<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>Fat-free mass (FFM) estimation has dramatic importance for body composition evaluation, often providing a basis for treatment of obesity and muscular dystrophy. However, current methods of FFM estimation have several drawbacks, usually related to either cost-effectiveness and equipment size (dual-energy X-ray absorptiometry (DEXA) scan) or model limitations. In this study, we present and validate a new FFM estimation model based on hand-to-hand bioimpedance analysis (BIA) and arm volume. Forty-two participants underwent a full-body DEXA scan, a series of anthropometric measurements, and upper-body BIA measurements with the custom-designed wearable wrist-worn impedance meter. A new two truncated cones (TTC) model was trained on DEXA data and achieved the best performance metrics of 0.886 ± 0.051 r<sup>2</sup>, 0.052 ± 0.009 % mean average error, and 6.884 ± 1.283 kg maximal residual error in FFM estimation. The model further demonstrated its effectiveness in Bland-Altman comparisons with the skinfold thickness-based FFM estimation method, achieving the least mean bias (0.007 kg). The novel TTC model can provide an alternative to full-body BIA measurements, demonstrating an accurate FFM estimation independently of population variables.</p></abstract>

Fat-free mass (FFM) estimation has dramatic importance for body composition evaluation, often providing a basis for treatment of obesity and muscular dystrophy. However, current methods of FFM estimation have several drawbacks, usually related to either cost-effectiveness and equipment size (dual-energy X-ray absorptiometry (DEXA) scan) or model limitations. In this study, we present and validate a new FFM estimation model based on hand-to-hand bioimpedance analysis (BIA) and arm volume. Forty-two participants underwent a full-body DEXA scan, a series of anthropometric measurements, and upper-body BIA measurements with the custom-designed wearable wrist-worn impedance meter. A new two truncated cones (TTC) model was trained on DEXA data and achieved the best performance metrics of 0.886 ± 0.051 r2, 0.052 ± 0.009 % mean average error, and 6.884 ± 1.283 kg maximal residual error in FFM estimation. The model further demonstrated its effectiveness in Bland-Altman comparisons with the skinfold thickness-based FFM estimation method, achieving the least mean bias (0.007 kg). The novel TTC model can provide an alternative to full-body BIA measurements, demonstrating an accurate FFM estimation independently of population variables.

Validation of non-empirical fat-free mass estimation model for a wrist-worn device

Journal of Electrical Bioimpedance

This work is licensed under the Creative Commons Attribution 4.0 International License.

Fat-free mass (FFM) estimation has dramatic importance for body composition evaluation, often providing a basis for treatment of obesity and muscular...

Validation of non-empirical fat-free mass estimation model for a wrist-worn device

Online veröffentlicht: 25. Juni 2022

Seitenbereich: 31 - 38

Eingereicht: 08. Apr. 2022

DOI: https://doi.org/10.2478/joeb-2022-0006

Schlüsselwörter
Bioelectrical impedance analysis, fat free mass, upper-body measurement, multiple linear regression, predictive modeling

© 2022 Aleksandr Polokhin, Anna Pronina, Andrey Boev, Stas Gorbunov, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Validation of non-empirical fat-free mass estimation model for a wrist-worn device

Online veröffentlicht: 25. Juni 2022

Seitenbereich: 31 - 38

Eingereicht: 08. Apr. 2022

DOI: https://doi.org/10.2478/joeb-2022-0006

SchlüsselwörterBioelectrical impedance analysis, fat free mass, upper-body measurement, multiple linear regression, predictive modeling

© 2022 Aleksandr Polokhin, Anna Pronina, Andrey Boev, Stas Gorbunov, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Schlüsselwörter
Bioelectrical impedance analysis, fat free mass, upper-body measurement, multiple linear regression, predictive modeling