NEW YORK (Reuters Health) – When ventilator-associated pneumonia (VAP) is suspected, the absence of Gram staining on a respiratory specimen reliably rules out the diagnosis. On the other hand, a positive Gram stain has a low predictive value for VAP and correlates only moderately with organisms recovered in culture, investigators report in Clinical Infectious Diseases online June 7.

“Therefore, a positive Gram stain should not be used to narrow anti-infective therapy until culture results become available,” advise Dr. John C. O’Horo, with Aurora Healthcare Metro in Milwaukee, Wisconsin, and colleagues.

They note that appropriate initial antibiotic therapy is essential to optimize outcomes of ventilator-associated pneumonia. Cultures of respiratory specimens helps confirm the diagnosis and tailor antibiotics, but results take 48-72 hours. Gram staining may provide rapid information, but its usefulness in this setting is unclear.

The team therefore conducted a meta-analysis of 24 relevant studies to determine the predictive value of respiratory specimen Gram stain in cases of suspected VAP, and to evaluate its utility in guiding anti-infective therapy.

The pooled sensitivity and specificity of Gram stain for diagnosis of VAP were 79% and 74%, respectively, according to the report.

The negative and positive predictive values (NPV and PPV) of the test varied with the prevalence of VAP, the authors explain. “For a VAP prevalence between 20-30%, which is a clinically common scenario, the NPV of Gram stain is over 90% and the PPV is 40%,” they report, “indicating that only 40% of positive Gram stains indicate VAP, while VAP can be excluded in 90% of cases when a Gram stain is negative.”

As for the correlation between Gram stain bacterial morphology and organisms recovered in culture, the pooled kappa statistic of 0.54 indicated only fair agreement, Dr. O’Horo and colleagues found.

They say that their results have important clinical implications. “First, empiric therapy for presumed VAP should comprise of antibiotics with a broad spectrum of activity directed against the common pathogens implicated in VAP. Presence of a single type of organism on Gram stain does not allow narrowing of initial therapy.”

“Second,” they continue,” a negative Gram stain was highly predictive of absence of VAP. Our data suggests that, in the absence of a high clinical pre-test probability for VAP, a negative Gram stain in a clinically stable patient with suspicion for infection should prompt a search for alternative sites of presumed infection.”

The authors conclude: “Ultimately, VAP remains a clinical diagnosis requiring the consideration of multiple laboratory, physical examination findings and radiographic features. No single factor can confirm or rule out a diagnosis of VAP.”

SOURCE:

Clin Infect Dis 2012.