Acute hypoxemic respiratory failure requires endotracheal intubation and invasive mechanical ventilation in approximately 30-40% of cases, due to severe hypoxemia and/or clinical signs of acute respiratory distress. The primary objectives of invasive mechanical ventilation are to reduce respiratory effort and improve oxygenation. However, this intervention is also associated with both direct and indirect adverse effects, mainly linked to the need for sedation and often neuromuscular blockade. These include hemodynamic compromise, neuromuscular weakness, ventilator-induced lung injury, and infectious complications.

An ideal intubation strategy would therefore strike a balance: avoiding the risks of delayed intubation-such as refractory hypoxemia, excessive respiratory effort, and patient self-inflicted lung injury (P-SILI)-while limiting complications associated with invasive mechanical ventilation by withholding it in patients who might otherwise recover without. To date, the optimal strategy for achieving this risk-benefit balance remains uncertain.

Clinical practice suggests a broad consensus on the necessity of intubation when so-called safety criteria are met: severe hypoxemia (SaO₂/FiO₂ ratio < 88), marked respiratory distress (use of accessory muscles, thoracoabdominal paradox, respiratory rate > 40/min), extra-respiratory manifestations of hypoxia (e.g., altered consciousness), and/or uncontrolled hemodynamic instability. Beyond these safety thresholds, however, debate persists. Some advocate for earlier intubation-a so-called liberal approach-triggered by predefined hypoxemia criteria (e.g., SpO₂/FiO₂ < 110), with the aim of limiting the deleterious consequences of sustained hypoxemia.

In routine practice, the criteria guiding intubation vary widely between clinicians and cannot be attributed to strong scientific evidence. This study therefore seeks to compare, in a randomized interventional design, the two main strategies currently applied across centers:

• Liberal intubation strategy: prioritizing the prevention of organ dysfunction related to hypoxemia (notably hypoxic cardiac arrest) and the risk of P-SILI.
• Restrictive intubation strategy: prioritizing the reduction of invasive mechanical ventilation use, with the goal of minimizing ventilation-related harm and its associated therapeutic burden.

Inclusion Criteria:

• Adult patient

• Patient admitted to intensive care less than 24 hours ago

• Acute respiratory failure with hypoxemia defined by either:

  • Oxygen therapy ≥ 10 L/min via high-concentration mask required for SpO2 ≥ 92%
  • High-flow oxygen therapy with FiO2 ≥ 50% required for SpO2 ≥ 92%

• Informed consent of the patient or a trusted relative (when the patient is unable to give consent)

Exclusion Criteria:

• Acute hypercapnic respiratory failure (defined by PaCO2 > 45 mmHg)
• Cardiogenic pulmonary edema
• Exacerbation of chronic respiratory disease
• Respiratory failure requiring long-term oxygen therapy
• Neuromuscular disease
• Glasgow Coma Scale score ≤ 12
• Decision to intubate immediately
• Invasive mechanical ventilation within the previous 7 days
• Treatment limitation decisions for intubation
• Person deprived of liberty by judicial or administrative decision : Person undergoing compulsory psychiatric care, person subject to legal protection measures, Pregnant, breastfeeding, or parturient patient