Ten Year Experience with Extended Criteria Cardiac Transplantation

Abstract

Background

Extended criteria cardiac transplant (ECCT) programs expand the transplant pool by matching donors and recipients typically excluded from the transplant process because of age or co-morbidity. There is a paucity of data examining long-term outcomes with this strategy.

Methods and Results

Between January 2000 and December 2009, adult patients undergoing isolated heart transplant were prospectively classified as ECCT based on pre-specified criteria. Baseline characteristics and outcomes were compared between ECCT and standard criteria cardiac transplant (SCCT) recipients. Two Cox proportional hazards models were developed. The first to identify clinical variables contributing to survival between the two groups, and the second to determine the additional risk associated with assignment to ECCT. Among the 454 patients who underwent heart transplant, 84 (18.5%) were ECCT. Compared to SCCT, ECCT patients were older (median 66.6 yrs vs. 53.2 yrs, p<0.001), with higher frequency of diabetes (46.4% vs. 24.6%, p<0.001) and chronic kidney disease (median eGFR 55 ml/min vs. 61.6, ml/min, p=0.001). After adjustment for baseline characteristics, SCCT survival was higher than ECCT at 1 (89% vs. 86%; p=0.18) and 5 (77% vs. 66%; p=0.035) years. In a multivariate model that included listing criteria, Cr (HR = 1.05 per 0.1 mg/DL, 95% CI: (1.02, 1.09), P=0.001) was a significant predictor of post-transplant mortality.

Conclusions

ECCT is an acceptable alternative for advanced heart failure therapy in select patients. Age and renal dysfunction are important determinants of long-term survival and post-transplant morbidity.

Orthotopic heart transplantation (OHT) remains the most effective therapy for end-stage heart failure (HF) with median survival rates that exceed eleven years¹. Despite improving outcomes, the overall impact of transplantation is limited by the significant shortage of suitable donor hearts. Each year in the United States, approximately 2,200 patients with advanced HF undergo cardiac transplantation, whereas nearly 150,000 patients may be in need of advanced therapies such as OHT². As a result, transplantation is typically limited to a highly select patient population that is younger with a lower burden of co-morbid illness. Conversely, the growing population of patients with advanced age and end-stage HF are often believed to be at high risk for complications after OHT and excluded from transplant listing.

Several US transplant centers have developed formal extended criteria cardiac transplantation (ECCT) strategies. The underlying premise of ECCT is to match patients who fall outside traditional listing criteria because of advanced age or co-morbid illness with donors or donor organs that have “high risk” features. Previous analyses of ECCT outcomes have demonstrated shortened mean survival with ECCT compared with standard criteria cardiac transplantation (SCCT)³. However, post-transplant survival in this cohort is superior to that anticipated with ongoing medical management with advanced systolic heart failure^4–10. A prior study at our center confirmed these acceptable outcomes and also demonstrated similar post-transplant morbidity for patients undergoing ECCT and SCCT ³.

With median OHT survival now extending beyond 10 years, re-examination of intermediate and long-term outcomes after ECCT is warranted. Furthermore, with the necessity of life-long immunosuppression using nephrotoxic agents, ascertaining the long-term impact of ECCT on renal function is clinically important. The purpose of this study is to describe our experience with ECCT over the last 10 years by comparing outcomes with ECCT and SCCT, to understand factors that influence survival, and to examine the impact of renal dysfunction in the ECCT population.

Methods

An extended-criteria protocol was developed and initiated at our center in January 2000. Patients were considered for ECCT if they were age 65 or older or they had significant co-morbidities including significant renal insufficiency, peripheral arterial disease (including carotid artery disease), or poorly controlled diabetes mellitus. Cytomegalovirus (CMV) seropositivity is not considered a risk factor for consideration of SCCT or ECCT. ECCT patients signed informed consent for the transplant team to utilize organs that may be deemed unsuitable for use in SCCT recipients. Unfavorable characteristics of these hearts included older donors, single vessel coronary artery disease, left ventricular (LV) hypertrophy, LV dysfunction, high-dose inotropic support or donors considered high risk by the United Network for Organ Sharing (UNOS). Pediatric patients (< 18 years of age), patients receiving multi-organ transplants and patients undergoing retransplantation were excluded from this analysis.

