DISCUSSION
To date, the majority of investigations of IMI in children have focused
on infections resulting from a specific organism, such as invasive
aspergillosis, or on very high-risk groups, such as HSCT recipients.
Despite substantial evidence supporting the use of antifungal
prophylaxis with mold activity in at-risk pediatric cancer populations,
few studies have assessed the impact of appropriate prophylaxis on IMI
incidence outside of a clinical trial, and none have detailed
risk-adapted strategies for guidelines implementation and the subsequent
impact on rates of IMI. In this 15-year single-institution study, the
overall incidence of IMI in children treated for hematopoietic
malignancy with chemotherapy alone was 4.6%, ranging from a low of
1.7% for a diagnosis of lymphoma to a high of 6.4% for a diagnosis of
AML. This rate approximates the frequency of IMI previously reported in
U.S.- and European-based childhood cancer
centers.10,11
Our results suggest a higher risk for IMI in children with Hispanic
ethnicity. Factors contributing to the higher rates of IMI observed in
Hispanics are likely multifactorial. A higher than expected incidence of
infections with Histoplasma, Coccidioides , and Blastomycesspecies has been observed in persons of African, Native American, and
Asian descent,12suggesting a role for genetic
predisposition.13Further investigation to elucidate the genetic and environmental factors
that underlie the excess IMI risk observed among Hispanic children is
needed.
We also demonstrate clinical benefit to implementation of a standardized
approach to antifungal prophylaxis in high risk childhood cancer
populations, evidenced by risk reduction for proven/probable IMI after
algorithm implementation. Our results validate and extend findings from
a prior, smaller single-institution study that also reported a reduction
in IMI incidence after applying risk-adapted IMI prophylaxis in children
with leukemia.14Antifungal prophylaxis is routinely prescribed to prevent IMI in
children with AML, who are at high risk of infection due to prolonged
periods of profound neutropenia induced by intensive chemotherapy. By
broadening prophylaxis to include anti-mold coverage, we observed a
reduction in the incidence of IMI in patients with AML treated at TCH by
60%. Some adult studies have observed an increase in Mucor spp.
infections with use of broadened, anti-mold
prophylaxis.15,16Although it is difficult to assess the effect of broadening prophylaxis
in a sample of this size, we did observe an increase in RhizopusIMI post-implementation, one of the species in the Mucorales order. Of
the proven IMI cases, Rhizopus or Mucor spp. were isolated
in five out of 40 cases from 2006-2015 (13%), vs. four out of 13 cases
between 2016-2020 (31%). Overall, concurrent to eliminating IMIs
related to Aspergillus , we saw an increase in IMIs due toRhizopus , Fusarium , andTrichosporon 17 that did not impact overall
survival, but that did correspond with a higher rate of admission to the
intensive care unit after 2016, and, for those who died, a higher rate
of active fungal disease at time of death. Proven or probable IMI had a
substantial impact on survival: only 27% of children with AML and 67%
of children with ALL and concurrent IMI survived, compared with current
average U.S. survival rates that approach 70% and 90%, for AML and ALL
respectively.18,19
The Children’s Oncology Group recommends primary antifungal prophylaxis
for children receiving HSCT as well as for patients with AML or
myelodysplastic syndrome, with moderate evidence also suggesting benefit
to primary prophylaxis for children with anticipated neutropenia greater
than 7 days duration.20As noted above, a recent systematic review produced consensus guidelines
for systemic antifungal prophylaxis in children with cancer and who are
treated with HSCT that extends these recommendations to include
mold-active antifungal prophylaxis for children treated for AML, and
consideration of mold-active antifungal prophylaxis in children with
newly diagnosed or relapsed ALL who are at high risk for fungal
infection.8 Though our
antifungal prophylaxis algorithm (Figure 1 ) was developed prior
to the publication of these consensus guidelines, the algorithm largely
adheres to the guideline principles. Specifically, according to our
algorithm, antifungal prophylaxis with agents that have anti-mold
activity is initiated for all patients diagnosed with AML, relapsed ALL
or AML, and initiated during all phases of ALL therapy when inpatient
admission through count recovery is recommended. Children with lymphoma
are not prescribed antifungal prophylaxis. In addition, the TCH
algorithm suggests appropriate anti-mold coverage to minimize potential
interactions with chemotherapy agents routinely employed in subsets of
ALL and AML, and considers variable absorption and pharmacokinetics of
voriconazole and posaconazole in
children.21 Exceptions
to the published guidelines include our use of antifungal agents that
are not mold-active (e.g. fluconazole) for specific moderate risk
populations, such as during intensive treatment courses for children
with Down syndrome associated ALL, infant ALL, and T cell ALL, as well
as during periods of hyperglycemia for patients who develop
steroid-induced diabetes during treatment. Importantly, the algorithm is
a ‘living’ document, modified as treatment protocols are retired or
initiated, as changes are made to the hospital formulary, or as new
categories of risk are identified.
Limitations to our study include the anticipated constraints of a
retrospective chart review, and the study restriction to a single
institution. The relative rarity of this outcome precluded our ability
to assess associations and outcomes for individual mold species. We also
had insufficient data available to compare antimicrobial resistance
patterns in isolated molds pre and post-algorithm implementation. Last,
while our antifungal prophylaxis algorithm is evidence-based and largely
aligned with current consensus guidelines, adaptations were made in
consideration of prevalent fungal pathogens at our institution,
additional populations perceived as high risk, and clinical practice
preferences for intravenous vs. oral antimicrobials depending on
inpatient or outpatient status. These adaptations should be considered
when assessing the generalizability of our approach.
As noted by Lehrnbecher et al., the practical application of clinical
guidelines requires an appreciation of local IFD epidemiology and the
appropriate engagement and education of key
stakeholders.8 Here, we
describe an approach for the clinical application of evidence-based
recommendations to provide antifungal prophylaxis in children with
hematologic malignancies that is informed by host and treatment factors,
local IMI epidemiology, and input from a multidisciplinary team.
Further, we demonstrate a reduction in IMI incidence as clinical
evidence of the efficacy of our approach. Ongoing efforts include
continued prospective monitoring of fungal species isolated from at-risk
populations to assess susceptibility patterns, outcomes, and the need
for further algorithm modifications.
Acknowledgements: The study team would like to thank the
patients and families who contributed data to this research. This work
was supported by a St. Baldrick’s Foundation Consortium Grant to Karen
R. Rabin (Reducing Ethnic Disparities in Acute Leukemia), and by an
American Society of Hematology Minority Medical Student Award to Ashley
Ikwuezunma (mentors Michael E. Scheurer and Maria M. Gramatges).
Data Sharing: De-identified data that support the findings of
this study are available on request from the corresponding author. The
data are not publicly available due to privacy or ethical restrictions.