Nonfermentative Gram-negative bacilli colonization was documented in one third of intubated patients before tracheotomy. Seventeen patients had A baumannii, and 15 others had P aeruginosa. Interestingly, persistent colonization by Pseudomonas was followed by pneumonia in four cases, but A baumannii colonization was followed by infection in one patient alone. Similar observations were found for E cloacae, coagulase-negative staphylococci, and C albicans. This suggests that these isolates should be considered nonpathogenic in patients with airway colonization.
The impact of VAP on outcome remains controversial. Our study suggests that the development of pneumonia in this series was not associated with increased mortality. Other reports have shown that mortality is associated with delayed antibiotic therapy rather than with development of pneumonia in critically ill patients; however, development of pneumonia prolonged ICU stay by 2 weeks, suggesting that the economic implications of this complication are important. This is consistent with otherи reports on the epidemiology of pneumonia in intubated patients.
Our study has several limitations. First, it was performed in a single ICU and the case-mix, treatments, indications for tracheotomy and clinical practices may not be generalizable to other settings. For example, all patients received antibiotic prophylaxis prior to the tracheotomy, and this may influence the findings. Second, the sample was relatively small, a fact that limits our ability to detect all possible differences between the study groups of interest. A type II error cannot be ruled out. Third, we used clinical criteria to establish the diagnosis of pneumonia. Though this may be considered a limitation, previous studies have demonstrated that clinically diagnosed pneumonia reflects the population of patients in clinical practice. Fourth, many patients receiving long-term ventilation through a tracheotomy acquire pneumonia weeks or even longer after the onset of tracheotomy. This subgroup of patients was out of the scope of this study. Consequently, implications for patients receiving long-term ventilation may be different. Such study and the epidemiology of recurrent episodes of pneumonia should be investigated in populations with long-term follow-up. Finally, no serial colonization culture samples were collected. It is evident that colonization may change, partly due to antibiotic use, but a recent study16 documented that routine surveillance may not be predictive to identify pathogens of VAP in intubated patients.
