Empirical antifungal treatment is the gold standard for patients who are neutropenic and have persistent fever under broad-spectrum antibiotics. The rational is that fungal infections are difficult to early diagnose, and are life-threatening. Historical trials have shown a small benefit of survival when this strategy is used. According to the drug usde for this strategy, safety and costs may be concerns. However, since this routine practice has been implemented in hematology, new non-invasive biological diagnostic methods are available to early diagnose fungal infections, such as galactomannan antigenemia for aspergillosis. The goal of our study is to show that limiting the administration of antifungals in this setting to patients with clinical foci of infection, or to patients with a positive galactomannan antigenemia reduces the risk of toxicity of the antifungal drug, and has no impact on the overall mortality of patients treated with chemotherapy for hematologic malignancies.

Patients are eligible if they have an hematologic malignancy, and receive chemotherapy with an expected neutropenic phase of > 10 days. Patients are randomized according to a 1:1 ratio to receive either the usual empirical strategy (antifungals are introduced if they have persistent fever after 4 days of broad-spectrum antibacterials) or the pre-empirical strategy (administration of antifungals is limited to patients with pneumonia, septic shock, sinusitis, grade 3 mucositis, aspergillus colonization, liver or splenic abscesses, or positive galactomannan antigenemia). The antifungals administered are deoxycholate amphotericin B or liposome amphotericin B, according to the creatinin clearance. This strategy is applied during the first 14 days of persistent fever, then the therapy is left at the discretion of the investigator. The primary endpoint is survival at neutrophil recovery, or, in case of persistent neutropenia, at day 60 at the latest. Secondary objectives are the incidence of invasive fungal infections (IFI), IFI-free survival, number of febrile days, and renal function at study completion.

Inclusion Criteria:

• Malignant Hemopathy
• Induction or consolidation phase of chemotherapy, with expected neutropenia (< 500/mm3) during at least 10 days
• Hospitalisation during aplasia

Exclusion Criteria:

• allogeneic haematopoietic stem cell transplants
• Previous fungal infection according to EORTC-MSG criteria
• Active fungal infection according to EORTC-MSG criteria
• Previous anaphylactic intolerance to polyenes
• known aspergillosis infection
• Sepsis
• Pneumopathy