Validation of bioelectrical impedance analysis for estimating limb lean mass in free-living Caucasian elderly people

Marina De Rui,Nicola Veronese,Francesco Bolzetta,Linda Berton,Sara Carraro,Giulia Bano,Caterina Trevisan,Simona Pizzato,Alessandra Coin,Egle Perissinotto,Enzo Manzato,Giuseppe Sergi
DOI: https://doi.org/10.1016/j.clnu.2016.04.011
Abstract:Background & aims: Aging is characterized by a loss of limb lean mass (LLM) that can lead to physical disability and death. Regional bioelectrical impedance analysis (BIA) may be a reliable method for estimating LLM, but no prediction equations are available for elderly Caucasian subjects. The aim of this study was to develop and validate a BIA-based equation for predicting LLM in healthy elderly Caucasians, taking dual X-ray absorptiometry (DXA) as the reference method. Methods: Using a cross-sectional design, 244 free-living healthy Caucasian subjects (117 men, 179 women) over 60 years of age were enrolled. LLM was measured with DXA (LLMDXA), and the resistance (Rz) and reactance (Xc) of each limb were measured with a regional bioelectrical impedance analyzer. A resistive index (RI) was calculated from stature in meters divided by Rz of each arm. A BIA-based multiple regression equation for predicting the lean mass (LM) of dominant and non-dominant limbs was developed using a double cross-validation technique. Results: Using the sample as a whole, cross-validation resulted in an equation specific for each limb, as follows, where sex equals 1 for males, and 0 for females: LM (kg) = -0.081 + (0.061*RI) + (0.010*body weight) + (0.299*sex) for the dominant arm; LM (kg) = -0.026 + (0.014*RI) + (0.009*body weight) + (0.352*sex) for the non-dominant arm; LM (kg) = -0.462 + (0.027*RI) + (0.047*body weight) + (0.639*sex) + (0.026*Xc) for the dominant leg; and for the non-dominant leg, LM (kg) = -0.522 + (0.029*RI) + (0.045*body weight) + (0.569*sex) + (0.025*Xc). The DXA-measured and BIA-predicted LLM for each limb did not differ significantly. Conclusion: Our newly-developed BIA equations seem to provide a valid estimation of LLM in older Caucasian adults.
What problem does this paper attempt to address?