Advertisement

Long-term transplant outcomes of donor hearts with left ventricular dysfunction

Open ArchivePublished:October 23, 2018DOI:https://doi.org/10.1016/j.jtcvs.2018.07.115

      Abstract

      Objective

      Despite small single-center reports demonstrating acceptable outcomes using donor hearts with left ventricular dysfunction, 19% of potential donor hearts are currently unused exclusively because of left ventricular dysfunction. We investigated modern long-term survival of transplanted donor hearts with left ventricular dysfunction using a large, diverse cohort.

      Methods

      Using the United Network for Organ Sharing database, we reviewed all adult heart transplants between January 2000 and March 2016. Baseline and postoperative characteristics and Kaplan–Meier survival curves were compared. A covariates-adjusted Cox regression model was developed to estimate post-transplant mortality. To address observed variation in patient profile across donor ejection fraction, a propensity score was built using Cox predictors as covariates in a generalized multiple linear regression model. All the variables in the original Cox model were included. For each recipient, a predicted donor ejection fraction was generated and exported as a new balancing score that was used in a subsequent Cox model. Cubic spline analysis suggested that at most 3 and perhaps no ejection fraction categories were appropriate. Therefore, in 1 Cox model we added donor ejection fraction as a grouped variable (using the spline-directed categories) and in the other as a continuous variable.

      Results

      A total of 31,712 donor hearts were transplanted during the study period. A total of 742 donor hearts were excluded for no recorded left ventricular ejection fraction, and 20 donor hearts were excluded for left ventricular ejection fraction less than 20%. Donor hearts with reduced left ventricular ejection fraction were from younger donors, more commonly male donors, and donors with lower body mass index than normal donor hearts. Recipients of donor hearts with reduced left ventricular ejection fraction were more likely to be on mechanical ventilation. Kaplan–Meier curves revealed no significant differences in recipient survival up to 15 years of follow-up (P = .694 log-rank test). Cox regression analysis showed that after adjustment for propensity variation, transplant year, and region, ejection fraction had no statistically significant impact on mortality when analyzed as a categoric or continuous variable. Left ventricular ejection fraction at approximately 1 year after transplantation was normal for all groups.

      Conclusions

      Carefully selected donor hearts with even markedly diminished left ventricular ejection fraction can be transplanted with long-term survival equivalent to normal donor hearts and therefore should not be excluded from consideration on the basis of depressed left ventricular ejection fraction alone. Functional recovery of even the most impaired donor hearts in this study suggests that studies of left ventricular function in the setting of brain death should be interpreted cautiously.

      Key Words

      Abbreviations and Acronyms:

