An emergency department–based randomized trial of nonbronchoscopic bronchoalveolar lavage for early pathogen identification in severe community-acquired pneumonia

Abstract

Study Objectives: Many patients with community-acquired pneumonia are treated empirically without an aggressive search for causative pathogens, an approach adopted largely because of the costs and difficulties encountered during efforts to identify the causative organisms. Blood and sputum cultures are not sensitive, and the more invasive techniques of broncoscopy and lung biopsy are generally time consuming and not cost-effective. The technique of nonbronchoscopic bronchoalveolar lavage (BAL) has been shown to accurately diagnose the causes of nosocomial pneumonia. The purpose of this study was to determine whether an emergency department–based BAL protocol would lead to more frequent isolation of pneumonia pathogens and result in more changes to tailored antibiotic therapy in comparison with standard care. Methods: We studied all adult patients admitted with a diagnosis of pneumonia who were tracheally intubated and who had obtainable familial consent in the ED of an urban county hospital from March 1998 to October 1999. Exclusions included antibiotic use within the past 5 days, pneumothorax, hemoptysis, or persistent hypoxia using 100% oxygen. Patients were randomized to standard care versus standard care plus BAL. Blood culture specimens were drawn from all patients before the initiation of antibiotics. All other diagnostic tests were ordered at the discretion of treating physicians. BAL fluid, sputum, and blood culture specimens were tracked, and patient antibiotic course was followed to assess any change in regimen. Results: Twenty-six of 64 patients evaluated for study participation met all eligibility criteria; 14 patients received standard care, and 12 patients received standard care plus BAL. Pneumonia pathogens were identified in 10 (83.3%) of 12 patients in the BAL group and in 4 (28.6%) of 14 patients in the standard care group (P =.007). Comparing BAL versus non-BAL groups, there was no significant difference in the likelihood of overall antibiotic regimen changes (P =.149), but there was a difference with regard to antibiotic changes made in patients with positive culture test results (P =.026). No major complications occurred with BAL catheterizations. Conclusion: ED–based BAL catheterization allows for early identification of pathogens in severe community-acquired pneumonia, which leads to changes in antibiotic therapy. [Rodriguez RM, Fancher ML, Phelps M, Hawkins K, Johnson J, Stacks K, Rossini T, Way M, Holland D. An emergency department–based randomized trial of nonbronchoscopic bronchoalveolar lavage for early pathogen identification in severe community-acquired pneumonia. Ann Emerg Med. October 2001;38:357-363.]