Early discontinuation of empirical antibiotic treatment in neutropenic patients with acute myeloid leukaemia and high-risk myelodysplastic syndrome

Introduction: Current guidelines advocate empirical antibiotic treatment (EAT) in haematological patients with febrile neutropenia. However, the optimal duration of EAT is unknown. In 2011, we have introduced a protocol, promoting discontinuation of carbapenems as EAT after 3 days in most patients and discouraging the standard use of vancomycin. This study assesses the effect of introducing this protocol on carbapenem and vancomycin use in high-risk haematological patients and its safety.

Methods: A retrospective before-after study was performed comparing a cohort from 2007 to 2011 (period I, before restrictive EAT use) with a cohort from 2011 to 2014 (period II, restrictive EAT use). Neutropenic episodes related to chemotherapy or stem cell transplantation (SCT) in patients with acute myeloid leukaemia (AML) or high-risk myelodysplastic syndrome (MDS) were analysed. The primary outcome was the use of carbapenems and vancomycin as EAT during neutropenia, expressed as days of therapy (DOT)/100 neutropenic days and analysed with interrupted time series (ITS). Also the use of other antibiotics was analysed. Safety measurements included 30-day mortality, ICU admittance within 30 days after start of EAT and positive blood cultures with carbapenem-susceptible microorganisms.

Results: Three hundred sixty-two neutropenic episodes with a median duration of 18 days were analysed, involving 201 patients. ITS analysis showed decreased carbapenem use with a step change of - 16.1 DOT/100 neutropenic days (95% CI - 26.77 to - 1.39) and an overall reduction of 21.6% (8.7 DOT/100 neutropenic days). Vancomycin use decreased with a step change of - 13.7 DOT/100 neutropenic days (95% CI - 23.75 to - 3.0) and an overall reduction of 54.7% (14.6 DOT/100 neutropenic days). The use of all antibiotics combined decreased from 155.6 to 138 DOT/100 neutropenic days, a reduction of 11.3%. No deaths directly related to early discontinuation of EAT were identified, also no notable difference in ICU-admission (9/116 in period I, 9/152 in period II) and positive blood cultures (4/116 in period I, 2/152 in period II) was detected.

Conclusion: The introduction of a protocol promoting restrictive use of EAT resulted in reduction of carbapenem and vancomycin use and appears to be safe in AML or high-risk MDS patients with febrile neutropenia during chemotherapy or SCT.