Determinação e concordância dos componentes da composição corporal obtidos a partir de diferentes métodos avaliativos

Abstract

Body composition has been shown to be an important indicator of health, since many chronic diseases are related to excess adiposity. In this context the appropriate assessment is required in order to obtain the correct diagnosis in terms of nutritional status. Thus, the aim of this study was to analyze the relationship between the components of body composition of college students in the city of Santa Maria, RS obtained from hydrostatic weighing (HW), bioelectrical impedance (BI) and anthropometry and agreement between shown diagnostics. The study included 164 college students (104 men and 60 women) aged between 18 and 30 years. For the determination of body composition, subjects were evaluated by means of HW, BI and anthropometry (weight, height, skinfolds thickness, body circumferences, and bone breadths), and analyzed 48 equations (32 generalized and 16 specific) for review the percentage of body fat (%BF), 7 equations of lean body mass (LBM), two of muscle mass (MM) and two bone mass (BM). Descriptive statistics, the Kolmogorov - Smirnov test, the t test for dependent samples, the Pearson correlation coefficient , the standard error of estimate, the total error and constant error of the scatter diagram was used as well, Bland - Altman (1986) and the kappa coefficient to verify the concordance of the results. The level of significance was set at 5%. Data analysis was performed with SPSS, version 14.0. The average %BF was found to be 16.3% in men and 26.3% women. As for the agreement, the generalized equations demonstrated greater power of the relative body fat, 16 equations (9 men and 7 for women) Petroski (1995) showed no significant difference data obtained using the method, as well as an equation Guedes and Guedes (1991) for women and BI for men. The kappa coefficient of concordance observed was low (≤0.44) having few overlapping diagnoses. MCM obtained by different equations differed from results obtained with the criterion method, as compared to BI (body composition analyzer), only the results obtained for men did not differ from the criterion method. The equations of MM Janssen et al. (2000) and Lee et al. (2000) showed a correlation coefficient of r = 0.72 for men and r = 0.76 for women, and no statistical difference was observed. BM already obtained by means of equations Von Döbeln modified by Rocha (1975) and Martin (1991) showed statistically different results. Based on the results we conclude that: a) for the determination of %BF equation Petroski (1995) that uses variables as the sum of nine skinfolds thickness and age, for men, and employing the sum of three skinfold thickness and age, for women were those that showed the best results in agreement with the reference method, however, the diagnostic showed up just the same; b) LBM provided by body composition analyzer proved acceptable for men, whereas for females the results were not satisfactory; c) the MM results presented good agreement between the equations for both genders; and d) the equations of BM demonstrate different results, both when analyzing the female group, as the male group.