Author + information
- Received March 25, 2017
- Revision received July 3, 2017
- Accepted July 10, 2017
- Published online August 28, 2017.
- Shivank Madan, MD, MHAa,
- Omar Saeed, MDa,
- Peter Vlismas, MDb,
- Ioanna Katsa, MDc,
- Snehal R. Patel, MDa,
- Julia J. Shin, MDa,
- William A. Jakobleff, MDd,
- Daniel J. Goldstein, MDd,
- Daniel B. Sims, MDa and
- Ulrich P. Jorde, MDa,∗ ()
- aDivision of Cardiology, Department of Medicine, Montefiore Medical Center/Albert Einstein College of Medicine, Bronx, New York
- bDepartment of Medicine, Montefiore Medical Center/Albert Einstein College of Medicine, Bronx, New York
- cDepartment of Medicine, Jacobi Medical Center/Albert Einstein College of Medicine, Bronx, New York
- dDepartment of Cardiovascular and Thoracic Surgery, Montefiore Medical Center/Albert Einstein College of Medicine, Bronx, New York
- ↵∗Address for correspondence:
Dr. Ulrich P. Jorde, Department of Medicine, Albert Einstein College of Medicine/Montefiore Medical Center, 111 East 210th Street, Bronx, New York 10467.
Background Left ventricular systolic dysfunction (LVSD) accounts for almost 25% of nonacceptance of potential donor hearts. Previous smaller, single-center studies showed that LVSD following brain death may be transient, and such hearts can be successfully resuscitated with resolution of LVSD, then transplanted.
Objectives This study evaluated outcomes of donor hearts with LVSD on initial transthoracic echocardiogram (TTE) that resolved during donor management.
Methods We reviewed echocardiograms of all cardiac donors in the United Network of Organ Sharing database that were transplanted from January 1, 2007, to September 30, 2015, and identified 472 donor hearts with LVSD (left ventricular ejection fraction [LVEF] ≤40%) on initial TTE that resolved (LVEF ≥50%) during donor management on a subsequent TTE. These patients comprised the improved donor LVEF group. These were compared with donor hearts with normal LVEF (LVEF ≥55%) on the initial TTE for recipient mortality, cardiac allograft vasculopathy (CAV), and primary graft failure (PGF).
Results There was no significant difference in recipient mortality at 30 days, 1 year, 3 years, and 5 years of follow-up, nor any difference in rates of PGF at 90 days and CAV at 5 years between recipients of donor hearts with improved LVEF and recipients of donor hearts with initially normal LVEF. Post-transplant length of stay was also similar between the 2 groups. Using propensity scores, 461 transplants in the improved-donor LVEF group were matched to 461 transplants in the normal-donor LVEF group. There was no significant difference in PGF at 90 days or recipient mortality after up to 5 years of follow-up.
Conclusions In the largest analysis of donor hearts with transient LVSD, we found that such hearts can be successfully resuscitated and transplanted without increasing recipient mortality, CAV, or PGF. These results underscore the importance of appropriate donor management and should help to increase utilization of donor hearts with transient LVSD.
- cardiac allograft vasculopathy
- ejection fraction
- length of stay
- primary graft failure
It is estimated that >20,000 patients in the United States may benefit from heart transplantation every year, yet only about 2,000 to 2,400 adult heart transplants are performed in the United States annually, largely due to limited donor organ availability (1,2). Although it is true that donor organ identification and consent rates need to increase, there is also a need to safely increase utilization of already available donor organs using evidence-based criteria (2,3). More than two-thirds of currently available donor organs are discarded due to stringent acceptance criteria that have not been rigorously tested in clinical or research settings (2). Left ventricular systolic dysfunction (LVSD) is 1 of the most common causes of rejecting potential donors, accounting for up to 25% of the nonuse of donor hearts (4). Even in the absence of previous cardiac disease, LVSD may occur in up to 42% of patients with brain death (5).
