Temporal trends and post-transplant outcomes in anthracycline-induced cardiomyopathy: A contemporary United Network for Organ Sharing analysis

Abstract

Background

Outcomes after heart transplantation for anthracycline-induced cardiomyopathy (AICM) remain incompletely characterized, particularly compared with idiopathic dilated cardiomyopathy (DCM) and ischemic cardiomyopathy (ICM).

Methods

We analyzed adult first-time isolated heart transplant recipients in the Organ Procurement and Transplantation Network/United Network for Organ Sharing registry from January 1, 2000, to September 25, 2025. Recipients were grouped as AICM, idiopathic DCM, or ICM by registry diagnosis. The primary outcome was all-cause post-transplant mortality. Secondary outcomes included graft failure, post-transplant dialysis, stroke, pacemaker implantation, treated rejection, and length of stay.

Results

Compared with idiopathic DCM, AICM was not associated with higher adjusted mortality (HR 1.11, 95% CI 0.98-1.25; p=0.111) or graft failure (HR 1.09, 95% CI 0.96-1.23; p=0.185) but was associated with post-transplant dialysis (OR 1.31, 95% CI 1.03-1.67; p=0.030) and longer index hospitalization (IRR 1.12, 95% CI 1.06-1.18; p<0.001). AICM was also associated with higher malignancy-related mortality compared with idiopathic DCM (sHR 2.27, 95% CI 1.66-3.09; p<0.001) and ICM (sHR 1.70, 95% CI 1.24-2.33; p<0.001).

Conclusions

Among selected adult heart transplant recipients, anthracycline-induced cardiomyopathy was not associated with worse post-transplant survival or graft failure compared with idiopathic DCM. However, AICM recipients had higher adjusted odds of post-transplant dialysis, longer index hospitalization, and a greater burden of malignancy-related death, highlighting the need for individualized oncologic and renal-risk assessment in transplant candidates with prior anthracycline exposure.

Although anthracycline use has declined, these agents remain widely used, with clinical utility limited by cumulative dose-dependent cardiotoxicity that may progress to overt heart failure.¹ As advances in cancer detection and treatment continue to expand the population of long-term survivors, anthracycline-induced cardiomyopathy (AICM) has emerged as an increasingly recognized contributor to advanced heart failure, accounting for approximately 2% to 3% of patients undergoing advanced heart failure therapies in the U.S.2, 3 In a subset of affected patients, AICM progresses to end-stage systolic dysfunction requiring durable mechanical circulatory support (MCS) or heart transplantation.4, 5, 6

Transplantation in patients with AICM requires risk assessment beyond standard heart failure criteria. This reflects a competing oncologic mortality risk not captured by cardiac evaluation alone, compounded by the potential for post-transplant immunosuppression to promote malignancy recurrence or de novo cancer through impaired tumor immune surveillance.⁷ Oncologic prognosis, cancer-free interval, recurrence risk, and chronic immunosuppression must therefore be considered alongside expected transplant benefit.5, 8, 9 Although prior studies suggest favorable post-transplant outcomes in selected patients with AICM, outcomes following the 2018 Organ Procurement and Transplantation Network/United Network for Organ Sharing (OPTN/UNOS) adult heart allocation policy change remain incompletely defined.4, 10, 11

Additionally, prior anthracycline exposure may confer residual organ toxicity, including subclinical renal and cardiac injury that could influence perioperative morbidity. However, registry-level data remain limited, restricting the ability to individualize risk assessment in this population. These considerations, combined with the distinct demographic profile of AICM patients, underscore the need for AICM-specific outcome data to inform transplant candidacy decisions. Early post-transplant morbidity, graft-failure patterns, and allocation-era interactions have not been comprehensively assessed in a contemporary national cohort. Therefore, we used the UNOS transplant registry to evaluate temporal trends, post-transplant survival, graft failure, and morbidity among adult heart recipients with AICM.

Methods

Data source

We performed a retrospective cohort study using Standard Transplant Analysis and Research files from the OPTN. The OPTN is overseen by the Health Resources and Services Administration (HRSA) of the US Department of Health and Human Services. The interpretation and reporting of these data are the responsibility of the authors and do not represent official policy of the Scientific Registry of Transplant Recipients, OPTN/UNOS, or the US government. Because the dataset was deidentified, the University of Miami determined that separate institutional review board approval was not required.

Study population

We identified adult recipients aged ≥18 years at listing who underwent first-time heart transplantation between January 1, 2000, and September 25, 2025. Group allocation was defined using the standardized diagnostic codes entered by the transplant centers at the time of listing and transplantation. The cohort was restricted to recipients with a center-reported primary diagnosis of AICM (code 1001), idiopathic DCM (code 1000), or ICM (code 1007). AICM was the exposure group. Idiopathic DCM served as the primary comparator since it is the closest nonischemic cardiomyopathy (NICM) group to AICM. Recipients were excluded if they had prior transplantation, underwent multiorgan transplantation, heterotopic heart transplantation, had a missing transplant date, or lacked sufficient follow-up to ascertain survival (defined as lacking a valid patient survival time, last known status date, or composite death date).

