Prone position (PP) has been proved to be effective in severe ARDS patients. On the other hand, High flow nasal cannula (HFNC) may prevent intubation in hypoxemic Acute respiratory failure (ARF) patients.

Our hypothesis is that the combination of PP and HFNC in patients with COVID19 induced ARDS may decrease the need of mechanical ventilation.

Primary outcome: Therapeutic failure within 28 days of randomization (death or intubation). Secondary outcomes: to analyze PP feasibility and safety in HFNC patients and to analyze effectiveness in terms of oxygenation.

Methods: multicentric randomized study including patients with COVID19 induced ARDS supported with HFNC. Experimental group will received HFNC and PP whereas observation group will received standard care.

Optimization of non-invasive respiratory management of COVID19 induced ARDS patients may decrease the need of invasive mechanical ventilation and subsequently ICU and hospital length of stay.

Inclusion Criteria:

• COVID-19 pneumonia according to the diagnostic criteria in effect at the time of inclusion or very strongly suspected.
• Patient treated by nasal high flow.
• Moderate or severe ARDS: bilateral radiological opacities not explained entirely by effusions, atelectasis or nodules; acute hypoxemia with worsening within the 7 previous days, not entirely explained by left ventricular failure; PaO2 / FiO2 ratio <300 mmHg (or SpO2 / FiO2 equivalent).
• Informed consent.
• Beneficiary or affiliated to a social security scheme.

Exclusion Criteria:

• Indication of immediate tracheal intubation
• Significant acute progressive circulatory insufficiency
• Impaired alertness, confusion, restlessness
• Body mass index> 40 kg / m2
• Chest trauma or other contraindication to prone position
• Pneumothorax
• Vulnerable person: safeguard of justice
• Pregnant or lactating woman

Data will be shared to perform a meta-trial including other studies that assess the effectiveness of prone position in COVID-19 patients supported with nasal high flow.

No patient identifying data will be transferred to the data repository. The data set will use pseudoanonymised data. The consent process for the study will highlight that this is a meta-trial and that both aggregated and individual data without personal identifying features will be used to formulate the results from the study. Local regulation for the management of patient data needs to be adhered to as part of agreement for involvement in the study.

Data will be shared during the study at two points, Time 1 An interim analysis at 200 patients in which aggregated outcome data for the primary outcome and severe adverse effects will be analyzed

Time 2 Final analysis outlining salient clinical feature of patients for each group, outcomes

Data will be shared when first 200 patients (counting the 5 studies participating in the meta-trial) will be included and at the end of inclusion