Near Universal National Uptake of Dexrazoxane in the Real-World Setting Associated with Reduced Cardiac ICU Care Requirements in Pediatric Acute Myeloid Leukemia

Background:
Cardiotoxicity is one of the most significant toxicities associated with pediatric acute myeloid leukemia (AML) treatment due to the high cumulative anthracycline doses required in the standard chemotherapy backbone. Dexrazoxane has proven to be an effective cardioprotectant to mitigate this toxicity, but its use has been historically limited with one 2013 study reporting <5% utilization in children with AML. The Children’s Oncology Group (COG) trial AAML1031 first presented data in 2018 demonstrating the benefit of dexrazoxane during AML therapy. As a result, COG subsequently mandated its use in pediatric AML trials starting in 2020. However, real-world uptake during this time and its potential impact on outcomes are unknown.

Methods:
We utilized a validated machine-learning algorithm that identified a cohort of pediatric AML patients from the Pediatric Health Information System (PHIS), newly diagnosed between 2010-2021 and representing 28 states. PHIS is a comparative database containing inpatient clinical and resource utilization data for 49 children’s tertiary hospitals across the United States. Descriptive statistics summarized dexrazoxane use over time. Primary analyses were performed at the patient-level and evaluated trends in the initiation of dexrazoxane with the first induction chemotherapy course (initiators vs. non-initiators). Secondary course-level analyses examined trends in dexrazoxane utilization at each anthracycline-containing chemotherapy cycle. Logistic regression models including individual cardiotoxicity risk factors and their interaction with year tested heterogeneity in temporal trends. We categorized time into three epochs based on dexrazoxane utilization: 2011-2015 (<25% dexrazoxane utilization), 2016-2018 (25-75% dexrazoxane utilization), 2019-2021 (>75% dexrazoxane utilization). Log binomial regressions assessed separately the impact of epoch and dexrazoxane exposure (initiators vs. non-initiators) on the following outcomes: cardiac ICU level of care utilization, beta blocker use, ACE-inhibitor use, and 6-month mortality (as a proxy for treatment-related mortality).

Results:
2211 pediatric patients (median age 9 [IQR 2-14], 23% Hispanic, 26% non-Hispanic non-white, 47% public-only insurance) contributed a total of 5129 anthracycline-containing courses to the analytic cohort. Overall dexrazoxane utilization increased rapidly from 11% in 2010-2015 to 44% in 2016-2018 to 92% in 2019-2021, with the most striking gains occurring from 2017-2020. Course-specific analyses suggested a practice shift starting in 2017 with consistent initiation of dexrazoxane at initial induction rather than waiting until later cycles with higher cumulative anthracycline burden. Analyses by region demonstrated the earlier adoption of dexrazoxane in the Northeast compared to other parts of the country (p<0.001). Trends over time in dexrazoxane uptake did not differ by known cardiotoxicity risk factors (age, p=0.29; sex, p=0.20; race/ethnicity, p=0.16). There were also no significant differences by presenting insurance status (p=0.50). With increasing dexrazoxane exposure over time, there was a significant decline in cardiac ICU level of care (33% in 2010-2015 vs. 18% in 2019-2021, p<0.001) and the use of ACE-inhibitors (20% in 2010-2015 vs. 12% in 2019-2021, p<0.001). Overall, dexrazoxane initiation was significantly associated with lower incidence of cardiac ICU level of care (RR 0.70, 95% CI 0.60-0.81, p<0.001) and ACE-inhibitor use (RR 0.67, 95% CI 0.55-0.82, p<0.001). There were no clear trends in the use of beta blockers over time. Six-month mortality was 6.9% for dexrazoxane initiators vs. 8.8% for non-initiators (p=0.21).

Conclusion:
In a national real-world cohort of pediatric AML patients, we observe a dramatic increase in dexrazoxane utilization to near universal uptake with an associated reduction in cardiotoxicity as proxied through reduced cardiac ICU level of care utilization and use of ACE-inhibitors. For a subset of this cohort, we are actively abstracting individual echocardiogram reports in addition to comprehensive clinical data. Future work will include analyses of dexrazoxane exposure with longitudinal measures of cardiac structure and function from the echocardiograms, durations of neutropenia, infectious complications, as well as event-free and overall survival beyond six months.