Title: Longitudinal DXA studies: minimum scanning interval for pediatric assessment of body fat Authors
|Shypailo, Roman -|
|Ellis, Kenneth -|
Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: April 15, 2008
Publication Date: July 1, 2008
Citation: Shypailo, R.J., Ellis, K.J. 2008. Longitudinal DXA studies: minimum scanning interval for pediatric assessment of body fat [Abstract]. 8th International Symposium on In Vivo Body Composition Studies. Paper No. 86. Technical Abstract: The increased prevalence of obesity in the United States has led to the increased use of Dual-energy X-ray absorptiometry (DXA) for assessment of body fat (TBF). The importance of early intervention has focused attention on pediatric populations. We used DXA precision analyses to determine suitable scanning intervals for the measurement of changes in TBF and lean mass (LBM) in children. Short-term precision was based on repeated measurements of 30 children (ages 7-14 y) using a Hologic Delphi (vers. 12.1). We calculated the root mean square standard deviation (RMS-SD), percent coefficient of variation (RMS-%CV), and the least significant change (LSC) at the 95% confidence interval. Long-term precision was based on 770 whole body phantom measurements spanning a 6-year period. The SD, %CV, and LSC for the phantom were calculated. Long-term LSC-SD (phantom) was 523 g for TBF and 454 g for LBM. The LSC-%CV values were 3.7% and 3.2%, respectively. Short-term LSC-SD (pediatric group) was slightly higher, with results of 680 g and 900 g for TBF and LBM, respectively, while LSC-%CV results were 2.7% and 3.5%. LSC-%CV results for DXA-calculated body weight were comparable (1.4% vs. 1.1% for phantom vs. subjects), whereas the pediatric LSC-SD (900 g) was higher than that for the phantom (400 g). Short-term precision results suggest that total TBF changes of at least 3% are needed in order to detect a true change at the 95% confidence level. Changes approaching 1 kg would also ensure that the change in TBF from baseline is significant. Similar changes are needed for LBM. These estimates should be applicable to DXA instruments manufactured by Hologic. Precision testing with other manufacturers' machines should be done in order to derive accurate LSC thresholds for TBF and LBM.