By Dean L. Winslow, MD, FACP, FIDSA

Professor of Medicine, Division of General Medical Disciplines, Division of Infectious Diseases and Geographic Medicine, Stanford University

Dr. Winslow reports no financial relationships related to this field of study.

SYNOPSIS: In contrast to blood cultures obtained on therapy in patients with Gram-positive bacteremia and endocarditis, follow-up blood cultures in patients with Gram-negative bacteremia seldom provide useful information.

SOURCE: Canzoneri CN, Akhavan BJ, Tosur Z, et al. Follow-up blood cultures in Gram-negative bacteremia: Are they needed? Clin Infect Dis 2017;65:1776-1779.

In a retrospective study conducted at a large hospital in Houston, researchers studied 500 episodes of bacteremia to determine the frequency of follow-up blood cultures (FUBC) and assess risk factors for persistent bacteremia. Of the 500 episodes of bacteremia, 383 (77%) had at least one FUBC drawn. This included 54% of patients with initial bacteremia due to Gram-positive cocci (GPCs), 37% with bacteremia due to Gram-negative rods (GNRs), and 8% with polymicrobial bacteremia. Persistent bacteremia (defined as positive blood culture for the original organism at least 24 hours after the initial blood culture was drawn) was more common in GPC bacteremia (21%) than in polymicrobial (10%) or GNR bacteremia (6%). Duration of bacteremia was similar between groups (2.7-2.9 days). Positive FUBCs were most commonly Staphylococcus aureus (31), coagulase-negative Staphylococcus (six), Enterococcus (four), Escherichia coli (five), and Klebsiella, Serratia, and Stenotrophomonas (one each).

For all patients in the case series, factors shown to be predictive for positive FUBCs included fever on the day the FUBC was drawn, presence of a central catheter, and ESRD on hemodialysis. When broken down by persistent GPC vs. GNR bacteremia, fever, presence of a central catheter, DM, and ESRD on hemodialysis were present for GPC bacteremia, but only the presence of fever at the time the FUBC was drawn was predictive of persistent GNR bacteremia (six of eight patients).

The source of bacteremia was known in 273 (71%) patients who had FUBCs drawn. Only 37 had positive FUBCs. Broken down by source, the rate of positive FUBCs was quite low for most sources (UTI 3%, severe skin infection 6%, intra-abdominal infection 10%, osteomyelitis 0%, but higher for central catheter [34%] and pneumonia [15%].)


At our institution, FUBCs are ordered commonly, and when physicians are questioned about this practice, they are surprised to learn that this is not considered standard of care. This relatively small study goes a long way toward illuminating that this is not a very helpful practice, especially in patients with GNR bacteremia who are doing well on appropriate antibiotics. (Overall, of the 140 patients with initial GNR bacteremia, it should be emphasized that only eight had positive FUBCs.) As the authors point out, not only does ordering routine follow-up blood cultures in patients with GNR bacteremia seldom produce helpful information, but common false-positive results can lead to longer length of stay, additional inappropriate antibiotic therapy, and increased healthcare costs.