Studies suggest that if it were possible to prevent, reverse, or safely ignore donor LVSD without adversely affecting recipient outcomes, more than 800 additional grafts (about 30% of the annual adult transplant volume in the United States) (2,6) would be available for transplantation each year (4). Small, single-center studies have suggested that the LVSD associated with the “autonomic storm” of brain death is transient, and that such donor hearts could be successfully resuscitated and transplanted (7,8). However, many transplant centers continue to reject donors with LVSD, largely due to a lack of consensus regarding the management of such donor hearts, as well as the absence of any robust post-transplant outcomes data. Hence, in the largest study done to date, we evaluated the outcomes of donor hearts with LVSD on the initial transthoracic echocardiogram (TTE) that improved on the following TTE during donor management.
Data source and study cohort
We used the nationwide, multicenter Organ Procurement and Transplant Network database, managed by the United Network of Organ Sharing (UNOS). Patient-level nonidentifiable information was retrieved for key characteristics of organ donors and transplant recipients. No separate informed consent was required, and the study was deemed exempt by the Institutional Review Board at Albert Einstein College of Medicine, Montefiore Medical Center, given that the UNOS data are deidentified, publicly available, and used in compliance with the UNOS data user agreement.
All adult (age ≥18 years) heart transplant recipients who received a transplant between January 1, 2007, and September 30, 2015, were eligible for this analysis. The start date was chosen to reflect the donor acceptance practices and transplant outcomes post-donor allocation changes introduced in 2006. A separate query was requested for detailed donor TTE (including multiple TTEs performed for the same donor during donor management) and donor inotropic data, and was then merged with the standard transplant analysis files. Main exclusion criteria included multiorgan transplants, repeat transplants, donor age >55 years, structural abnormality or significant coronary artery disease in the donor heart, missing TTE information, or left ventricular ejection fraction (LVEF) <55% (if there was only 1 donor TTE). From all donors who had multiple TTEs during donor management, we identified donor hearts with LVSD on the initial TTE (defined as LVEF ≤40%) that resolved during donor management (LVEF improved to ≥50%) based on the follow-up TTE. This constituted the improved donor LVEF group. Transplants using improved donor LVEF hearts were compared to transplants using donor hearts with initially normal LVEF (≥55%) on the single TTE that was submitted to UNOS (normal donor LVEF group). Donor hearts with “low-normal” or “borderline” LVEF (50% to <55%) were not included in the normal donor LVEF group (9,10).
Definitions and outcomes
Size mismatch was defined as donor weight ≥30% below the recipient weight (11), and sex mismatch was defined as transplant from a female donor to a male recipient (12). The primary outcome of the study was all-cause mortality during up to 5 years of follow-up. Secondary outcomes assessed included primary graft failure (PGF) up to 90 days, cardiac allograft vasculopathy (CAV) up to 5 years, and post-transplant hospital length of stay (LOS). Similar to previous UNOS analyses, PGF was defined in terms of hard outcomes, such as “death or retransplantation within 90 days of the index cardiac transplant due to graft failure that was not related to infection, rejection, or surgical technical issues” (3,13). CAV diagnosis was based on self-reporting to UNOS by transplant centers at yearly patient follow-up. Recipients with missing or unknown CAV status were excluded from CAV analysis.
Baseline recipient and donor characteristics for the 2 groups (improved donor LVEF vs. normal donor LVEF) were compared using the chi-square test for categorical variables, and unpaired Student t test and Wilcoxon rank sum test for continuous variables, with a normal and non-normal distribution, respectively. Recipient mortality, PGF, and CAV in the 2 groups were compared using unadjusted and adjusted Cox proportional hazards regression models and Kaplan-Meier analysis.
Due to a difference in sample size of the 2 study groups and the large number of variables with a potential to confound association between improved donor LVEF and outcomes, we also used a propensity-score analysis to identify a cohort of recipient-donor matches with similar baseline characteristics. The propensity score is a probability of having an exposure (improved donor LVEF heart vs. normal-donor LVEF heart), conditioned on a set of baseline characteristics (14). The propensity score was estimated using a multivariable, logistic-regression model, with “use of improved donor LVEF heart” as the treatment status regressed on baseline recipient and donor characteristics. Matching was performed using a 1:1 matching algorithm (greedy type matching) without replacement, with a caliper width of 0.2 of the SD of the logit of the propensity score. Post-match covariate balance was checked using standardized differences. A standardized difference <10% indicated good balance between the groups (15,16). Recipient mortality in the propensity-matched cohorts was compared using Cox models and Kaplan-Meier analysis. All statistical analyses were performed using Stata 13 software (StataCorp LLC, College Station, Texas), and 2-sided p values <0.05 were considered significant.