Outcomes and covariates

The primary outcome was all-cause mortality, measured from the date of transplantation to death or last available follow-up. Secondary outcomes included post-transplant dialysis during the hospitalization, stroke before discharge, permanent pacemaker (PPM) implantation before discharge, treated rejection within 1 year, graft failure, and length of stay. Graft failure was evaluated as a secondary time-to-event outcome using the OPTN/UNOS graft-status indicator and corresponding graft follow-up time. Follow-up began on the date of transplantation and ended at the registry-recorded graft-failure date or censoring at the last available graft-status follow-up. Mortality and graft failure were additionally evaluated with follow-up censored at 1, 5, and 10 years.

Cause-of-death patterns were summarized descriptively; categories included graft failure, infection, cardiovascular or cerebrovascular death, hemorrhage, pulmonary causes, malignancy, other organ failure, and other or unknown causes.

The primary adjusted mortality and secondary outcome models used a prespecified adjustment set that included recipient age at listing, sex, race and ethnicity, body mass index (BMI), diabetes, pretransplant dialysis, recipient serum creatinine at transplantation, donor age, donor sex, donor serum creatinine, ischemic time, and year of transplantation.

Statistical analysis

Baseline characteristics are presented as number and percentage, mean ± standard deviation (SD), or median with interquartile range. Categorical variables were compared using Pearson chi-square tests; one-way ANOVA was used for normally distributed continuous variables and Kruskal-Wallis tests for skewed continuous variables. Standardized mean differences (SMDs) were calculated to assess baseline differences, with the primary SMD comparing AICM with DCM.

Unadjusted survival was estimated using Kaplan-Meier and compared with the log-rank test. Pairwise log-rank comparisons were performed with Bonferroni-adjusted p values. Cox proportional hazards regression was used to estimate hazard ratios (HRs) for mortality and graft failure. The proportional hazards assumption, reported in Supplementary Table S10, was assessed using Schoenfeld residuals. Mortality and graft failure were also assessed using horizon-specific adjusted Cox models at 1, 5, and 10 years.

Missing covariates were handled using multiple imputation by chained equations with 30 imputed datasets and pooled using Rubin's rules. Details of missingness and imputation are shown in Supplementary Tables S1 and S2.

Secondary outcomes were analyzed using logistic regression and reported as odds ratios (ORs) with 95% confidence intervals (CI). Length of stay was analyzed using negative binomial regression and reported as an incidence rate ratio (IRR) with 95% CI. Sensitivity analyses included complete-case analysis, adjustment for prior malignancy, restriction to transplants performed from 2010 through 2025, and propensity score overlap weighting. Covariate balance after overlap weighting was assessed using SMDs. For secondary outcomes, we fitted expanded models additionally adjusting for hospitalization status at transplant, mechanical ventilation, MCS support, inotropic support, and cardiac output (CO).

At-transplant immunosuppression was derived from medication-specific OPTN/UNOS fields indicating whether each agent was used for induction or maintenance at the time of transplantation.

Malignancy-related death was evaluated using Fine-Gray competing-risk regression, adjusted for recipient age, sex, and year of transplantation, and reported as subdistribution hazard ratios. To evaluate whether immunosuppression management explained the association with malignancy-related death, at-transplant immunosuppression patterns were summarized by diagnosis group. An exploratory Fine-Gray sensitivity model for malignancy-related death was additionally adjusted for at-transplant induction therapy and maintenance tacrolimus, mycophenolate, corticosteroid, and mTOR inhibitor use.

Allocation era was defined using the October 18, 2018, adult heart allocation policy change. Baseline characteristics were compared before and after the policy change. Era-stratified Cox models used complete-case data within each era and adjusted with relevant covariates for model stability. For the allocation-era analysis, LVAD use was evaluated separately using a strict LVAD definition, defined as isolated LVAD or LVAD plus RVAD support at transplantation. Diagnosis-by-era interaction was assessed using logistic regression models including diagnosis group, allocation era, and a diagnosis group-by-era interaction term.

Temporal trends in listing and transplant volumes were described. These trend analyses were used to contextualize changes in transplant volume and were not restricted to the final analytic transplant cohort used for post-transplant outcome analyses. Annual transplant counts from the final analytic cohort were summarized separately. The registry snapshot extended through September 2025, which was treated as an incomplete final year in descriptive trend displays.

All tests were 2-sided, and p values <0.05 were considered statistically significant. Analyses were performed using Stata version 18.0 and Python.

Results

Study cohort and baseline characteristics

The final cohort included 38,897 adult recipients of first-time isolated heart transplantation, of whom 748 had AICM, 19,809 had idiopathic DCM, and 18,340 had ICM (Figure 1). P values in Table 1 reflect overall 3-group comparisons; pairwise AICM-versus-idiopathic DCM comparisons are reported in Supplementary Table S3. The median post-transplant follow-up for the entire analytic cohort was 5.9 years (IQR 2.0-10.8 years). Median follow-up was 6.7 (IQR 2.9-11.0) years for AICM, 5.6 (2.0-10.6) years for idiopathic DCM, and 6.0 (2.0-10.9) years for ICM. Compared with idiopathic DCM, recipients with AICM were younger (50.7 ± 13.3 vs 51.8 ± 12.5; p = 0.037) and more likely to be female (71.7% vs 27.0%). AICM recipients had lower BMI (25.5 ± 5.0 vs 27.6 ± 5.1 kg/m²; p < 0.001) and a lower prevalence of diabetes (19.4% vs 25.2%; p < 0.001). In the overall 3-group comparison, recipients with ICM were the oldest (58.7 ± 8.2 years), predominantly male, and had the highest prevalence of diabetes mellitus (37.4%).