      EF (ejection fraction), LV (left ventricular), LVEF (left ventricular ejection fraction), UNOS (United Network for Organ Sharing)
      Figure thumbnail fx1
      No statistically significant differences in Kaplan–Meier survival curves were found among groups stratified by LVEF.
      Hearts with diminished LVEF should not be excluded from consideration of transplantation on the basis of depressed LV function alone.
      Despite evidence that approximately 20% of donor hearts are not used for transplantation exclusively because of depressed LV function, our study finds that donor hearts with even markedly diminished LV function may be transplanted with long-term survival equivalent to hearts with normal LV function.
      See Commentaries on pages 1876, 1878, and 1880.
      Transplant rates have remained static over the past 15 years at approximately 2000 per year in the United States,
      • Lund L.H.
      • Edwards L.B.
      • Kucheryavaya A.Y.
      • Benden C.
      • Christie J.D.
      • Dipchand A.I.
      • et al.
      The registry of the International Society for Heart and Lung Transplantation: thirty-first official adult heart transplant report–2014.
      and donor use rates are trending downward.
      • Khush K.K.
      • Zaroff J.G.
      • Nguyen J.
      • Menza R.
      • Goldstein B.A.
      National decline in donor heart utilization with regional variability: 1995-2010.
      This has provoked recent reexamination of donor hearts previously excluded from consideration, including use of hearts with a history of resuscitated cardiac arrest
      • Galeone A.
      • Varnous S.
      • Lebreton G.
      • Barreda E.
      • Hariri S.
      • Pavie A.
      • et al.
      Impact of cardiac arrest resuscitated donors on heart transplant recipients' outcome.
      • Trivedi J.R.
      • Gallo M.
      • Slaughter M.S.
      Use of resuscitated donor hearts for transplantation: same results either side of the Atlantic Ocean.
      and hearts from older donors.
      • Prieto D.
      • Correlia P.
      • Baptista M.
      • Antunes M.J.
      Outcome after heart transplantation from older donor age: expanding the donor pool.
      We sought to examine another source of underused donor hearts: those with diminished left ventricular (LV) function. Among the reasons that hearts are declined for transplantation, ventricular dysfunction (often a sequel of the pathophysiologic environment caused by brain death) is among the most common and potentially tractable. Previous work from our group has documented that 19% of potential donor hearts (>1300 hearts per year) are not accepted for transplantation based entirely on LV dysfunction without evidence of structural disease.
      • Tryon D.
      • Hasaniya N.W.
      • Jabo B.
      • Razzouk A.J.
      • Bailey L.L.
      • Rabkin D.G.
      Effect of left ventricular dysfunction on utilization of donor hearts.
      This figure has not been affected by somewhat small, mostly single-center clinical studies
      • Smits J.M.
      • De Paum M.
      • de Vries E.
      • Rahmel A.
      • Meiser B.
      • Laufer G.
      • et al.
      Donor scoring system for heart transplantation and the impact on patient survival.
      • Khush K.K.
      • Menza R.
      • Nguyen J.
      • Zaroff J.G.
      • Goldstein B.A.
      Donor predictors of allograft utilization and recipient outcomes after heart transplantation.
      • Rao S.D.
      • Watson A.
      • Lo P.
      • Jabbour A.
      • Hayward C.S.
      • Keogh A.M.
      • et al.
      Successful orthotopic heart transplantation using donors with left ventricular systolic dysfunction: Evidence for brain death-induced Takotsubo cardiomyopathy.
      demonstrating equivalent short- and medium-term outcomes of dysfunctional donor hearts when compared with donor hearts with normal LV function or by experimental studies demonstrating functional resuscitation of hearts damaged by the pathophysiology of brain death.
      • Galinañes M.
      • Hearse D.J.
      Brain death-induced impairment of cardiac contractile performance can be reversed by explantation and may not preclude the use of the hearts for transplantation.
      • Mikhova K.M.
      • Don C.W.
      • Laflamme M.
      • Kellum J.A.
      • Mulligan M.S.
      • Verrier E.D.
      • et al.
      Effect of cytokine hemoadsorption on brain death-induced ventricular dysfunction in a porcine model.
      Brain death has been associated not only with impaired donor organ function, an effect that is pronounced in the heart, but also with postoperative complications, including primary graft dysfunction, increased immunogenicity resulting in a higher incidence of both acute and chronic rejection, and worse long-term outcomes.
      • Wilhelm M.J.
      • Pratschke J.
      • Beato F.
      • Taal M.
      • Kusaka M.
      • Hancock W.W.
      • et al.
      Activation of the heart by donor brain death accelerates acute rejection after transplantation.
      • Pratschke J.
      • Wilhelm M.J.
      • Laskowski I.
      • Kusaka M.
      • Beato F.
      • Tullius S.G.
      • et al.
      Influence of donor brain death on chronic rejection of renal transplants in rats.
      The mechanism of brain death–induced ventricular dysfunction is not well understood and likely multifactorial. Ischemic injury,
      • Halejcio-Delophont P.
      • Siaghy E.M.
      • Devaux Y.
      • Richoux J.P.
      • Bischoff N.
      • Carteaux J.P.
      • et al.
      Consequences of brain death on coronary blood flow and myocardial metabolism.
      direct catecholamine or cytokine-induced myocardial injury,
      • Mikhova K.M.
      • Don C.W.
      • Laflamme M.
      • Kellum J.A.
      • Mulligan M.S.
      • Verrier E.D.
      • et al.
      Effect of cytokine hemoadsorption on brain death-induced ventricular dysfunction in a porcine model.
      • White M.
      • Wieschmann R.J.
      • Roden R.L.
      • Hagan M.B.
      • Wollmering M.M.
      • Port J.D.
      • et al.
      Cardiac ß-adrenergic neuroeffector systems in acute myocardial dysfunction related to brain injury; evidence for catecholamine-mediated myocardial damage.
      and impaired ß-adrenergic receptor signaling
      • D'Amico T.A.
      • Meyers C.H.
      • Koutlas T.C.
      • Peterseim D.S.
      • Sabiston D.C.
      • Van Trigt P.
      • et al.
      Desensitization of myocardial ß-adrenergic receptors and deterioration of left ventricular function after brain death.
      have all been implicated. Although both regional and global LV dysfunction are common in the setting of brain death, neither is correlated with pathologic changes, indicating that the dysfunction is irreversible.
      • Dujardin K.S.
      • McCully R.B.
      • Wijdicks E.F.M.
      • Tazelaar H.D.
      • Seward J.B.
      • McGregor C.G.A.
      • et al.
      Myocardial dysfunction associated with brain death: clinical, echocardiographic, and pathologic features.
      • Mehra M.R.
      • Ventura H.O.
      Brain injury and ventricular dysfunction: insights into reversible heart failure.
      Given the experimental and available clinical evidence supporting the possibility of functional resuscitation of donor hearts with impaired ventricular function in the context of brain death, we believed further investigation was warranted. The current study is an effort to provide more long-term outcome data of donor hearts with LV dysfunction in the modern era of heart transplantation and determine whether consideration of expansion of the donor pool to include some of these organs is appropriate.

