Usefulness of initial blood cultures in patients admitted with pneumonia from an emergency department in Japan.

Guidelines recommend obtaining blood cultures for all patients admitted with pneumonia. However, recent American studies have reported the low impact of these cultures on antibiotic therapy. Our aim was to investigate the incidence of bacteremia and change of therapy in admitted pneumonia patients from whom blood cultures were obtained in the emergency department (ED). A retrospective, observational, cohort study was conducted on consecutive patients (age >/=12 years) with pneumonia hospitalized through the ED between January 1 and December 31, 2006, in an urban teaching hospital in Japan. Data were collected on antibiotic sensitivities, empirical antibiotics, and changes of antibiotic management. Blood cultures were classified as positive, negative, or contaminant, based on previously established criteria. Out of 164 consecutive cases, blood cultures were positive in 6 patients (3.7%; 95% confidence interval [CI], 0.8%-6.6%), contaminated in 6 (3.7%), and negative in 152 (92.7%). Of the 6 bacteremic patients, 2 cases were likely to have been caused by concomitant diseases. Blood culture results altered therapy for 4 patients (2.4% of 164; 95% CI, 0.7%-6.1%), of whom 2 patients (1.2%; 95% CI, 0.1%-4.3%) had their coverage narrowed, 1 patient (0.6%; 95% CI, 0.0%-3.4%) had coverage broadened, and 1 patient had altered therapy before the drug sensitivities were reported. Considering cost and workload, the overall total annual cost was <euro>758 631 (<euro>107 = 1 $US in June 2008). Blood cultures could identify organisms in only a few patients with pneumonia and rarely altered antibiotic management even in patients with positive cultures. It may not be necessary to obtain blood cultures for patients admitted with pneumonia.