Figure 1.

Patient selection flowchart. Sequential cohort restrictions from OPTN/UNOS registry records to the final analytic cohort. AICM, anthracycline-induced cardiomyopathy; DCM, dilated cardiomyopathy; ICM, ischemic cardiomyopathy; OPTN/UNOS, Organ Procurement and Transplantation Network/United Network for Organ Sharing

Table 1.

Baseline Characteristics of the Final Analytic Cohort by Cardiomyopathy Etiology

Characteristic	AICM (n=748)	Idiopathic DCM (n=19,809)	ICM (n=18,340)	pvalue
Recipient characteristics
Age at transplant, y	50.7±13.3	51.8±12.5	58.7±8.2	<0.001
Female sex	536 (71.7)	5356 (27.0)	2500 (13.6)	<0.001
Race/ethnicity				<0.001
White	499 (66.7)	10,891 (55.0)	14,152 (77.2)
Black	150 (20.1)	6231 (31.5)	1919 (10.5)
Hispanic	62 (8.3)	1839 (9.3)	1399 (7.6)
Asian	28 (3.7)	584 (2.9)	662 (3.6)
Other	9 (1.2)	264 (1.3)	208 (1.1)
Body mass index, kg/m²	25.5±5.0	27.6±5.1	27.6±4.5	<0.001
Serum creatinine, mg/dL	1.0 [0.8-1.3]	1.2 [0.9-1.4]	1.2 [1.0-1.5]	<0.001
Total bilirubin, mg/dL	0.7 [0.5-1.1]	0.7 [0.5-1.2]	0.7 [0.5-1.1]	<0.001
Diabetes mellitus	145 (19.4)	4970 (25.2)	6826 (37.4)	<0.001
Pretransplant dialysis	15 (2.0)	438 (2.2)	449 (2.5)	0.213
Functional status at transplant				<0.001
Independent	101 (14.8)	2352 (13.4)	2421 (15.3)
Moderate limitation	191 (28.0)	5487 (31.2)	5730 (36.3)
Severe limitation	390 (57.2)	9766 (55.5)	7655 (48.4)
Hospitalization status at transplant				<0.001
ICU	288 (38.5)	8107 (40.9)	6472 (35.3)
Hospitalized, not ICU	125 (16.7)	3034 (15.3)	2709 (14.8)
Not hospitalized	335 (44.8)	8662 (43.7)	9155 (49.9)
Prior cardiac surgery	120 (17.5)	3560 (20.7)	4271 (28.3)	<0.001
Mechanical ventilation	9 (1.2)	274 (1.4)	420 (2.3)	<0.001
Waitlist duration, d	48 [14-196]	59 [16-213]	72 [19-244]	<0.001
Pretransplant support and hemodynamics
Extracorporeal membrane oxygenation	13 (1.7)	535 (2.7)	378 (2.1)	<0.001
Intra-aortic balloon pump	97 (13.0)	2611 (13.2)	2080 (11.3)	<0.001
Inotropic support	363 (48.5)	8400 (42.4)	6910 (37.7)	<0.001
Ventricular assist device support	258 (35.8)	8498 (46.1)	7170 (43.0)	<0.001
VAD type among recipients with VAD/TAH support				<0.001
LVAD	217 (84.1)	7371 (86.7)	6140 (85.6)
RVAD	0 (0.0)	44 (0.5)	42 (0.6)
TAH	2 (0.8)	132 (1.6)	68 (0.9)
LVAD + RVAD	22 (8.5)	439 (5.2)	295 (4.1)
VAD/TAH unspecified	17 (6.6)	512 (6.0)	625 (8.7)
Pulmonary capillary wedge pressure, mm Hg	18.0±8.4	18.6±9.1	17.9±8.8	<0.001
Mean pulmonary artery pressure, mm Hg	27.7±9.7	28.1±10.2	27.4±10.2	<0.001
Cardiac output, L/min	3.8 [3.0-4.8]	4.3 [3.5-5.3]	4.5 [3.7-5.4]	<0.001
Donor characteristics
Donor age, y	31.4±11.7	32.3±11.2	32.8±11.6	<0.001
Female donor	367 (49.1)	5486 (27.7)	4632 (25.3)	<0.001
Donor body mass index, kg/m²	24.7 [21.8-28.4]	26.5 [23.4-30.6]	26.4 [23.3-30.3]	<0.001
Donor serum creatinine, mg/dL	0.9 [0.7-1.3]	1.0 [0.8-1.4]	1.0 [0.8-1.4]	<0.001
Donor left ventricular ejection fraction, %	61.6±6.9	61.6±7.1	61.6±7.3	0.783
Donor total bilirubin, mg/dL	0.7 [0.4-1.1]	0.7 [0.5-1.2]	0.7 [0.5-1.2]	<0.001
Donation after circulatory death	21 (2.8)	807 (4.1)	614 (3.3)	<0.001
Donor-recipient matching / transplant characteristics
Ischemic time, h	3.3 [2.6-3.9]	3.3 [2.6-3.9]	3.3 [2.6-4.0]	0.002
Calendar year of transplant	2014 [2009-2019]	2016 [2008-2021]	2013 [2006-2020]	<0.001
Transplant era				<0.001
2000-2009	204 (27.3)	5855 (29.6)	6895 (37.6)
2010-2017	304 (40.6)	5870 (29.6)	5350 (29.2)
2018-2025	240 (32.1)	8084 (40.8)	6095 (33.2)
Donor-to-transplant center distance, nautical miles	159 [22-366]	140 [24-339]	127 [21-324]	<0.001
Donor-recipient weight ratio	1.0 [0.9-1.2]	1.0 [0.8-1.1]	1.0 [0.9-1.1]	<0.001
Donor-recipient height ratio	1.0 [1.0-1.1]	1.0 [1.0-1.0]	1.0 [1.0-1.0]	<0.001
Predicted heart mass mismatch, %	8.4 [−3.7 to 26.3]	0.4 [−8.9 to 12.5]	−0.5 [−9.3 to 10.5]	<0.001
Donor-recipient sex mismatch	271 (36.2)	4466 (22.5)	4062 (22.1)	<0.001

