Can empiric broad-spectrum antibiotics for ventilator-associated pneumonia be narrowed based on Gram's stain results of bronchoalveolar lavage fluid.

BACKGROUND: Although Gram's stain (GS) of bronchoalveolar lavage fluid is routinely obtained, its usefulness remains unclear. Our purpose was to assess the value of the GS in diagnosing ventilator-associated pneumonia (VAP) and to determine whether early antibiotic narrowing was feasible.

METHODS: Retrospective study of critically ill surgical patients with clinical and quantitative evidence of VAP. GS and culture data were collated and analyzed.

RESULTS: GS results correlated with cultured organisms in 60% of cases. Overall accuracy of GS in diagnosing VAP for any organism was 58%. Gram-negative rod staining yielded like pneumonia at a statistically higher rate than Gram-positive cocci. A lack of bacteria on GS demonstrated no growth on culture 26.5% of the time, revealing the poorest correlation of all.

CONCLUSIONS: GS cannot be a reliable surrogate for quantitative culture. Altering antibiotic therapy based on GS could lead to inappropriate therapy and, presumably, greater morbidity and mortality.