ABSTRACT
Total daily energy expenditure (TDEE) can be assessed in children using several methods: the double-labelled water technique (DLW); indirect whole-body calorimetry; activity diary; accelerometry and heart rate (HR) monitoring. This last, low-cost and convenient technique has been validated against the DLW technique that is normally considered the reference method. The main advantages of HR monitoring are its social acceptability, and the examination of the time spent in specific activities or intensive exercise sessions. The aim of this study was to assess a new method for computing energy expenditure (EE) with the HR monitoring technique in children, using an open-circuit ventilated hood indirect calorimetry system. Eleven healthy children participated in this study, with a mean age of age 8.9+/-3 years. Seven of them were studied again 6 months later, so that 17 measurements were available for analysis. The calibration period was used to collect simultaneous data from HR and EE measured by indirect calorimetry (IC). Measurements were made when the children were resting (REE=Resting Energy Expenditure), during 30 min of a post-prandial period, and during two different activity levels on a cyclergometer (two periods of 15 min at 20 and 60 W output). Results of EE during a longer calibration period (2.5 h) were compared with those of a shorter period (1.5 h), for a group of six subjects. Results of nightly EE measurements of five subjects were recorded using IC and compared with the HR predicted method. Results of EE measurements estimated with different regression equations were compared to EE measured by IC. The short calibration period (1.5 h) gave similar results to a longer period (2.5 h), with a mean difference of less than 3%. The polynomial third-order relationship between HR and EE gave the highest correlation coefficient, the smallest mean square error and the best agreement (Bland-Altman method). EE predicted with this model gave the best results during the total calibration period, the post-prandial period and the medium and high activity periods. Comparison between measured and predicted nightly EE showed that REE - REE/10 was the best formula to assess nightly EE, and the mean difference was less than 4%. This study demonstrates the validity of a shortening of the calibration period and the reliability of the REE - REE/10 formula to compute nightly EE. The best model for computing daily EE is a polynomial third-order regression equation.
