Can bioelectric impedance monitors be used to accurately estimate body fat in Chinese adults?

Asia Pac J Clin Nutr. 2007;16(1):66-73.

Abstract

Many laboratory-based methods exist to estimate body fat, yet few can be rapidly and easily applied to field studies. Bioelectric impedance analysis (BIA) has developed to include portable foot-to-foot systems that can estimate body fat during field studies, but it is unclear if the data they provide are comparable to anthropometric methods traditionally used in large epidemiological fieldwork. This study analysed the reliability and validity of three BIA devices (low, medium, and high cost), from duplicate measures of mass and percentage body fat (%BF) from 20 young Chinese. Comparisons were made to reference values of %BF derived from 38 duplicated anthropometric measurements and the mean of at least 7 regression equations. All three BIA devices were reliable, with intraclass correlation coefficients never below 0.999, whilst both technical errors of measurement and coefficients of variation (expressed as percentages) were below 1%. Validity analysis revealed all three devices significantly overestimated %BF using the standard measurement setting (no correction for athletic status) compared to the reference method: UM-022 (+3.2%, p <0.01), BF-350 (+2.6%, p <0.01), and TBF-410 (+2.1%, p <0.01). When %BF was corrected for athletic status, neither the BF-350 (+0.3%, p =0.72), nor the TBF-410 (-0.2%, p =0.86) produced a %BF that differed significantly from the reference method. It was concluded that these three BIA devices were reliable and could be recommended as valid field measures of mass and %BF in this sample population provided the device allows a correction for athletic status.

Publication types

  • Comparative Study

MeSH terms

  • Adipose Tissue / physiology*
  • Adult
  • Anthropometry / instrumentation
  • Anthropometry / methods*
  • Body Composition / physiology*
  • Costs and Cost Analysis*
  • Electric Impedance*
  • Female
  • Humans
  • Male
  • Reference Values
  • Regression Analysis
  • Reproducibility of Results
  • Sensitivity and Specificity