Utility of blood cultures in pneumonia patients admitted through the emergency department

Abstract

Study objectives: The American Thoracic Society and the American Society for Infectious Diseases recommend 2 sets of blood cultures as part of the evaluation of hospitalized pneumonia patients. Several studies have questioned the utility of blood cultures in community-acquired pneumonia, suggesting positive cultures seldom lead to important therapeutic changes. Our institution uses a pneumonia pathway requiring 2 sets of blood cultures with antibiotic guidelines for different clinical situations. Our objective is to identify risk factors that would enable emergency department (ED) clinicians to reduce or eliminate blood cultures in pneumonia patients. Methods: York Hospital institutional review board approval was obtained. A retrospective chart review using a standardized data collection form was conducted by the 3 investigators. The study was conducted in a community teaching hospital with an ED residency program and an annual census of 60,000 patients. Patients admitted through the ED during calendar years 2001 and 2002 with an admission diagnosis of pneumonia (diagnosis-related group 96, 97) were reviewed. Patients younger than 18 years or with a hospital discharge diagnosis other than pneumonia were excluded. Antibiotics were started in the ED according to a pneumonia pathway order set. The following risk factors were evaluated for positive blood culture results: pneumonia severity index (PSI), recent hospitalization, immunocompromised state, nursing home residency, and chronic obstructive pulmonary disease. Sensitivity to standard pneumonia pathway antibiotics was reviewed. Data were analyzed by descriptive statistics. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Results: Six hundred eighty-six charts were reviewed. Eighty-two (12%) were excluded because of nonpneumonia discharge diagnoses, which resulted in 604 eligible charts. Of 44 (7.3%) positive blood cultures identified, 25 (4.1%) were found to be contaminants, and 19 (3.1%) were true positives ( Streptococcus pneumoniae , group B streptococcus, Staphylococcus aureus, Escherichia coli, Haemophilus influenzae ). All organisms were sensitive to standard pneumonia pathway antibiotics. The antibiotic was changed from Levofloxacin to penicillin in 1 positive culture for S pneumoniae. PSI level V (OR 3.2, 95% CI 1.306 to 7.850) was the only statistically significant risk factor for a positive blood culture result. All other PSI scores, individually or combined, had ORs less than 1. Other risk factors examined did not reach statistical significance: recent hospitalization (OR 0.805, 95% CI 0.232 to 2.797), nursing home resident (OR 0.560, 95% CI 0.162 to 1.938), chronic obstructive pulmonary disease (OR 0.480, 95% CI 0.158 to 1.455), and immunocompromised state (OR 1.112, 95% CI 0.319 to 3.880). Conclusion: Only PSI level V patients had a statistically significant OR for positive blood culture results. All positive blood culture results were sensitive to pneumonia pathway antibiotics. The only culture-driven antibiotic change was to an antibiotic with a narrower spectrum. In our community hospital population, the addition of blood cultures did not affect clinical management, suggesting that cultures could be eliminated in many patients. We recommend drawing blood cultures and ordering their completion for PSI level V patients.