Values are presented as mean ± SD, median [IQR], or n (%). p values compare the 3 diagnosis groups. Percentages were calculated among recipients with nonmissing data for each variable. AICM, anthracycline-induced cardiomyopathy; ECMO, extracorporeal membrane oxygenation; IABP, intra-aortic balloon pump; ICU, intensive care unit; ICM, ischemic cardiomyopathy; Idiopathic DCM, idiopathic dilated cardiomyopathy; IQR, interquartile range; LVAD, left ventricular assist device; RVAD, right ventricular assist device; SD, standard deviation; TAH, total artificial heart; VAD, ventricular assist device

At the time of transplantation, AICM recipients were more likely than idiopathic DCM recipients to be receiving inotropic support (48.5% vs 42.4%; p < 0.001) but less likely to have VAD support (35.8% vs 46.1%; p < 0.001).

Compared with idiopathic DCM, AICM recipients were more likely to receive hearts from females (49.1% vs 27.7%) and from younger donors (31.4 ± 11.7 vs 32.3 ± 11.2 years). In the overall 3-group comparison, donor age was 32.8 ± 11.6 years in the ICM cohort (p < 0.001). Additional transplant and donor-recipient matching characteristics are reported in Table 1.

At transplant immunosuppression consisted predominantly of calcineurin inhibitors, mycophenolate, and corticosteroids across all diagnosis groups. Tacrolimus maintenance was recorded in 82.2% of AICM recipients, 81.0% of idiopathic DCM recipients, and 75.7% of ICM recipients, while mycophenolate maintenance was recorded in 93.2%, 92.6%, and 90.3%, respectively. Complete at-transplant immunosuppression patterns are shown in Supplementary Table S14.

Post-transplant survival and graft failure

During follow-up, death occurred in 270 AICM recipients, 7171 idiopathic DCM recipients, and 9228 ICM recipients.

Kaplan-Meier analysis demonstrated significant differences in post-transplant survival across the 3 diagnosis groups (log-rank p<0.001) (Figure 2). In pairwise comparisons, survival did not differ significantly between AICM and idiopathic DCM (log-rank p=0.678). In contrast, AICM recipients had better unadjusted survival than ICM recipients (log-rank p<0.001) (Supplementary Table S4).

Figure 2.

Kaplan-Meier survival comparison by phenotype. Kaplan-Meier curves comparing post-transplant survival among patients with AICM, idiopathic dilated cardiomyopathy (idiopathic DCM; labeled IDCM in the figure), and ICM. AICM, anthracycline-induced cardiomyopathy; ICM, ischemic cardiomyopathy; Idiopathic DCM, idiopathic dilated cardiomyopathy

AICM was not significantly associated with all-cause mortality compared with DCM in unadjusted (HR 0.97, 95% CI 0.86-1.10; p=0.678) or adjusted analysis (HR 1.11, 95% CI 0.98-1.25; p=0.111) (Table 2). Findings were consistent at 1 year, 5 years, and 10 years. In contrast, compared with ICM, AICM was associated with significantly lower adjusted all-cause mortality (HR 0.86, 95% CI 0.76-0.98; p=0.025) (Supplementary Table S5).

Table 2.

