Measured Versus Predicted Total Energy Expenditure: Evaluating Methods To Approach Daily Energy Requirements Across Physical Activity Levels.

In healthy, weight-stable adults, total energy requirements are equivalent to total energy expenditure (TEE). Variations in physical activity energy expenditure (PAEE) can make accurate prediction of TEE challenging. The present study aimed to determine the validity of developed approaches to predict TEE in individuals varying in PA levels. PURPOSE: To compare the validity of different models developed to predict TEE on a sample of adults varying in PA levels. METHODS: Forty-seven healthy, weight-stable adults (20 female and 27 male; 20-55 yrs) with PA levels ranging from sedentary to ultra-endurance training were recruited. Resting metabolic rate (RMR) and body composition were measured by indirect calorimetry and DXA, respectively. Accelerometry was used to quantify physical activity (ACC PAEE). Four days of food intake records (FIR) and three 24-hour dietary recalls (24hDR) were used to estimate energy intake. TEE was measured via doubly labeled water (DLW). We applied fourteen different methods to predict TEE: daily energy intake by FIR and by 24hDR; ACC PAEE added to RMR with a standard thermic effect of food (TEF); ACC PAEE added to RMR and the TEF from the energy intake reports; four equations published by Plucker et al. (2018); three equations published by Pontzer et al. (2021); and three equations published by Vinken et al. (1999). One-way repeated-measures ANOVA with Bonferroni post-hoc tests were used to compare means. Bland-Altman plots, mean difference (%), root mean square error (RMSE), and accuracy percentage were used to evaluate the agreement between measured and predicted TEE. RESULTS: All the applied methods underestimated measured TEE, although the Plucker equation based on age, body mass, height, and RMR performed best. There was no significant difference between measured and predicted TEE with mean bias of 218.64 ± 583.36 kcal, mean difference% of 5.82, and RMSE% of 20.56. Large limits of agreement, high RMSE values, and lower percentages of accuracy reflect a sizeable amount of error at the individual level. CONCLUSIONS: The Plucker et al. (2018) equation may have the potential to accurately estimate TEE in a population varying in PA, although caution must be used when individual analyses are performed. Future studies are needed to develop accurate methods for the prediction of TEE. OPN: VT Presidential Postdoctoral Fellowship; KRH: VT Translational Obesity Research Interdisciplinary Graduate Education Predoctoral Fellowship; GZR: NextGeneration EU funds Margarita Salas Postdoctoral Fellowship.