An evaluation of the Gram stain in protected bronchoalveolar lavage fluid for the early diagnosis of ventilator-associated pneumonia.

We investigated the usefulness and reliability of the Gram stain value versus quantitative cultures in the early diagnosis of ventilator-associated pneumonia (VAP) using the protected bronchoalveolar lavage (PBAL). One hundred four mechanically ventilated patients (age = 52 +/- 19; SAPS II = 38 +/- 15) with a strong suspicion of VAP were consecutively included. One hundred sixteen PBAL were performed and mini-bronchoalveolar lavage were analyzed using the Gram stain standard method and the conventional quantitative culture technique. VAP diagnosis was based on a positive quantitative culture of mini-bronchoalveolar lavage fluid (cutoff > or = 10(3) CFU/mL). A final diagnosis of VAP was established in 67 patients and there was no infection in 49 cases. Regarding detection of bacteria using the Gram stain, we found a sensitivity of 76.2%, a specificity of 100%, a positive predictive value of 100% and a negative predictive value of 75.4%. There was a good agreement with the final diagnosis (kappa statistic 0.73; concordance 86.2%). The degree of qualitative agreement between Gram stain and quantitative cultures was analyzed in the VAP group: the correlation was complete in 39% (26 of 67 VAP), partial in 28% (19 of 67 VAP) and there was no correlation in 33% (22 of 67 VAP). We conclude that despite its overall "good agreement," the Gram stain is of limited use for the rapid diagnosis of VAP and unreliable for the early adaptation of empirical antimicrobial therapy when using the noninvasive PBAL procedure.