Data were abstracted from the medical record by retrospective chart review. Recipient baseline characteristics were obtained including demographics, etiology of cardiomyopathy, pre-transplant mechanical circulatory support (MCS), laboratory values, and UNOS listing at the time of OHT. Donor data was collected from the United Network for Organ Sharing (UNOS) database. Variables extracted included ischemic time, donor age, LV ejection fraction, LV posterior wall thickness, septal wall thickness, the number of inotrope medications at the time of organ procurement and UNOS classification of the donor as “high risk” (suggesting a potential for disease transmission from donor to recipient). The sequence number at which the donor organ was accepted by our center for a suitable recipient was also extracted in addition to the length of donor “downtime” (presumed or observed pulselessness) and need for cardiopulmonary resuscitation (CPR).

Endpoint definitions

The primary end-point for this study was all-cause mortality. Pre-defined short-, intermediate-, and long-term outcomes were defined as measures of post-transplant morbidity Short-term outcomes at one year post-OHT included cellular rejection episodes (ISHLT Grade ≥ 2R/3A cellular rejection); re-hospitalization for any cause; presence of cardiac allograft vasculopathy (CAV, defined as a ≥ 75% lesion in any epicardial coronary artery by angiography); renal function measured by serum creatinine (Cr) and estimated glomerular filtration rate [eGFR, calculated using the Modification of Diet in Renal Disease (MDRD) formula] and chronic kidney disease (CKD) stage. We do not routinely screen for antibody mediated/humoral rejection and these data were not extracted. Intermediate-term (5-year) outcomes included presence of CAV and renal function.

Statistical Methods

Continuous variables are expressed as the median and 25^th/75^th percentiles; categorical variables are expressed as number and percentage. Continuous and ordinal variables were compared using the Wilcoxon rank-sum test and categorical variables were compared using the Pearson’s chi-square test or exact test when appropriate. Kaplan-Meier methods were used to estimate survival in both ECCT and SCCT groups and a Cox proportional hazards model compared time to mortality by listing criteria. A multivariate Cox proportional hazards model was constructed using available baseline clinical characteristics to determine the impact of these variables on post-transplant survival. Incremental risk associated with assignment to the ECCT list was determined by a second Cox proportional model that included listing criteria. Proportional hazards and linearity assumptions were checked for all models and transformations applied when necessary. Adjusted survival curves by listing criteria were estimated using the corrected group prognosis method¹¹.

Non-mortality post-transplant outcomes including renal function, presence of cardiac allograft vasculopathy, cellular rejection episodes, and rehospitalization are described using the median and 25^th/75^th percentiles or percentage. All analyses were conducted using SAS statistical software version 9 (SAS Inc., Cary, NC). A p value <0.05 was considered statistically significant. The study protocol was approved by the Institutional Review Board at Duke University Medical Center.

Results

From January 1, 2000 through December 31, 2009 454 patients qualified for inclusion and underwent primary cardiac transplantation at our institution. We identified 84 ECCT patients (age range 43–75 years) and 370 SCCT patients (age range 18–68 years). Baseline characteristics are described in Table 1. ECCT patients were significantly older (66.6 vs. 53.2 years; p < 0.001) than SCCT patients and a greater proportion were white, had a history of DM, and had ischemic cardiomyopathy (ICM). Although the proportion of SCCT patients supported with pre-transplant left ventricular assist device (VAD) was greater than that of the ECCT patients, there was no significant difference in the proportion of patients supported with an intra-aortic balloon pump (IABP) at the time of cardiac transplantation.

Table 1.

