Development and validation of bioelectrical impedance prediction equations estimating regional lean soft tissue mass in middle-aged adults

• Published in: European journal of clinical nutrition Vol. 77; no. 2; pp. 202 - 211
• Main Authors: Sardinha, Luís B., Rosa, Gil B., Hetherington-Rauth, Megan, Correia, Inês R., Magalhães, João P., Silva, Analiza M., Lukaski, Henry
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 01.02.2023; Nature Publishing Group
• Subjects: 692/700/1421/2109; 692/700/2814; 9/10; Absorptiometry; Absorptiometry, Photon; Adult; Adults; Bioelectricity; Body Composition; Clinical Nutrition; Cross-Sectional Studies; Dual energy X-ray absorptiometry; Electric Impedance; Epidemiology; Humans; Impedance; Internal Medicine; Linear Models; Mathematical models; Medicine; Medicine & Public Health; Metabolic Diseases; Middle age; Middle Aged; Parameters; Prediction models; Public Health; Regional analysis; Regional development; Reproducibility of Results; Soft tissues
• ISSN: 0954-3007; 1476-5640; 1476-5640
• DOI: 10.1038/s41430-022-01224-0

Background/Objectives Bioelectrical impedance (BIA) whole-body and regional raw parameters have been used to develop prediction models to estimate whole-body lean soft tissue (LSTM), with less attention being given to the development of models for regional LSTM. Therefore, we aimed to develop and validate BIA-derived equations predicting regional LSTM against dual x-ray absorptiometry (DXA) in healthy adults. Subjects/Methods 149 adults were included in this cross-sectional investigation. Whole-body and regional LSTM were assessed by DXA, and raw bioelectrical parameters of distinct body regions were measured using a 50 kHz phase sensitive BIA analyzer. BIA-derived equations were developed using a stepwise multiple linear regression approach in 2/3 of the sample and cross-validated in the remaining sample. Results Slopes and intercepts of predicted LSTM and DXA measured LSTM did not differ from 1 and 0, respectively, for each region ( p  ≥ 0.05), with the exception for the trunk ( p  < 0.05). The BIA-derived equations exhibited a strong relationship ( p  < 0.001) between the predicted and measured LSTM for each of the following body regions: right and left arms (R = 0.94; R = 0.96), right and left legs (R = 0.88; R = 0.88), upper body (R = 0.96), lower body (R = 0.89), right and left sides of the body (R = 0.94; R = 0.94), and trunk (R = 0.90). Agreement analyses revealed no associations between the differences and the means of the predicted and DXA-derived LSTM. Conclusion The developed BIA-derived equations provide a valid estimate of regional LSTM in middle-aged healthy adults, representing a cost-effective and time-efficient alternative to DXA for the assessment and identification of LSTM imbalances in both clinical and sport-specific contexts.