HAL will be down for maintenance from Friday, June 10 at 4pm through Monday, June 13 at 9am. More information
Skip to Main content Skip to Navigation
Master Thesis

Early de-escalation of the empirical antibiotic in high-risk febrile neutropenia with bloodstream infection: a retrospective study

Abstract : Infections are the main cause of morbidity and mortality for haematological patients. Fever will occur in above 80% of the high-risk neutropenia episodes and require immediate broad spectrum antibiotics. European Conference on Infections in Leukaemia-4 (ECIL-4) recommandations advocate to realise a de-escalation “according to the organism identified using narrower-spectrum agents”. However, they are usually not applied by clinicians and antibiotic stewardship programs (ASP) in haematology are not often implemented. The objectives of our study were to evaluate the rate of de-escalation within 96 hours after bacteremia in an haematological unit and to measure the impact of an ASP, implemented at the beginning of 2019, regarding the safety of de-escalation and antibiotic consumption. Every patient with positive blood culture sampled in the Department of Haematologic Intensive Care Unit from January 1st 2018 to December 31st 2019 was retrospectively included. De-escalation was defined by either discontinuation of one of the antibiotics in case of combination (e.g. anti-Methicillin-resistant Staphylococcus aureus (MRSA) antibiotic) or switch for an antibiotic with a narrower spectrum of activity. 135 bacteremia were analysed, among them 110 (81%) corresponded to neutropenic patients at the time of de-escalation. The most frequent causes of neutropenia were hematopoietic stem cell transplantation (HSCT) conditioning regimens (n=44, 33%), induction chemotherapy (n= 32, 24%) and autologous HSCT conditioning regimens -for myeloma or lymphoma- (n = 21, 16%). De-escalation was carried out for 22% (n= 30) of the patients, including 16% (n= 18) of the neutropenic patients. The de-escalation was considered optimal in 83% (n= 25), corresponding to 100% (n= 12) and 72% (n= 13), respectively of the non-neutropenic and neutropenic patients. The safety of patients was not statistically different for those who underwent de-escalation : fever recurrence occurred in 27% (n= 8) and 28% (n= 29) respectively for patients with and without de-escalation (p = 1). A new episode of bacteremia occurred in respectively 1 and 8 patients with and without de-escalation (p = 0.25). Clinical complications within seven days of bacteremia required intensive care unit (ICU) admission for 10% (n= 3) of the patients with de-escalation and 3% (n= 3) of the patients without de-escalation, the de-escalation was not the cause of the complication. No difference concerning the ratio of de-escalation was observed between 2018 (22%, n= 17) and 2019 (23%, n= 13). When realized, the de-escalation for neutropenic patients was classified as optimal for respectively 60% (n= 6) and 88% (n= 7) patients in 2018 and 2019. The median duration and interquartile range of the antibiotic prescription for the bacteremia in days of therapy (DOT) were respectively 15 [12-23] and 13 [10-19] (p = 0,33) in 2018 and 2019. The average and standard deviation of DOT for carbapenem were respectively 4,9 (9,7) and 3,7 (6,3) in 2018 and 2019 (p = 0,88). De-escalation is possible and safe for selected patients with high risk febrile neutropenia with bloodstream infections under certain conditions, especially fever regression and the absence of focal infection, including any digestive symptoms and stage 3 or 4 mucositis. The common work between haematologist and IDS provides a more regulated de-escalation but sustained attempts are required and drastic changes in prescription can not be expected so quickly.
Document type :
Master Thesis
Complete list of metadata

Contributor : Jean-Hugues Morneau Connect in order to contact the contributor
Submitted on : Tuesday, September 28, 2021 - 9:17:20 AM
Last modification on : Thursday, September 30, 2021 - 3:24:06 AM
Long-term archiving on: : Wednesday, December 29, 2021 - 6:05:48 PM


Files produced by the author(s)


  • HAL Id : dumas-03355728, version 1


Mathieu Simonet. Early de-escalation of the empirical antibiotic in high-risk febrile neutropenia with bloodstream infection: a retrospective study. Human health and pathology. 2021. ⟨dumas-03355728⟩



Record views


Files downloads