Post-Transplant Mortality and Graft-Failure Analyses Comparing AICM With Idiopathic DCM

Analysis	No. analyzed	Events	HR (95% CI)	pvalue
Mortality analyses
Unadjusted	20,557	7441	0.97 (0.86-1.10)	0.678
Multivariable adjusted	20,557	7441	1.11 (0.98-1.25)	0.111
1-year mortality	20,557	1769	1.15 (0.89-1.47)	0.279
5-year mortality	20,557	3685	1.07 (0.90-1.27)	0.460
10-year mortality	20,557	5389	1.10 (0.96-1.27)	0.178
Graft-failure analyses
Unadjusted	20,557	7445	0.98 (0.87-1.11)	0.771
Multivariable adjusted	20,557	7445	1.09 (0.96-1.23)	0.185
1-year graft failure	20,557	1844	1.15 (0.90-1.46)	0.260
5-year graft failure	20,557	3837	1.06 (0.90-1.26)	0.493
10-year graft failure	20,557	5522	1.11 (0.96-1.27)	0.156

Values are hazard ratios with 95% confidence intervals comparing AICM with idiopathic DCM. Multivariable models adjusted for recipient age, sex, race/ethnicity, body mass index, diabetes mellitus, pretransplant dialysis, recipient serum creatinine at transplantation, donor age, donor sex, donor serum creatinine, ischemic time, and calendar year. Estimates were pooled across 30 multiply imputed datasets using Rubin's rules

Graft failure occurred in 275 AICM, 7170 idiopathic DCM, and 8833 ICM recipients. AICM was not associated with graft failure in unadjusted (HR 0.98, 95% CI 0.87-1.11; p=0.771) or adjusted Cox regression (HR 1.09, 95% CI 0.96-1.23; p=0.185). Graft-failure analyses at 1 year, 5 years, and 10 years were not significant (Table 2). Compared with ICM, AICM was not significantly associated with graft failure (HR 0.88, 95% CI 0.78-1.00; p=0.059) (Supplementary Table S5).

Secondary outcomes

Secondary outcomes are summarized in Table 3. AICM had higher odds of post-transplant dialysis compared with idiopathic DCM (OR 1.31, 95% CI 1.03-1.67; p=0.030), which persisted in an expanded sensitivity model that adjusted for hospitalization status, mechanical ventilation, MCS support, inotropic support, and CO (Supplementary Table S9B). Replacing serum creatinine with CKD-EPI eGFR, AICM remained associated with higher odds of post-transplant dialysis compared with DCM (OR 1.32, 95% CI 1.04-1.69; p=0.022) (Supplementary Table S9D).

Table 3.

Post-Transplant Secondary Outcomes Comparing AICM With Idiopathic DCM

Outcome	AICM (n=748)	Idiopathic DCM (n=19,809)	No. analyzed	Events	Unadjusted OR/IRR (95% CI)	Unadjustedpvalue	Adjusted OR/IRR (95% CI)	Adjustedpvalue
Post-transplant dialysis	89/744 (12.0%)	2271/19,620 (11.6%)	20,364	2360	1.04 (0.83-1.30)	0.746	1.31 (1.03-1.67)	0.030
Post-transplant stroke	28/744 (3.8%)	555/19,592 (2.8%)	20,336	583	1.34 (0.91-1.98)	0.137	1.30 (0.87-1.94)	0.197
Permanent pacemaker before discharge	17/744 (2.3%)	551/19,599 (2.8%)	20,343	568	0.81 (0.50-1.32)	0.393	0.86 (0.52-1.41)	0.552
Treated rejection within 1 year	152/639 (23.8%)	3696/16,059 (23.0%)	16,698	3848	1.04 (0.87-1.26)	0.650	0.93 (0.77-1.13)	0.458
Index transplant length of stay, d	15 (10-22)	15 (11-22)	20,336		1.10 (1.04-1.16)	<0.001	1.12 (1.06-1.18)	<0.001

Values are shown as n/N (%) for binary outcomes and median [IQR] for length of stay. Adjusted secondary-outcome models included recipient age at listing, sex, race/ethnicity, body mass index, diabetes mellitus, pretransplant dialysis, recipient serum creatinine at transplantation, donor age, donor sex, donor serum creatinine, ischemic time, and calendar year of transplantation. Estimates were pooled across 30 multiply imputed datasets using Rubin's rules. Percentages for secondary outcomes were calculated among recipients with nonmissing outcome data

There were no significant differences in stroke (OR 1.30, 95% CI 0.87-1.94; p=0.197), PPM implantation (OR 0.86, 95% CI 0.52-1.41; p=0.552), or treated rejection within 1 year (OR 0.93, 95% CI 0.77-1.13; p=0.458). Although median length of stay was similar between groups, AICM was associated with longer hospitalization after adjustment in the negative binomial model (IRR 1.12, 95% CI 1.06-1.18; p<0.001). Secondary outcomes for the AICM versus ICM comparison are shown in Supplementary Table S6

Cause-of-death patterns

Cause-of-death counts are shown in Figure 3 and in Supplementary Table S12.

Figure 3.