      Materials and Methods

      Our study design was submitted to the Institutional Review Board at Loma Linda University Medical Center, which determined that the study did not meet the definition of human subject research because it did not involve identifiable information, no data or specimens were collected, and there was no direct intervention or interaction. Therefore, the Institutional Review Board decided our study did not require review or approval.
      All isolated, adult (recipient aged ≥18 years), first-time heart transplants in the United States between January 1, 2000, and March 31, 2016, were analyzed. Data on donor hearts and post-transplant outcomes were obtained from the United Network for Organ Sharing (UNOS) as the contractor for the Organ Procurement and Transplantation Network provided by way of Standard Transplant Analysis and Research files. Hearts that were transplanted in conjunction with other organs were excluded from this analysis. This work was supported in part by Health Resources and Services Administration Contract 234-2005-37011C. The content is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.
      Donor hearts with left ventricular ejection fraction (LVEF) less than 20% were excluded from the study because we found the values to be implausible. Donor hearts without recorded LVEF were also excluded. The overwhelming majority of LVEF calculations were made by echocardiogram (95.1%), although left ventriculogram during left heart catheterization was used in 2.5% of cases and multi-gated acquisition scan was used in 0.03%. Baseline characteristics of both donor hearts and recipients were compared, as were recipient postoperative complications and overall survival.

