Blood Cultures Taken During the First 72 Hours of Antibiotic Therapy Are a Waste of Time and Money

Abstract & Commentary

Synopsis: There appears to be no diagnostic value of blood cultures obtained in the first 72 hours after institution of antibiotic therapy.

Source: Grace CJ, et al. Usefulness of blood culture for hospitalized patients who are receiving antibiotic therapy. Clin Infect Dis. 2001;32:1651-1655.

Grace and colleagues reviewed the notes of 1446 patients to identify 139 who had been admitted to the hospital for suspected infection, who had had 2 sets of blood cultures taken before antibiotic therapy (preantibiotic blood cultures), who had been given antibiotic therapy within 24 hours, and who had had a further 2 sets of blood cultures taken during the first 72 hours of therapy (antibiotic blood cultures). Most of the 785 patients were excluded from the study because of an inadequate number of blood cultures—133 because of antibiotic use before admission, 72 because no therapy was given, 187 because of admission to the ICU, and the remaining 130 because they were neutropenic. Of the 598 blood culture sets obtained, 272 (45%) were taken before antibiotic therapy was started. Blood cultures were negative or contaminated in 83 cases and remained so during the first 72 hours of antibiotic therapy. Staphylococci were isolated from the preantibiotic blood cultures of 25 patients and the same isolate from the antibiotic blood cultures of 19 (76%) patients (Staphylococcus aureus was involved in 18 cases), streptococci from 14 preantibiotic blood cultures with the same isolate from the antibiotic blood cultures of 5 (36%) patients, and Gram-negative bacilli from the preantibiotic blood cultures of 17 patients and the same isolate from the antibiotic blood cultures of 2 (12%) patients. The antibiotic blood cultures of only 1 patient yielded a new bacterium, Bacteroides fragilis in addition to the Escherichia coli identified before therapy. Having been admitted initially for the evaluation of fever, this patient underwent a CT scan of the abdomen on day 3 anyway, which revealed sigmoid diverticulitis. The rates of positive blood cultures on day 1, 2, and 3 of therapy were 16% (26/162 sets), 15% (15/99 sets), and 20% (13/65 sets), respectively. Grace et al conclude that the results of blood cultures taken during the first 72 hours of antibiotic therapy can be predicted from the preantibiotic blood cultures, and they advise physicians to wait for the results of the preantibiotic blood cultures before deciding to order more.

Comment by J. Peter Donnelly, Phd

This is a straightforward study with a simple message: taking blood cultures during the first 72 hours of antibiotic therapy is not going to reveal any new information, so why not simply wait for the preantibiotic culture results before deciding on ordering more? Are they justified in their conclusion? The answer is yes if the purpose of taking blood cultures during therapy is to uncover the cause of persisting fever. Grace et al clearly show that no useful new information is to be gained. They remind us that it is not uncommon anyway for fever to persist for 72 hours when a patient has an infection, and they can find no justification for taking blood cultures as a reflex action to persisting fever. However, if one is trying to assess the efficacy of therapy in eradicating the cause of bacteremia, the question then becomes "when is the appropriate time for taking follow-up cultures to assess microbiological efficacy?" In the current study, S aureus and, to a lesser extent, the streptococci, seemed to have persisted despite 72 hours of therapy. This would make any physician uncomfortable and want to modify treatment if he or she was aware of this. However, this information wouldn’t usually be known until 48 hours after the blood cultures were taken simply because of the turnaround time. So, by day 4 of therapy (ie, after a full 72 hours of therapy) only the results of the preantibiotic cultures and any taken on day 1 and perhaps day 2 would be known making it impossible to judge the microbiological efficacy of treatment in eradicating the pathogen simply because this would be considered too short for treatment to have worked. The question of eradication is usually only posed after 72 hours of therapy so the blood cultures taken on day 3 would be taken too soon although they would become known on day 5. Hence, the question of microbiological efficacy is also not addressed by taking cultures during the first 72 hours of therapy. Therefore, the 55% of blood cultures taken during therapy were of no clinical value and hence were a waste of time and resources. If these results were echoed throughout the hospital including the ICU and neutropenic patient, there would be considerable savings because the number of blood cultures taken could be reduced by half.

I doubt if anyone really knows why taking blood cultures on a daily basis ever became commonplace. The arguments advanced in favor of this practice are not evidence-based but seem appealing and plausible. For instance, the more often one takes blood cultures, the less likely a causative agent will be missed. This would only be true if antibiotic therapy was not being given since the presence of antibiotics in blood cultures reduces their yield. Another argument is that persistent fever might represent constant bacteremia or repeated bouts of bacteremia. This may be true in endocarditis, or perhaps catheter-related infections, but little else. Instead, like many of the customary practices in medicine, what seems reasonable may turn out to be little more than a ritual that should actually be considered a relic of the past and abandoned altogether.