Precision and accuracy of bioelectrical impedance analysis devices in supine versus standing position with or without retractable handle in Caucasian subjects

Abstract

Background & aims: Bioelectrical impedance analysis (BIA) could be facilitated in subjects who are able to stand by using scales without (BIAstd4) or with a retractable handle (BIAstd8), provided that they are as precise as BIA devices commonly used in the supine position in the hospital setting (BIAsup). This observational prospective cross-sectional study aimed to compare the precision and accuracy of BIAstd4, BIAstd8 and BIAsup in a Caucasian population.

Methods: Fat mass percentage (FM%) was measured in 160 healthy Caucasian subjects (80 men/80 women) aged 20-60 years, with a body mass index (BMI) ≥18.5 and < 30 kg/m2, using the HAGRID Body Fat Scales (Huawei Technologies Co., Ltd., China) in BIAstd4 or BIAstd8 mode, and the Nutriguard-M (Data Input GmbH, Germany) as BIAsup. Intra-unit and inter-unit precisions of each device were evaluated by calculating the coefficients of variation (CV%) of 3 measurements with 3 different units of each device. Inter-device precisions were evaluated with Pearson correlations, Bland-Altman plots, and repeated measures ANOVA followed by post-hoc Bonferroni tests. Accuracy of these BIA devices was estimated in a subgroup of 16 subjects, using comparison with dual-energy X-ray absorptiometry (DXA).

Results: The study population was 40 ± 12 years old, with a body height and weight of 171 ± 10 cm and 72.2 ± 11.5 Kg, respectively. All three devices were very precise with intra-unit CV% of 0.5%, 0.9%, and 0.3% and inter-unit CV% of 0.5%, 1.1%, and 0.4% for BIAstd4, BIAstd8 and BIAsup, respectively. Inter-device precision was ±2.1% for BIAstd4/BIAsup, ±1.9% for BIAstd8/BIAsup, and ±1.3% for BIAstd8/BIAstd4. Bland-Altman plots showed bias ±1.96 SD of 0.3 ± 5.2% for BIAstd4/BIAsup, -0.4 ± 4.5% for BIAstd8/BIAsup and -0.6 ± 3.1% for BIAstd8/BIAstd4. Compared to DXA, all three devices tended to underestimate FM% in men with low BMI, while only BIAstd4 and BIAstd8 tended to overestimate FM% in women with high BMI. FM% measurement accuracy was ±2.6% for BIAsup/DXA, ±3.3% for BIAstd4/DXA, and ±3.4% for BIAstd8/DXA.

Conclusions: Both BIAstd4 and BIAstd8 show a good intra- and inter-unit precision close to BIAsup, making them suitable for rapid body composition assessment in non-bedridden subjects. However, all these three devices should not be used interchangeably, because BIAstd4 and BIAstd8 tend to accentuate FM% changes during body composition monitoring compared to BIAsup and DXA.

Trial registration: ClinicalTrial.gov no. NCT04504799.

Keywords: Bioelectrical impedance analysis; Body composition; Dual-energy X-ray absorptiometry; Fat mass; Nutritional assessment.

Conflict of interest statement

Declaration of competing interest The authors declare no competing interests.

Copyright © 2021 The Author(s). Published by Elsevier Ltd.. All rights reserved.