[Energy expenditure of healthy Brazilian adults: a comparison of methods]

There are many available methods to measure the human energy consumption. Due to the high cost and small number of laboratories able to perform it, the direct calorimetry and doubly labeled water technique are rarely used to measure the energy expenditure. As consequence, there is an increase use of the indirect calorimetry and bioelectrical impedance analysis methods by researchers and predictive equations by physicians to calculate the total energy requirements in research and in clinic.

To evaluate the basal energy expenditure (GEB) and total energy expenditure (GET) of healthy subjects comparing results of indirect calorimetry (CI), bioelectrical impedance analysis (BIA) and predictive equations such as estimated energy requirements (EER), Fleish, FAO/WHO/UNU 1985 and 2001.

It was evaluated the energy expenditure in 81 subjects of both gender by CI and BIA. Of these, the GEB by predictive equations FAO/WHO/UNU 1985 and 2001 and the GET by EER, Fleisch and FAO/WHO/UNU 1985 and 2001 were calculated. The data was analyzed with the software Sigma Stat for Windows version 2.03 and Dunn's test was used for comparison of mean obtained with methods and equations (p < 0.05).

It was verified that 60.5% of the participants were men. The majority of subjects were eutrofic. BIA showed significant differences between men and women. The differences between the values of GEB, obtained by BIA and CI, were statistically significant only for women (p < 0.05) with an underestimation of the GEB results obtained by BIA in relation to those obtained by CI. Significant difference was found between the predictive equations and CI for both sexes, also demonstrating an underestimation of the equations in relation to CI. It was observed an overestimation of GET with the predictive equations of FAO/WHO/UNU/85 and 2001 in relation to EER and Fleisch for both sexes.

The estimation of GEB can be carried out by both BIA and CI, although BIA must be used with criteria since it underestimates in a significant way for women. In relation to the predictive equations, overestimation can develop positive energy balance in populations with risks for overweight and obesity. Thus, it is necessary comparisons to identify the limitations of each method.