Equações antropométricas e impedância bioelétrica: validade para estimar a composição corporal de pacientes com HIV/AIDS

Abstract

Considering that there is a large number of people living with HIV/AIDS in Brazil and being a disease that has not been cured so far, such a situation has triggered the need to deepen the issues that cover topics involving this audience. Since issues related to body composition are prominent in this audience, it becomes relevant to investigate the accuracy of methods for assessing body composition by testing their applicability. Thus, the objective of this study was to analyze the cross-validity of anthropometric and bioelectrical impedance (BI) equations for the estimation of relative body fat (% BF) and muscle mass (MM) in a sample of HIV/ IDS patients from city of Santa Maria - RS, using hydrostatic weighing as the criterion method (HW). 58 individuals with HIV / AIDS (48 men and 10 women) participated in this study, aged between 15 and 66 years. It is noteworthy that due to the reduced number of female subjects with results in some evaluation methods, these will be partially presented and discussed in this study, the exception will be in the characterization of the sample and in the analysis of MM estimate. To determine body composition, the subjects were evaluated using HW, BI and anthropometry, and the cross-validity of 40 anthropometric and BI equations was analyzed. For the equations that estimate body density (Bd) and lean body mass (LBM), the %BF was deducted. The analyzes were performed using the SPSS statistical package, version 23.0, adopting a significance level of 5%. The cross-validation criteria suggested by Lohman (1992), the graphic analysis of the dispersions in relation to the mean, using the method of Bland and Altman (1986) and the diagnostic analysis (kappa) were used. The mean %BF was 20.72% in men and 26.19% in women. The differences between the estimates of the conversion equations for Bd and LBM in %BF were not significant, except for the equations (PM13 to PM16) by Petroski (1995) and the 3 male equations by Carvalho-Pires Neto (1998a). Regarding agreement, Petroski's equations showed significant results, which ranged from r = 0.73 to r = 0.81. On the other hand, the Carvalho - Pires Neto, BI and the Wilmore & Behnke (WeB) (1969) equation presented agreement between low and moderate. Regarding the standard error of estimate, all equations did not present acceptable values for cross-validation. The observed kappa coefficient of agreement was low, with few coincident diagnoses. For the MM estimate, the equation by Janssen et al. (2000) did not differ statistically from the criterion measure, which did not occur with the equation of Lee et al. (2000). The estimated% MM results showed a moderate correlation coefficient. Considering the categories for interpreting the% MM levels, most of the individuals evaluated presented results between high and very high for Lee's equation. As for the Janssen equation, the predominance was of subjects in the range considered normal. As for the kappa coefficient, the values found indicate a small (Eq. Janssen) and discrete (Eq. Lee) agreement. Based on the results obtained, it was possible to conclude that the equations explored in this study, in general, were not efficient for the estimation of body components of patients living with HIV/ AIDS.