Cause-of-death distribution by diagnosis group and time horizon. Cause-of-death categories are shown among recipients who died within 1, 5, and 10 years after heart transplantation, stratified by cardiomyopathy etiology. Percentages reflect the proportion of deaths within each diagnosis group and time horizon. Idiopathic DCM is labeled IDCM in the figure for readability. AICM, anthracycline-induced cardiomyopathy; CV, cardiovascular; ICM, ischemic cardiomyopathy

At 1 year, 68 (9.1%) AICM recipients had died. Early deaths were most commonly attributed to cardiovascular or cerebrovascular causes at 32.4%, followed by infection at 19.1%, other organ failure at 14.7%, and graft failure at 13.2%. Malignancy-related death was uncommon within the first year, accounting for 2.9% of AICM deaths.

By 5 years, 137 AICM recipients had died. Cardiovascular or cerebrovascular causes remained the most common cause of death in AICM, accounting for 20.4%. At 5 years, malignancy accounted for 11.7% of deaths among AICM recipients, compared with 5.1% among idiopathic DCM recipients and 8.5% among ICM recipients.

By 10 years of follow-up, 210 (28.1%) AICM recipients had died. The most common cause was other or unknown causes, followed by cardiovascular or cerebrovascular death and malignancy-related death. Malignancy accounted for 16.7% of AICM deaths at 10 years, compared with 7.7% of idiopathic DCM deaths and 11.9% of ICM deaths.

In exploratory Fine-Gray analyses, the unadjusted cumulative incidence of malignancy-related death was higher in AICM than in idiopathic DCM during the full follow-up period and at 5 and 10 years, but not at 1 year. After adjustment for recipient age, sex, and transplantation year, the cumulative incidence was higher in AICM during the full follow-up period and at 1, 5, and 10 years. Compared with ICM, AICM did not have a higher unadjusted cumulative incidence of malignancy-related death at any time point. After adjustment, however, AICM was associated with a higher cumulative incidence during the full follow-up period and at 10 years, but not at 1 or 5 years. Full results are shown in Table 4.

Table 4.

Fine-Gray Competing-Risk Models for Malignancy-Related Death

Time horizon	Comparison	Malignancy deaths	Competing deaths	Adjusted sHR (95% CI)	pvalue
Overall follow-up	AICM vs idiopathic DCM	661	6780	2.27 (1.66-3.09)	<0.001
	AICM vs ICM	1211	8287	1.70 (1.24-2.33)	<0.001
1 year	AICM vs idiopathic DCM	22	1747	4.83 (1.03-22.59)	0.045
	AICM vs ICM	39	2040	2.50 (0.52-12.11)	0.255
5 years	AICM vs idiopathic DCM	198	3487	2.36 (1.38-4.04)	0.002
	AICM vs ICM	359	3798	1.65 (0.96-2.85)	0.072
10 years	AICM vs idiopathic DCM	436	4953	2.41 (1.68-3.46)	<0.001
	AICM vs ICM	796	5786	1.77 (1.23-2.57)	0.002

Subdistribution hazard ratios (sHRs) were estimated using Fine–Gray competing-risk regression, with death from causes other than malignancy treated as a competing event. Models were adjusted for recipient age, sex, and transplant year. The AICM versus idiopathic DCM analysis included 20,557 recipients, and the AICM versus ICM analysis included 19,088 recipients. AICM, anthracycline-induced cardiomyopathy; CI, confidence interval; DCM, dilated cardiomyopathy; ICM, ischemic cardiomyopathy; sHR, subdistribution hazard ratio

Additional adjustment for at-transplant induction therapy and maintenance tacrolimus, mycophenolate, corticosteroid, and mTOR inhibitor use yielded results consistent with the primary analysis of malignancy-related death (Supplementary Table S15).

Among AICM recipients with malignancy-related death, 11 of 47 patients (23.4%) had registry-recorded recurrence of a pretransplant malignancy before death, whereas 43 of 47 patients (91.5%) had registry-recorded de novo malignancy or post-transplant lymphoproliferative disorder before death. These categories were not mutually exclusive because some recipients had both recurrent and de novo malignancy recorded. These findings are presented descriptively in Supplementary Table S16.

Sensitivity analyses

Sensitivity analyses were consistent with the primary findings. Compared with idiopathic DCM, AICM was not significantly associated with all-cause mortality (HR 1.10, 95% CI 0.97-1.25), after adjustment for prior malignancy (HR 1.09, 95% CI 0.94-1.27), after restriction to transplants performed after 2010 (HR 1.08, 95% CI 0.91-1.28), or in the overlap-weighted model (HR 1.06, 95% CI 0.93-1.20). Similarly, AICM was not significantly associated with graft failure (HR 1.08, 95% CI 0.95-1.22), after adjustment for prior malignancy (HR 1.06, 95% CI 0.92-1.23), after restriction to post-2010 (HR 1.06, 95% CI 0.90-1.26), or in the overlap-weighted model (HR 1.04, 95% CI 0.91-1.18). Sensitivity analyses are summarized in Supplementary Tables S9 and S11 and Supplementary Figures S1 and S3.

Temporal trends and allocation era

Annual listing and transplant volumes are shown in Figure 4 and Supplementary Table S13. In the descriptive trend, annual AICM transplants increased from 7 in 2000 to 49 in 2017 and 47 in 2019, then were lower at 16 in 2025. Over the same period, idiopathic DCM transplant volume increased from 619 in 2000 to 1202 in 2024, whereas ICM transplant volume increased from 779 to 926.

