Ventilator-associated pneumonia (VAP) remains a major challenge. Controversy exists regarding the role for invasive approaches (e.g., bronchoscopy) for the diagnosis of VAP. It is unclear if the method of diagnosis affects mortality in this disease.
We searched multiple databases to identify randomized, controlled trials (RCTs) of invasive strategies for the diagnosis of VAP. We had no language restrictions and only included studies that reported the impact of diagnostic approach on mortality. RCTs were graded for quality and 2 reviewers independently extracted data regarding: invasive method employed, organisms recovered, antibiotic (abx) treatment regimens, and initial rate of adequacy of abxs given. Adequacy of abx was defined as the prescription of a specific regimen that had in vitro activity against the culprit pathogen. Endpoints included mortality and the effect of invasive cultures on subsequent abx utilization.
We identified 4 RCTs which included 628 patients. Only one trial was multicenter and none were blinded. The overall quality of these studies was moderate (median Jadad score of 5). VAP was confirmed bronchoscopically in 44% to 69% of participants randomized to an invasive approach. The most frequently isolated pathogens were P. aeruginosa and S. aureus. Generally most subjects (90.3%) received adequate abxs; however, in one trial there was a significant difference between the invasive and non-invasive arms with respect to this factor. Overall, an invasive approach did not alter mortality (odds ratio 0.89, 95% CI: 0.56-1.41). Invasive testing affected antibiotic utilization (odds ratio for change in antibiotic management after invasive sampling: 2.85, 95% CI: 1.45-5.59).
Despite its impact on antibiotic management, results from bronchoscopy do not alter mortality in VAP. Differences in the rates of initially adequate abxs likely explain this apparent discordance.
Physicians should develop methods for improving the rate of initially adequate abx in VAP. Invasive testing may represent a means for altering antibiotic utilization but this approach does not confer a survival advantage.
J.H. Sherner, None.