Modelos matemáticos para estimativa da gordura corporal de adolescentes utilizando dobras cutâneas, a partir da absorciometria de raios-X de dupla energia

Abstract

Introduction: Studies have found a transition from obesity of the adult population to children and adolescents, which in turn, can lead to clinical manifestations, such as: coronary diseases, type 2 diabetes, and psychosocial complications increasingly early. However, methods for evaluating nutritional status for this age group, mainly involving low cost techniques such as skinfold thickness measurements (ST), are imprecise in Brazilian studies. Factor which can lead to a mistaken interpretation of the body composition of the evaluated ones. Objective: To develop new mathematical models, based on DC measurements, based on dual energy X-rays absorptiometry (DXA), to estimate fat mass (G) in adolescents. Methods: This was an exploratory descriptive study in which 416 male adolescents aged 12 to 17 years were evaluated, 42 of whom were separated to compose the study validation sample. Measurements of total body mass, stature, waist and hip circumference were obtained, nine anatomical points based on ST: biceps, triceps, subscapular, pectoral, mid axillary, abdominal, suprailiac, thigh and calf muscles, as well as G and bone mineral density (BMD) measured with DXA technology. For the development of the equations, a multiple linear regression model was used by the ordinary least square (OLS) method. Results: The group had a mean body mass index (BMI) of 21.25± 4.12 kg / m² and %G = 20.57 ± 5.80%. From %G, the prevalence of excess fat was verified in 38.3% of adolescents. The impact of fat on adolescent BMD indicated an association in the order of r = -0.358; P <0.005, with BMD reduction up to 14% for the spine region in adolescents with obesity compared to eutrophic. The development of new mathematical models that meet criteria of high coefficient of determination (R²), low standard error of estimation (SEE), control of colinearity, residue normalities, homoscedasticity and practicality, allowed the presentation of three options with R² = 0.932 and SEE 1.79; R² = 0.912 and SEE = 1.78; R² = 0.850 and SEE = 1.87, respectively. In all the options, the variables age and height were employed, as well as triceps and subscapular ST. Conclusion: The results obtained evidenced the possibility of developing new mathematical models for the evaluation of body fat in adolescents with results superior to the existing models in the literature.