Outcome of postoperative pneumonia in the Eole study

Objectives: Prognosis factors of ventilator-associated pneumonia (VAP) have been largely investigated, while the data concerning postoperative pneumonia (POP) are scarce. The aim of this multicenter, prospective study was to evaluate the predictive factors of mortality due to POP and the impact of initial antibiotic therapy on outcome. Methods: Two hundred centers were included. Diagnosis of POP was assessed on clinical and laboratory criteria, chest X-ray changes and microbiological criteria, when possible. Outcomes of the patients were noted. An independent committee made a retrospective assessment of appropriateness of antimicrobial therapy. Results: The overall mortality among the 556 cases of POP was 23% (126 patients). Five parameters were independently associated with mortality: American Society of Anesthesiology (ASA) grade 3 or more (p<0.001), age 64 years or more (p<0.01), time to onset of pneumonia more than 3 days (p<0.01), mottling (p<0.05) and hypotension (p<0.05). Among the 322 microbiologically confirmed cases of pneumonia, 92 received inappropriate antibiotic (AB) therapy (29%). No difference in mortality was observed between the patients receiving inappropriate and appropriate AB therapy (22.8 vs 16.9%). In this subgroup, three parameters remained independently associated with mortality: ASA grade 3 or higher (p<0.001), time to onset of pneumonia more than 3 days (p<0.05) and hypotension (p<0.05). Inappropriate initial AB did not modify the model (p=0.22). Conclusions: Five independent predictive factors for mortality of POP were identified. Despite a trend toward decreased mortality with appropriate initial antimicrobial therapy, no difference was observed between the groups. Polymicrobial pneumonia or non-fermenting Gram-negative bacilli appeared to be a risk factor for inappropriate AB.