Reliability, biological variability, and accuracy of multi-frequency bioelectrical impedance analysis for measuring...

Area of Science:

• Physiology
• Biomedical Engineering
• Body Composition Analysis

Background:

• Bioelectrical impedance analysis (BIA) systems are increasingly utilized in research and fitness assessments due to technological advancements, affordability, and ease of use.
• Multifrequency BIA (MF-BIA) offers potential improvements in accuracy and precision, particularly for segmental analyses through octopolar contacts.

Purpose of the Study:

• To evaluate the reliability, biological variability, and accuracy of component measures (total body water, mass, and composition) using a commercially available multifrequency BIA system (InBody 770).

Main Methods:

• Fourteen healthy adults underwent duplicate MF-BIA assessments over three weeks under controlled conditions.
• Systematic error (test-retest reliability) and biological variability (day-to-day reliability) were assessed using intraclass correlation coefficients (ICC).
• Accuracy of MF-BIA body composition measurements was compared against dual-energy x-ray absorptiometry (DXA).

Main Results:

• Very high test-retest reliability (ICC ≥ 0.999) was observed for whole-body water and mass, and high reliability (ICC 0.973-1.000) for regional measures.
• Minor biological variability was noted, with slightly higher day-to-day differences in body water and mass compared to same-day measurements.
• Compared to DXA, MF-BIA showed a -4.0% bias for % body fat, overpredicted fat-free mass by 2.8 kg, and underpredicted fat mass by -2.9 kg.

Conclusions:

• The InBody 770 MF-BIA system exhibits high methodological reliability and stable day-to-day measurements for major body composition components in healthy individuals.
• A consistent ~3% body fat offset compared to criterion methods was confirmed.
• The precision and consistency of the InBody 770 support its potential as a military standard method and its suitability across a broad range of body fat percentages.