Resting Metabolic Rate Prediction Equation Accuracy in Structural Firefighters

It remains unclear whether predictive resting metabolic rate (RMR) equations accurately predict RMR in firefighters. The purpose of this study was to examine the accuracy of six RMR prediction equations (Cunningham, De Lorenzo, Harris-Benedict, Mifflin-St Jeor, Nelson, and Jagim) in firefighters. Male firefighters (n=26; [mean ± SD] age: 38.2 ± 7.6 y; height: 180.9 ± 6.8 cm; body mass: 92.0 ± 15.6 kg; BMI: 28.1 ± 4.4 kg· m^-2) participated in annual fitness and health evaluations including RMR determination and body composition assessment. A repeated measures ANOVA with Bonferroni post hoc analyses was selected to determine mean differences between measured and predicted RMR. Linear regression analysis was used to assess the accuracy of each RMR prediction method (p<0.05) and to determine standard error of the estimate (SEE). All prediction equations significantly underestimated RMR (all, p<0.001), except the Jagim equation, which significantly overestimated RMR (p<0.001). Equations with the closest agreement to measured RMR were the Harris-Benedict (R² = 0.696, p = 0.004, root mean square prediction error (RMSE) = 314 kcals, %RMSE = 14.2%) and the DeLorenzo (R² = 0.675, p<0.001, RMSE = 242 kcals, %RMSE = 10.9%). The Nelson equation yielded the highest RMSE (412 kcals, %RMSE = 18.6%). The variance in equations ranged from an SEE = 173 kcal· d^-1 (Harris-Benedict) to an SEE = 215 kcal· d^-1 (Cunningham), accounting for 70% and 53% of the variance in RMR, respectively. RMR prediction equations underestimate the energy requirements of firefighters; thus, caution should be exercised when interpreting values.