Have a personal or library account? Click to login
The effect of postural orientation on body composition and total body water estimates produced by smartwatch bioelectrical impedance analysis: an intra- and inter-device evaluation Cover

The effect of postural orientation on body composition and total body water estimates produced by smartwatch bioelectrical impedance analysis: an intra- and inter-device evaluation

Journal of Electrical Bioimpedance
Volume 15 (2024): Issue 1 (January 2024)
By: Anabelle Vallecillo-Bustos,  Abby T. Compton,  Sydney H. Swafford,  Megan E. Renna,  Tanner Thorsen,  Jon Stavres and  Austin J. Graybeal  
Open Access
|Aug 2024

Figures & Tables

Fig 1a-d.

Bland-Altman analyses for body composition and total body water estimates measured by SWBIA and the criterion DXA and BIS in the combined sample. Bland-Altman plots for BF% (a), FM (b), FFM (c), and TBW (d) estimates by SWBIA body position (rows) are displayed. The upper and lower dashed lines on each plot represents the 95% limits of agreement, and the middle-dashed lines represent the mean difference between SWBIA and the criterion DXA or BIS. The solid blue line represents the regression line and the surrounding shading represents the 95% confidence interval of the regression line. SWBIA: smartwatch bioelectrical impedance analysis; DXA: dual-energy X-ray absorptiometry; BIS: bioelectrical impedance spectroscopy; BF%: body fat percentage; FM: fat mass; FFM: fat-free mass; TBW: total body water; LOA: 95% limits of agreement. *, significant proportional bias at p ≤ 0.026.
Bland-Altman analyses for body composition and total body water estimates measured by SWBIA and the criterion DXA and BIS in the combined sample. Bland-Altman plots for BF% (a), FM (b), FFM (c), and TBW (d) estimates by SWBIA body position (rows) are displayed. The upper and lower dashed lines on each plot represents the 95% limits of agreement, and the middle-dashed lines represent the mean difference between SWBIA and the criterion DXA or BIS. The solid blue line represents the regression line and the surrounding shading represents the 95% confidence interval of the regression line. SWBIA: smartwatch bioelectrical impedance analysis; DXA: dual-energy X-ray absorptiometry; BIS: bioelectrical impedance spectroscopy; BF%: body fat percentage; FM: fat mass; FFM: fat-free mass; TBW: total body water; LOA: 95% limits of agreement. *, significant proportional bias at p ≤ 0.026.

Fig 2a-d.

Bland-Altman analyses for body composition and total body water estimates measured by SWBIA and the criterion DXA and BIS in females. Bland-Altman plots for BF% (a), FM (b), FFM (c), and TBW (d) estimates by SWBIA body position (rows) are displayed. The upper and lower dashed lines on each plot represent the 95% limits of agreement, and the middle-dashed lines represent the mean difference between SWBIA and the criterion DXA or BIS. The solid blue line represents the regression line, and the surrounding shading represents the 95% confidence interval of the regression line. SWBIA: smartwatch bioelectrical impedance analysis; DXA: dual-energy X-ray absorptiometry; BIS: bioelectrical impedance spectroscopy; BF%: body fat percentage; FM: fat mass; FFM: fat-free mass; TBW: total body water; LOA: 95% limits of agreement. *, significant proportional bias at p ≤ 0.003.
Bland-Altman analyses for body composition and total body water estimates measured by SWBIA and the criterion DXA and BIS in females. Bland-Altman plots for BF% (a), FM (b), FFM (c), and TBW (d) estimates by SWBIA body position (rows) are displayed. The upper and lower dashed lines on each plot represent the 95% limits of agreement, and the middle-dashed lines represent the mean difference between SWBIA and the criterion DXA or BIS. The solid blue line represents the regression line, and the surrounding shading represents the 95% confidence interval of the regression line. SWBIA: smartwatch bioelectrical impedance analysis; DXA: dual-energy X-ray absorptiometry; BIS: bioelectrical impedance spectroscopy; BF%: body fat percentage; FM: fat mass; FFM: fat-free mass; TBW: total body water; LOA: 95% limits of agreement. *, significant proportional bias at p ≤ 0.003.

Fig 3a-d.

