Recommendations from the European Respiratory Society are to use high-flow oxygen therapy rather than conventional oxygen therapy and non-invasive ventilation (NIV) in cases of acute hypoxaemic respiratory failure.

Studies have shown that high-flow oxygen therapy can reduce the need for intubation and the initiation of NIV without reducing mortality. In addition, by improving oxygenation, high-flow oxygen therapy can create a feeling of security which may delay endotracheal intubation.

Two studies have suggested that mortality is higher in patients receiving delayed intubation (≥ 48 hours) after failure of high flow oxygen therapy. Another study suggests that delayed endotracheal intubation after high-flow oxygen therapy ≥ 48 hours increases the risk of patient mortality independently of comorbidities and severity of illness on admission. Finally, other results suggest that the increased risk of mortality may be significant after 36 hours of high-flow oxygen therapy. It is therefore difficult to determine the right time for intubation and invasive ventilation.

A diagnostic decision aid could be of great use to the attending physician in optimising the patient's respiratory assistance strategy, whether invasive or non-invasive. Consequently, any clinical marker that can provide early detection of altered respiratory status on high-flow oxygen therapy deserves to be evaluated.

Respiratory variability is synonymous with respiratory 'good health'. A decrease in this same variabilitý is pathological and indicates an increase in the level of loads imposed on the respiratory system. It means that the measurement of respiratory variability indices may be one of these markers

There are devices for non-invasive, continuous monitoring of respiratory variability, with automated frequency analysis of thoracic movement. In this study, respiratory variability indices will be measured in patients admitted to intensive care under high-flow oxygen therapy using an external sensor called REVAMODE.

Recommendations from the European Respiratory Society are to use high-flow oxygen therapy rather than conventional oxygen therapy and non-invasive ventilation (NIV) in cases of acute hypoxaemic respiratory failure.

Studies have shown that high-flow oxygen therapy can reduce the need for intubation and the initiation of NIV without reducing mortality. In addition, by improving oxygenation, high-flow oxygen therapy can create a feeling of security which may delay endotracheal intubation. Delaying intubation would expose the patient to respiratory effort, leading to self-inflicted lung injury by a mechanism similar to ventilation-induced lung injury. Self-inflicted lung injury may increase mortality in acute respiratory distress syndrome (ARDS).

Two studies have suggested that mortality is higher in patients receiving delayed intubation (≥ 48 hours) after failure of high flow oxygen therapy. Another study suggests that delayed endotracheal intubation after high-flow oxygen therapy ≥ 48 hours increases the risk of patient mortality independently of comorbidities and severity of illness on admission. Finally, other results suggest that the increased risk of mortality may be significant after 36 hours of high-flow oxygen therapy. It is therefore difficult to determine the right time for intubation and invasive ventilation.

A diagnostic decision aid could be of great use to the attending physician in optimising the patient's respiratory assistance strategy, whether invasive or non-invasive. Consequently, any clinical marker that can provide early detection of altered respiratory status on high-flow oxygen therapy deserves to be evaluated.

Respiratory variability is synonymous with respiratory 'good health'. A decrease in this same variabilitý is pathological and indicates an increase in the level of loads imposed on the respiratory system. For this reason, this study is based on the hypothesis that the measurement of respiratory variability indices may be one of these markers.

There are devices for non-invasive, continuous monitoring of respiratory variability, with automated frequency analysis of thoracic movement. In this study, respiratory variability indices will be measured in patients admitted to intensive care under high-flow oxygen therapy using an external sensor called REVAMODE.

Inclusion Criteria:

• Patient, male or female, aged ≥ 18 years
• Patient admitted to intensive care with acute hypoxaemic respiratory failure
• Patient treated with high-flow oxygen therapy

Exclusion Criteria:

• Patient admitted to intensive care with invasive ventilation
• Patient with therapeutic abstention for intubation
• Pregnant or breast-feeding women