COMPARAÇÃO ENTRE UM MODELO DE PREDIÇÃO BASEADO NA TOPOGRAFIA DE SUPERFÍCIE 3D E A BIOIMPEDÂNCIA PARA O CÁLCULO DA COMPOSIÇÃO CORPORAL

Abstract

Introduction: Body composition (BC) analysis is a fundamental part of the assessment carried out by nutritionists and other health professionals. Alterations in body composition have been associated with various diseases, such as heart disease and diabetes, and allow for the diagnosis and monitoring of conditions such as visceral obesity and sarcopenia. Among the methods for assessing WC, we can mention DEXA as the gold standard and Bioimpedance (BIA) as the most widely used clinically. However, both have limitations, such as the cost and difficulty of access to DEXA and the high sensitivity of BIA in relation to changes in body hydration, as well as the variety of devices used. Assessments of WC using three-dimensional (3D) scanning are promising, but there are still no specific studies on the Brazilian population. Objective: To compare a body composition prediction model based on 3D scanning with the results obtained by BIA in healthy adults living in the city of Campo Grande / MS. Methodology: 101 individuals aged between 19 and 57, of both sexes, were assessed using tape-measure anthropometry, 3D scanning, BIA and an evaluation form. The scan data was used to build a regression model to predict WC from BIA (3D Regression) and to feed a regression formula obtained from a North American sample (3D Ng Regression). Statistical analysis: The sample selection method for the linear regression was Stepwise Backward. Student's t-test for independent samples was used to compare circumferences measured by tape measure and scanner, and to compare WC variables obtained by scanner (3D regression), BIA and 3D regression (Ng). A significance level of 5% was adopted. Results: 3D scanning was able to correctly predict fat percentage, fat mass and fat-free mass for men and women when compared to BIA. The 3D Regression (BK Ng et al, 2019), constructed from a foreign sample, was not able to predict the percentage of fat and fat mass for men and percentage of fat for women, in the sample of this study. The 3D circumference measurement was statistically the same as the thigh tape, both for the general population and stratified by BMI ˃25 (n=62) and BMI ≤ 25 (n=39). Waist measurement was statistically equal only for BMI ≤ 25 (n=39). Conclusion: Regression based on 3D scanning was able to correctly predict all WC variables when compared to BIA. The differences found between the regression used in the North American population and the BIA results in our sample reinforce the need to use studies that reflect the characteristics of the Brazilian population and raise awareness of the use of studies with other populations.