Development and validation of a new prediction equation for energy expenditure in Mexican adults with overweight and obesity

Background Accurate predictive equations of resting energy expenditure (REE) are crucial in devising nutritional strategies to fight overweight/obesity, especially in countries where these are highly prevalent. Objective To develop and validate a new equation for predicting REE in Mexican adults with overweight and obesity. A second objective was to compare the impact of the use of urea nitrogen (UN) to measure REE. Methods We conducted a cross-sectional study, including 410 men and women with overweight and obesity (age range: 20-60 yrs.). Participants were randomly assigned to the development or validation groups; the development group included 200 subjects (Body Mass Index (BMI) = 31.6 ± 5.4 kg/m2, mean ± SD) and the validation group 210 subjects (BMI = 31.7 ± 4.78 kg/m2). The new predictive equation was derived by stepwise multiple linear regression analysis in the development group and then, validated in the other group. The accuracy of the new equation was compared to several predictive existing equations. REE was measured by indirect calorimetry (IC) and body composition by bioimpedance analysis. For the second objective, 30 subjects (10 normal weight, 10 with overweight and 10 with obesity) were involved to measure UN in a 24-hour urine collection. The value of UN was used in the Weir equation to obtain REE and it was compared with the REE obtained by the simplified Weir equation. Results Two predictive equations were developed in this study. The first equation (EQ1) included LBM, weight, sex and age variables in the model (R2= 0.518, bias= 0.014 kJ/day), since lean body mass (LBM) was the best single variable to predict REE (R2= 0.440). When LBM is not an available parameter, a second equation was developed (EQ2) based on only weight, sex and age (R2= 0.513, bias= -0.025 kJ/day). Compared with others predictive equations, the new developed equations showed the lowest bias (Kj/day): EQ1= 9.92, EQ2= 25.7, Harris-Benedict= -67.4, Valencia= 130, WHO/FAO/UNU= 270, Mifflin-St Jeor= 309, Owen= -808, Carrasco= -1097. When we compared REE by Weir equations, using or not UN, there was not statistical significant differences. Conclusion The present equations had the highest predictive accuracy in subjects with overweight or obesity compared with the other equations previously derived from different populations. Also when body composition is not available, the alternative equation demonstrated a comparably predictive accuracy for its use in nutritional treatment in subjects with overweight/obesity.