Accuracy of the Resting Metabolic Rate Equations

Abstract

Background The resting metabolic rate (RMR) is recommended to be measured via indirect calorimetry. Since an indirect calorimeter may not be available in all settings, predictive equations are used worldwide. Given that the accuracy of the predictive equations may depend on ethnicity, we aimed to evaluate the most accurate prediction equation in Turkish adults. Methods We included 57 voluntary participants (68.4% women) who were normal in body weight and overweight or obese (63.2%). Harris-Benedict; Owen; WHO/FAO/UNU weight; WHO/FAO/UNU weight and height equations; and Mifflin-St. Jeor equations were used to compare the predicted RMRs and the indirect calorimeter measurements. The root mean square error, mean percentage difference, and concordance correlation coefficient were calculated to indicate the predictivity of the equations and to show the deviation of the differences between the predicted and measured RMRs. Results Although it was not within the agreement of ± 10% of the measured RMR, which is acceptable for accuracy, the Harris-Benedict equation provided the highest accuracy (61.9%), whereas the Owen equation had the lowest accuracy (19%) for normal body weight. The WHO/FAO/UNU weight equation had the highest accuracy (58.3%), whereas the Owen and WHO/FAO/UNU weight and height equations had 25% accuracy for overweight and obese individuals. Conclusions Although none of the five widely used equations showed good accuracy in terms of our RMR measurements via indirect calorimetry, the Harris-Benedict equation was recommended for normal body weight, and the WHO/FAO/UNU weight equation was recommended for overweight and obese individuals until a new, population-based formula was developed.