Objectives: Information on clinical practice regarding the diagnosis of pneumonia in European intensive care units is limited. The aim of this study was to describe the spectrum of actual diagnostic practices in a large sample of European intensive care units.
Design: Prospective, observational, multicenter study.
Setting: Twenty-seven intensive care units of nine European countries.
Patients: Consecutive patients requiring invasive mechanical ventilation for an admission diagnosis of pneumonia or receiving mechanical ventilation for >48 hrs irrespective of admission diagnosis.
Interventions: None.
Measurements and main results: A total of 2,436 patients were evaluated; 827 were admitted with or developed nosocomial pneumonia (hospital-acquired pneumonia [HAP], 27.1%; ventilator-associated pneumonia [VAP], 56.2%; very early onset VAP, 16.7%). Mean age was 59.4 +/- 18.1 yrs, 65.0% were men, and mean admission Simplified Acute Physiology Score II was 46.7 +/- 17.1. Worsening oxygenation (76.8%), purulent/changing respiratory secretions (72.1%), and new temperature elevation (69.2%) were the most frequent clinical signs of nosocomial pneumonia. Etiological diagnosis was based on noninvasive respiratory specimens in 74.8% of episodes. Bronchoscopy was performed in 23.3% of episodes. Bronchoscopy performance, after adjustment by severity of illness, age, and type of hospital, were predicted by worsening oxygenation (odds ratio 2.03; 95% confidence interval, 1.27-3.24) and male sex (odds ratio 1.77; 95% confidence interval, 1.19-2.65). Definite cause was documented in 69.5% of nosocomial pneumonia cases. The most common isolates were Staphylococcus aureus (16.3% methicillin-sensitive S. aureus and 16.0% methicillin-resistant S. aureus), Pseudomonas aeruginosa (23.1%), and Acinetobacter baumannii (19.1%). Presence of nosocomial pneumonia significantly prolonged mean length of mechanical ventilation (10.3 days, p < .05) and mean intensive care unit length of stay (12.2 days, p < .05) in intensive care unit survivors. Mortality rate was 37.7% for nosocomial pneumonia vs. 31.6% for patients without pneumonia (p < .05).
Conclusions: Etiological diagnosis of nosocomial pneumonia in a large sample of European intensive care units was based mainly on noninvasive techniques. However, there was high variability in bronchoscopy use between the participating intensive care units.