The study is a nationwide, multicenter, open label, randomized controlled trial.

A target population of 220 children in treatment for cancer with neutropenic fever and a neutrophil count below 0.5 × 10⁹ cells/L with expected duration for more than 7 days will be recruited during the first 48 hours of antibiotic treatment (24 months inclusion period). They will be randomized 1:1 as follows:

• Experimental group: Discontinuation of antibiotics, despite neutrophil count below 0.5 × 10⁹ cells/L, after 48 hours of apyrexia and clinical stability
• Control group: Discontinuation of antibiotics when neutrophil count is equal to or above 0.5 × 10⁹ cells/L and the child is afebrile and clinically stable (up to maximum of 14 days after apyrexia and clinical stability).

Primary endpoint is the number of days without antibiotic treatment in 28 days after treatment initiation. Secondary endpoints are crude mortality, severe adverse events, days with relapsing fever, and alterations of the microbiome.

BACKGROUND: In children with cancer, infections are a major cause of morbidity and mortality due to chemotherapy-related immunosuppression, comprising up to 70% of all treatment-related deaths. The relationship between neutropenia and life-threatening infection has been well established, and any signs of infection including fever has to be treated promptly with broad spectrum antibiotics. However, the duration of antibiotic therapy in pediatric cancer patients with febrile neutropenia without a known source is unknown. Traditionally, antibiotic therapy is continued until neutrophil recovery, due to fear of relapse of infection potentially as life-threatening sepsis, but scientific evidence supporting this strategy is limited. It has been debated that the strategy may be over-cautious, leading to unnecessarily prolonged antibiotic treatments with important side effects, including extended hospital admissions, adverse drug events, super-infection with fungi and multi-resistant bacteria, and thus affecting the children on both short and long term.

Studies in adults have shown that withdrawal of empirical antimicrobial therapy based on clinical assessment, despite severe neutropenia, can reduce antibiotic treatment significantly without compromising patient safety. This approach is warranted in children with cancer, but no randomized studies have been conducted in high risk children with prolonged neutropenia. A few studies have documented that early discontinuation of empirical antibiotics in low-risk children with short-lasting neutropenia is safe, but as stressed by recent international guidelines, the question of optimal duration of empirical antibiotics for high-risk children with prolonged bone marrow suppression continues to be a major research gap.

HYPOTHESIS: Empirical antibiotic therapy of febrile neutropenia can safely be discontinued after 48 hours of apyrexia and clinical stability, despite severe neutropenia, in high-risk children with expected prolonged neutropenia.

AIM: To establish evidence-based treatment strategies for children with febrile neutropenia by performing a randomized controlled trial investigating the efficacy and safety of early termination of empirical antibiotic therapy.

METHODS: Children treated at one of the 4 pediatric oncology departments in Denmark who are admitted with neutropenic fever will be evaluated according to current routines, including complete physical examination, assessment of severity signs and source, blood samples (biochemistry and hematology), blood cultures (collected from the central venous catheter), and additional samples from infected sites as clinically indicated. After obtaining cultures, empiric antibiotic therapy for neutropenic fever is started. The type of antibiotics will depend on current local guideline. Children who remain febrile despite empirical antibiotic therapy will be treated and investigated according to local and international guidelines.

Eligible subjects will be screened and included as described.

Included subjects will be followed for 28 days after initiation of empirical antibiotic treatment or until neutrophile recovery, whichever is longest.

They will be assessed at the following protocolled time points:

1. At 48 hours of apyrexia
2. At recovery of neutropenia
3. At 28 (±2) days (final assessment)
4. In the event of relapsing fever after the initial 48 hours of apyrexia during the follow-up period.

Each assessment will include a clinical interview, targeted physical examination and review of patient chart including biochemistry and hematology. A fecal and a pharyngeal swap will be collected and stored at the final assessment.

Children receiving antibiotics with no additional need for in-hospital treatment can receive their antibiotic treatment by continuous infusion on a portable pump and discharged to outpatient therapy, according to local department guidelines.

STATISTICS: The number of days free of antibiotics will be compared between the groups by linear regression analysis. A multivariate linear regression analysis will be carried out adjusting for age, gender, neutropenia duration, and underlying malignant disease and other significant risk factors.

Secondary outcomes will be compared by a non-inferiority assumption with an inferiority margin of 10%. All tests will be two-sided, with p values of 0.05 considered significant. An interim analysis will be conducted when 50% of the patients have been recruited to assess the safety of the study. A worst-case imputation method will be used for missing data in an exploratory sensitivity analysis.

Inclusion Criteria (for each episode):

1. patient with cancer aged 18 years or below,
2. absolute neutrophil count below 0.5x10⁹/L with expected duration of neutropenia for more than 7 days, and
3. a single temperature of at least 38.5°C, or a temperature above 38.0°C sustained over a 1-hour period (auricular, oral or rectal).

Exclusion Criteria (for each episode):

1. known etiology of fever, defined by a clinically significant positive culture (blood, urine, tracheal) collected during the feverish episode or a clinically documented focal infection despite negative cultures (e.g. pneumonia and cellulitis),
2. obvious non-infectious causes of fever (e.g. drug and transfusion related),
3. children requiring prophylactic antibacterial antibiotics according to protocol besides co-trimoxazole for Pneumocystis jirovecii after the febrile episode.