Impact of Early Antibiotic De-Escalation on Mortality in Febrile Neutropenia in Patients with Hematologic Malignancy Receiving Treatment: A Meta-Analysis

Background

Patients with hematologic malignancies are at an elevated risk of infection due to their immunocompromised status. Infection is a leading cause of morbidity and mortality, particularly in those undergoing treatments like chemotherapy and hematopoietic stem cell transplantation (HSCT). Notably, about 80-90% of HSCT recipients develop febrile neutropenia (FN), which significantly increases their susceptibility to infections.

There is no consensus on the duration of broad-spectrum antibiotic (BSA) use for febrile neutropenia (FN) in current guidelines. Due to limited clinical data on the feasibility and safety of de-escalation of antibiotic therapy, we conducted a meta-analysis of retrospective studies to compare the mortality rate of early antibiotic de-escalation in patients with hematological malignancies who underwent treatment and developed FN before hematopoietic recovery versus those who continued BSA until clinical signs of hematopoietic recovery.

Methods

A comprehensive literature review was conducted up to January 2024. The search algorithm included a combination of MeSH terms, Emtree synonyms, and free words to ensure a thorough identification of studies comparing early de-escalation of BSA with standard use in the treatment of FN in patients with hematologic malignancies.

The inclusion criteria were: studies involving patients with malignancies post-treatment (e.g., chemotherapy, bone marrow transplant) with neutropenia (WBC < 3000 or ANC < 1500) and a febrile condition (101°F or 38.3°C). The interventions studied included antimicrobial stewardship practices focusing on the de-escalation or discontinuation of antibiotics. The primary outcome is mortality. Study designs included retrospective and prospective observational studies, and randomized controlled trials (RCTs) up to January 2024.

Odds Ratios (ORs) and Risk Ratios (RRs) with 95% confidence intervals (CIs) were extracted and pooled using random-effects and fixed-effects models. Heterogeneity was assessed using the Cochrane Q test and I² statistic. Sensitivity analyses and trial sequential analysis (TSA) were performed to assess robustness and required information size (RIS).

Results

Eight studies were included in this meta-analysis. The patient populations included adult patients with hematological malignancies, with AML being the most prevalent diagnosis in most studies. Patients received either HSCT or chemotherapy, and all developed FN of unknown origin during their treatment course. The timing of de-escalation varied among studies, with most implementing the strategy within 2-5 days of initial treatment, while the comparison groups continued the BSA until hematopoietic recovery.

The RR for mortality favored early de-escalation (RR 0.21, 95% CI 0.12-0.37), indicating a significant reduction in mortality risk. However, there was significant heterogeneity observed between the included studies (I²=59%, p=0.02). A sensitivity test was done excluding Bansal 2023 study with the result remaining significant (RR 0.39, 95% CI 0.20-0.75). A TSA was also conducted to assess the effect of early de-escalation on mortality risk in patients with hematologic malignancies, which revealed a significant protective effect of the intervention before reaching the Required Information Size (RIS) of 2745 participants. The cumulative Z-curve crossed the lower boundary of the trial sequential monitoring boundaries, yielding a pooled risk ratio of 0.19 (95% CI: 0.26-0.82, p=0.0083).

Conclusion

Previous studies have suggested that early de-escalation before hematopoietic recovery in HSCT patients might be feasible. However, most studies were small, despite showing the feasibility of early de-escalation. Previous studies suggested that the reduction of overall mortality was mainly due to the reduction in fatal respiratory and fungal infections.

Our study, with the strength of a larger sample size, suggested that early de-escalation of broad-spectrum antibiotics in hematologic cancer patients with FN significantly reduces mortality compared to prolonged BSA use. These findings advocate for revisiting current guidelines to incorporate de-escalation strategies, enhancing antibiotic stewardship while improving patient outcomes.

No relevant conflicts of interest to declare.