       Statistical Analysis

      All data were analyzed with SAS 9.4 (SAS Institute Inc, Cary, NC). Kolmogorov–Smirnov analysis demonstrated that continuous variables were not normally distributed; therefore, data are reported as median with selected percentiles (5th, 25th, 75th, and 95th) above and below the median and compared using the Kruskal–Wallis test. The chi-square test was used to compare categoric variables. All reported P values were 2 sided. A univariate, restricted, cubic spline regression was performed using a least-squares model between donor LVEF and survival. We based the smoother on a restricted spline basis with 5 equal knots placed at the 1/6, 2/6, 3/6, 4/6, and 5/6 quantiles. Knots 1 to 5 corresponded with the following donor ejection fractions (EFs): 17.33%, 33.67%, 50.0%, 66.33%, and 82.67%, respectively. This model indicates a statistically significant fit (F = 22.60, P < .0001) with Knot 5 being statistically different from Knot 2 (P = .011), Knot 3 (P = .006), and Knot 4 (P = .040). Knot 1 was not identified as being statistically different from Knot 5 (P = .269). Because of the appearance of the cubic spline curve (Figure 1), we thought that at most 3 groups (group 1: EF ˂40, group 2: EF 40-≤60, and group 3: EF >60) and perhaps no groups would be most appropriate.
      Figure thumbnail gr1
      Figure 1Univariate, restricted cubic spline regression using a least-squares model between donor LVEF and survival. LV, Left ventricular.
      Kaplan–Meier curves with donor hearts grouped by LV function, as indicated by the spine analysis, were generated to compare the overall survival of the groups. Given the imbalance in sample number and resultant events among the groups, Kaplan–Meier curves were truncated for less than 10 patients at risk in an attempt to attenuate wide confidence intervals.
      A covariates-adjusted Cox regression model with robust standard errors was developed to estimate post-transplant mortality hazard ratios with 95% confidence interval for LV dysfunction. Covariates considered for inclusion in the model included all available donor and recipient data (Tables 1 and 2), as well as UNOS region and transplant era (we divided the 17 years of the study into 4 different periods: 2000-2003, 2004-2007, 2008-2011, and 2012-2016). The Cox model assumes that hazard ratios should be constant over time (proportional hazards). We assessed each categoric covariate for whether the assumption of proportional hazards was reasonable. Where nonproportional hazards were present, we considered a stratified model. We assessed proportional hazards with a plot of standardized score residuals over time. We concluded that if the residuals were unusually large at any time point, the proportional hazards assumption was violated. In that case, we produced plots of randomly generated score processes to allow for graphic assessment of the observed residuals. To address observed variation in patient profile across donor EF, a propensity score was built using Cox predictors as covariates in a generalized multiple linear regression model.
      • Rosenbaum P.R.
      • Rubin D.B.
      The central role of the propensity score in observational studies for causal effects.
      All the variables in the original Cox model were included. For each recipient, a predicted donor EF was generated and exported as a new balancing score that was used in a subsequent Cox model. As noted earlier, Cubic spline analysis suggested that at most 3 groups and perhaps no groups were appropriate. We evaluated donor EF separately as a grouped variable and then as a continuous variable in the Cox model.
      Table 1Comparison of baseline heart donor characteristics
      Group 1 (n = 124)Group 2 (n = 9203)Group 3 (n = 21,623)P value
      Age
      Age is reported as median years with 5th, 25th, 75th, and 95th percentiles.
      21.0 (13.0, 17.0, 28.0, 43.0)28.0 (16.0, 21.0, 39.0, 52.0)30.0 (16.0, 22.0, 42.0, 53.0)<.001
      Gender (%)<.001
       Male76.673.269.8
       Female23.426.830.2
      Race (%)
       White75.066.766.6.002
       Black16.115.314.1
       Hispanic8.8715.316.2
       Asian0.001.311.82
       Indigenous
      American Indian, Alaska Native, or Native Hawaiian/Other Pacific Islander.
      0.000.610.69
       Multiracial0.000.670.57
       Unknown0.000.050.01
      Body mass index
      Body mass index is reported as median kg/m2 with 5th, 25th, 75th, and 95th percentiles.
      23.7 (18.3, 21.6, 26.8, 33.5)25.4 (19.6, 22.7, 29.1, 36.7)26.0 (19.7, 23.1, 29.8, 37.7)<.001
      Cause of death (%)
       Anoxia19.417.316.2<.001
       Stroke12.920.724.5
       Head trauma66.159.356.5
       Brain tumor0.000.680.86
       Other1.612.081.88
      Inotropes at procurement (%)51.753.551.8.091
      Triiodothyronine at procurement4.031.961.40<.001
      Ischemic time
      Ischemic time is reported as median hours with 5th, 25th, 75th, and 95th percentiles.
      3.23 (1.55, 2.47, 3.77, 4.60)3.13 (1.53, 2.42, 3.82, 4.85)3.20 (1.57, 2.48, 3.87, 4.95).211
      LVEF
      LVEF is reported as median % with 5th, 25th, 75th, and 95th percentiles.
      33.0 (19.0, 30.0, 35.0, 37.0)55.0 (45.0, 51.0, 55.0, 58.0)65.0 (60.0, 60.0, 70.0, 75.0)<.001
      LVEF, Left ventricular ejection fraction.
      Age is reported as median years with 5th, 25th, 75th, and 95th percentiles.
      American Indian, Alaska Native, or Native Hawaiian/Other Pacific Islander.
      Body mass index is reported as median kg/m2 with 5th, 25th, 75th, and 95th percentiles.
      § Ischemic time is reported as median hours with 5th, 25th, 75th, and 95th percentiles.
      LVEF is reported as median % with 5th, 25th, 75th, and 95th percentiles.
      Table 2Comparison of baseline heart recipient characteristics
      Group 1 (n = 124)Group 2 (n = 9203)Group 3 (n = 21,623)P value
      Baseline
       Age
      Age is reported as median years with 5th, 25th, 75th, and 95th percentiles.
      54.5 (25.0, 44.5, 62.0, 67.0)55.0 (26.0, 46.0, 62.0, 68.0)55.0 (26.0, 46.0, 62.0, 68.0).556
       Gender (%).341
      Male74.275.874.8
      Female25.824.225.2
       Race (%)
      White71.870.670.2.006
      Black16.119.217.9
      Hispanic10.56.67.8
      Asian1.612.542.88
      Indigenous0.000.520.71
      Multiracial0.000.480.52
       Body mass index
      Body mass index is reported as median kg/m2 with 5th, 25th, 75th, and 95th percentiles.
      26.5 (18.6, 22.7, 29.5, 33.2)26.5 (19.5, 23.4, 29.9, 35.3)26.6 (19.5, 23.4, 30.2, 35.3).983
       Ventilator
      At time of transplant (%).
      4.031.521.64<.001
       ECMO
      At time of transplant (%).
      0.810.55.59.681
       LVAD
      At time of transplant (%).
      9.2113.514.0.115
       Inotropes
      At time of transplant (%).
      54.743.541.7<.001
       Total days on waiting list
      Median days with 5th, 25th, 75th, and 95th percentiles.
      81.5 (4, 22.5, 232.5, 810)89 (4, 26, 251, 830)89 (4, 27, 252, 853).024
       Days status 1A
      Median days with 5th, 25th, 75th, and 95th percentiles.
      15 (1, 6, 35, 107)21 (0, 0, 28, 98)21 (2, 8, 40, 133).441
       Days status 1B
      Median days with 5th, 25th, 75th, and 95th percentiles.
      22.5 (2, 7, 96, 282)49 (2, 13, 136, 419)48 (2, 13, 132, 411).011
       Days status 2
      Median days with 5th, 25th, 75th, and 95th percentiles.
      115 (1.5, 25, 408, 916)95 (3, 27, 287, 940)96 (3, 26, 287, 928).009
      Postoperative
       ECMO at 72 h (%)0.810.550.59.681
       Length of stay
      Median days with 5th, 25th, 75th, and 95th percentiles.
      13 (7, 10, 24, 83)14 (7, 10, 22, 53)14 (7, 10, 21, 53)<.001
       New dialysis (%)9.6810.3510.19<.001
       Pacemaker (%)3.233.103.50<.001
       New stroke (%)0.002.482.39<.001
       Treated for rejection in 1 y (%)20.1619.8819.76<.001
       LVEF at 1 y
      LVEF reported as median percent with 5th, 25th, 75th, and 95th percentiles.
      60 (45, 55, 64, 70)60 (45, 55, 65, 70)60 (50, 55, 65, 71)<.001
      ECMO, Extracorporeal membrane oxygenation; LVAD, left ventricular assist device; LVEF, left ventricular ejection fraction.
      Age is reported as median years with 5th, 25th, 75th, and 95th percentiles.
      Body mass index is reported as median kg/m2 with 5th, 25th, 75th, and 95th percentiles.
      At time of transplant (%).
      § Median days with 5th, 25th, 75th, and 95th percentiles.
      LVEF reported as median percent with 5th, 25th, 75th, and 95th percentiles.