Bland-Altman analyses for body composition and total body water estimates measured by SWBIA and the criterion DXA and BIS in males. Bland-Altman plots for BF% (a), FM (b), FFM (c), and TBW (d) estimates by SWBIA body position (rows) are displayed. The upper and lower dashed lines on each plot represents the 95% limits of agreement, and the middle-dashed lines represent the mean difference between SWBIA and the criterion DXA or BIS. The solid blue line represents the regression line and the surrounding shading represents the 95% confidence interval of the regression line. SWBIA: smartwatch bioelectrical impedance analysis; DXA: dual-energy X-ray absorptiometry; BIS: bioelectrical impedance spectroscopy; BF%: body fat percentage; FM: fat mass; FFM: fat-free mass; TBW: total body water; LOA: 95% limits of agreement. *, significant proportional bias at p ≤ 0.045.
Bland-Altman analyses for body composition and total body water estimates measured by SWBIA and the criterion DXA and BIS in males. Bland-Altman plots for BF% (a), FM (b), FFM (c), and TBW (d) estimates by SWBIA body position (rows) are displayed. The upper and lower dashed lines on each plot represents the 95% limits of agreement, and the middle-dashed lines represent the mean difference between SWBIA and the criterion DXA or BIS. The solid blue line represents the regression line and the surrounding shading represents the 95% confidence interval of the regression line. SWBIA: smartwatch bioelectrical impedance analysis; DXA: dual-energy X-ray absorptiometry; BIS: bioelectrical impedance spectroscopy; BF%: body fat percentage; FM: fat mass; FFM: fat-free mass; TBW: total body water; LOA: 95% limits of agreement. *, significant proportional bias at p ≤ 0.045.

Agreement between SWBIA and BIS-derived estimates of total body water by body position_

Mean (95%CI)MD (95%CI)R2RMSECCC
Combined (N = 106)
BIS37.0 (35.4, 38.5)
SWBIASeated37.2 (35.7, 38.7).08 (-.46, .62).882.80.94
SWBIAStand37.2 (35.6, 38.7).05 (-.49, .59).892.79.94
SWBIASupine37.1 (35.5, 38.6)*,†-.05 (-.59, .49).892.79.94
Female (N = 62)
BIS33.4 (31.9, 34.9)
SWBIASeated32.6 (31.3, 33.8)-.78 (-1.43, -.14)‡.822.64.89
SWBIAStand32.5 (31.3, 33.8)-.85 (-1.49, -.20)‡.822.66.89
SWBIASupine32.4 (31.1, 33.7)†-.96 (-1.61, -.31)‡.812.72.88
Male (N = 44)
BIS42.1 (39.6, 44.5)
SWBIASeated43.9 (41.7, 46.1)1.29 (.46, 2.13)‡.883.02.92
SWBIAStand43.9 (41.7, 46.0)1.32 (.50, 2.13)‡.892.96.93
SWBIASupine43.8 (41.6, 46.0)1.23 (.43, 2.04)‡.892.88.93

Participant characteristics_

Total (n = 117)Female (n = 69)Male (n = 48)
Mean ± SDMean ± SDMean ± SD
Age (y)21.4 ± 3.021.2 ± 1.821.7 ± 4.1
Height (cm)168.2 ± 8.9163.4 ± 5.9*175.2 ± 7.8
Weight (kg)71.9 ± 18.168.2 ± 17.6*77.1 ± 17.7
BMI (kg/m2)25.3 ± 5.725.5 ± 6.425.0 ± 4.5
Body fat (%)28.6 ± 10.333.6 ± 8.9*21.5 ± 7.6
Fat mass (kg)21.2 ± 12.324.1 ± 13.2*17.2 ± 9.6
FFM (kg)50.7 ± 12.044.2 ± 6.7*60.0 ± 11.8
TBW (L)37.0 ± 8.133.4 ± 6.0*42.1 ± 7.9
N (%)N (%)N (%)
Race
White42 (35.9%)29 (42.0%)13 (27.1%)
Black37 (31.6%)25 (36.2%)12 (25.0%)
Asian32 (27.4%)10 (14.5%)22 (45.8%)
Hispanic6 (5.1%)5 (7.2%)1 (2.1%)
Weight Status
Normal66 (56.4%)40 (58.0%)26 (54.2%)
Overweight33 (28.2%)18 (26.1%)15 (31.3%)
Obese18 (15.4%)11 (15.9%)7 (14.6%)

Agreement between SWBIA and DXA-derived body composition estimates by body position_