Figure 4.

Annual listing and heart transplant volumes by cardiomyopathy etiology. Annual listing and heart transplant counts are shown for anthracycline-induced cardiomyopathy, idiopathic dilated cardiomyopathy, and ischemic cardiomyopathy from 2000 through 2025. These descriptive trend counts were derived from a broader listed-vs-transplanted trend source and are not restricted to the final analytic transplant cohort used for post-transplant survival and outcome analyses. The vertical reference line marks the October 18, 2018 OPTN/UNOS adult heart allocation policy change. The 2025 registry year was incomplete because the data snapshot extended through September 25, 2025

After implementation of the 2018 adult heart allocation policy, 204 AICM recipients underwent transplantation. The post-policy AICM cohort was older, had higher BMI, and had a greater prevalence of diabetes compared with the pre-policy cohort. ECMO use increased from 0.7% to 4.4%, and IABP use increased from 6.2% to 30.9%; durable VAD support remained similar. Median waitlist duration decreased from 60 to 25 days. Pre- and post-allocation characteristics are shown in Supplementary Table S7, and calendar-year trends in pretransplant support are shown in Supplementary Figure S2.

Allocation-era analyses showed no evidence that the association between AICM and post-transplant outcomes differed before vs after the policy change. For AICM vs DCM, adjusted HRs for mortality and graft failure demonstrated no statistically significant association, with no diagnosis-by-era interaction (P interaction=0.399). Similarly, for AICM vs ICM, there was no significant interaction by allocation era for mortality or graft failure. Allocation-era stratified mortality and graft-failure analyses are shown in Supplementary Table S8.

In an era-stratified analysis, the prevalence of LVAD support at transplantation was similar among AICM recipients before and after the 2018 allocation policy change (30.9% vs 34.8%; p = 0.306), whereas it was higher in the post-allocation era among idiopathic DCM recipients (35.6% vs 46.0%; p <0.001) and ICM recipients (31.0% vs 44.8%; p <0.001). Complete LVAD analysis among era stratification is shown in Supplementary Table S17.

Discussion

In this national cohort of 38,897 adult first-time isolated heart transplant recipients, AICM was not significantly associated with worse post-transplant survival or graft failure when compared with idiopathic DCM. In contrast, AICM was associated with higher odds of post-transplant dialysis and longer hospitalization, whereas stroke, PPM implantation, and treated rejection within 1 year did not differ significantly. These findings suggest that, among patients who reach transplantation, AICM is not a marker of inferior patient or graft survival, but it does identify a distinct transplant phenotype with greater early post-operative resource use.

Compared with recipients with idiopathic DCM, recipients with AICM were more often women, had lower BMI, lower serum creatinine, less diabetes, lower CO, and more frequent inotropic support at transplantation. Despite this profile, durable VAD was used less often. This pattern is consistent with prior analyses of AICM, in which bridge-to-transplant LVAD use was lower than in other cardiomyopathy groups despite contemporary growth in MCS.⁴ The reasons are likely to be multifactorial; smaller body size, right-sided reserve, prior cancer-related comorbidity, bleeding risk, frailty, and center-level selection practices may all influence bridge strategy in patients with AICM.

Earlier INTERMARCS data suggested a greater need for right ventricular support among patients with AICM undergoing durable MCS.¹⁰ More contemporary continuous-flow LVAD data, however, have shown survival in AICM comparable to ischemic and idiopathic DCM, although major bleeding may be more frequent than in DCM.¹¹ Similarly, patients with a history of cancer undergoing durable LVAD implantation appear to have higher bleeding risk and lower transplantation rates, while survival among those with cancer therapy-induced heart failure is not inferior to patients without cancer.¹² Thus, lower LVAD use in AICM should not be interpreted as evidence that durable MCS is inappropriate. Rather, it underscores the need for individualized assessment of biventricular physiology, bleeding risk, oncologic prognosis, and transplant timing.

The absence of a survival difference after transplantation is consistent with prior registry experience, including that in the contemporary allocation era.4, 13 Comparable survival among AICM and NICM recipients has been reported, despite higher post-transplant infection and malignancy rates in AICM.¹⁴ In our study, although AICM survival was higher when compared to ICM, these recipients were older, male, and had a higher prevalence of diabetes. Even with adjustment, residual confounding is likely. This comparison should be interpreted in the clinical context, rather than as evidence that AICM intrinsically confers lower post-transplant risk. The present analysis differs by focusing on adult recipients with AICM compared directly with idiopathic DCM, incorporating donor and recipient adjustment, testing allocation-era interaction, and extending through the post-2018 allocation era. Prior literature and the present data support heart transplantation as an appropriate therapy for selected patients with AICM.

