[Estimate equation of body composition of the elderly male].

OBJECTIVES This study aims to develop and validate a predictive body composition equation in elderly men. METHODOLOGY Correlational study using a sample of 85 men with 69.5±5,8 years from Vitoria, Brazil. The sample was divided in two groups: the regression group (n=54) was used to develop the predictive equation and a validation group (n=31). The following measurements: skinfolds (S), bone diameter (BD) and body circumferences (BC), body fat percent (%BF) measure by DXA. The predictive equations were developed using multiple linear regressions and validated through the stepwise method. Equations were compared using Student's T test for paired samples. Residual scores were analyzed using the Blant and Altman method. RESULTS The regression group had 68.5±4,5 years, 78,5±14,4 kg; 1,6±0,6 m, and % BF 30.4±6,8. The validation group had 71.2±7.4 years, body weight 75.2±11 kg, height 1.6±0.5 m, and % BF 29.4±7.7. The resulting predictive equation is % body fat =23.371+0.337 (BC abdomen)-1.464 (BC left arm)+0.501 (S calf)+0.175 (body weight)-0.741 (BD acromial)-0.415 (BC biiliocristal)+0.374 (Cleft tight) + 0.370 (BC neck). The best-fit plot was significant due to the correlation coefficient r=0.8. The r2 was 0.8, meaning that the independent variables are able to explain 87 % of the estimate. The EPE was 2.6 %. CONCLUSION In light of the current results, the equation was considered adequate to estimate body composition in elderly men. The equation can be a simple and accurate alternative to estimate body fat in this population.