Mean (95%CI)MD (95%CI)R2RMSECCC
Combined (N = 117) BF%
DXA28.6 (26.7, 30.5)
SWBIASeated29.0 (27.4, 30.7).43 (-.38, 1.23).824.40.90
SWBIAStand28.9 (27.2, 30.6).30 (-.50, 1.10).824.37.90
SWBIASupine29.2 (27.5, 30.9)*.62 (-.17, 1.42).824.37.90
FM (kg)
DXA21.2 (19.0, 23.5)
SWBIASeated21.6 (19.5, 23.6).31 (-.28, .89).943.18.96
SWBIAStand21.5 (19.4, 23.5).22 (-.37, .80).933.19.96
SWBIASupine21.7 (19.6, 23.8)*.45 (-.12, 1.03).943.15.96
FFM (kg)
DXA50.7 (48.5, 52.8)
SWBIASeated50.3 (48.3, 52.4)-.31 (-.89, .28).933.18.96
SWBIAStand50.4 (48.4, 52.5)-.22 (-.80, .37).933.19.96
SWBIASupine50.2 (48.1, 52.3)*-.45 (-1.03, .12).933.15.96
Females (n = 69) BF%
DXA33.6 (31.4, 35.7)
SWBIASeated34.2 (32.4, 35.9).56 (-.35, 1.47).823.81.89
SWBIAStand33.9 (32.2, 35.7).34 (-.60, 1.27).823.88.89
SWBIASupine34.4 (32.7, 36.1)*.81 (-.11, 1.73).833.89.88
FM (kg)
DXA24.1 (20.9, 27.2)
SWBIASeated24.3 (21.4, 27.3).26 (-.35, .87).972.54.98
SWBIAStand24.2 (21.3, 27.1).12 (-.50, .74).972.57.98
SWBIASupine24.5 (21.6, 27.5)*.44 (-.17, 1.05).972.55.98
FFM (kg)
DXA44.2 (42.6, 45.8)
SWBIASeated43.9 (42.3, 45.5)-.26 (-.87, .35).862.54.93
SWBIAStand44.0 (42.5, 45.6)-.12 (-.74, .50).862.57.93
SWBIASupine43.7 (42.1, 45.3)*-.44 (-1.05, .17).862.55.93
Males (n = 48) BF%
DXA21.5 (19.3, 23.7)
SWBIASeated21.7 (19.9, 23.5).24 (-1.26, 1.74).565.13.72
SWBIAStand21.7 (19.9, 23.5).25 (-1.22, 1.72).565.00.73
SWBIASupine21.8 (20.0, 23.7).35 (-1.11, 1.81).574.98.74
FM (kg)
DXA17.2 (14.4, 20.0)
SWBIASeated17.5 (15.0, 20.1).37 (-.77, 1.52).833.93.91
SWBIAStand17.5 (15.0, 20.0).36 (-.79, 1.50).833.90.91
SWBIASupine17.6 (15.1, 20.2).47 (-.65, 1.59).843.85.91
FFM (kg)
DXA60.0 (56.5, 63.4)
SWBIASeated59.6 (56.6, 62.6)-.37 (-1.52, .77).903.93.94
SWBIAStand59.6 (56.6, 62.6)-.36 (-1.50, .79).893.90.94
SWBIASupine59.5 (56.5, 62.5)-.47 (-1.59, .65).903.85.94
DOI: https://doi.org/10.2478/joeb-2024-0010 | Journal eISSN: 1891-5469
Journal RSS Feed
Language: English
Page range: 89 - 98
Submitted on: Jun 3, 2024
Published on: Aug 5, 2024
Published by: University of Oslo
In partnership with: Paradigm Publishing Services
Publication frequency: 1 issue per year
Keywords:
Wearables,
smartwatch,
digital health,
sensors,
body composition,
body water,
digital anthropometrics
Related subjects:
Engineering,
Bioengineering and biomedical engineering,
Biomedical electronics,
Life sciences,
Biophysics,
Medicine,
Biomedical engineering,
Physics,
Spectroscopy and metrology

© 2024 Anabelle Vallecillo-Bustos, Abby T. Compton, Sydney H. Swafford, Megan E. Renna, Tanner Thorsen, Jon Stavres, Austin J. Graybeal, published by University of Oslo
This work is licensed under the Creative Commons Attribution 4.0 License.

Previous articleVolume 15 (2024): Issue 1 (January 2024)Next article