• PaO2/FiO2 ratio [ Time Frame: every 6 hours during the first 48 hours after randomization ]
• Respiratory rate [ Time Frame: every 6 hours during the first 48 hours after randomization ]
• Dyspnea score [ Time Frame: every 6 hours during the first 48 hours after randomization ]
  Dyspnea score : +2 = significant improvement; +1 = slight improvement; 0 = no change; -1 = slight deterioration ; -2 = significant deterioration
• ICU and hospital length of stay [ Time Frame: Up to 14 days after randomization ]
• ICU or in-hospital mortality [ Time Frame: Up to 14 days after randomization ]
• Number of ventilator free-days [ Time Frame: Up to 14 days after randomization ]
  Days alive and without invasive or non-invasive mechanical ventilation

In blunt chest trauma patients without immediate life-threatening conditions, delayed respiratory failure and need for mechanical ventilation may still occur in 12 to 40% of patients, depending on the severity of the trauma, the preexisting conditions and the intensity of initial management.

In this context, non-invasive ventilation (NIV) is recommended in hypoxemic chest trauma patients, defined as a PaO2/FiO2 ratio < 200 mmHg. However, there is a large heterogeneity among studies regarding the severity of injuries, the degree of hypoxemia and the timing of enrollment. The interest of a preventive strategy during the early phase of blunt chest trauma, before the occurrence of respiratory distress or severe hypoxemia, is not formally established in the literature. Moreover, high-flow nasal oxygen therapy (HFNC-O2) appears to be a reliable and better tolerated alternative to conventional oxygen therapy (COT), associated with a significant reduction in intubation rate in hypoxemic patients.

Two NIV strategies are compared:

1. In the experimental strategy, NIV is performed after inclusion in patients with moderate hypoxemia, defined by a PaO2/FiO2 ratio < 300 mmHg. The minimally required duration of NIV was 4 hours per day for at least 2 calendar days.
2. In the control group, patients receive oxygen from nasal cannula or high concentration oxygen mask according to the FiO2 needed to achieve SpO2 > 92%. NIV is initiated only in patients having PaO2/FiO2 ratio < 200 mmHg under COT.

Investigators hypothesized that an early strategy associating HFNC-O2 and preventive NIV in hypoxemic blunt chest trauma patients may reduce the need for mechanical ventilation compared to the recommended strategy associating COT and late NIV.

• Chest Injuries
• Respiratory Failure

• Combination Product: Preventive strategy
  In the experimental strategy, NIV is performed after inclusion in patients with moderate hypoxemia, defined by a PaO2/FiO2 ratio < 300 mmHg. The minimally required duration of NIV was 4 hours per day for at least 2 calendar days. The daily duration of NIV can be increased at the discretion of the physician in patients with signs of delayed respiratory failure under HFNC-O2 and improving under NIV. Beyond the first 48 hours, NIV and HFNC-O2 can be stopped and the patient switched to COT if respiratory rate < 25/min and SpO2 > 92% under FiO2 < 30% for at least 6 hours.
• Combination Product: Standard of care
  In the control group, patients receive oxygen from nasal cannula or high concentration oxygen mask according to the FiO2 needed to achieve SpO2 > 92%. NIV is initiated only in patients having PaO2/FiO2 ratio < 200 mmHg under COT. A trial of curative NIV is allowed at the discretion of the physician in patients who have signs of delayed respiratory failure and no other organ dysfunction. The non-improvement of respiratory conditions after 1 hour of NIV, the NIV-dependence (≥ 12 consecutive hours) or NIV-intolerance should be considered as criteria for endotracheal intubation.

• Experimental: An "early" NIV strategy associated with HFNC-O2
  Intervention: Combination Product: Preventive strategy
• Active Comparator: A "late" NIV strategy associated with COT
  Intervention: Combination Product: Standard of care

Inclusion Criteria:

• Patient admitted in intensive care unit within 48 hours after a high-risk blunt chest trauma, defined by a TTS (Thorax Trauma Severity) score ≥ 8.
• Hypoxemia defined by a PaO2/FiO2 ratio < 300, and the absence of hypercapnia (PaCO2 < 45 mmHg).
• Without indication of endotracheal intubation at inclusion.
• Affiliated person or beneficiary of a social security scheme.
• Free, informed and written consent signed by the participant and the investigator (at the latest on the day of inclusion and before any examination required by the research).

Exclusion Criteria:

• Criteria relating to formal indication to NIV: Exacerbation of underlying chronic respiratory disease, cardiogenic pulmonary edema, severe neutropenia.
• Criteria relating to contraindications to NIV: Hemodynamic instability, Glasgow Coma Scale score ≤ 12 or excessive agitation, or other contraindications to non-invasive ventilation (active gastrointestinal bleeding, low level of consciousness, multiorgan failure, airway patency problems, lack of cooperation or hemodynamic instability).
• Associated traumatic lesions entailing particular risks: severe brain injury, complex facial trauma, tetraplegia, tracheobronchial or esophageal injuries, thoracic or abdominal trauma with indication for surgery by thoracotomy or laparotomy.
• Criteria relating to the regulation: A do-not-intubate order and a decision not to participate, persons placed under judicial protection, persons participating in another research including a period of exclusion still in course, severely altered physical and/or psychological health which, according to the investigator, could affect the participant's compliance of the study.