Prediction of fat-free body mass from bioelectrical impedance and anthropometry among 3-year-old children using DXA

TY - JOUR

T1 - Prediction of fat-free body mass from bioelectrical impedance and anthropometry among 3-year-old children using DXA

AU - Ejlerskov, Katrine Tschentscher

AU - Jensen, Signe Marie

AU - Christensen, Line B

AU - Ritz, Christian

AU - Michaelsen, Kim F.

AU - Mølgaard, Christian

N1 - CURIS 2014 NEXS 075

PY - 2014/1/27

Y1 - 2014/1/27

N2 - For 3-year-old children suitable methods to estimate body composition are sparse. We aimed to develop predictive equations for estimating fat-free mass (FFM) from bioelectrical impedance (BIA) and anthropometry using dual-energy X-ray absorptiometry (DXA) as reference method using data from 99 healthy 3-year-old Danish children. Predictive equations were derived from two multiple linear regression models, a comprehensive model (height 2 /resistance (RI), six anthropometric measurements) and a simple model (RI, height, weight). Their uncertainty was quantified by means of 10-fold cross-validation approach. Prediction error of FFM was 3.0% for both equations (root mean square error: 360 and 356â€...g, respectively). The derived equations produced BIA-based prediction of FFM and FM near DXA scan results. We suggest that the predictive equations can be applied in similar population samples aged 2-4 years. The derived equations may prove useful for studies linking body composition to early risk factors and early onset of obesity.

AB - For 3-year-old children suitable methods to estimate body composition are sparse. We aimed to develop predictive equations for estimating fat-free mass (FFM) from bioelectrical impedance (BIA) and anthropometry using dual-energy X-ray absorptiometry (DXA) as reference method using data from 99 healthy 3-year-old Danish children. Predictive equations were derived from two multiple linear regression models, a comprehensive model (height 2 /resistance (RI), six anthropometric measurements) and a simple model (RI, height, weight). Their uncertainty was quantified by means of 10-fold cross-validation approach. Prediction error of FFM was 3.0% for both equations (root mean square error: 360 and 356â€...g, respectively). The derived equations produced BIA-based prediction of FFM and FM near DXA scan results. We suggest that the predictive equations can be applied in similar population samples aged 2-4 years. The derived equations may prove useful for studies linking body composition to early risk factors and early onset of obesity.

U2 - 10.1038/srep03889

DO - 10.1038/srep03889

M3 - Journal article

C2 - 24463487

SN - 2045-2322

VL - 4

JO - Scientific Reports

JF - Scientific Reports

M1 - 3889

ER -