Low Rates of Vancomycin Use in Febrile Neutropenia: A Multicenter Study

Introduction: The prompt initiation of a betalactam antibiotic in febrile neutropenia (FN) is standard of care in patients with hematologic malignancies and hematopoietic cell transplant (HCT) recipients. The use of a glycopeptide is recommended by guidelines in situations such as suspected catheter-related, skin and soft tissue infection, pneumonia or hemodynamic instability, but the level of these recommendations is weak. The aim of this study was to evaluate the use of vancomycin in febrile neutropenia (FN) in four Brazilian centers in which more restrictive criteria for using vancomycin have been applied.

Methods: We retrospectively reviewed all episodes of FN from 2013-2019 in four hospitals: 2 teaching public, 1 public and 1 private hospital, and looked at reasons for giving vancomycin and outcomes. Criteria for using vancomycin was infection by methicillin-resistant Gram-positive bacteria, skin or soft tissue infection and shock.

Results: We recorded 461 episodes of FN in 312 patients. Median age was 50 years (17 - 86) and 61% were male. FN occurred after chemotherapy in 70% of cases and after HCT in 30% (allogeneic in 11.5%). The most frequent underlying diseases were acute myeloid leukemia (28%), non-Hodgkin's lymphoma (25%), acute lymphoid leukemia (16%) and multiple myeloma (13.4%). Colonization by methicillin-resistant Staphylococcus aureus (MRSA) was present in only 3%, and quinolone prophylaxis was given in 32% of episodes. According to international guidelines, vancomycin should have been started at first fever in 120 episodes (26%) but it was given in only 28 (6%). Reasons for use at first fever were skin or soft tissue infection (n=10), hypotension or shock (n=8), catheter-related infection (n=3), mucositis (n=1), and no apparent reason (n=6). Vancomycin was added after day 1 in 69 episodes (15%), for the following reasons: Gram-positive bacteremia (n=36), skin or soft tissue infection (n=7), mucositis (n=7), hypotension or shock (n=6), pneumonia (n=3) and no apparent reason (n=10). Bacteremia was documented in 143 episodes (31%): 81 by Gram-negative and 62 by Gram-positive bacteria. Bacteremia due to Staphylococcus aureus occurred in only 4 episodes (1 MRSA). The 15- and 30-day mortality was 5.4% and 9.8%, respectively, with no significant differences among private (7.6%) and public (11.3%) hospitals.

Conclusions: Despite the lack of benefit of empiric vancomycin in FN either in terms of fever duration or mortality, vancomycin is still widely used in to manage FN in hematologic patients and HCT recipients, potentially increasing toxicity. In our hospitals vancomycin was given in a very low proportion of episodes, despite having classic indications for its use as per guidelines. Our findings highlight the importance of caution in adopting recommendations from guidelines without looking at local epidemiology especially if the level recommendation is weak. These measures have the potential to decrease coast and toxicity.