Objective: This study was designed to determine if bronchoalveolar lavage (BAL) quantitative culture results can be used confidently for the diagnosis of bacterial ventilator-associated pneumonia (VAP) without taking dilution into account.
Design: Prospective observational cohort study.
Setting: A 12-bed medical ICU in a teaching hospital.
Patients: A total of 241 BAL (three 50-mL aliquots) were performed in 127 patients presenting a suspicion of VAP.
Interventions: All consecutive adults who were ventilated more than 48 h were included if VAP was clinically suspected. A dilution factor, k, was developed according to the formula: dilution factor k = concentration of urea in plasma/concentration of urea in lavage fluid recovered. Using this dilution factor, the quantitative bacterial counts were interpreted accordingly with a corrected positive threshold at 10(5) colony forming unit (CFU) mL(-1).
Measurements and results: Eighty-nine BAL with at least one micro-organism > or = 10(4) CFU mL(-1) were identified (37%). In 176 BAL (73%), k ranged from 10 to 100. Median k was 24.4 (9.7-40.2) in VAP group and 24.6 (13.1-57.8) in patients without pneumonia (NS). Among the 25 BAL with micro-organism counts of 10(4) CFU mL(-1), 3 had a dilution factor lower than 10, resulting in corrected counts below the threshold of 10(5) CFU mL(-1). Two out of 15 patients with micro-organism counts of 10(3) CFU mL(-1) had corrected micro-organism counts of 10(5) CFU mL(-1). Finally, only five BAL (2.1%) were misclassified when the dilution correction factor was applied.
Conclusions: Using urea as dilution factor, we showed that BAL dilution variations did not alter the interpretation of BAL quantitative bacterial culture when administrating three aliquots of 50 mL of saline.