There were 17,584 adult heart transplants during the study period, and 15,853 met the initial inclusion/exclusion criteria. Of the 4,449 donor hearts with at least 2 or more TTEs performed during donor listing and management, we identified 472 donor hearts with LVSD on initial TTE (LVEF ≤40%) that improved to LVEF ≥50% on a follow-up TTE (the improved donor LVEF group). The normal-donor LVEF group consisted of 11,223 donor hearts with normal LVEF (≥55%) on the single TTE done during donor listing and management (Figure 1). The breakdown of 4,449 heart transplants with multiple donor TTEs is shown in Online Table 1. The median number of TTEs for all 4,449 donors with multiple TTEs and for the 472 donors who met the final study criteria for the improved donor LVEF group in both cases was 2 (interquartile range [IQR]: 2 to 3).
The baseline recipient characteristics of the improved- and normal-donor LVEF groups were similar (Table 1). Compared with the normal-donor LVEF group, the donors in the improved-donor LVEF group were younger (age 25 years [IQR: 20 to 32 years] vs. 30 years [IQR: 22 to 41 years]; p < 0.001), had a lower LVEF at transplant (57% [IQR: 55% to 60%] vs. 61% [IQR: 60% to 65%]; p < 0.001), and had a higher prevalence of brain anoxic injury as a cause of death (35.38% vs. 17.78%; p < 0.001) (Table 2).
Kaplan-Meier estimates of 30-day, 1-year, 3-year, and 5-year recipient survival for the overall cohort were 96.0%, 90.1%, 83.5%, and 76.8% respectively. Using Cox models, there was no significant difference in mortality between the normal- and improved-donor LVEF groups at 30 days (4.0% vs. 4.1%), 1 year (9.9% vs. 10.6%), 3 years (16.5% vs. 16.9%), and 5 years (23.2% vs. 24.4%) of follow-up. The results were consistent in Cox models adjusted for differences in baseline characteristics and for recipient and donor characteristics known to affect post-transplant outcomes (Table 3). Kaplan-Meier analysis of recipient survival also showed no significant difference in recipient mortality after up to 5 years of follow-up (p = 0.630 for log-rank test) (Figure 2A). We also stratified our analysis based on donor age (<30 and ≥30 years) and again found no significant difference in recipient mortality between the groups (Online Table 2). Causes of death in the improved LVEF group post-transplant are shown in Online Table 3.
Recipients of hearts with improved-donor LVEF had similar rates of PGF leading to death or retransplantation at up to 90 days of follow-up compared with recipients of hearts with normal-donor LVEF using unadjusted and adjusted Cox models (Table 3). Kaplan-Meier analysis also showed similar rates of PGF for the 2 groups (p = 0.573 for log-rank test) (Figure 2B). Additionally, we evaluated post-transplant LOS, as immediate post-transplant conditions like PGF, infection, rejection, and so on would be expected to increase post-transplant LOS. There was no significant difference in the median post-transplant LOS between improved- and normal-donor LVEF groups (14 days [IQR: 10 to 23 days] vs. 14 days [IQR: 10 to 21 days]; p = 0.4518).
Follow-up information on CAV was available for 7,972 heart transplants (normal-donor LVEF n = 7,663; improved-donor LVEF n = 309) and was included in the CAV analysis. Transplant recipients of improved-donor LVEF hearts were not at increased risk of developing CAV, with up to 5 years of follow-up compared with recipients of donor hearts with normal LVEF, using unadjusted and adjusted Cox models (Table 3). Kaplan-Meier analysis showed similar rates of CAV for the 2 groups at 5 years (improved LVEF vs. normal LVEF: 33.0% vs. 36.4%; p = 0.123 for log-rank test) (Figure 2C).
