Evaluation of systematic administration of oral prednisolone (1mg/Kg/day) as an add on therapy in Chronic Obstructive Pulmonary Disease (COPD) patients admitted to intensive care unit (ICU) for severe exacerbation of COPD. Patients with pneumonia are excluded.

Randomization is stratified according to ventilatory support: non invasive or conventional ventilation.The major outcome is the ICU mortality rate in overall population and stratified according to ventilatory mode (noninvasive ventilation (NIV) versus conventional).

Secondary outcomes are superinfection necessitating a new antibiotic course, Length of mechanical ventilation (MV) (and ventilatory free days), Length of ICU stay, The frequency of gastric bleeding episodes that of frequency of hyperglycemic episodes.

Inclusion Criteria:

• All COPD patients (according to the ATS definition) experiencing acute exacerbation originating in acute respiratory failure and requiring ICU admission will be included in the study.

COPD exacerbation is defined by the increased frequency of cough, volume and purulence of sputum and that of wheeze.

Acute respiratory failure is defined by the presence of hypercapnia with PaCO2 >45mmHg associated with pH > 7.35 and signs of respiratory muscle fatigue (contraction of accessory respiratory muscles, thoracoabdominal swinging ,..).

Exclusion Criteria:

• Asthmatic patients defined by a reversible obstructive disease following nebulized bronchodilators,
• Patients with uncontrolled left heart failure,
• AECOPD patients with a radiologically documented pneumonia,
• Systemic corticotherapy within 30 days before screening,
• contra-indication to corticosteroids (active gastroduodenal ulcer, uncontrolled sepsis, etc. ..)

There is no plan to IPD so far