Background: Although recommended in patients with acute myeloblastic leukaemia (AML) after induction chemotherapy, real-life use of antifungal prophylaxis (AFP) is different among centres.

Materials and methods: This is an ancillary study to a randomized trial on intensive induction chemotherapy in AML patients (ALFA-0702/NCT00932412), where AFP with posaconazole was recommended. IFIs were graded by investigators and by central reviewers according to the revised EORTC definitions. Experts conclusions were compared to the investigators' ones.

Results: A total of 677 patients were included. Four AFP strategies were reported: Group-1: no AFP (n = 203, 30%), Group-2: posaconazole (n = 241, 36%), Group-3: posaconazole with other AFP (n = 142, 21%), Group-4: other AFP (n = 91, 13%). Experts graded more IFI than investigators: proven/probable IFI, 9.0% (n = 61) versus 6.2% (n = 42). The cumulative incidence at day60 of probable/proven IFI was 13.9% (Group-1); 7.9% (Group-2); 5.6% (Group-3); and 6.6% (Group-4). IFI onset was 26 (19-31) days after induction in Groups 2-3, versus 16 (9-25) days in Group 1 and 20 (12-24) days in Group 4 (P< .001). After a median follow-up of 27.5 months (0.4-73.4), the mortality rate was 38.3%, with 5.4% attributed to IFI. In multivariate analysis, IFI occurrence was an independent risk of death (HR5.63, 95%-CI 2.62-12.08, P< .001). EORTC recommendations were applied in only 57% of patients. In patients without IFI, the rate of AML complete remission was higher.

Conclusions: In AML patients, AFP delayed the onset of IFI in addition of decreasing their rate. The frequent misidentification of IFI impacts their appropriate management according to recommendations. hematological remission was more frequent in patients without IFI.