Baseline Characteristics and Lab Measures by Listing Criteria

PARAMETER	STANDARD (SC)	EXTENDED (EC)	p-VALUE
Total Number of Patients	370	84
Age (N=370 SC, 84 EC)	53.2 (43.8, 59.4)	66.6 (59.2, 69.0)	< 0.001
Female Sex	88/370 (23.8%)	19/84 (22.6%)	0.82
Black Race	106/370 (28.6%)	14/84 (16.7%)	0.025
History of Diabetes	91/370 (24.6%)	39/84 (46.4%)	< 0.001
Ischemic Etiology	160/370 (43.2%)	59/84 (70.2%)	< 0.001
BUN (mg/dL)   (N=369 SC, 84 EC)	19 (14, 26)	25 (21, 36)	< 0.001
eGFR (ml/min/1.73m2)   (N=369 SC, 84 EC)	61.6 (48.2,76.5)	55.0(40.3,66.3)	0.001
Serum Creatinine (mg/dL) (N=369 SC, 84 EC)			0.027
CKD ≥3	178/369 (48.2%)	52/84 (61.9%)	0.024
VAD	110/370 (29.7%)	16/84 (19.0%)	0.048
IABP†	29/368 (7.9%)	8/84 (9.5%)	0.62
UNOS status at transplant			0.035
1A	122/370 (33.0%)	23/84 (27.4%)
1B	162/370 (43.8%)	30/84 (35.7%)
2	86/370 (23.2%)	31/84 (36.9%)

BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate; CKD, chronic kidney disease stage; VAD, ventricular assist device; IABP, intra-aortic balloon pump; UNOS, united network for organ sharing. Continuous covariates are tested using the Wilcoxon Rank-Sum test. Categorical covariates are tested using a Pearson Chi-Square test.

Continuous covariates are presented as Median (25^th, 75^th) percentiles; Binary covariates are presented as proportion (%). Continuous covariates are tested using the Wilcoxon Rank-Sum test. Categorical covariates are tested using a Pearson Chi-Square test.

Donor Data

ECCT donors were older (39.5 vs. 33.0 years; p < 0.001) and the median cold ischemic time was longer (3.7 vs. 3.4 hours; p < 0.001) than in SCCT. The number of donors older than 35 years was significantly higher in the ECCT group (71% vs 47%, p<0.001), and 38% of the donors for ECCT were greater than 45 years of age (p=0.06 compared with SCCT). Data was available on the majority of donors but as shown in Table 2 certain data elements were not available in the UNOS data files. Donor LV ejection fraction, and number of inotrope medications administered were similar between the cohorts. Further, there was no difference in the proportion of donors classified as UNOS high risk, or the proportion of donors who were reported to have cardiac arrest requiring CPR or “downtime” before CPR was initiated (Table 2). The cause of death was not different between ECCT and SCCT donors but median sequence number for the ECCT group recipients was higher than the SCCT group (27 vs 6, p<0.001).

Table 2.

