Prospective Comparison of Bronchoalveolar Lavage and Quantitative Deep Tracheal Aspirate in the Diagnosis of Ventilator Associated Pneumonia

Abstract

Ventilator associated pneumonia (VAP) is common in trauma patients, and accurate diagnosis of VAP may improve survival. With the risk of development of bacterial resistance, we also strive to minimize the use of unnecessary antibiotics. Recent studies suggest that quantitative deep endotracheal aspirate (QDEA) is adequate in VAP diagnosis. We currently use bronchoalveolar lavage (BAL) diagnosis. The purpose of this study was to examine the accuracy of QDEA as compared with BAL in diagnosing VAP in trauma patients. We prospectively compared the results of BAL and QDEA in intubated patients suspected of having VAP during an 8-month period. Indication for BAL was pulmonary infiltrate, systemic inflammatory response syndrome, and C-reactive protein >17 mg/dL at > or =48 hours after admission. Study patients underwent QDEA immediately before BAL, and quantitative cultures were compared for both specimens. The techniques differ in that QDEA involves the direct culture of sputum suctioned from the distal trachea, whereas BAL involves lavage of the bronchoalveolar tree with sterile saline, which is then cultured. VAP was diagnosed on BAL if > or =10(5) cfu/mL was present on culture. The ability of QDEA to diagnose pneumonia was examined at cutoffs of > or =10(5) cfu/mL and > or =10(4) cfu/mL, as compared with BAL at > or =10(5). Sixty-one patients underwent BAL during this period, and 39 of these underwent both BAL and QDEA for the study. Of the 39 studied patients between March 16, 2002, and November 4, 2002, 20 (51%) were found to have VAP by BAL (> or =10(5) cfu/mL). Using this cutoff for QDEA, 18 of 20 (90%) would have been correctly diagnosed. Using > or =10 cfu/mL for QDEA, the rate of correct diagnosis would increase to 19 of 20 (95%). However, of the 19 who did not have pneumonia according to BAL, 6 (31%) would have been incorrectly diagnosed with VAP using the QDEA cutoff of > or =10(5) cfu/mL. A QDEA cutoff of > or =10 (4) cfu/mL would result in the even higher false-positive rate of 8 of 19 patients (42%). Whereas most patients with pneumonia by BAL would have been diagnosed by QDEA, use of QDEA in treatment decisions would have led to needless antibiotic administration in 31% of VAP-negative patients at a cutoff of > or =10(5) cfu/mL and 42% at > or =10(4) cfu/mL. The use of QDEA in VAP diagnosis is limited because of the rate of over-diagnosis. With the increasing problems associated with excess antibiotic use, we believe these results support the use of BAL over QDEA in the diagnosis of VAP in the ventilated trauma patient.