Validation of resting metabolic rate prediction equations for teenagers

Resumen

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 p<0.05, the average of constant error (with a less than 5% discrepancy) and the graphic dispersion analysis proposed by Bland and Altman (1986). The majority of RMR´s prediction equations of the in girls did not present signifi cant difference (p>0.05) 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<0.05) in relation to the standard measure. The equations proposed by Schofi eld (1985), Tverskaya et al. (1998) and the model of Müller et al. (2004) had standard error variations above 5%. It is concluded that, in such a way, in both the sexes, none of the prediction equations possess validity in the determination of RMR in adolescents with the characteristics of the studied groups.