Development and Validation of Ultrasound-based Prediction Equations for Fat and Lean Mass in Adults With Obesity

This is the authors' abstract. We don't add key points for this paper.

Background & aims Obesity treatment is measured by total weight loss or BMI loss, neither of which accurately reflects fat mass. Ultrasound (US) offers a non-invasive, portable alternative for measuring subcutaneous adipose tissue (SAT) thickness. The aim of this study was to (1) assess the reliability of US-derived SAT measurements, and (2) develop and validate prediction equations for fat mass (FM) and lean mass (LM) based on SAT thickness at multiple sites. Methods Two separate observational studies were performed. To investigate reliability, the intra- and interrater reliability of the SAT measurements was assessed. For the development and validation of prediction equations, linear regression was used to generate a seven- and three-point model for the estimation of both lean and fat mass. Validity was assessed in the validation cohort using Bland-Altman plots, intraclass correlation coefficients (ICC), and linear regression to detect proportional bias. Results US measurements showed high intra- and good interrater reliability (resp. ICC=0.948; ICC=0.859). All models demonstrated a high predictive accuracy for DXA-based body composition (R 2 >0.71 for fat mass assessment, R 2 >0.90 for lean mass assessment). There was a strong agreement with DXA across all models (ICC > 0.846). Conclusion Ultrasonography-based measurement of SAT provides a cheap, easy-to-use, and reliable tool to assess fat and lean mass in a population with class II/III obesity.