Main hypothesis: microbiological diagnossis off severe community acquired pneumonia can be performed by non invasive or semi invasive microbiological tools, semi invasive tools including protected distal bronchial samplings by the mean of Fiber optic bronchoscopy (FOB). A microbiological diagnosis could improve antibiotic therapy efficacy and improve patient's outcome.

These Two strategies have never been prospectivally evauated.

• Aim of the study: To evaluate 2 diagnostic strategies: non invasive or semi invasive including protected distal bronchial samplings by the mean of Fiber optic bronchoscopy (FOB)for the care of patients admitted in intensive care for severe community acquired pneumonia and receiving an empirical antibiotic therapy as recommanded by 2001 American thoracic guidelines
• Type of study randomized multicentric controlled open study

Type of study randomized multicentric controlled open study

Number of patients: 200 patients (100 patients in each group)admitted in ICU for SCAP

• Number of center : 5
• Clinical strategies under evaluation :
• Group A: semi-invasive strategy: FOB within 24 hours after ICU admission
• Group B: non-invasive strategy: no FOB within 24 hours after ICU admission
• Evaluation criteria:

Main criteria: Changes in initial empirical antibiotic treatment according to microbiological results obtained in each group.

others

• FOB related complications
• Microbiological results
• Type of antibiotic treatment modifications
• ICU outcome and outcome at day 28
• Duration of antibiotic therapy and nimber of days without antibiotics at day 28 after inclusion

Inclusion Criteria:

New radiological infiltrate SCAP clinical diagnosis admitted in ICU Admitted in hospital for less than 24 hours Aged 18 or older Signed informed consent

Exclusion Criteria:

Pregnancy Muribond patients Previuos hospitalisation within One month Neutropenia None HIV status