Study objective: We retrospectively examined the clinical utility of obtaining routine blood cultures before the administration of antibiotics in certain nonimmunosuppressed patients with community-acquired pneumonia (CAP) admitted to the hospital during 1991.
Design: Retrospective review.
Setting: Grady Memorial Hospital (a county hospital primarily serving inner-city Atlanta).
Patients or participants: Hospital discharge diagnosis listings identified 1,250 adults ( > or = 18 years old) with pneumonia. From this group of patients, we selected patients admitted to the hospital with (1) respiratory symptoms and a lobar infiltrate on chest radiograph that were present at the time of hospital admission, (2) two or more sets of blood cultures obtained within 48 h of hospital admission, and (3) absence of defined risk factors: HIV-related illness, malignancy, recent chemotherapy, steroid therapy, sickle cell disease, nursing home residence, or hospital stays within the past 14 days.
Measurements and results: Five hundred seventeen patients (mean age, 52 years;: age range, 18 to 103 years) qualified. Of these 517 patients, 25 patients (4.8%) had growth in blood cultures considered contaminants while 34 (6.6%) had blood cultures positive for the following pathogens: 29 Streptococcus pneumoniae, 3 Haemophilus influenzae, and 1 Streptococcus pyogenes, 1 Escherichia coli. Antibiotic therapy was changed for 7 of the 34 patients with positive blood cultures (1.4% of study patients). Antibiotic regimens were altered in 48 additional patients based on sputum culture, poor clinical response, and allergic reactions.
Conclusions: Few blood cultures were positive for likely infecting organisms in adult patients with CAP without defined underlying risk factors. Furthermore, a total of $34,122 was spent on blood cultures at $66 per patient. In this carefully defined group of patients, blood cultures may have limited clinical utility and questionable cost-effectiveness.