Background & aims: Bioelectrical impedance analysis (BIA) could be facilitated in subjects who are able to stand by using scales without (BIA std4 ) or with a retractable handle (BIA std8 ), provided that they are as precise as BIA devices commonly used in the supine position in the hospital setting (BIA sup ). This observational prospective cross-sectional study aimed to compare the precision and accuracy of BIA std4 , BIA std8 and BIA sup in a Caucasian population. Methods: Fat mass percentage (FM%) was measured in 160 healthy Caucasian subjects (80 men/80 women) aged 20e60 years, with a body mass index (BMI) !18.5 and < 30 kg/m 2 , using the HAGRID Body Fat Scales (Huawei Technologies Co., Ltd., China) in BIA std4 or BIA std8 mode, and the Nutriguard-M (Data Input GmbH, Germany) as BIA sup . 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, BlandeAltman 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 BIA std4 , BIA std8 and BIA sup , respectively. Inter-device precision was ±2.1% for BIA std4 /BIA sup , ±1.9% for BIA std8 /BIA sup , and ±1.3% for BIA std8 /BIA std4 . Bland eAltman plots showed bias ±1.96 SD of 0.3 ± 5.2% for BIA std4 /BIA sup , À0.4 ± 4.5% for BIA std8 /BIA sup and À0.6 ± 3.1% for BIA std8 /BIA std4 . Compared to DXA, all three devices tended to underestimate FM% in men with low BMI, while only BIA std4 and BIA std8 tended to overestimate FM% in women with high BMI. FM% measurement accuracy was ±2.6% for BIA sup /DXA, ±3.3% for BIA std4 /DXA, and ±3.4% for BIA std8 /DXA. Conclusions: Both BIA std4 and BIA std8 show a good intra-and inter-unit precision close to BIA sup , making them suitable for rapid body composition assessment in non-bedridden subjects. However, all these three devices should not be used interchangeably, because BIA std4 and BIA std8 tend to accentuate FM% changes during body composition monitoring compared to BIA sup and DXA. Trial registration: ClinicalTrial.gov no. NCT04504799.

No abstract