With the use of propensity scores, 461 transplants with improved-donor LVEF hearts were matched to 461 transplants with normal-donor LVEF hearts. After matching, there were no significant differences in the baseline recipient and donor characteristics, and all standardized differences were <10%, indicating good balance between the groups (Tables 4 and 5). Using Cox models, there was no significant difference in mortality between the recipients of improved- and normal-donor LVEF hearts at 30-day, 1-year, 3-year, and 5-year follow-up (Table 6). In addition, there was no significant difference between the 2 groups in recipient mortality using Kaplan-Meier analysis (p = 0.690 for log-rank test) (Figure 2D). Similarly, rates of PGF did not differ between the 2 groups, whether donor hearts had improved or normal LVEF (1.3% vs. 1.3%, respectively; p = 0.993 for log-rank test; Cox hazard ratio: 0.99 [95% confidence interval: 0.32 to 3.11]; p = 0.994).
Changes in LVEF during donor management
For the 472 donors in the improved LVEF group, the median initial donor LVEF at listing was 35% (IQR: 25% to 40%), improving to 57% (IQR: 55% to 60%) pre-transplant. The number of donors with initial LVEF ≤25%, 26% to 35%, and 36% to 40% were 130, 184, and 158, respectively. The median time difference between the initial TTE with LVSD and the subsequent TTE showing improved left ventricular function was 18.0 h (IQR: 11.2 to 43.1 h). Compared with those in the normal LVEF group, donors in the improved LVEF group had greater use of dobutamine (19.3% vs. 3.9%; p < 0.001), epinephrine (8.5% vs. 4.2%; p < 0.001), and norepinephrine (63.3% vs. 54%; p < 0.001), but similar use of dopamine and neosynephrine during donor management. In the improved LVEF group, the peak doses were higher for dobutamine (3 mcg/kg/min [IQR: 2 to 5 mcg/kg/min] vs. 1 μg/kg/min [IQR: 0.5 to 2.5 μg/kg/min]; p < 0.001) and norepinephrine (0.20 μg/kg/min [IQR: 0.07 to 0.40 μg/kg/min] vs. 0.14 μg/kg/min [IQR: 0.07 to 0.31 μg/kg/min]; p <0.001) (Table 7). Furthermore, compared with donors in the normal LVEF group, donors in the improved LVEF group had greater use of pre-recovery triiodothyronine (2.33% vs. 0.90%; p = 0.007), but similar use of pre-recovery thyroxine (77.5% vs. 73.6%; p = 0.143), pre-recovery diuretics (69.9% vs. 68.7%; p = 0.656), and pre-recovery steroids (73.1% vs. 75.6%; p = 0.385). In the improved LVEF group, 43 donors (9.1%) had spontaneous recovery without needing inotropes or vasopressors at any point during donor management, and 166 donors (35.2%) required inotropes or vasopressors only transiently during donor management and were not on any inotropes by the time of organ procurement.
Peak donor troponin I levels were available for 430 donor hearts in the improved LVEF and 9,137 donor hearts in the normal LVEF group. Compared with donor hearts with normal LVEF, those with improved LVEF had higher peak troponin I levels (1.68 ng/ml [IQR: 0.48 to 4.35 ng/ml] vs. 0.25 ng/ml [IQR: 0.06 to 0.99 ng/ml]; p < 0.001) and were more likely to be transplanted out of the UNOS region from where the donor offer originated (33.26% vs. 24.09%; p < 0.001). We also found a significant trend for an increasing number of donor hearts with improved LVEF being transplanted annually, from 27 in 2007 to 75 in 2014 (p < 0.001 for linear trend) (Figure 3).
Low donor LVEF and donor acceptance
In a separate analysis, we found that between November 1, 1999 (approximately when UNOS regularly started reporting donor LVEF), and September 30, 2015, information on donor LVEF and donor heart disposition was available for 67,690 donors. Of these, 31,995 donor hearts were not transplanted, and their disposition was as follows: consent not requested (n = 397), consent not obtained (n = 344), organ not recovered (n = 21,677), organ was recovered but not for transplant (n = 9,247), and organ recovered for transplant but not transplanted (n = 330). Of these 31,995 donor hearts that were not transplanted, donor LVEF was ≤40% in 8,448 hearts (26.4%). Although it is possible that these donor hearts (with LVEF ≤40%) might have been rejected for other reasons as well (e.g., older donors, structural abnormalities, coronary disease, elevated cardiac biomarkers, and so on), low donor LVEF could have been a major contributory reason for nonprocurement of these donor hearts. This is consistent with previously published studies (4) and underscores the potential for increasing donor heart utilization.
