Breakthrough Invasive Fungal Infections (bIFI) Are Uncommon in Patients with Newly Diagnosed Acute Leukemia Receiving Primary Antifungal Prophylaxis

Introduction: Mold-active primary antifungal prophylaxis (PAP) is widely recommended in neutropenic patients (pts) with newly diagnosed acute myeloid leukemia (AML) and high-risk myelodysplastic syndrome (MDS) who undergo remission-induction chemotherapy (RIC). Posaconazole (PCZ) prophylaxis resulted in fewer invasive fungal infections (IFIs) when compared to fluconazole and was associated with a survival advantage in this population (Cornely et al, 2007). Similarly, pts with acute lymphoblastic leukemia (ALL) undergoing RIC are also at risk of IFI due to prolonged neutropenia. Although PCZ is the preferred agent for PAP, the incorporation of targeted agents into acute leukemia therapy calls for more individualized choices in PAP. Other mold-active agents including voriconazole (VCZ) and isavuconazole (ISA) or the echinocandins are alternatives which may be preferred in individual settings due to variations in toxicity, patient co-morbidities, drug interactions, and cost. Little contemporary data exists to compare the incidence of breakthrough IFI (bIFIs) in pts with AML or ALL receiving PCZ, VCZ, or ISA as prophylaxis during RIC.

Methods: We reviewed the medical records of all consecutive pts with newly diagnosed AML/MDS or ALL treated at our institution from 3/2016-7/2019. Included pts received high-intensity chemotherapy, or a lower-intensity venetoclax (VEN)-containing regimen, for RIC. Therapy with high-dose (> 1g/m²/day) cytarabine (HiDAC), continuous cytarabine plus an anthracycline (3+7), or HyperCVAD was considered high-intensity therapy. Patients receiving PCZ, VCZ, or ISA for > 5 days beginning during induction therapy were included. Baseline evidence of prior mold infection and treatment with a concomitant echinocandin were not allowed. Echinocandin use preceding mold-active PAP was allowed, however prior use of an amphotericin B product was not. bIFI were defined according to ECMM criteria (Cornely et al, 2019).

Results: We identified 232 pts with AML/MDS (n=186), ALL (n=43), and biphenotypic leukemia (n=3). Among the AML/MDS pts, 31% (n=57) received a lower-intensity VEN-containing regimen, while 69% (n=129) received a high-intensity regimen with or without VEN. Nearly all pts with ALL and biphenotypic leukemia received high-intensity RIC with HyperCVAD or a HiDAC containing regimen. Of the 232 pts, PCZ (n=111), VCZ (n=84), or ISA (n=37) were used as PAP, respectively (Table 1). Most pts (n=157; 68%) received an echinocandin for a median of 6 days (range, 0-38), prior to transitioning to a mold-active triazole. Ten (4.3%) pts had a bIFI (6 proven, 1 probable, 3 possible) during induction therapy while receiving PAP (Table 2) including 9 (4.8%) pts with AML/MDS and 1 (2.3%) patient with ALL. An equal number of pts with bIFI were receiving lower-intensity, VEN-based therapy, or high-intensity therapy. Among the 84 pts receiving VCZ, 4 (4.8%) had a bIFI (4 proven); 3 pts (2.7%) receiving PCZ had a bIFI (2 proven, 1 possible); 3 pts (8.1%) receiving ISA had a bIFI (1 probable, 2 possible). C. glabrata (n=3), and C. krusei, Cryptococcus spp. and Fusarium spp. (one each) accounted for the 6 proven bIFIs. One patient had both C. glabrata and C. krusei fungemia. The probable bIFI was pneumonia with a positive Aspergillus GM from BAL. The 3 possible bIFI were pneumonia (n=2) and sinusitis (n=1). Eight pts (80%) were neutropenic (ANC < 500 cells/mm³) for >14 days at the time of bIFI, 1 pt was neutropenic for >7 days, and 1 pt had ANC > 500 cells/mm³. Seven pts with bIFI received a prior echinocandin for a median of 3 days (range, 0-15) prior to initiation of triazole PAP. Seven pts were neutropenic for < 7 days (n=2), 7-14 days (n=3), or > 14 days (n=2) at the time of azole initiation. bIFI occurred after a median of 20 days (range, 5-72) of azole PAP and a median of 24 days (range, 12-71) from the initiation of RIC. One patient with bIFI deceased within 42 days of starting RIC and did not achieve a response after RIC (bIFI-related mortality: 0.44%).

Conclusion: The incidence (<5%) and mortality (< 0.5%) due to bIFI in a contemporary cohort of pts with newly diagnosed acute leukemia receiving PAP is low. bIFI occurred late in induction therapy and most often in pts with > 14 days of neutropenia. Prophylaxis with VCZ, PCZ, or ISA, with or without a prior echinocandin, appear to be comparable options for PAP in pts with newly diagnosed AML or ALL undergoing RIC.