Comparação entre os percentuais de gordura obtidos por meio da avaliação corporal por dobras cutâneas, bioimpedância, ultrassom portátil e DXA em indivíduos eutróficos, com sobrepeso e com obesidade submetidos ou não à cirurgia bariátrica

Abstract

Introduction: Body composition includes water, fat, proteins, and minerals, with fat being the most variable component. Although Body Mass Index (BMI) is widely used to assess nutritional status, it does not accurately reflect fat distribution, requiring more precise methods. Objective: To compare body fat percentage (%BF) obtained through Dual-Energy X-ray Absorptiometry (DXA), Bioelectrical Impedance Analysis (BIA), Portable Ultrasound (US), and Skinfold Thickness (SF). Methods: A total of 251 adults were evaluated (81.67% women, median age 47), including individuals with normal weight, overweight, and obesity, both with and without bariatric surgery. Participants were recruited from the Nutrition and Obesity Clinic at HC-UFMG and through social media, with a minimum age of 18 years. All underwent BIA and US assessments; SF was applied to 87 individuals with normal weight or overweight; DXA was performed on a subgroup of 24 participants. Jackson and Pollock equations (3, 4, and 7 skinfolds)and Siri’s equation were used to calculate %BF. Statistical analysis included Intraclass Correlation Coefficient (ICC) to assess agreement between methods, and Pearson’s correlationto evaluate linear relationships. A mixed-effects linear regression was used to compare bariatric and non-bariatric groups. Results: Mean %BF values were: BIA 36.26%, US 36.50%, SF (7 folds) 23.39%, and DXA 37.77%. Overall agreement between BIA, US, and SF was good to excellent (ICC 0.795–0.921), higher in men (ICC 0.752–0.901) than in women (ICC 0.546–0.783). In individuals with obesity, agreement between BIA and US was low (ICC 0.306–0.469), but good to excellent in normal weight and overweight groups. Compared to DXA, BIA and US showed moderate to good agreement (BIA: ICC 0.548–0.708; US: ICC 0.635–0.776), while SF showed low agreement (ICC 0.302–0.464). No significant difference in %BF was found between bariatric and non-bariatric patients after adjusting for BMI, sex, and age. Conclusion: In conclusion, BIA and US showed satisfactory concordance, particularly in normal-weight and overweight individuals, representing feasible alternatives to DXA. However, US technical limitations in obese subjects and methodological differences from the reference standard require careful interpretation.