In this largest analysis to date of donor hearts with LVSD on an initial TTE that improved during donor management, when comparing donor hearts with normal versus improved LVEF, there was: 1) no significant difference in recipient mortality at 30-day, 1-year, 3-year, or 5-year follow-up; 2) no significant difference in CAV at up to 5 years of follow-up; 3) no significant difference in rates of PGF leading to death or retransplantation at up to 90 days of follow-up; and 4) no difference in post-transplant LOS. Finally, we found similar results using propensity-score matching analysis (i.e., there was no significant difference in either recipient mortality with up to 5 years of follow-up or PGF between the 2 donor LVEF groups). These results provide concrete evidence that donor hearts with LVSD on the initial echocardiogram, if appropriately managed, can be successfully resuscitated and have good short- and intermediate-term outcomes. Considering that LVSD accounts for 25% to 30% of nonuse of donor hearts, the findings of this study should be encouraging and help to increase donor utilization rates.
Several pathophysiological mechanisms for myocardial dysfunction following brain death have been proposed (Central Illustration). Brain death leads to a catecholamine surge (“sympathetic storm”) that initially causes hypertension, tachycardia, and increased cardiac output. This is followed by a subsequent catecholamine depletion and desensitization of myocardial beta-adrenergic receptors, leading to a decrease in myocardial contraction, loss of sympathetic tone, and reduction in systemic vascular resistance (17). This vasodilation and afterload reduction also leads to decreased coronary perfusion, triggering further myocardial dysfunction and cardiovascular collapse (18). Szabó et al. (19) elegantly showed in situ that if coronary perfusion pressure was decoupled from aortic pressure and elevated to pre-brain death levels, then myocardial contractility following brain death could be reversed to baseline levels. An early pilot study found that of 13 potential donors with LVSD on the initial TTE, 12 had improvement in left ventricular systolic function with intense donor management, and they showed a 92% survival at an average follow-up of 16 months (7). Borbely et al. (20) found that of 29 brain-dead heart donors with cardiac dysfunction on the initial TTE, 15 (52%) demonstrated resolution of cardiac dysfunction over time.
In our analysis of 472 donors with LVSD (LVEF ≤40%) on the initial TTE that resolved on a subsequent TTE (LVEF ≥50%) during donor management, we found that such hearts could be successfully resuscitated and transplanted with excellent outcomes in terms of recipient mortality, CAV, PGF, and post-transplant LOS. Furthermore, in a stratified analysis, we found no difference in recipient mortality between normal- and improved-donor LVEF groups in both younger and older donors. The International Society of Heart and Lung Transplant recommends against the routine use of donor hearts requiring excessive inotropic support (e.g., dopamine at 20 μg/kg/min or similar doses of other adrenergic agents) despite aggressive optimization of pre- and after-load (11). It is interesting to note that although the use of some vasopressors or inotropes (dobutamine, epinephrine, and norepinephrine) and their peak doses (dobutamine and norepinephrine) were higher in the improved LVEF group, the median peak doses of the vasopressors or inotropes were much lower than the International Society of Heart and Lung Transplant–recommended thresholds, suggesting that transplant centers across the United States continue to avoid using donor hearts requiring moderate- or high-dose inotropic support. We also found that improved LVEF donors were more likely to be transplanted out of the UNOS region from where the donor offer originated initially. Additionally, there has been a gradual increase in the number of improved LVEF donor hearts being transplanted in the United States each year (Figure 3). This is likely because of continued shortage of donors and due to increased experience among transplant centers in the perioperative care of such transplants.
