Body composition in prepubertal girls: comparison of six methods.

Objective: To compare estimates of fat mass (FM), fat-free mass (FFM) and percentage body fat (%fat) by six different methods in prepubertal girls.

Design: Cross-sectional study.

Subjects: Normal-weight, multi-ethnic, prepubertal girls (age=8.5+/-0.4 y, n=101).

Measurements: Body composition was measured in each child by anthropometry (skinfold thickness using Slaughter equation), dual-energy X-ray absorptiometry (DXA), total body potassium (TBK), isotope dilution for total body water measurement (TBW), multifrequency bioelectrical impedance spectroscopy (BIS), and total body electrical conductivity (TOBEC).

Results: TOBEC and skinfold thickness yielded the lowest values of FM followed by DXA, TBK, TBW and BIS, with BIS giving the highest value of FM. All methods were significantly different for FFM, FM and %fat (P<0.001), except FFM by DXA and TBK. The Bland-Altman limits of agreement among the methods reveal that they are not directly interchangeable for FM, FFM, or %fat. The largest mean difference for FM was between TOBEC and BIS (-2.90 kg), whereas the smallest mean difference was between TOBEC and skinfold thickness (-0.14 kg). For FFM, the largest mean difference was also between TOBEC and BIS (2.83 kg), but the smallest mean difference for FFM was between DXA and TBK (-0.03 kg). For %fat, the mean differences were larger, -10.5% for TOBEC and BIS and+9.7% for skinfold thickness and BIS. The closest two techniques for %fat were TOBEC and skinfold thickness (mean difference of -0.62%) and DXA and TBK (-1.81%).

Conclusions: We found that estimates of body composition in prepubertal 8-y-old girls are highly method-dependent and that the six methods studied (DXA, TBK, TBW, TOBEC, BIS and anthropometry) are not directly interchangeable.