The deleterious effects of alveolar hyperoxia and (severe) hyperoxaemia are well described. Achieving safe and efficient oxygenation may be challenging. The hypothesis of the present study is that closed-loop oxygen control in intubated and mechanically ventilated critically ill patients improves oxygen administration compared with standard manual oxygen titrations. A single-blind, randomised crossover clinical trial assessing the efficacy and safety of the use of a closed-loop oxygen control versus manual oxygen titration in patients receiving mechanical ventilation has been designed. Patients will be randomised to receive first either closed-loop oxygen control (CLOC) or manual oxygen titration (MOT). The percentage of time spent in optimal and sub-optimal SpO2 (oxygen saturation by pulse oximetry) ranges will be calculated for each period.

The deleterious effects of alveolar hyperoxia and (severe) hyperoxaemia are well described. A safe and efficient oxygenation should be achieved. However, from a practical point of view, this may require frequent manual adjustments of the inspired oxygen, which might be unfeasible, especially in high-demand periods. Several automatic oxygen systems have been tested in non-intubated patients. The hypothesis of the present study is that closed-loop oxygen control in intubated and mechanically ventilated critically ill patients improves oxygen administration compared with standard manual oxygen titrations. Thus, the objective of the study is to compare the percentage of time spent in a predefined optimal SpO2 range between closed-loop oxygen control and manual oxygen titration.

A single-blind, randomised crossover clinical trial assessing the efficacy and safety of the use of a closed-loop oxygen control versus manual oxygen titration in patients receiving mechanical ventilation has been designed. Once included, patients will be randomised to receive first either closed-loop oxygen control (CLOC) or manual oxygen titration (MOT). Every period will last for 4h. Clinical and respiratory variables will be recorded. A memory box will be connected to the patient's ventilator. This device will record ventilator parameters, including FiO2 and SpO2. To assess the outcome, the percentage of time spent in optimal and sub-optimal SpO2 ranges will be calculated for each period.

Inclusion Criteria:

• Patients admitted to the ICU who are expected to require invasive mechanical ventilation for at least 8 hours and are expected to stay respiratory stable´ in the upcoming 8 hours.
• Acute respiratory failure with PaO2/FiO2 <300 with FiO2 ≥ 0.4 at inclusion
• Age older than 18 years old.
• Written informed consent signed and dated by the patient or one relative in case that the patient is unable to consent, after full explanation of the study by the investigator and prior to study participation.

Exclusion Criteria:

• Low quality on the SpO2 measurement using finger and ear sensor (quality index <60%).
• Hemodynamic instability defined as a need of continuous infusion of epinephrine or norepinephrine > 1 mg/h.
• Severe acidosis (pH ≤ 7.30).
• Chronic or acute dyshemoglobinemia: methemoglobin, CO poisoning, sickle cell disease.
• Patient under guardianship or deprived of liberties.
• Impossibility to give informed consent by both patient and family (i.e. language barrier).
• Patient included in another interventional research study under consent with similar outcome.