PLoS By Category | Recent PLoS Articles

Science Policy

Use of Fat Mass and Fat Free Mass Standard Deviation Scores Obtained Using Simple Measurement Methods in Healthy Children and Patients: Comparison with the Reference 4-Component Model
Published: Friday, May 17, 2013
Author: Rachel R. Atherton et al.

by Rachel R. Atherton, Jane E. Williams, Jonathan C. K. Wells, Mary S. Fewtrell


Clinical application of body composition (BC) measurements for individual children has been limited by lack of appropriate reference data.


(1) To compare fat mass (FM) and fat free mass (FFM) standard deviation scores (SDS) generated using new body composition reference data and obtained using simple measurement methods in healthy children and patients with those obtained using the reference 4-component (4-C) model; (2) To determine the extent to which scores from simple methods agree with those from the 4-C model in identification of abnormal body composition.


FM SDS were calculated for 4-C model, dual-energy X-ray absorptiometry (DXA; GE Lunar Prodigy), BMI and skinfold thicknesses (SFT); and FFM SDS for 4CM, DXA and bioelectrical impedance analysis (BIA; height2/Z)) in 927 subjects aged 3.8–22.0 y (211 healthy, 716 patients).


DXA was the most accurate method for both FM and FFM SDS in healthy subjects and patients (mean bias (limits of agreement) FM SDS 0.03 (±0.62); FFM SDS -0.04 (±0.72)), and provided best agreement with the 4-C model in identifying abnormal BC (SDS =-2 or =2). BMI and SFTs were reasonable predictors of abnormal FM SDS, but poor in providing an absolute value. BIA was comparable to DXA for FFM SDS and in identifying abnormal subjects.


DXA may be used both for research and clinically to determine FM and FFM SDS. BIA may be used to assess FFM SDS in place of DXA. BMI and SFTs can be used to measure adiposity for groups but not individuals. The performance of simpler techniques in monitoring longitudinal BC changes requires investigation. Ultimately, the most appropriate method should be determined by its predictive value for clinical outcome.