Indirect calorimetry is the gold standard to measure energy expenditure. In fact it is not always available and inconstantly feasible. Various equations for predicting energy expenditure based on body weights have been created. This study aims at determining the best suitable predictive strategy unless indirect calorimetry is available.

Several clinical studies have shown that energy deficit as well as overfeeding lead to an increased risk of complications, especially infections, and increased mortality. The gold standard for determining energy expenditure in intensive care patients is indirect calorimetry. This method is expensive and requires a trained team for its use. In addition the reliability of the measurements depends on the clinical situation and may not be feasible (e.g. inspiratory O2 fraction >60%, end expiratory pressure > 9cmH2O, presence of bronchial gap, etc.). Moreover, there is no longer on the market valid indirect calorimeter for clinical use in mechanically ventilated patients, and the maintenance of the old valid one (Deltatrac II®) becomes increasingly problematic because the lack of spare parts. Intensivists have no more choice and use prediction equations for energy expenditure which are based on imprecise anthropometric data (height, weight). Especially for obese or malnourished patients in the intensive care the body weight represents not a reliable data. Similarly, secondary water inflation due to metabolic stress and resuscitation complicates the determination of the real body weight. The difficulty is to know which weight to use in predictive equations. Due to the paucity in literature on this subject, there is currently no consensus on the reference weight to use in the determination of nutritional needs and medication doses. So each prescriber calculates the energy target by taking a reference weight based on his own convictions. This study is part of a quality process of care and practices harmonization, aiming to identify the reference weight to be used and the best suited predictive equation, to predict energy expenditure for patients who cannot benefit from an indirect calorimetry.

Inclusion Criteria:

• Patients admitted to the Geneva ICU
• Length of stay > 72 hours
• Mechanical ventilation with: FiO2 < 60%; positive end expiratory pressure < 9 cmH2O; no air leaks, absence of pulmonary multi-resistant bacteria
• Without mechanical ventilation: no claustrophobia; no oxygen dependence

Exclusion Criteria:

• All patients without inclusion criteria

ICU ventilated patients of a tertiary mixed ICU population