We limited our CIDCM cohort to the dilated phenotype because: (1) DCM is the most common morphological manifestation of chemotherapy-induced cardiotoxicity and is frequently responsive to advanced HF therapies, including VAD support, a key focus of the present study; (2) other phenotypes such as restrictive physiology frequently seen after chest irradiation or in mixed cancer therapy exposures carry distinct pathophysiological characteristics and outcome trajectories [14], which would hamper direct comparison with the NIDCM cohort and not amenable to VAD support; (3) prior UNOS analyses of adriamycin-associated cardiomyopathy similarly excluded hypertrophic and restrictive forms [13]. However, the UNOS database lacks detailed oncologic variables such as cancer stage, specific chemotherapeutic agents, cumulative dosing, radiation exposure, and remission duration prior to transplantation. As a result, distinguishing among specific cardiotoxic mechanisms (e.g., anthracyclines, HER2-targeted therapies, immune checkpoint inhibitors) was not feasible, and some degree of misclassification may have occurred.

We extracted demographic and clinical characteristics including age, sex, race/ethnicity, weight, body mass index (BMI), listing status, wait-list duration, prior oncologic diagnosis, Ebstein Bar Virus (EBV) and cytomegalovirus (CMV) serostatus, blood type, use of mechanical circulatory support (MCS) (including percutaneous and durable VADs), ventilatory/inotropic support, baseline hemodynamics at listing, induction immunosuppression at transplant and maintenance immunosuppression at discharge. Post-transplant outcomes were assessed using UNOS event and cause fields, specifically capturing instances of treated allograft rejection and de novo or recurrent malignancy.

Primary post-transplant outcomes included (1) post-transplant mortality, defined as time from transplant to death or graft failure requiring re-transplantation, censored at last follow-up; (2) incidence of treated allograft rejection; and (3) occurrence of new or relapsed malignancy following HT.

Continuous variables were summarized as medians with interquartile ranges (IQRs) and compared using the Wilcoxon–Mann–Whitney test. Categorical variables were presented as counts and percentages and compared using the Chi-square test or Fisher’s exact test, as appropriate. Post-transplant survival, and freedom from treated rejection were analyzed using Kaplan–Meier (K–M) curves to depict post-transplant mortality and rejection-free probabilities. Group comparisons were assessed using log-rank tests and Cox proportional hazards models, with hazard ratios and 95% confidence intervals reported. Covariates for multivariable models were selected a priori based on established associations with post-transplant outcomes and clinical relevance, including demographic factors (age, sex, race/ethnicity), markers of illness severity at transplant (mechanical circulatory support and/or inotrope use), and transplant-related variables. Analyses employed complete-case methodology; missing data were not imputed. Select laboratory and serologic variables, such as EBV and CMV serostatus, had higher rates of missingness in the pediatric cohort and were included only in age-appropriate models when biologically relevant and sufficiently complete. Multivariable analyses were performed using Cox proportional hazards regression models to estimate adjusted hazard ratios (HRs) with corresponding 95% confidence intervals (CIs). All tests were two-tailed, and a p-value 0.05 was considered statistically significant. Statistical analyses were performed using Stata version 19 (Stata Corp, College Station, TX, USA). Cause-specific mortality analyses were descriptive; competing-risks regression was not performed due to limited cause-specific event counts and the study’s primary focus on all-cause mortality. No formal adjustment for multiple comparisons was performed, as the primary analyses were prespecified and hypothesis-driven, focusing on clinically distinct outcomes (mortality, rejection, and malignancy). Results are therefore interpreted in the context of effect size, consistency across analyses, and biological plausibility.

Supplementary Fig. 1 illustrates the CONSORT-style flow diagram of patient selection from the UNOS registry. After application of predefined inclusion and exclusion criteria, the final study population comprised 52 pediatric and 475 adult recipients with CIDCM, and 2240 pediatric and 26,046 adult recipients with NIDCM. Table 1 summarizes the baseline characteristics of the pediatric cohorts. Compared with NIDCM, CIDCM recipients were older (median 10.9 vs. 6.1 years; p 0.001), more often female (66% vs. 49%; p = 0.04), and heavier (39.6 vs. 27.3 kg; p = 0.005). Ventilatory support was less common in CIDCM (26% vs. 46%; p = 0.018). Other listing characteristics, hemodynamics, and maintenance/induction immunosuppression were similar.

