Validity of non-invasive methods for body composition measurements in older adults
Date
Authors
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Journal Issue
Is Version Of
Versions
Series
Department
Abstract
ABSTRACT
Purpose: The aim of this study was to investigate the validity of bioelectrical impedance analysis (BIA) and air-displacement plethysmography (ADP) in relation to dual energy X-ray absorptiometry (DXA) (i.e. the reference standard). Methods: Sixty-three older adults aged 60-96 years (40 men, 23 women). Body percent fat (%BF) was estimated by BIA, ADP and DXA. Single frequency (50Hz) BIA that measures whole body impedance was used and Kyle's equation was applied to estimate fat-free mass. Paired sample t-test, absolute percent errors and Cohen's d were used to evaluate differences among the 3 different methods. Methods agreement was assessed by Pearson's correlation, regression analysis and Bland-Altman plots. Classification agreement of obesity was evaluated using Kappa statistics. Results: ADP and BIA significantly overestimated %BF relative to DXA by 3.3% and 3.1%, yielding absolute errors of 14.1% and 12.4%, respectively. However, ADP (Cohen's d=0.35) had better agreement with DXA and BIA (Cohen's d=0.40). Regression analysis indicated smaller individual variations of ADP (SEE=3.23) compared to BIA (SEE=4.78). In addition, ADP (Kappa=0.58) showed better obesity classification agreement relative to DXA in comparison with BIA (Kappa = 0.35). However, Bland-Altman plots showed a positive proportional pattern (Slope= 0.24, R2=0.24, p<0.05) of biases in ADP, while no systematic pattern of biases was observed for BIA. A gender difference was also detected, indicating a better agreement in males than females. Conclusion: Given that both BIA and ADP had acceptable agreement with DXA in estimating %BF of older adults, ADP showed relatively better agreement in body composition measurement (i.e. %BF) and obesity classification in comparison to BIA. However, practitioners and/or researchers should be aware of the potential biases when using ADP to estimate %BF in older populations.