BACKGROUND: In routine practice, assessment of the nutritional status of critically ill patients still relies on traditional methods such as anthropometric measurements, biochemical markers, and predictive equations. OBJECTIVE: To compare resting energy expenditure measured by indirect calorimetry (REEIC) with REE calculated by using the Harris-Benedict equation with 3 different sources of body weight (from bed scale, REEHB1; ideal body weight, REEHB2; and predicted body weight, REEHB3). METHODS: This study included 205 critically ill patients (115 men, 90 women) evaluated within the first 48 hours of admission and undergoing mechanical ventilation. REE was measured by indirect calorimetry for 30 minutes and calculated by using the Harris-Benedict equation with the 3 sources of body weight. Data were compared by the Bland-Altman method. RESULTS: The values based on ideal and predicted body weight (REEHB2 and REEHB3) did not agree with REEIC. Bland-Altman analysis showed that the limits of agreement varied from +796.1 kcal/d to -559.6 kcal/d for REEHB2 and from +809.2 kcal/d to -564.7 kcal/d for REEHB3. REEIC and REEHB1 (body weight determined by bed scale) agreed the best; the bias was -18.8 kcal/d. However, REEHB1 still overestimated REEIC by +555.3 kcal/d and underestimated it by -593.0 kcal/d. CONCLUSION: For measuring REE in critically ill patients undergoing mechanical ventilation, calculation via the Harris-Benedict equation, regardless of the source of body weight, cannot be substituted for indirect calorimetry.