Thus, it is probable that the volume of fluid plated with the two techniques reflects a similar concentration of bacteria in lung secretions. This hypothesis is supported by the strong correlation between TPC and BAL quantitative cultures we found. Furthermore, our data are comparable with studies in mechanically ventilated patients, studies in baboons, and a study comparing the two techniques in healthy human volunteers. In our opinion, a BAL fluid culture with more than 103 cfu/ml should be considered diagnostic in patients with CAP provided that they have not been treated with antibiotics.
Our hypothesis was to test the diagnostic value of BAL in comparison with TPC cultures in patients with CAP. We have demonstrated that BAL is as reliable as TPC in this setting. Since these techniques are highly operator dependent, the choice of one of them will rely on the pulmonologists skills. Considering that patients with pneumonia are frequently dehydrated and have few bronchial secretions, making it very difficult sometimes to place the brush in the proper site, obtaining a good specimen is easily accomplished by BAL. However, we do not advocate routine bronchoscopy in all patients with pneumonia admitted to the hospital. We believe (and it is our practice) that in the moderately ill patient with CAP, the initial investigation must be noninvasive. However, in patients with severe illness or when a bacterial etiology other than S pneumoniae is suspected, bronchoscopy with BAL provides a rapid and accurate method of identifying the causative agent. The diagnostic value of BAL cultures in patients with CAP and therapy failure remains to be defined in future prospective studies.