We project to test the effect of prone position on mortality in severe ARDS patients (PaO2/FiO2<150 with FiO2 > or = 0.6 and positive end-expiratory pressure > or = 5 cmH2O). Ventilator will be set in accordance with current standards aiming at protecting the lungs from VALI. Patients will be randomized into two arms: prone group in which proning will be realized for at least 16 hours a day and supine group in which patients will stay in a semi-recumbent position. Primary end-point is 28-day mortality of all causes. Secondary end-points are 90-day mortality of all causes and incidence of ventilator-acquired pneumonia. Study sample was calculated to detect ability of proning to reduce mortality from 60 to 45% percent, at one-tailed alpha error of 5% and power of 90% and 230 patients are needed in each arm.

The acute respiratory distress syndrome (ARDS) is an acute respiratory failure characterized by bilateral pulmonary infiltrates on frontal chest radiography, arterial oxygen tension (Pao2) to fraction of inspired oxygen (FiO2) ratio of 200 or less, and absence of evidence for left atrial hypertension. In-hospital mortality remains high: from 40 to 60%. Typical pathological findings include disruption of the alveolar epithelium, diffuse alveolar damage, hyaline membranes and pulmonary inflammation. Gas exchange is severely altered making of invasive mechanical ventilation required in most of the cases. Mechanical ventilation should be judiciously set in order to protect the lung from ventilator- associated lung injury (VALI). VALI is commonly seen as the expression of two main alveolar strains. The well documented one to date is over distension of alveoli when high tidal volumes are used. A North-American randomized controlled trial (RCT) has demonstrated that using low tidal volumes (6 ml per kilo of predicted weight) results in a 22% decrease in mortality when compared with high tidal volumes (12 ml per kilo of predicted weight). As suggested by an ancillary study of this trial, a reduced pulmonary inflammation is one of the contributing factors of this result. Therefore, minimizing VALI is a primary therapeutic goal. Low lung volume VALI is the second potential mechanism, though less well documented. It is expected to result from repetitive collapse and reopening of terminal respiratory units. Experimental data suggest that proning can reduce VALI by homogenizing lung strains. In ARDS patients, proning is mainly used due to its ability to improve gas exchange in 60 to 70% of patients. Nonetheless, 3 RCTs failed to demonstrate a significant effect on mortality of proning in three clinical situations: 1: acute lung injury, 2: hypoxemic acute respiratory failure, 3: ARDS. Post hoc analysis of one of them has suggested an effect of proning on mortality for the most hypoxemic sub-group. Studies limitations can explain lack of statistical significance as lack of power, heterogeneity of patients, insufficient duration of proning, high tidal volumes. Only one RCT was specifically designed for ARDS. Mortality was 25% lowest in the prone position arm, but this difference was not significant due to lack of power. So, the capability of proning to improve patient survival in ARDS is still an open question.

Inclusion Criteria:

1. Adult patients aged 18 years or more, both gender

2. Receiving endotracheal intubation and mechanical ventilation for less than 36 hours

3. Presenting with a Severe ARDS :

   • Acute onset
   • PaO2/FiO2 ≤ 200 (≤ 26.7 kPa)
   • Bilateral pulmonary infiltrates on frontal chest radiography
   • Capillary wedge pressure < 18 mm Hg if measured or no evidence for left atrial hypertension

4. During since 12-24 hours

5. AND with severity criteria : PaO2/FiO2 < 150 mm Hg (< 20 kPa) under FiO2 ≥ 0.6, PEEP ≥ 5 cm H2O and a tidal volume equal to 6 ml/kg PBW.

6. Written consent from patient's next of kin

Exclusion Criteria:

1. contra-indication to the prone position

   • Intracranial pressure > 30 mm Hg or cerebral perfusion pressure < 60 mm Hg
   • Massive haemoptysis needing urgent surgical or radiological treatment
   • Tracheal or thoracic surgery in the last 15 days
   • Facial trauma or surgery in the last 15 days
   • Deep venous thrombosis or pulmonary embolism treated in the last 2 days
   • Cardiac pace maker implantation in the last 2 days
   • Unstable bone dislocations of rachis, femur, rib cage, pelvis
   • Mean systolic arterial pressure less than 70 mm Hg despite vasopressive therapy
   • Pregnancy
   • Bronchopleural fistula treated with a single anterior chest tube (risk of kinking in prone position)

2. Exclusion criteria for respiratory raisons

   • Necessity to use NOi or almitrine
   • Extra body circulation for oxygenation and/or CO2 epuration

3. Exclusion criteria linked to clinical context

   • Pulmonary transplantation
   • Burns ≥ 20% of body surface
   • Chronic respiratory insufficiency under home oxygen therapy or non invasive ventilation
   • Chronic disease supposed to be fatal in less than 1 year
   • Non invasive ventilation for more than 24 hours before enrollment

4. Other exclusion criteria

   • Decision to stop life-supportive therapies before enrollment
   • Enrollment in an other trial testing an intervention and with mortality as primary end point in the last 30 days
   • Prior enrollment in the same study
   • Prone position before enrollment