Donor Data

Donor Characteristic	STANDARD (SC)	EXTENDED (EC)	p-VALUE
Total Number of Patients	370	84
Sequence Number   (N=318 SC, 69 EC)	6 (2, 24)	27 (6, 92)	<0.001
Ischemic Time (Hours)   (N= 356 SC, 82 EC)	3.4 (2.8, 3.9)	3.7 (3.3, 4.4)	<0.001
UNOS High Risk*	27/218 (12.4%)	5/42 (11.9%)	0.93
Female	114/368 (31.0%)	31/84 (36.9%)	0.29
Age (Years) (N= 368 SC, 84 EC)	33 (22, 46)	40 (32, 51)	<0.001
Age ≥ 35 (Years)	171/368 (46.5%)	60/84 (71.4%)	<0.001
Age ≥ 45 (Years)	102/368 (27.7%)	32/84 (38.1%)	0.060
History of DM	18/365 (4.9%)	4/82 (4.9%)	>0.99
History of Insulin Dependence	10/16 (62.5%)	1/4 (25.0%)	0.28
History of Cigarette Use	122/367 (33.2%)	35/80 (43.8%)	0.074
Recent Cigarette Use	110/118 (93.2%)	32/35 (91.4%)	0.72
Heavy Alcohol Use	33/182 (18.1%)	10/39 (25.6%)	0.28
Cocaine Use	77/364 (21.2%)	22/81 (27.2%)	0.24
Recent Cocaine Use	43/64 (67.2%)	13/21 (61.9%)	0.66
Other Drug Use	117/361 (32.4%)	29/81 (35.8%)	0.56
Recent Other Drug Use	81/103 (78.6%)	19/26 (73.1%)	0.54
Smoking, Alcohol, Cocaine, Other Drug Use	202/367 (55.0%)	53/83 (63.9%)	0.14
Cardiac Arrest with Downtime	26/105 (24.8%)	7/21 (33.3%)	0.41
CPR Following Down Time	28/107 (26.2%)	7/21 (33.3%)	0.50
LVEF (N= 357 SC, 79 EC)	60 (55, 65)	60 (55, 65)	0.58
Number of Inotropes Administered			0.65
0	204/370 (55.1%)	50/84 (59.5%)
1	151/370 (40.8%)	32/84 (38.1%)
2	15/370 (4.1%)	2/84 (2.4%)
Cause of Death			0.50
Anoxia	56/370 (15.1%)	14/84 (16.7%)
Bacterial Meningitis	1/370 (0.3%)	0/84 (0.0%)
Stroke	113/370 (30.5%)
CNS Tumor	3/370 (0.8%)	0/84 (0.0%)
Head Trauma	187/370 (50.5%)	35/84 (41.7%)
Other/Unknown	10/370 (2.7%)	1/84 (1.2%)
Donor PFO/ASD Closure at Txp	54/366(14.8%)	13/84 (15.5%)	0.87
Donor CABG at Txp	2/366 (0.5%)	6/84 (7.1%)	0.001
Donor Valve Repair at Txp	3/366 (0.8%)	0/84 (0%)	>0.99
Gender Mismatch b/w Donor and Recipient (Female-Male / Male- Female)	86/370 (23.2%)	24/84 (28.6%)	0.30
Female Donor/Male Recipient (denominator=Male Recipients)	56/282 (19.9%)	18/65 (27.7%)	0.17

UNOS, united network for organ sharing; LVEF, Left ventricular ejection fraction; DM, diabetes mellitus; LV, left ventricle

^*UNOS high risk refers to risk of potential viral disease transmission

Continuous variables are expressed as median (Q1, Q3), categorical variables are expressed as proportion (%)

Continuous covariates are tested using the Wilcoxon Rank-Sum test. Categorical covariates are tested using a Pearson Chi-Square test.

Unadjusted Survival

Unadjusted Kaplan-Meier survival estimates at 1- and 5-years were 90% and 78% for SCCT patients and 82% and 58% for ECCT patients (Figure 1). By univariate analysis, ECCT was associated with a two-fold risk for increased mortality (HR = 2.07, 95% CI: (1.42, 3.03), p < 0.001).

Figure 1. Unadjusted Kaplan-Meier curves presenting estimated survival probabilities for each listing criteria over time.

*P-value is determined based on the unadjusted hazard ratio estimated in a proportional hazards model.

Risk Factor Analysis

To identify recipient factors that influence survival following OHT, multivariate analysis was performed using the following candidate variables: age, race, sex, presence of VAD, presence of IABP, ICM, baseline Cr, and UNOS status at time of transplant. After adjusting for all other variables only increasing age greater than 50 (HR = 1.04 per year, 95% CI: (1.01, 1.07), P=0.014) and Cr (HR = 1.05 per 0.1 mg/dL increase, 95% CI: (1.02, 1.09), P =0.001) were significant predictors of mortality (Table 3). Other parameters that showed a trend towards reduced survival included black race, female sex, and ICM (p=0.051, p=0.056, and p=0.051, respectively) (Table 3).

Table 3.