In a recent analysis of 127 donor hearts with a singly noted LVEF <40%, Chen et al. (21) showed 1-year survival to be 84.6% versus 91.3% for donor hearts with normal LVEF (>50%). Although this did not reach statistical significance (p = 0.125), it revealed that donors with an isolated reduced LVEF could have lower 1-year survival. In contrast, we identified a subset of donor hearts with LVSD on initial TTE that resolved with donor management on a subsequent TTE and had excellent short- and intermediate-term outcomes compared with donor hearts with initially normal LVEF.
Our study strongly suggests that transplant centers should not reject potential donor hearts based on LVSD on a single TTE alone. TTE should be repeated in such hearts after hormonal and hemodynamic optimization. If LVSD resolves on the subsequent echocardiogram without relying on high doses of inotropes, such hearts can be successfully transplanted with good outcomes. This also means that physicians in intensive care units (cardiologists, intensivists, and neurologists) should not exclude patients as potential heart donors based on a single TTE with LVSD alone, especially if the donor is young and in the setting of brain death. Such patients should be considered “possible donors,” and local organ procurement organizations should be involved so that there is a chance of donor optimization and reassessment (Figure 4).
Thus, the results of this analysis should not only help to increase the use of existing donor hearts, but also should help increase the actual number of patients who are considered potential heart donors. From the transplant physician’s perspective, the current study should help in making informed decisions when evaluating donor hearts with transient LVSD.
The results of this study should be interpreted in the context of several important limitations. First, due to the retrospective and registry-based nature of this analysis, we were unable to delineate specific donor management practices (i.e., more granular details about relationship of donor hemodynamics to various interventions) that were associated with improvement in donor heart function. Future multicenter prospective studies regarding donor management of such marginal hearts are required given that current donor management practices vary considerably across U.S. transplant centers. Second, we were unable to determine the rates of PGF requiring inotropes or temporary mechanical circulatory support, as such information is not captured by UNOS; this may have underestimated the true prevalence of PGF. Third, as UNOS does not have CAV information on all recipients, transplants with missing information on CAV follow-up were excluded from the CAV analysis. Fourth, there is a possibility of an inherent selection bias in this study, as we only report the outcomes of donor hearts with improved LVSD that were transplanted. Regrettably, we cannot determine the outcomes of donor hearts with improved LVSD that were not transplanted and/or why those hearts might not have been used. Fifth, the outcomes of donor hearts with transient LVSD in combination with other multiple donor risk factors (e.g., older donors or longer ischemic time) are unknown and need to be established. Additionally, long term (10-year) outcomes of donor hearts with improved LVEF remain unknown. Finally, the suggested algorithm to evaluate candidacy for donor hearts with LVSD is based on a retrospective analysis and needs to be interpreted in the setting of this limitation.
Donor hearts with LVSD on the initial echocardiogram that resolves during donor management on a subsequent echocardiogram can be successfully transplanted without increasing short- and intermediate-term recipient mortality, PGF, or CAV. Our results underscore the importance of donor management and should provide confidence to transplant centers to consider hearts with transient LVSD. Finally, physicians in the intensive care units should not exclude their patients as potential heart donors based on LVSD on a single echocardiogram.
COMPETENCY IN PATIENT CARE: Because LVSD may be transient after brain death, cardiac function should be re-evaluated after appropriate management for suitability as donor hearts.
TRANSLATIONAL OUTLOOK: Improvement in donor management strategies could increase organ availability for cardiac transplantation.
For supplemental tables, please see the online version of this article.
All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- cardiac allograft vasculopathy
- length of stay
- left ventricular systolic dysfunction
- primary graft failure
- transthoracic echocardiogram
- United Network of Organ Sharing
- Received March 25, 2017.
- Revision received July 3, 2017.
- Accepted July 10, 2017.
- 2017 American College of Cardiology Foundation
- Madan S.,
- Saeed O.,
- Shin J.,
- et al.
- Tsao C.W.,
- Lyass A.,
- Larson M.G.,
- et al.
- Fonarow G.C.,
- Hsu J.J.
- Weiss E.S.,
- Allen J.G.,
- Patel N.D.,
- et al.
- Borbely X.I.,
- Krishnamoorthy V.,
- Modi S.,
- et al.
- Chen C.W.,
- Sprys M.H.,
- Gaffey A.C.,
- et al.