Bronchoscopy in ventilator-associated pneumonia: agreement of calibrated loop and serial dilution

Abstract

Rationale: Although the serial dilution technique for quantitative culture of bronchoalveolar fluid is considered to be the gold standard for the diagnosis of ventilator-associated pneumonia, it is more labor intensive than the calibrated loop technique.

Objective: We sought to determine the agreement between the calibrated loop and serial dilution techniques in the diagnosis of ventilator-associated pneumonia.

Methods: We prospectively measured bacterial colony counts by the serial dilution and calibrated loop techniques in 121 bronchoalveolar lavage samples of 104 patients with suspected ventilator-associated pneumonia.

Measurements and main results: At the time of bronchoscopy, patients had received mechanical ventilation for a median of 8 d. Patients were receiving antibiotics when 90 of the 121 (74.4%) bronchoalveolar samples were obtained. The colony counts of 13 bacterial isolates were too numerous to count by the calibrated loop technique; by serial dilution technique, their counts ranged from 4.70 to 6.74 log10 cfu/ml. Fifty other bacteria had paired colony counts measured by each of the two techniques: the bias (95% confidence interval) between the two techniques was -0.380 (-0.665 to -0.095) log10 cfu/ml, with precision of 1.002 log10 cfu/ml and 95% limits of agreement of -2.344 to 1.584 log10 cfu/ml. Using the threshold of 4 log10 cfu/ml as a criterion for the diagnosis of ventilator-associated pneumonia, there was discordance only for one bacterial organism between the two techniques.

Conclusions: The calibrated loop technique can be used for the diagnosis of ventilator-associated pneumonia using bronchoalveolar lavage fluid.