The purpose of the study is to evaluate, in a prospective way, the clinical, the manometric an gazometric benefit of non-invasive ventilation (VNI), (INFANTFLOW [EME, Brighton, England]) during bronchiolitis of the infant compared to a conventional managment.

1. Experimental plan: The infants (0 to 6 month) admitted in pediatric intesive care unit for respiratory distress with bronchiolitis but not requiring mechanical ventilation are randomized in two groups: group "A" is treated with a non-invasive ventilation device(Infantflow) during 6 hours, group "B" is treated conventionally by inhalation of a Air/O2 mixture during 6 hours.
2. Outcome: Primary outcome: clinical score of respiratory distress at H0 and H6. Secondary outcome: respiratory and cardiac frequency, average blood pressure. Manometric: Variation of esophageal pressure at H0, H6. Gasometric: Minimal FiO2 necessary to reach an oxygen saturation between 94 and 98%, transcutaneus PCO2, PaO2/FiO2
3. Analyze results: Intention to treat analysis after having checked the randomization by comparing clinical gravity in the two arms (clinical score).
4. Calendar of the study: October 2006 at April 2008
5. Follow-up of the patients: the study proceeds during the hospitalization of the infants during 6 hours and will not impose any specific follow-up.

Inclusion Criteria:

• Infant from 0 to 6 months admitted in the PICU
• Clinical diagnosis: bronchiolitis
• Signs of respiratory distress evaluated with a clinical score of respiratory distress > 4

Exclusion Criteria:

• Cardiopathy, cystic fibrosis or a neuromuscular desease
• Need for mechanical ventilation