Background: Methods to calculate energy expenditure (EE) based on CO2 measurements (EEVCO2) have been proposed as a surrogate to indirect calorimetry. This study aimed at evaluating whether EEVCO2 could be considered as an alternative to EE measured by indirect calorimetry. Methods: Indirect calorimetry measurements conducted for clinical purposes on 278 mechanically ventilated ICU patients were retrospectively analyzed. EEVCO2 was calculated by a converted Weir's equation using CO2 consumption (VCO2) measured by indirect calorimetry and assumed respiratory quotients (RQ): 0.85 (EEVCO2_0.85) and food quotient (FQ; EEVCO2_FQ). Mean calculated EEVCO2 and measured EE were compared by paired t test. Accuracy of EEVCO2 was evaluated according to the clinically relevant standard of 5% accuracy rate to the measured EE, and the more general standard of 10% accuracy rate. The effects of the timing of measurement (before or after the 7th ICU day) and energy provision rates (<90 or ≥90% of EE) on 5% accuracy rates were also analyzed (chi-square tests). Results: Mean biases for EEVCO2_0.85 and EEVCO2_FQ were -21 and -48 kcal/d (p = 0.04 and 0.00, respectively), and 10% accuracy rates were 77.7 and 77.3%, respectively. However, 5% accuracy rates were 46.0 and 46.4%, respectively. Accuracy rates were not affected by the timing of the measurement, or the energy provision rates at the time of measurements. Conclusions: Calculated EE based on CO2 measurement was not sufficiently accurate to consider the results as an alternative to measured EE by indirect calorimetry. Therefore, EE measured by indirect calorimetry remains as the gold standard to guide nutrition therapy.