Risk Factors and Predictors of Outcomes after Invasive Fungal Infection Among Patients with Acute Myeloid Leukemia

Introduction: Patients (pts) with acute myeloid leukemia (AML) incur an elevated risk of infections related to myelosupressive therapy. Invasive fungal infections (IFIs) are amongst the most serious encountered, given the associated high rate of mortality (45% for candidiasis/aspergillosis and up to 70% for fusariosis and mucormycosis)(Mishra et al., Indian J Hematol Blood Transfus. 2020). In this study, we sought to describe the rate of IFI in this population, compare outcomes between pts with AML based on particular factors relating to IFI, antifungal prophylaxis (AFP), and identifying predictors of more favorable outcome.

Methods: We conducted a single-center, retrospective, observational study of pts diagnosed with AML at Yale Cancer Center between 2013-2024. IFIs were characterized based on consensus definitions from the Mycoses Study Group Education and Research Consortium (Donnelly et al., Clin Infect Dis. 2019). Chi-square analysis and Wilcoxon/Kruskal-Wallis rank-sum tests were used for categorical and continuous variables, respectively. Survival analysis was performed with the Kaplan-Meier method. Statistical significance was established with a p-value <0.05.

Results: Among the 330 pts included, the median age was 62 (interquartile range [IQR] 52-70), with 58.2% being male and 79.7% being white. IFIs were identified in 72 pts (21.8%). When comparing pts with and without IFI, there were no differences in any baseline characteristics including age, sex, ethnicity, cardiac/pulmonary comorbidity, frontline AML regimen, or initial AFP agent chosen between groups. Additionally, there were no differences in disease biology including AML subtype, ELN 2022 risk, karyotype, and all single-gene mutations analyzed with the lone exception of PHF6, which was much more common amongst IFI pts (7.1% vs 0.4%, p=0.001).

When examining which AFP was at steady-state at IFI diagnosis the plurality of pts were on voriconazole (n=28) with n=20 on no AFP, n=12 on isavuconazole, and n=10 on anidulafungin. Additionally, only 32 pts in the IFI group were on a single AFP agent during the 30 days preceding IFI. Within the IFI group, half of pts (n=34) had possible IFI, with n=11 probable IFI and n=27 proven IFI. A trend was noted between IFI characterization and increasing days of neutropenia preceding IFI diagnosis, with proven IFIs averaging 34.9 days ANC <0.2 before diagnosis relative to 25.3 and 14.1 days for probable and possible IFI, respectively (p=0.096). Amongst cases of IFI, Aspergillus was the most common pathogen (n=13, 18%), with 10 (14%) cases of Candida, 5 (7%) cases of Fusarium, and 6 (8%) cases of Mucorales. 40 IFI patients (55.6%) did not have an infectious agent identified, including all possible IFIs. Mucorales was associated with the longest period of neutropenia (ANC <0.2) before diagnosis when compared to all other IFIs (16 days (14-49) vs 10 days (1-18), p=0.199).

IFI diagnosis was associated with early mortality, specifically 29% 30-day mortality and 50% 90-day mortality. Pts developing IFI had shorter OS when compared with pts not developing an IFI (13.8 months [95% CI: 10.4-17.5] vs 18 months [13.6-23.7], p=0.095), although this did not meet statistical significance and was not impacted by censoring at time of allogeneic stem cell transplantation (p=0.68). The higher rate of 30-day mortality in the non-IFI group (8.1% vs 4.2%) may have influenced these findings. However, the rates of mortality were not different based on IFI characterization by consensus definition (p=0.87) or by particular species amongst the 27 pts with proven IFI and microbiologic diagnosis (p=0.53). When analyzing risk factors for mortality within 90 days, relapsed/refractory disease (rrAML) was associated with poorer post-IFI survival compared to those diagnosed peri-induction or post-alloSCT (p=0.048). This may be related to an enrichment for non-adverse disease risk by ELN 2022 that was associated with better overall post-IFI survival (8.5 months (1.3-34.9) vs 1.8 months (IQR 0.5-5.4), p=0.0005).

Conclusions: IFIs are a cause of significant mortality in pts with AML apparently independent of any patient-, disease-, or treatment-specific characteristics, except sub-optimal time on AFP. Adverse ELN 2022 risk and rrAML were highly predictive of mortality after IFI diagnosis. These results support that AFP effectively prevents IFI amongst patients with deep, protracted neutropenia.