Baseline Patient Characteristics, Univariate and Multivariate Predictors of Mortality

	Univariate			Multivariate
	HR	95% CI	P-value	HR	95% CI	P-value
Age < 50 years	0.98	0.95, 1.01	0.14	0.98	0.95, 1.01	0.14
Age ≥ 50 years	1.04	1.01, 1.07	0.005	1.04	1.01, 1.07	0.014
Black Race	1.36	0.95, 1.96	0.098	1.50	1.00, 2.24	0.051
Female	1.23	0.84, 1.80	0.28	1.49	0.99, 2.23	0.056
VAD	0.89	0.60, 1.34	0.58	1.07	0.59, 1.95	0.82
Diabetes	1.47	1.03, 2.10	0.035	1.33	0.90, 1.97	0.15
Creatinine (per 0.1 mg/dL)	1.06	1.03, 1.09	<0.001	1.05	1.02, 1.09	0.001
ICM	1.40	1.00, 1.98	0.053	1.48	1.00,2.21	0.051
IABP	2.01	1.13, 3.59	0.018	1.91	0.82, 4.46	0.13
UNOS 1B (reference=1A)	0.89	0.59, 1.33	056	1.14	0.62, 2.13	0.67
UNOS 2 (reference=1A)	0.84	0.53, 1.33	0.46	1.14	0.56, 2.33	0.71

VAD, ventricular assist device; ICM, ischemic cardiomyopathy; IABP, intra-aortic balloon pump;, UNOS, United Network for Organ Sharing

P-values derived from Cox proportional hazards model

Adjusted Survival

A second Cox proportional hazards model was created to incorporate recipient variables impacting survival (from Model 1) with the use of ECCT donors. In this model, ECCT listing and Cr were associated with survival (Table 4). After adjustment for baseline characteristics, ECCT was associated with increased risk of mortality (HR = 1.62, 95% CI: (1.02, 2.58), p=0.042) and SCCT survival was higher than ECCT at 1- (89% vs. 86%) and 5-years (77% vs. 66%). Increased Cr (HR = 1.05 per 0.1 mg/DL, 95% CI: (1.02, 1.09), P=0.001) was also significantly associated with increased risk of post-transplant mortality after adjusting for other variables, including transplant listing status. Adjusted survival curves are shown in Figure 2.

Table 4.

Baseline Patient Characteristics Including Listing Criteria, Univariate and Multivariate Predictors of Mortality

	Univariate			Multivariate
	HR	95% CI	P-value	HR	95% CI	P-value
ECCT	2.07	1.42, 3.03	< 0.001	1.62	1.02, 2.58	0.042
Age < 50 years	0.98	0.95, 1.0	0.14	0.98	0.95, 1.01	0.21
Age ≥ 50 years	1.04	1.01, 1.07	0.005	1.02	0.99, 1.06	0.21
Black Race	1.36	0.95, 1.96	0.098	1.49	0.99, 2.24	0.053
Female	1.23	0.84, 1.80	0.28	1.44	0.96, 2.15	0.078
VAD	0.89	0.60, 1.34	0.58	1.06	0.58, 1.93	0.84
Diabetes	1.47	1.03, 2.10	0.035	1.26	0.85, 1.88	0.25
Creatinine (per 0.1 mg/dL)	1.06	1.03, 1.09	<0.001	1.05	1.02, 1.09	0.001
ICM	1.40	1.00, 1.98	0.053	1.44	0.97,2.15	0.070
IABP	2.01	1.13,3.59	0.018	1.75	0.75,4.12	0.20
UNOS 1B (reference=1A)	0.89	0.59, 1.33	0.57	1.14	0.61, 2.12	0.69
UNOS 2 (reference=1A)	0.84	0.53, 1.33	0.46	1.09	0.53, 2.23	0.82

VAD, ventricular assist device; ICM, ischemic cardiomyopathy; IABP, intra-aortic balloon pump;, UNOS, United Network for Organ Sharing

P-values derived from Cox proportional hazards model

Figure 2. Adjusted Kaplan-Meier Survival Curves by Listing Criteria.

*P-value is determined based on the hazard ratio estimated in a proportional hazards model, adjusted for baseline variables.

