Study objectives: Although it is considered standard of care to obtain blood cultures for patients hospitalized for pneumonia, several retrospective studies have questioned the utility and cost-effectiveness of this practice. The objective of this study was to determine the impact of emergency department (ED) blood cultures on antimicrobial therapy for patients with pneumonia. Methods: We performed a prospective, observational, cohort study of consecutive adult (age >18 years) patients treated at an urban university ED between February 1, 2000, and February 1, 2001. Eligible patients met the following criteria: radiographic evidence of pneumonia, clinical evidence of pneumonia, and blood culture obtained. We prospectively reviewed medical records and scored blood culture results as positive, negative, or contaminant according to previously established criteria. Additionally, data were collected on the antimicrobial sensitivities, empiric antibiotic therapy, and changes to and reasons for changing the empiric therapy. Results: During the 1-year study period, 3,926 patients had blood cultures obtained for any reason, of which 3,762 (96%) were captured. Of these, 415 (11%) of 3,762 patients had radiograph-proven pneumonia, and of these 415 patients, 29 (7.0%) had true bacteremia. Three bacteremic patients died before blood culture results were available and were excluded from the analysis. Blood culture results altered therapy for 16 (3.9%) patients, of whom 12 (2.9%) had their coverage narrowed, and only 4 (1.0%) had their coverage broadened because of resistance to empiric therapy. For the 10 patients with bacteremia whose therapy was not altered, culture results actually supported narrowing therapy in 7 cases, but this was not done. Conclusion: Blood cultures rarely altered therapy for patients presenting to the ED for pneumonia. More discriminatory blood culture use could potentially reduce resource utilization.