Baseline characteristics are summarized in Table 2 for the adult cohort. CIDCM recipients were younger (52.0 vs. 54.2 years; p 0.001), predominantly female (74% vs. 24%; p 0.001), with lower weight and BMI (both p 0.001). Listing urgency, blood type, and use of inotropes or ventilatory support were comparable. For all outcomes, p-values were additionally adjusted using the Benjamini–Hochberg false discovery rate (FDR) procedure; results are presented in Supplementary Table 1. Key findings were consistent after FDR adjustment. Sensitivity analyses demonstrated consistent effect estimates across model specifications (Supplementary Table 2).

MCS utilization was categorized as durable VAD therapy or temporary support. Among adults, durable VAD implantation was significantly less frequent in patients with CIDCM compared to those with NIDCM (39% vs. 49%, p = 0.001). In the pediatric cohort, durable VAD use was comparable between CIDCM and NIDCM groups (approximately 44% vs. 53%, p = 0.940). Temporary percutaneous MCS with intra-aortic balloon pump (IABP) was used equally among adult CIDCM and NIDCM patients (15% each), whereas IABP use in pediatric patients was rare. Extracorporeal membrane oxygenation (ECMO) was less commonly employed in adults with CIDCM (1%) compared to NIDCM (3%, p = 0.040), and among pediatric patients, ECMO was not utilized in the CIDCM group but was used in 3% of those with NIDCM.

Post-transplant survival outcomes are illustrated in Fig. 1. In the pediatric cohort, long-term survival did not differ significantly between patients with CIDCM and those with NIDCM, with 1-, 5-, and 10-year survival rates of 92%, 86%, and 76% for CIDCM versus 95%, 82%, and 68% for NIDCM (log-rank p = 0.951; hazard ratio [HR] 0.92, 95% confidence interval [CI] 0.76–1.34). In contrast, adult patients with CIDCM demonstrated significantly better survival compared to those with NIDCM, with respective 1-, 5-, and 10-year survival rates of 92%, 82%, and 68% versus 91%, 79%, and 59% (log-rank p = 0.018; HR 0.78, 95% CI 0.64–0.96).

Fig. 1.

Kaplan–Meier survival analysis of pediatric and adult heart transplant recipients with CIDCM and NIDCM. NIDCM, non-ischemic dilated cardiomyopathy; CIDCM, chemotherapy-induced dilated cardiomyopathy.

Multivariable Cox proportional hazards models evaluating all-cause mortality are shown in Fig. 2. In pediatric recipients, increased mortality risk was independently associated with age greater than 10 years at listing (HR 1.12 per year, p 0.0001), female sex (HR 1.23, p = 0.0081), Black race (HR 1.72, p 0.001), Hispanic ethnicity (HR 1.34, p = 0.0099), and use of a VAD at the time of transplant (HR 1.29, p = 0.003). Cardiomyopathy subtype (CIDCM vs. NIDCM) was not associated with mortality in pediatric recipients. Among adults, increased mortality risk was associated with age $\ge$50 years (HR 1.21, p 0.01), female sex (HR 1.20, p = 0.04), and Black race (HR 2.45, p 0.001). Pre-transplant inotrope use, VAD support, and cardiomyopathy subtype were not independently associated with mortality.

Fig. 2.

Multivariable analysis of predictors of mortality.

Cause-of-death distributions are summarized descriptively in Table 3A,3B. Overall, the relative proportions of death attributable to graft failure, infection, malignancy, and other causes were similar between cardiomyopathy subtypes. Among adult recipients, cerebrovascular deaths were significantly less frequent in the CIDCM group compared to those with NIDCM (12% vs. 21%, p = 0.0140), whereas no significant differences were observed in the pediatric cohort. These analyses were descriptive and not intended to estimate cumulative incidence in the presence of competing risks.

Treated rejection-free survival is illustrated in Fig. 3. In the pediatric cohort, rejection-free survival did not differ significantly between patients with CIDCM and those with NIDCM (log-rank p = 0.846; HR 0.92, 95% CI 0.41–2.07). Among adult transplant recipients, overall rejection-free survival was similar between groups, although a trend favoring improved outcomes in the CIDCM cohort was observed (log-rank p = 0.067; HR 0.83, 95% CI 0.68–1.01).

