Am Fam Physician. 2003;68(4):721-724
Clinical Question: How useful are blood cultures in patients with community-acquired pneumonia (CAP)?
Setting: Inpatient (any location)
Study Design: Cohort (prospective)
Synopsis: Although the value of routine blood cultures for patients with CAP has long been questioned, they are still recommended by all of the major clinical practice guidelines. This is one of the largest and best designed studies of this practice, and it is part of a larger clinical trial comparing strict use of a practice guideline at nine hospitals with usual care at 10 hospitals.
The authors identified 2,804 patients with suspected CAP, of whom 1,743 met inclusion criteria; 716 were cared for at “guideline hospitals” and 1,027 at “conventional care hospitals.” Patients were included in the trial if they had at least two signs or symptoms of CAP (e.g., fever, productive cough, chest pain, shortness of breath, rales) and chest radiography was consistent with CAP. Patients with immunodeficiency; those who were very ill and required direct admission to the intensive care unit; patients with alcoholism; and those who were pregnant, nursing, or in renal failure were excluded. The guideline called for blood cultures to be obtained before antibiotics were administered; patients in the conventional care hospitals had blood cultures obtained at the discretion of their physician. The primary outcome was whether the use of blood cultures led to a clinically important change in management.
Of the 1,743 patients who met criteria for inclusion in the emergency department, 1,022 were admitted to the hospital. Of these, 760 (74.4 percent) had blood cultures drawn, and 44 (5.7 percent) had a significant organism (Streptococcus pneumoniae, 66 percent; Enterobacteriaceae, 16.3 percent; Staphylococcus aureus, 11.4 percent; Haemophilus influenzae, 2.3 percent; and various other bacteria). The likelihood that a blood culture would be positive did not differ between conventional and intervention groups and also was similar among groups with different levels of severity according to Fine's validated Pneumonia Severity Index (4.6 to 8.0 percent).
After receiving the results, some physicians switched nine patients to a broader spectrum antibiotic during the course of the illness, but this change was supported by blood culture results in only three cases. The three patients had penicillin-resistant S. pneumoniae, methicillin-sensitive S. aureus, and methicillin-resistant S. aureus; however, the latter patient died despite the change in therapy. Switching patients to a less expensive, narrower spectrum, or less toxic drug was a more rational response and was supported by blood culture results in 12 of 14 cases. The regimen remained unchanged in 20 patients, despite the fact that in 17 of these patients the blood culture result supported a change in therapy.
Bottom Line: In patients presenting with CAP, blood culture results prompt a change in antibiotic therapy in 2 percent of patients, but in only 0.4 percent is this change likely to improve the patient's outcome. At a cost of approximately $60 for a pair of blood cultures at the typical American hospital, that comes to $15,200 per potentially useful change. This study provides further support for the argument that routine blood cultures in patients with CAP add little to patient care and can be omitted in many cases. It may still be prudent to order cultures for patients with more severe disease (Fine Pneumonia Severity Index Class III to V). (Level of Evidence: 2c)