<b>Validation of resting metabolic rate prediction equations for teenagers</b>

The resting metabolic rate (RMR) can be defi ned as the minimum rate of energy spent and represents the main component of the energetic outlay. The purpose of this study is to validate equations to predict the resting metabolic rate in teenagers (103 individuals, being 51 girls and 52 boys, with age between 10 and 17 years) from Florianópolis – SC – Brazil. It was measured: the body weight, body height, skinfolds and obtained the lean and body fat mass through bioimpedance. The nonproteic RMR was measured by Weir’s equation (1949), utilizing AeroSport TEEM-100 gas analyzer. The studied equations were: Harry and Benedict (1919), Schofi eld (1985), WHO/FAO/UNU (1985), Henry and Rees (1991), Molnár et al. (1998), Tverskaya et al. (1998) and Müller et al. (2004). In order to study the cross-validation of the RMR prediction equations and its standard measure (Weir 1949), the following statistics procedure were calculated: Pearson’s correlation (r ≥ 0.70), the “t” test with the signifi cance level of p0.05) in relation to the standard measure, with exception of the equations suggested for Tverskaya et al. (1998), and the two models of Müller et al (2004). Even though there was not a signifi cant difference, only the models considered for Henry and Rees (1991), and Molnár et al. (1995) had gotten constant error variation under 5%. All the equations analyzed in the study in girls had not reached criterion of correlation values of 0.70 with the indirect calorimetry. Analyzing the prediction equations of RMR in boys, all of them had moderate correlation coeffi cients with the indirect calorimetry, however below 0.70. Only the equation developed for Tverskaya et al. (1998) presented differences (p ABSTRACT0,05) em relação à medida padrão (Weir 1949), com exceção das equações sugeridas por Tverskaya et al. (1998) e os dois modelos de Müller et al (2004). Mesmo não havendo diferença signifi cativa, somente os modelos propostos por Henry e Rees (1991) e Molnár et al. (1995) obtiveram variação de erro constante inferior a 5%. Todas as equações analisadas no estudo em meninas não alcançaram valor critério de correlação de 0,70 com a calorimetria indireta. Analisando as equações de predição da TMR em meninos, todas tiveram correlação moderada com a calorimetria indireta, porém ficando abaixo do coefi ciente 0,70. Somente a equação desenvolvida por Tverskaya et al. (1998) teve diferença (p