BackgroundAn accurate estimation of athletes’ energy needs is crucial in diet planning to improve sport performance and to maintain an appropriate body composition. This study aimed to develop and validate in elite athletes new equations for estimating resting energy expenditure (REE) based on anthropometric parameters as well as bioimpedance analysis-(BIA) derived raw variables and to validate the accuracy of selected predictive equations.MethodsAdult elite athletes aged 18-40 years were studied. Anthropometry, indirect calorimetry and BIA were performed in all subjects. The new predictive equations were generated using different regression models. Prediction accuracy of the new equations was assessed and compared with the one of five equations for estimating REE in normal-weight subjects and three athletes-specific predictive formulas as suggested in the literature.ResultsOne-hundred and twenty-six male athletes from different sport specialties were randomly assigned to the calibration (n=75) or validation group (n=51). REE was directly correlated with individual characteristics, except for age, and raw BIA variables. Most of the equations from the literature were reasonably accurate at the population level (bias ±5%). The new equations showed a mean bias -0.3% (Equation A based on anthropometric parameters) and -0.6% (Equation B based on (BIA) derived raw variables). Individual accuracy was ~75% in six out of eight of the selected equations and was even higher for Equation A (82.4%) and equation B (92.2%).ConclusionIn elite athletes, BIA-derived phase angle is a significant predictor of REE. The new equations have a very good prediction accuracy at both group and individual levels. The use of PhA as predictor of REE requires further research with respect to different sport specialties, training programs and training level.