Title: Lipometer: a potential field method for body fat assessment in children Authors
|Shypailo, Roman - BAYLOR COLLEGE OF MED|
Submitted to: International Journal of Body Composition Research
Publication Type: Abstract Only
Publication Acceptance Date: June 1, 2005
Publication Date: September 7, 2005
Citation: Ellis, K.J., Shypailo, R.J. 2005. Lipometer: A potential field method for body fat assessment in children [abstract}. International Journal of Body Composition Research. 3(3):83. Technical Abstract: The prevalance of childhood obesity has significantly increased worldwide during the last 20 years. A number of school-based behavioral, dietary, and/or physical-activity interventions have been proposed to reverse this trend. Hence, a safe, quick, noninvasive body composition method for assessing body fatness in children that is also portable and precise is desirable. We have examined the performance of a hand-held, computerized, optical-based device (Lipometer) designed to measure the thickness of the subcutaneous adipose tissue layer. Lipometer measurements, taking less than 5 minutes to perform, were obtained at 15 sites over the body. A whole DXA scan was also obtained. A total of 278 children (150 females), ages 3-18 years, in three ethnic groups (145 white, 75 black, 58 hispanic) were measured. The Liopmeter values for the 15 sites were intra-correlated, varied with age, were gender-dependent, and had minor ethnic differences. Correlations between Lipometer values and whole body fat mass and percentage body fat (% Fat) by DXA were significant (r=0.3 - 0.9) and gender dependent. Although the initial Lipometer estimates of % Fat were significantly correlated with the DXA values, the prediction accuracy (+/- 5% for % Fat) was poor. With the subjects randomally assigned to a calibration (n=124) or validation (n=154) group, a revised pediatric-based Lipometer prediction equation for % Fat was obtained, with the DXA values as the reference. For the validation group, there was significant improvement in the correlation(r=0.91, p <0.001) and prediction error +/- 2.5%) for whole-body % Fat. Bland-Altman analysis showed that the revised % Fat-Lipometer values had a mean bias is 1% with 95% limits of agreement of 6% compared with the DXA-based values. These preliminary findings suggest that the Lipometer may provide an acceptable field method for monitoring subcutaneous fat thickness and its relative body distribution in growing children and for assessing a child's whole-body adiposity status.