Evaluation of a Safe Neutrophil Count for Cessation of IV Antibiotics and Early Hospital Discharge in Stable, Afebrile Patients Recovering after Acute Myeloid Leukaemia (AML) Therapy or an Autograft

Background

Currently there are no guidelines on a safe neutrophil count for intravenous antibiotic (IVAB) cessation and hospital discharge in haematology patients recovering after myelosuppressive chemotherapy complicated by febrile neutropenia (FN).

Aims

We assessed the safety in stable afebrile patients after recent FN of (i) appropriately stopping IVAB and (ii) early hospital discharge respectively within 24 hrs of absolute neutrophil count (ANC) recovery to ≥0.2x10^9/L. Appropriate cessation required a minimum of 3 days of IVAB and no focus of unresolved infection. Safety was defined as the absence of (i) fever recurrence after IVAB cessation and (ii) readmission in the 10-days post discharge for non-line related bacterial sepsis. Barriers to early discharge were also evaluated.

Methods

A retrospective, single centre audit on adult haematology patients admitted with AML (n=73 courses of induction or consolidation; 27 patients) or for an autograft (n=68 admissions;65 patients) between 2017-2019 and 2016-2017 respectively. Exclusion criteria included patients with secondary AML, reduced intensity AML chemotherapy (low dose cytarabine or azacitidine) and outpatient IVAB use post-discharge. Patients who continued on oral antibiotics as inpatients or on discharge were included in the analysis. Data were analysed with Mann Whitney U test, Chi square test and Fisher's exact test where appropriate.

Results

Both cohorts had a median age of 59 years. Autograft conditioning consisted mostly of high dose melphalan alone (57%) or with busulfan (7%) and BEAM (19%). All of the AML regimens contained either intermediate or high-dose cytarabine and/or an anthracycline. Most admissions (n=128; 91%) were complicated by FN, 32% (n=41) with positive blood cultures. Nearly half (n=61; 48%) of FN episodes ceased IVAB appropriately with 22/61 (36%) transitioned to oral antibiotics; another 19 (15%) episodes ceased IVAB prior to ANC ≥0.2. None of these 80 episodes had recurrent fever requiring IVAB resumption. IVAB were continued in the remaining 48 (37%) episodes, due to (a) unresolved fever (n=21) (b) recent bacteraemia or unresolved infective focus (n=17) or (c) empirically at physician discretion (n=10), for a median (range) duration of 3 (1-10), 3 (1-15) and 2 (1-5) days respectively. Thirty-seven (29%) of all FN episodes were discharged on oral antibiotics. Discharge within 24h from ANC≥0.2 occurred in 47% overall; more frequently in AML (60%) vs autograft (32%; p=0.001) patients, predominantly due to less unresolved gastrointestinal toxicity (5% vs 59% respectively). Other barriers to early discharge with an incidence of >5% included IVAB use, persistent fever, analgesia for mainly mucositis and transfusion requirement. Unplanned readmission rates were low (6% autograft, 3% AML) with none due to confirmed non-line related bacteraemia; these did not differ according to discharge ANC (≤0.5 or >0.5; p=0.217).

Conclusion

In afebrile, clinically stable AML and autograft patients without medico-social barriers to early discharge, IVAB can be ceased and hospital discharges safely done within 24h from recovery ANC≥0.2.