSTRACT
The primary aim of this study was to evaluate the accuracy of skinfold thickness (SFT) measurements for the estimation of %Fat when compared to dual energy X-ray absorptiometry (DXA) in individuals with Down syndrome (DS). The secondary aim was to develop a new SFT-based body fat equation (SFTNICKERSON). SFT-based %Fat was estimated using a body fat equation from González-Agüero (SFTG-A) and body density conversion formulas from Siri (SFTSIRI) and Brozek (SFTBROZEK). Criterion %Fat was measured via DXA. SFTG-A, SFTSIRI, and SFTBROZEK were significantly lower than DXA (mean differences ranged from -7.59 to -13.51%; all p < 0.001). The SEE values ranged from 3.47% (SFTBROZEK) to 8.60% (SFTG-A). The 95% limits of agreement were greater than ±10% for all comparisons. Mid-axilla and suprailium were significant predictors of %Fat (both p < 0.05). %Fat SFTNICKERSON = 10.323 + (0.661 × mid-axilla) + (0.712 × suprailium). Age and all other skinfold sites were not statically significant in the regression model (all p > 0.05). Current findings indicate that SFTG-A, SFTSIRI, and SFTBROZEK erroneously place an individual with excessive adiposity in a normal healthy range. Accordingly, the current study developed a new equation (SFTNICKERSON) that can easily be administered in people with DS in a quick and efficient time frame. However, further research is warranted in this area.
Subject(s)
Down Syndrome , Humans , Adipose Tissue/diagnostic imaging , Body Composition , Skinfold Thickness , Absorptiometry, Photon/methods , AnthropometryABSTRACT
The purpose of this investigation was to determine the agreement between the body adiposity index (BAI) and dual energy X-ray absorptiometry (DXA) for measuring BF% in adults with Down syndrome (DS). Twenty adults (male: n=10; female: n=10) with Down syndrome volunteered to participate in this study. Criterion BF% was determined by DXA and predicted BF% was estimated by the BAI method. There was a significant mean difference (p<0.001) between DXA BF% (39.94±10.80%) and the BAI BF% (42.60±8.19%). The correlation between the two BF% variables was large and significant (r=0.73, p<0.001). However, the standard error of the estimate and total error was 7.79% and 7.86%, respectively. Additionally, the 95% limits of agreement ranged from 12.21% below to 17.52% above the constant error of 2.65%. Our findings suggest that on average, the BAI significantly overestimated BF% when compared to DXA values. Though there was a strong correlation between both methods, the wide limits of agreement suggest there is large amount of individual error when estimating BF% via the BAI. Therefore, the use of the BAI for individuals with DS does not appear to be accurate for estimating BF%.