Major Morbidity

There were no differences between ECCT and SCCT in the index LOS or number of acute allograft rejection episodes during the first year (Table 5). A greater proportion of patients in the ECCT group were hospitalized ≥ 2 times during the first year. At 5-years following transplant, the median eGFR of SCCT patients was 42.8 ml/min/m² compared to 37.1 ml/min/m² in the ECCT cohort. Median serum Cr in the SCCT and ECCT groups was 1.70 mg/dL and 1.90 mg/dL, respectively. When classified into chronic kidney disease (CKD) stages, 75.6% and 88.2 % of SCCT and ECCT had CKD ≥ stage 3 at 1 year post-OHT, while 81.0% and 100% of SCCT and ECCT patients alive at 5 years following transplantation developed CKD ≥ 3 (Table 5). 4.1% of SCCT and 9.4% of ECCT patients alive at 5 years following transplant required permanent hemodialysis (Table 5).

Table 5.

Follow-up Outcomes by Listing Criteria

PARAMETER	STANDARD (SC)	EXTENDED (EC)	p-value
Total Number of Patients	370	84
Days to Discharge Post Transplant   (N=355 SC, 79 EC)	10 (8, 14)	11 (9, 18)	0.028
Rejections in the first year			0.69
0	162/328 (49.4%)	38/68 (55.9%)
1	99/328 (30.2%)	20/68 (29.4%)
2	44/328 (13.4%)	7/68 (10.3%)
3+	23/328 (7.0%)	3/68 (4.4%)
Rehospitalizations in the first year			0.069
0	155/328 (47.3%)	22/68 (32.4%)
1	80/328 (24.4%)	19/68 (27.9%)
2	40/328 (12.2%)	15/68 (22.1%)
3+	53/328 (16.2%)	12/68 (17.6%)
BUN (mg/dL)
1 year (N=328 SC, 68 EC)	26 (20, 33)	31 (24, 37)	0.001
5 year (N=174 SC, 32 EC)	27 (20, 37)	32 (25, 42)	0.022
eGFR (ml/min/1.73m2)
1 year (N=328 SC, 68 EC)	44.6 (35.3, 57.7)	42.6 (30.9, 53.3)	0.096
5 year (N= 174 SC, 32 EC)	42.8 (29.7, 54.8)	37.1 (25.0, 42.4)	0.034
Serum Creatinine (mg/dL)
1 year (N= 328 SC, EC 68)	1.6 (1.3, 2.0)	1.6 (1.3, 1.9)	0.52
5 year (N= 174 SC, 32 EC)	1.7 (1.3, 2.1)	1.9 (1.5, 2.3)	0.051
Chronic Kidney Disease (stage ≥3)
1 year	248/328 (75.6%)	60/68 (88.2%)	0.023
5 year	141/174 (81.0%)	32/32 (100.0%)	0.007
Dialysis at 5 years	7/169 (4.1%)	3/32 (9.4%)	0.20

BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate.

Continuous covariates are presented as Median (25^th, 75^th percentiles). Binary outcomes are presented as percentage.

Discussion

The concept underlying our ECCT program is to transplant carefully selected patients who fall outside traditional listing criteria utilizing organs that are commonly not utilized due to older donor age, UNOS high risk classification or an imperfection in the donor heart. The current analysis extends a prior analysis of this program with incorporation of data from an additional 34 ECCT patients transplanted between 2006 and 2010 and with longer-term follow-up of the entire cohort³. We show that ECCT patients have significantly lower risk-adjusted survival at both 1-and 5-years compared to SCCT. However, it appears that ECCT survival remains superior to that of a similarly ill patient population treated medically⁵, ¹², particularly patients requiring inotropic support. This risk-benefit ratio is less well-defined for the 36% of patients in our ECCT cohort that were listed as UNOS 2 at the time of transplant, however accurate estimates of mortality in ambulatory advanced heart failure are lacking. In a recent collaborative study from three transplant centers, one-year estimated survival using the Seattle Heart Failure Model (SHFM) for UNOS 2 recipients was 89% at the time of transplant¹³. However, the mean age of those in this analysis was 53 years (similar to our SCCT cohort) and the majority of high risk patients required urgent rather than elective transplantation. Thus it appears very likely that older UNOS 2 patients remain at high risk for death or need for urgent transplantation. Our survival data are also consistent with reported 1-year survival rates after ECCT from other high volume centers (78–83%) ^{6, 14} and suggests that ECCT remains an important strategy for the management of end stage heart failure. Moreover, given the increasing lifespan of patients undergoing OHT in the current era, our study highlights important clinical characteristics such as recipient age and renal function that are associated with long term outcomes after ECCT.

