De-escalation and discontinuation strategies in high-risk neutropenic patients: an interrupted time series analyses of antimicrobial consumption and impact on outcome

Abstract

Febrile neutropenia (FN) is the main reason for antibiotic prescription in hematology wards where, on the other hand, antibiotic stewardship (AS) is poorly explored. The objectives of the present study were to evaluate (1) the impact of an AS intervention on antibiotic consumption and (2) the applicability and acceptance rate of the intervention and its clinical impact. A persuasive AS intervention based on European Conference on Infection in Leukaemia (ECIL) guidelines for FN was implemented in a high-risk hematology ward in a tertiary referral public university hospital. This included the creation and diffusion of flow charts on de-escalation and discontinuation of antibiotics for FN, and the introduction in the team of a doctor dedicated to the implementation of flow charts and to antibiotic prescription revision. All consecutive patients receiving antibiotics during hospitalization were included. A segmented linear regression model was performed for the evaluation of antibiotic consumption, taking into account 1-year pre-intervention period and 6-month intervention period. Overall, 137 consecutive antibiotic prescriptions were re-evaluated, 100 prescriptions were for FN. A significant reduction of the level of carbapenem consumption was observed during the intervention period (level change (estimate coefficient ± standard error) = - 135.28 ± 59.49; p = 0.04). Applicability and acceptability of flow charts were high. No differences in terms of intensive care unit transfers, bacteremia incidence, and mortality were found. A persuasive AS intervention in hematology significantly reduced carbapenem consumption without affecting outcome and was well accepted. This should encourage further applications of ECIL guidelines for FN.

Keywords: Antibiotic consumption; Antimicrobial stewardship; Carbapenem; De-escalation; Febrile neutropenia; Interrupted time series.