Fig. 3.

Kaplan-Meier analysis of rejection free survival of pediatric and adult heart transplant recipients with CIDCM and NIDCM. NIDCM, non-ischemic dilated cardiomyopathy; CIDCM, chemotherapy-induced dilated cardiomyopathy.

The observed cumulative incidence of post-transplant malignancy was low and did not differ significantly between groups during available follow-up; however, these findings primarily reflect early- to mid-term outcomes. Among pediatric CIDCM recipients, the most frequently documented antecedent malignancies were leukemia (15%) and lymphoma (12%). In adult CIDCM recipients, prior breast cancer (42.5%) and leukemia (25.3%) were the predominant diagnoses (Table 4A,4B). Post-transplant malignancy rates are presented in Fig. 4. In the pediatric cohort, the incidence of malignancy was comparable between patients with CIDCM and those with NIDCM, occurring in 5.4% and 8.3% of cases, respectively (p = 0.449). Among adults, malignancy rates were also similarly distributed, with 14.4% in the CIDCM group and 15.7% in the NIDCM group (p = 0.458). Person-year analyses reflected consistent patterns (Table 5A,5B,5C). Post-HT malignancy types were detailed in Table 6A,6B. Among pediatric recipients with CIDCM, post-transplant malignancies were predominantly post-transplant lymphoproliferative disorder (PTLD) or of unknown histology. In contrast, adult CIDCM recipients developed most skin cancers, followed by PTLD and other solid organ malignancies. Risk factors for post-transplant malignancy are illustrated in Fig. 5. In the pediatric cohort, age greater than 10 years at the time of listing was associated with a significantly increased risk of post-transplant malignancy (HR 1.50, p = 0.04), as was EBV seronegativity at the time of transplant (HR 2.10, p = 0.01). No other covariates reached statistical significance in this group. Among adult recipients, age $\ge$50 years was independently associated with elevated malignancy risk (HR 1.80, p = 0.02), while pre-transplant malignancy history, sex, and race did not demonstrate significant associations.

Fig. 4.

Probability of post-transplant malignancy in pediatric and adult patients with CIDCM vs. NIDCM (hazard ratios with 95% confidence intervals derived from Cox proportional hazards models). NIDCM, non-ischemic dilated cardiomyopathy; CIDCM, chemotherapy-induced dilated cardiomyopathy.

Fig. 5.

Predictors of post-transplant malignancy in pediatric and adult patients with CIDCM. CIDCM, chemotherapy-induced dilated cardiomyopathy; HT, heart transplantation.

Historically, cancer-related cardiomyopathy was associated with higher post-transplant risk due to concerns regarding recurrence, infection, and restrictive physiology [15]. Our findings, limited to the dilated phenotype and anthracycline chemotherapy, more accurately reflect patients who progress through contemporary VAD and HT pathways [16, 17, 18]. Improved outcomes likely result from advances in oncologic therapy, risk-based HF surveillance, and refined immunosuppression [19, 20]. The European Society of Cardiology (ESC) Cardio-Oncology Guidelines emphasize the role of multidisciplinary cardio-oncology teams, early HF detection, and careful surveillance in survivors receiving cardiotoxic therapy [21]. The International Society for Heart and Lung Transplantation (ISHLT) recommends individualized risk assessment in candidates with a history of malignancy, in collaboration with oncology specialists, to evaluate cancer-related survival and recurrence risk in the context of immunosuppression [17]. The Heart Failure Society of America (HFSA) statement similarly highlights integrated cardio-oncology care and the expanding role of advanced HF therapies for cancer survivors [5]. These guideline-driven shifts may have facilitated safer transplantation practices in this population. Durable VAD and ECMO use were lower in the CIDCM cohort compared to NIDCM, potentially reflecting earlier referral or more favorable pre-transplant status. This contrasts with earlier INTERMACS data (2006–2011), which reported higher rates of biventricular VAD use among CIDCM patients, likely due to a greater burden of right ventricular dysfunction [22]. However, in our analysis, we did not observe increased use RVADs in the CIDCM population. More recent analyses suggest similar outcomes between CIDCM and matched DCM cohorts, potentially reflecting improvements in management and risk assessment [23]. In alignment with the 2023 ISHLT guidelines, patients with a history of treated malignancy who are in sustained remission or considered disease-free may be eligible for VAD support either as destination therapy or as a bridge to transplant following comprehensive oncologic evaluation to assess recurrence risk and disease trajectory [16]. Differences in pre-transplant MCS utilization should be interpreted cautiously. Although adult recipients with CIDCM had lower rates of durable VAD support at transplantation, the UNOS registry does not capture referral timing, HF trajectory, or center-specific listing practices. As such, we cannot directly determine whether lower VAD utilization reflects earlier referral, differences in clinical severity, or provider decision-making. Observed differences may instead represent treatment-channeling and selection effects inherent to observational registry data rather than causal relationships.