Our analysis demonstrates the importance of carefully evaluating recipient factors when selecting patients for ECCT. Decreased ECCT survival may not be unexpected, given the ECCT cohort was older, with a greater burden of comorbidities including DM and renal dysfunction. The mean age at transplantation of our ECCT cohort was 66 years compared to 53 years in the SCCT cohort. Recipient age greater than 50 years was an important predictor for post-transplant survival in the overall cohort, which remained predictive in a multivariate model. When listing criteria was used as a covariate (Model 2), age was no longer predictive for post-transplant survival reflecting our practice of assigning older patients to ECCT. Careful selection of older ECCT candidates is also reflected in the severity of illness prior to transplant. We observed that a greater proportion of ECCT patients were transplanted as UNOS status 2 and fewer were supported with a VAD at the time of transplantation. Our general approach has been to utilize durable mechanical circulatory support as destination therapy (DT) in hemodynamically unstable patients, thus leaving the ambulatory, medically treated advanced heart failure population for ECCT. This strategy has allowed us to treat older patients with advanced therapies without progression of end organ dysfunction which might become preclusive while awaiting a suitable organ.

Existing end organ dysfunction and other chronic underlying conditions are important contributors to the differences seen in post-transplant outcomes with ECCT. Prior work has demonstrated that renal function is a major risk factor for decreased survival after cardiac transplantation ^{1, 15}. Baseline renal function has been identified as an important risk factor for post-transplant survival in a recently published recipient risk score ¹⁶. In transplant recipients over the age of 65 years, lower Cr was associated with improved 5-year survival in recipients ¹⁷. In the current analysis we show that baseline renal function is a potent predictive factor for patients undergoing OHT with each 0.1 mg/dL increase in serum creatinine associated with a 5% increase in mortality (Table 4). In addition to the association with survival, post-transplant renal dysfunction was more prominent in the ECCT group. Of patients with five years alive and followed up, the proportion with ≥ stage 3 CKD was 81.0% and 100% for SCCT and ECCT, respectively. Of patients with 5 years of follow up, 4.1% of SCCT and 9.4% of ECCT patients developed end-stage renal disease requiring permanent hemodialysis. The increased proportion in the ECCT group can be attributed to worse baseline renal function and a higher proportion of subjects with DM at the time of transplant. Further, the adverse impact of immunosuppression may be greater in this older cohort whose renal impairment is underestimated by using serum Cr alone ¹⁶.

Another important outcome that was different between the two groups was the rate of rehospitalization during the first year following transplant. Our group has previously shown that the rate of rehospitalizations was significantly higher in ECCT³. In the current analysis we confirm the higher rate of rehospitalizations with the larger ECCT cohort. A greater proportion of ECCT patients required admission following transplantation (64.3% ECCT, 52% SCCT). Common causes for post-OHT hospitalizations include acute allograft rejection, infection, complications of CAV, renal failure, and malignancy ¹. We did not find any differences between the two groups in terms of rejection episodes or the development of CAV, but previous work did not demonstrate differences in CMV infection rates³.

