• blood culture;
  • community-acquired pneumonia

Objectives: To determine if blood cultures identify organisms that are not appropriately treated with initial empiric antibiotics in hospitalized patients with community-acquired pneumonia, and to calculate the costs of blood cultures and cost savings realized by changing to narrower-spectrum antibiotics based on the results.

Methods: This was a retrospective observational study conducted in an urban academic emergency department (ED). Patients with an ED and final diagnosis of community-acquired pneumonia admitted between January 1, 2001, and August 30, 2003, were eligible when the results of at least one set of blood cultures obtained in the ED were available. Exclusion criteria included documented human immunodeficiency virus infection, immunosuppressive illness, chronic renal failure, chronic corticosteroid therapy, documented hospitalization within seven days before ED visit, transfer from another hospital, nursing home residency, and suspected aspiration pneumonia. The cost of blood cultures in all patients was calculated. The cost of the antibiotic regimens administered was compared with narrower-spectrum and less expensive alternatives based on the results.

Results: A total of 480 patients were eligible, and 191 were excluded. Thirteen (4.5%) of the 289 enrolled patients had true bacteremia; the organisms isolated were sensitive to the empiric antibiotics initially administered in all 13 cases (100%; 95% confidence interval = 75% to 100%). Streptococcus pneumoniae and Haemophilus influenzae were isolated in 11 and two patients, respectively. The potential savings of changing the antibiotic regimens to narrower-spectrum alternatives was only 170.

Conclusions: Appropriate empiric antibiotics were administered in all bacteremic patients. Antibiotic regimens were rarely changed based on blood culture results, and the potential savings from changes were minimal.