Confounding by indication should be considered when interpreting differences in pre-transplant support strategies between CIDCM and NIDCM recipients. Patients with CIDCM are often younger, more frequently female, and may be listed earlier during advanced HF due to close oncology–cardiology follow-up and heightened surveillance, potentially reducing the need for bridging support. Conversely, clinicians may be more conservative with durable VAD implantation in cancer survivors because of perceived risks related to prior chemotherapy exposure (e.g., cytopenias), infection, bleeding, frailty, or uncertainty regarding cancer recurrence—factors that are not fully captured in UNOS. In addition, CIDCM can be accompanied by treatment-related comorbidities (e.g., renal dysfunction, pulmonary toxicity) that influence both listing urgency and device selection. Although we adjusted for VAD use and other key covariates in multivariable models, residual confounding related to unmeasured clinical factors and center-level practice variation cannot be excluded.

The comparable rates of post-HT malignancy between CIDCM and NIDCM groups are reassuring, suggesting that prior anthracycline chemotherapy exposure does not independently increase malignancy risk in modern practice. Oncogenesis after HT remains multifactorial, driven by immunosuppression intensity, duration, viral co-infections (EBV, CMV), and age-related susceptibility [20]. In our pediatric cohort, EBV seronegativity and age $>$10 years independently predicted post-transplant malignancy, consistent with established PTLD risk profiles [24, 25]. In adults, only age $\ge$50 years was predictive. Our findings, consistent with prior studies [26, 27], reinforce AHA and ISHLT recommendations favoring tailored, risk-based post-transplant surveillance rather than categorical exclusion of patients with a history of malignancy, while underscoring the need for improved early detection strategies [5, 17].

Women comprised 74% of adult CIDCM recipients, with breast cancer as the leading antecedent diagnosis—consistent with prior reports identifying breast cancer as the most common malignancy associated with chemotherapy-induced cardiomyopathy [28]. Up to 7.6% of breast cancer survivors exposed to anthracyclines or HER2-targeted agents develop HF, most often among young adults who remain candidates for advanced therapies [29]. In this context, HT represents a potential treatment for end-stage HF, although long-term oncologic and graft outcomes require further study. In contrast, among pediatric transplant recipients, female sex has been independently associated with worse post-transplant survival. Prior studies across both congenital and acquired heart diseases have similarly reported poorer outcomes among adolescent females, potentially reflecting a multifactorial interplay of higher rates of medication nonadherence, illness-related anxiety, and psychosocial stressors that impair self-management and complicate transitions of care [30, 31]. Although the UNOS registry lacks data on adherence, psychosocial variables, and hormonal influences, existing literature suggests that adolescent females may face greater mental health burdens and transition-related challenges, which could contribute to adverse outcomes [32]. These divergent findings between adults and children likely reflect age-specific psychosocial and developmental factors rather than intrinsic biological risk, underscoring the need for prospective, mechanistic studies to clarify sex-based vulnerabilities across the lifespan.

Our study demonstrated safety and efficacy of HT in CIDCM underscore the need to reconsider historical barriers to transplantation in cancer survivors. The ISHLT consensus statement now endorses individualized listing decisions, advocating multidisciplinary collaboration to assess recurrence risk and disease-free intervals rather than a fixed five-year remission requirement [17]. This approach promotes equitable access to HT and aligns with observed outcomes in our CIDCM cohort, in whom oncologic remission duration, younger age, and favorable health status likely contributed to superior survival and lower rejection rates.