In patients with acute hypoxemic respiratory failure (AHRF), High Flow Nasal Therapy (HFNT) improves oxygenation, tolerance, and decrease work of breathing as compared to standard oxygen therapy by facemask. Current guidelines recommend adjusting oxygen flow rates to keep the oxygen saturation measured by pulse oximetry (SpO2) in the target range and avoid hypoxemia and hyperoxemia. The hypothesis of the study is that closed loop oxygen control increases the time spent within clinically targeted SpO2 ranges and decreases the time spent outside clinical target SpO2 ranges as compared to manual oxygen control in ICU patients treated with HFNT.

Inclusion Criteria:

• Adult (>17yo) patients admitted to the ICU treated with HFNT for at least 8h
• Requiring FiO2 ≥ 30% to keep SpO2 in the target ranges defined by the clinician
• 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:

• Patient with indication for immediate continuous positive airway pressure (CPAP), noninvasive ventilation (NIV), or invasive mechanical ventilation
• Hemodynamic instability defined as a need of continuous infusion of epinephrine or norepinephrine > 1 mg/h
• Low quality on the SpO2 measurement using finger and ear sensor (quality index below 60% on the Massimo SpO2 sensor, which is displayed by a red or orange colour bar)
• Severe acidosis (pH ≤ 7.30)
• Pregnant woman
• Patients deemed at high risk for need of mechanical ventilation within the next 8 hours
• Chronic or acute dyshemoglobinemia: methemoglobin, carbon monoxide (CO) poisoning, sickle cell disease
• Tracheotomised patient
• Formalized ethical decision to withhold or withdraw life support
• Patient under guardianship
• Patient deprived of liberties
• Patient included in another interventional research study under consent
• Patient already enrolled in the present study in a previous episode of acute respiratory failure