Typically ECCT programs utilize donor organs considered unacceptable for SCCT. Our ECCT program utilizes cardiac allografts from older donors but the majority of organs did not show evidence of diminished EF or high inotropic support requirements (Table 2). Ischemic time in the ECCT group was significantly longer reflecting our willingness to utilize organs declined for recipients in neighboring UNOS regions (we do not select organs with an anticipated ischemic time > 4 hours). This is supported further by the higher median sequence number of the donors utilized in the ECCT group, which reflects use of organs from outside our local region or high risk features of the donor/organ not captured by the UNOS data. Both donor age and ischemic time are associated with increased risk of death after transplant particularly within the first year post transplant ¹⁶. This increased risk is often due to primary graft dysfunction (PGD)¹⁸ defined broadly as severe uni- or bi-ventricular dysfunction severe enough to not meet the circulatory requirements of the recipient within 24 hours of cardiac transplantation. Thus the delayed (> 1 year) mortality associated with ECCT is more likely related to recipient factors including age and renal dysfunction rather than impaired donor organ quality ¹⁹. Implicit in this finding is our acknowledgement of the likely interaction between donor age and ischemic time and their combined impact on short-term outcomes. While we have no pre-defined upper limit for donor age, we are unlikely to select an older donor with an anticipated ischemic time that is also prolonged to avoid the risk of PGD and early mortality.

Indeed, the cause of death for the majority of patients transplanted at an advanced age is typically unrelated to allograft function. Of eleven ECCT patient deaths from our group’s previous work, only one was due to rejection and one was potentially related to the function of the allograft ³. Another study of 62 ECCT patients with a 2-year survival of 68% reported 21 deaths over a median follow-up of 3.25 years. 7 of 21 deaths occurred at least 1 year after transplant; none of these deaths resulted from graft failure or rejection . In a large study examining the outcomes of marginal donors from the UNOS registry, infection but not PGD nor rejection was a leading cause of morbidity and mortality both at 1 year and 5 years after transplantation²⁰. Furthermore, a composite report from the International Society for Heart and Lung Transplantation shows that graft failure is the cause of death in less than 20% of heart transplant patients 5 years following transplantation ¹. Therefore, decisions regarding ECCT listing are appropriately based primarily on recipient risk factors. In addition, our analysis suggests careful attention to immunosuppressive regimens in older ECCT recipients that may experience a greater number of adverse effects associated with these agents (i.e. infection, malignancy and renal failure). We currently do not withhold induction therapy from ECCT recipients and have similar maintenance regimens for all patients transplanted at our center. Our results suggest a more personalized approach to immunosuppression in this high risk cohort may further improve outcomes.

These results highlight the clinical trade-off that must be considered when determining treatment strategies for the end-stage heart failure patient. In the current era, OHT provides the best long-term survival, especially for patients eligible for SCCT ¹. Although ECCT patients experience reduced survival and considerable long-term morbidity associated with renal dysfunction, there is little data to suggest that continued medical therapy without OHT provides a reasonable alternative. Alternative surgical strategies for patients ineligible for SCCT such as LVAD continue to demonstrate improved outcomes in the context of clinical trials²¹ and post-market analyses. Indeed, recent analyses suggest survival rates with “destination therapy” LVAD in excess of 60 % at 2 years²². These rates may improve with LVAD utilization in patients who are less sick and not in critical cardiogenic shock at the time surgery and this hypothesis is being tested in ongoing clinical trials. The survival, quality of life, and adverse event data from these studies will provide further insight into the appropriate utilization of transplantation and durable mechanical circulatory support in patients ineligible for SCCT. However, long-term follow up data with contemporary assist devices will not be available for a number of years and there have been few comparisons of ECCT with destination therapy LVAD ²³. Our results suggest that ECCT continues to have an important role in providing treatment options for a growing population of patients with end-stage HF. Optimal outcomes can be achieved with this strategy with careful patient selection based on common clinical variables such as renal function. Further, significant post-transplant morbidity may be related to the toxic effects of chronic immunosuppression and careful attention should be paid to tailoring immunosuppressive medications for this older population.

Limitations

Our study has a number of important limitations. This is a retrospective review of a single center experience using ECCT that focuses on recipient rather than donor characteristics. We chose to disregard individual donor factors in the multivariable models to prevent “over-fitting,” and therefore, we were unable to control for donor-specific factors that affect outcomes. In an attempt to correct for this, our analysis was built to determine the effects of recipient factors on survival as well as to define the additional risk associated with ECCT listing once recipient features are accounted for. Finally, as a result of inconsistencies and missing medical records we are not able to accurately and reliably report recipient cause of death.