The central concern in AICM is not anthracycline exposure alone, but the oncologic context in which transplantation occurs. In Fine-Gray models, with nonmalignancy deaths treated as competing events, AICM was associated with a higher cumulative incidence of malignancy-related death. This finding suggests a cause-specific mortality pattern among AICM recipients rather than an excess in post-transplant mortality. Similarly, contemporary data suggest that more than 10% of heart transplant recipients develop de novo malignancy, and that recipients with prior malignancy have a higher post-transplant malignancy risk.6, 15

In addition, among cancer survivors undergoing heart transplantation, subsequent malignancies may reflect de novo rather than recurrence of the original malignancy.⁸ In the present cohort, de novo malignancy or PTLD was more prevalent than recurrence of pretransplant malignancy among AICM recipients, and the majority of those with malignancy-related death had de novo malignancy or PTLD reported. While these descriptive findings cannot establish a direct causal relationship between a specific malignancy and death, they suggest that the excess malignancy-related mortality in AICM may be driven predominantly by de novo malignancies rather than recurrence of the original cancer. OPTN/UNOS does not adjudicate whether a specific type of malignancy, de novo or recurrence, directly caused death or whether recurrence represented the original anthracycline-treated malignancy. Accordingly, these findings should be interpreted as descriptive. These data support the approach emphasized in contemporary International Society for Heart and Lung Transplantation guidelines: transplant candidacy in cancer survivors should be individualized and multidisciplinary, integrating oncologic prognosis, recurrence risk, and functional status.¹⁶

The higher odds of post-transplant dialysis in AICM warrant further attention. This association was observed despite lower baseline creatinine and diabetes prevalence among AICM, suggesting that pretransplant renal function alone does not explain the difference. Potential contributors include lower CO, inotrope dependence, perioperative hemodynamic instability, prior exposure to nephrotoxic therapies, and frailty or nutritional reserve. This is clinically relevant because post-transplant dialysis has been associated with adverse outcomes.¹⁷ The longer hospital stay in AICM similarly suggests greater post-operative complexity despite preserved survival. Whether these findings reflect AICM biology, prior cancer treatment, perioperative acuity, or residual selection effects cannot be determined from registry data. They do, however, support closer renal-risk assessment and perioperative planning in AICM candidates.

Our analysis also provides insight into AICM transplantation after the 2018 OPTN/UNOS adult heart allocation policy change. We found no statistical evidence of effect modification by allocation era, although precision was limited by the small number of post-policy AICM recipients. The revised system expanded adult heart allocation from 3 to 6 medical urgency statuses to better stratify candidates by acuity.¹⁸ National analyses have shown that the post-policy era is characterized by shorter wait times, greater use of temporary MCS, reduced LVAD bridging, and higher-acuity transplantation.19, 20, 21 In the present cohort, both AICM and DCM recipients were transplanted at higher acuity after the policy change, with increased ICU transplantation and MCS use. These findings suggest that the 2018 allocation policy changed the pathway to transplantation without a differential survival disadvantage for AICM.

A recent review compiling the available data on the 2018 allocation change reported that higher rates of post-transplant dialysis and increased length of stay have been noted.¹⁹ Our findings are similar to those previously reported; nonetheless, whether these were truly influenced by this policy or whether it reflects the capturing of higher acuity patients is yet to be determined.

Limitations

Several limitations should be acknowledged. First, the retrospective design precludes causal inference, and residual confounding may persist despite multivariable adjustment and overlap weighting. Second, exposure groups may be undercoded or misclassified within broader cardiomyopathy categories, a limitation inherent to administrative data. Third, oncology-specific variables, important for risk stratification and candidate selection, such as cumulative anthracycline dose, cancer stage, or remission duration, were unavailable. Furthermore, immunosuppressive drug doses, trough levels, or cumulative immunosuppressive exposure are not available in the UNOS database, limiting the analysis of the relationship between immunosuppressive strategy and malignancy-related outcomes. In addition, while the registry captures malignancy as a cause of death, it lacks the oncological data required to adjudicate whether these deaths were driven by recurrence or the development of a de novo malignancy. Although OPTN/UNOS provides national post-transplant follow-up, long-term cause-specific mortality, particularly malignancy-related mortality, may be incompletely captured. Cause-of-death reporting is registry-based and not centrally adjudicated. Therefore, malignancy-related death analyses should be interpreted with caution. Finally, this cohort includes only patients who were ultimately transplanted; outcomes among patients with AICM who were never referred or declined during listing remain poorly characterized.

Conclusion

After adjustment for recipient, donor, and transplant characteristics, AICM was not significantly associated with worse all-cause mortality or graft failure after heart transplantation compared with DCM. However, AICM recipients had higher odds of post-transplant dialysis, longer hospitalization and higher risk of malignancy-related death, whereas stroke, PPM implantation, and treated rejection within 1 year were not significantly different. These findings support heart transplantation as a reasonable treatment strategy for selected patients with AICM while highlighting the need for individualized oncologic assessment, renal-risk evaluation, and perioperative planning.

Data availability statement

The data underlying this article is available from the OPTN, subject to applicable data use agreements and OPTN/HRSA policies. Analytic code may be made available from the corresponding author upon reasonable request, subject to data use restrictions.

Funding

Dr. Wever Pinzon is supported by the NIH-NHLBI grant number K23HL150322.

Declaration of Competing Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Omar Wever-Pinzon reports financial support was provided by National Institutes of Health. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Supplementary material