• Intubation rate
  

• -Physiological outcome at 1,3 and 6 hours (arterial blood gases, respiratory rate)
  
• -Intolerance of NIV
  
• - Frequency of VAP
  
• - Duration of total MV duration
  
• - ICU LOS
  
• - Hospital LOS
  
• - Mortality (ICU and hospital)
  

Comparison of humidification devices during non invasive ventilation, in acute respiratory failure.

The hypothesis was, due to dead space, lower humidification, a reduced efficiency of the technique (NIV) when HME are used in comlparison with HH Several physiological previously performed showed that 1) with HME, because of its working principles, humidification was reduced because of leaks 2) work of breathing was increased with HME because of dead space and 3) alveolar ventilation was reduced with HME because of additional technical dead space.

For these reasons, the hypothesis was an improvement of NIV tolerance, of efficiency to clear te CO2 with HH in comparison with HME and finally to reduce the intubation rate with HH.

The patients were included when requiring NIV (see inclusion and exclusion criteria) and randomisation was performed with stratification according to presence or absence of respiratory acidosis.

The number of patients to include was baszd on the hypothesis that intubation rate would be reduced from 40 with HME to 25% with HME. A total of 250 patients was required with alpha risk of 0,05 and beta risk of 0,2 (power 80%). Intubation criteria were predefined according to the litterature. The expected duration was 18 months. The official support was institutional (DRRC of AP-HP). HH were furnished free of charge, as well as HME. Masks used were the same in both groups.