QUANTITATIVE CULTURE OF BRONCHOALVEOLAR LAVAGE FLUID FOR THE DIAGNOSIS OF BACTERIAL PNEUMONIA

Citation
De. Cantral et al., QUANTITATIVE CULTURE OF BRONCHOALVEOLAR LAVAGE FLUID FOR THE DIAGNOSIS OF BACTERIAL PNEUMONIA, The American journal of medicine, 95(6), 1993, pp. 601-607
Citations number
21
Categorie Soggetti
Medicine, General & Internal
ISSN journal
00029343
Volume
95
Issue
6
Year of publication
1993
Pages
601 - 607
Database
ISI
SICI code
0002-9343(1993)95:6<601:QCOBLF>2.0.ZU;2-Y
Abstract
PURPOSE: A prospective study to determine the usefulness of quantitati ve bacterial cultures of fluid obtained via fiberoptic bronchoscopy an d bronchoalveolar lavage as an aid in the diagnosis of bacterial pneum onia. PATIENTS AND METHODS: All patients undergoing fiberoptic broncho scopy with bronchoalveolar lavage during a 6 1/2-month period. Presenc e of pneumonia was determined using clinical, radiographic, laboratory , and histologic data. Quantitative bacterial cultures of bronchoalveo lar lavage fluid were determined using a 1-muL culture loop. RESULTS: Quantitative bacterial cultures of bronchoalveolar lavage (BAL) fluid were sensitive and specific predictors of bacterial pneumonia. Using 1 03 colony-forming units (cfu)/mL as the threshold value for a positive culture, we determined the sensitivity and specificity to be 90% and 97%, respectively. The data were also analyzed for the subgroups of pa tients who were intubated or were receiving antibiotics. The sensitivi ty and specificity were 78% and 96% for the group of patients receivin g antibiotics and 100% and 82% for the group of patients intubated for more than 24 hours at the time of BAL. Values for the area under the receiver operating characteristic curve for the 3 groups were 0.94, 0. 88, and 0.96, respectively. CONCLUSIONS: Quantitative bacterial cultur es of BAL fluid are sensitive and specific in the diagnosis of bacteri al pneumonia. The use of antibiotics at the time of BAL reduces the se nsitivity of the test, and prolonged intubation reduces the specificit y of the test.