• Thiago G Barbosa-Silva, Maria Cristina Gonzalez, Renata M Bielemann, Leonardo P Santos, Caroline Dos S Costa, Menezes Ana Maria B AMB Post-Graduate Program in Epidemiology, Federal University of Pelotas, Pelotas, RS, Brazil., and COCONUT Study Group, Brazil.
• Post-Graduate Program in Epidemiology, Federal University of Pelotas, Pelotas, RS, Brazil. Electronic address: tgbsilva@hotmail.com.
• Nutrition. 2021 Mar 1; 83: 111056.

ObjectiveThe aim of this study was to develop new appendicular lean mass (ALM) prediction models based on ultrasound and anthropometric measurements.MethodsThis was a cross-sectional assessment of a subsample from a population-based study (COMO VAI?), conducted with community-dwelling individuals ≥60 y of age living in a southern Brazilian city. ALM was measured by dual-energy x-ray absorptiometry (DXA). Muscle thickness (MT) from supine participants was assessed by ultrasound on the anterior aspect of both upper and lower limbs. Such measures, along with anthropometric data, were used to develop prediction models (multivariable linear regression) through the backward stepwise method.ResultsThe study included 190 participants composed mainly of women, white, and middle-class individuals. The best ALM predictive performance was achieved by a model based on two "lengths" (height and arm length), two circumferences (dominant arm and thigh), and two ultrasound-measured MTs (dominant arm and thigh): R2 = 0.90, limits of agreement: ±2.36 kg. Closely satisfactory results were also achieved by an "abbreviated" model composed by the two aforementioned "lengths" and MTs (R2 = 0.89, limits of agreement: ±2.51 kg). ALM estimates from both equations were unbiased and similar to DXA measurements (P = 0.13 and 0.09, respectively). Bootstrap analysis favorably suggested the validity of the models.ConclusionsBased on two ultrasound assessments and a few anthropometric measurements, the developed equations produced accurate and unbiased ALM estimates in the sample. Hence: 2 MTs + 2 lengths (+ 2 circumferences) = 4 limbs' muscle mass. Such models might represent promising alternatives for muscle assessment among older individuals.Copyright © 2020 Elsevier Inc. All rights reserved.

16 keywords...

hide…