Initiation of ACE inhibitors or beta-blockers during frontline therapy for pediatric Acute Myeloid Leukemia and the impacts on cardiotoxicity incidence and survival outcomes: A target trial emulation

Abstract

Abstract Background Anthracycline (ATC)-related cardiotoxicity is a well-recognized complication in pediatric acute myeloid leukemia (AML) and has been associated with worse survival outcomes. ATCs exert their cardiotoxic effects through multiple mechanisms, including induction of oxidative stress, DNA damage, and maladaptive cardiac remodeling, ultimately leading to progressive left ventricular systolic dysfunction (LVSD). Angiotensin-converting enzyme inhibitors (ACEi) and beta-blockers (BB) may mitigate these effects through antioxidant effects and neurohormonal regulation. However, data on the effectiveness of early ACEi/BB in preventing LVSD and the associated impacts on survival outcomes in pediatric AML are limited. Using a trial emulation approach, we evaluated whether initiation of ACEi and BB during frontline chemotherapy may reduce the incidence of LVSD and improve survival outcomes. Methods We used data from the REAL–AML Cohort, a retrospective cohort of pediatric AML patients aged <19 years at initial diagnosis who were treated at 17 U.S. institutions from 2011 onward. Information on ACEi and BB administrations during frontline AML therapy was obtained via a merge with the Pediatric Health Information System database (PHIS). We emulated an open-label, randomized, parallel-group trial comparing three treatment strategies: (1) initiation of BB, (2) initiation of ACEi, and (3) no initiation of either ACEi or BB during frontline chemotherapy. Medications were required to be initiated before evidence of CTCAE grade 2 or higher LVSD, defined as ejection fraction <50% or shortening fraction <24%. Patients with de novo AML were followed from the initiation of chemotherapy until the earliest occurrence of the outcome of interest, last contact, or 5-years of follow-up. The primary outcome was the development of LVSD based on CTCAE definitions, and secondary outcomes included overall survival (OS) and relapse-free survival (RFS). ACEi/BB treatment assignment was based on the first medication received. We applied clone-censor-weight emulation methodology to address confounding and avoid immortal time bias. Specifically, each patient was cloned into three records, corresponding to the three treatment strategies, resulting in identical baseline characteristics across arms. Clones were then censored at the time they deviated from their assigned treatment strategy. We examined the effects of ACEi and BB using marginal structural models (MSM) via pooled logistic regression to approximate hazard ratios (aHR) and to estimate 5-year risks. The MSM incorporated inverse-probability-of-censoring weights to account for artificial censoring introduced by the aforementioned cloning procedure, adjusting for both time-fixed (age, sex, race/ethnicity, cytomolecular risk, insurance, presentation acuity, trial enrollment, dexrazoxane use in induction I) and time-varying confounders (blood pressure, bacteremia, ICU-level requirements, empiric/definitive anti-infective use). The 95% confidence interval (CI) was constructed using non-parametric bootstrap. Results A total of 908 patients were included in the analytic cohort, with median follow-up of 51.8 months (IQR: 25.3–60.0). During frontline therapy and prior to the onset of grade 2+ LVSD, 8.4% initiated ACEi and 7.4% initiated BB. Grade 2+ LVSD occurred in 164 patients (18.1%), relapse in 285 (31.4%), and all-cause mortality in 255 (28.1%). In adjusted analyses, patients who initiated BB had a 15% lower hazard of LVSD compared to those unexposed to either ACEi or BB (aHR = 0.85, 95% CI: 0.68 to 0.93), with a 5-year risk difference (RD) of -4.2% (95% CI: -6.8 to -1.7). Patients who initiated BB during frontline therapy also had superior OS (aHR of 0.58, 95% CI: 0.36 to 0.96) corresponding to an 11% absolute reduction in 5-year mortality. However, no significant effect of BB was found for RFS. Additionally, there were no meaningful differences in the three outcomes with initiation of ACEi compared to no exposure. Conclusion Our findings suggest a potential benefit of early initiation of BB in preventing cardiotoxicity and improving OS among pediatric AML patients. In contrast, ACEi showed no measurable benefit across any of the evaluated outcomes. These results support further investigation into the role of BB as a cardioprotective strategy during frontline chemotherapy, including evaluation of optimal initiation timing and the comparative effectiveness of different BB agents.