This randomised controlled trial will determine the non-inferiority of stopping empiric antibiotics prior to absolute neutrophil count (ANC) recovery (Early Stopping) versus stopping antibiotics upon ANC recovery (Standard of Care/ Late Stopping) , in children with cancer and high-risk febrile neutropenia (FN).

Febrile neutropenia (FN) is a common complication of childhood cancer treatment and a leading cause of hospital admission and antibiotic exposure. Management typically involves broad-spectrum antibiotics until resolution of fever and absolute neutrophil count (ANC) recovery >500 cells/mm3. However, despite the frequency with which FN occurs, evidence to guide duration of antibiotics is limited to observational studies and small randomised controlled trials.Current international clinical guidelines provide conflicting recommendations on when to cease empiric antibiotics for FN. Early cessation of antibiotics in FN may translate to reduced antibiotic exposure and limit potential harms including drug side-effects, antimicrobial resistance, Clostridioides difficile infection and microbiome disruption. This randomised controlled trial will use a composite endpoint of fever recurrence, physiological instability, new bacteremia, intensive care admission and death to determine the non-inferiority of stopping antibiotics prior to ANC recovery compared with standard of care (SOC), in children with cancer and high-risk FN. Adopting a health informatics approach, patient identification, consent, randomisation and reporting of outcomes will be embedded within the electronic medical record (EMR). Children with high-risk FN who have been afebrile and clinically stable for at least 48 hours will be randomised to cease antibiotics or continue SOC. Data on primary outcomes, antibiotic duration, length of stay, C. difficile infection and antimicrobial resistance will be automatically collected by the EMR. This is the first study of its kind in children with high-risk FN and adopts a novel embedded trial design. Results will inform optimal antibiotic duration in FN, potentially reducing unnecessary antibiotic exposure.

Inclusion Criteria:

1. Diagnosis of

   • Acute myeloid leukemia (AML) or acute lymphoblastic leukaemia (ALL) in dose-intensive phases of induction/re-induction, intensification or consolidation or
   • ALL or acute lymphoblastic lymphoma patients on a TOT17 protocol or
   • Any disease within 100 days of allogeneic or autologous HSCT

2. Neutropenia (<500 cells/mm3)

3. Afebrile (temperature <38.0°C) period for at least 48 hours and no more than 96 hours after at least one temperature measured by axillary or tympanic thermometer (≥38.0°C)

4. Commenced on empiric FN antibiotics (any of piperacillin-tazobactam, cefepime, ceftazidime or vancomycin and ciprofloxacin)

Exclusion Criteria:

1. Prolonged febrile neutropenia (documented daily temperature ≥38.0°C for ≥5 days)
2. Documented positive blood culture since onset of FN episode and prior to randomisation
3. Documented other infection (microbiologically or clinically documented) requiring antibiotic treatment since onset of FN episode and prior to randomisation
4. Admitted to the ICU at the time of randomisation
5. Clinical instability (One or more conscious state, respiratory rate, blood pressure, heart rate or oxygen saturations in MET criteria OR two or more respiratory rate, blood pressure, heart rate or oxygen saturations simultaneously (+/- 4 hrs) in the clinical review criteria in 48 hours prior to randomisation)
6. Within 28 days of last randomisation

Beginning 6 months following analysis and publication of the primary outcome, data will be made available long-term for use by future researchers from a recognised research institution whose proposed use of the data has been ethically reviewed and approved by the Murdoch Children's Research Institute's (MCRI's) independent data use review committee (not including trial sponsor-investigator) and who accept MCRI's conditions, under a collaborator agreement, for accessing all of the available participant data collected during the trial (after full deidentification).

6 months following analysis and publication of the primary outcome

Requests for access to previously published anonymised datasets by the scientific community will be reviewed and determined by an independent committee within the MCRI and in accordance with institute policy.