Objective: Invasive diagnostic procedures such as bronchoalveolar lavage (BAL) with quantitative microbiological cultures are currently recommended for the diagnosis of nosocomial pneumonia. Commonly, in clinical practice, a threshold of > or =10 colony forming units/mL is used for therapeutic decisions. The use of these measurements in daily practice assumes that their repeatability is acceptable. However, many variations among the positive results have been noted. One of the most important is dilution of BAL, which may influence the quantitative results by minimizing bacterial counts. Knowledge of the extent of dilution may increase dramatically the value of quantitative cultures. The aim of this study was to determine to what extent specimens are diluted in BAL by measuring urea in BAL and blood. Furthermore, the impact of a potential dilution effect on the diagnosis of ventilator-associated pneumonia was studied.
Patients and setting: A total of 47 patients with ventilator-associated pneumonia in two medical intensive care units at the Vienna General Hospital, a university-affiliated facility.
Design: Prospective study performed between January 2001 and July 2002.
Methods: BAL fluid was divided immediately into two samples: one for direct microscopic examination of cytocentrifuge preparations for Gram staining to determine percentages of cells containing intracellular bacteria and one for quantitative cultures according to the Cumitech 7A guidelines. Epithelial lining fluid volume was calculated using urea as a marker of dilution and correlated with colony forming units per milliliter.
Results: Nineteen out of 47 patients (40%) revealed significant bacterial growth (> or =10 colony forming units/mL). Eight additional patients (17%) would have reached the cutoff level after correction of the dilution effect, which varied between 1.8- and 130-fold.
Conclusions: Data suggest a great variation of dilution during BAL procedures, which influences quantitative results. Using urea to determine the dilution quotient could increase the value of bacterial thresholds in the diagnosis and therapeutic decision of ventilator-associated pneumonia.