Reproducibility of blind protected bronchoalveolar lavage in mechanically ventilated children.

Blind protected bronchoalveolar lavage (BAL) could be an interesting tool in the diagnosis of ventilator-associated pneumonia in intubated children, but its reproducibility has never been evaluated. This study evaluates the reproducibility, feasibility, and safety of blind protected BAL in mechanically ventilated children. Two blind protected BAL were done, at a 2-hour interval, in 30 patients. The reproducibility of microbiologic and cytologic results was studied. A total of 60 BALs was analyzed. Bacterial growth was present in 26 of 60 BAL (43%). Reproducibility for the presence of bacteria on quantitative cultures was excellent (concordance, 93%; kappa [kappa], 0.86). Concordance for the type of bacteria isolated was 86% and for the number of bacteria was 79%. Reproducibility for the presence of neutrophils containing bacteria was perfect (concordance, 100%; kappa, 1) although only a few BALs had a positive result (8/60). Blind protected BAL was feasible in all patients and all samples were considered adequate for analysis. Complications were mostly benign and transitory except in two cases: one pneumothorax and one significant increase in intracranial pressure. Overall, blind protected BAL is a reproducible test in mechanically ventilated children, is easily feasible, and is usually well tolerated.