      Results

      During the study period, 31,712 isolated, first-time heart transplants were performed in recipients older than 18 years of age. Of these, 124 were in group 1 (EF <40%), 9203 were in group 2 (EF 40%-59.9%), and 21,623 were in group 3 (≥60%). Median LVEF for groups 1 to 3 was 33.0%, 55.0%, and 65.0%, respectively (P < .001). A total of 742 donor hearts were excluded for not having a recorded LVEF, and 20 donor hearts were excluded because of LVEF less than 20%. Mean follow-up was 1890 ± 1564 (standard deviation) days.
      Comparison of baseline heart donor characteristics is shown in Table 1. Median age varied inversely with ventricular function, and differences were statistically significant (P < .001). Group 3 had the highest proportion of female donors (30.2%), and group 1 had the lowest (23.4%) (P < .001). Donors in group 1 were significantly more likely to be white (P = .002). Group 1 had the lowest median body mass index at 23.7 kg/m2, whereas group 3 had the highest at 26.0 kg/m2 (P < .001). Donors in group 1 were more likely to have died of anoxia and head trauma and less likely to have died of stroke (P < .001). They were also more likely to be on triiodothyronine at the time of procurement. Comparison of recipient characteristics is shown in Table 2. There were no statistically significant differences in median age, gender, median body mass index, or use of mechanical circulatory support before transplant. Recipients of donor hearts from group 1 were significantly more likely to be on ventilator and inotropic support at the time of transplant.
      Kaplan–Meier survival curves demonstrated no statistically significant differences in survival among the 3 groups at up to 15 years of follow-up (P = .694) (Figure 2). There were no significant differences in the need for extracorporeal membrane oxygenation at 72 hours after transplant, and differences in length of stay, new indication for dialysis, new pacemaker, and treatment for rejection within 1 year of transplant were not clinically significant. Differences in LVEF at approximately 1 year after transplantation were statistically but not clinically significant because all LVEF groups had normalized (Table 2).
      Figure thumbnail gr2
      Figure 2Kaplan–Meier survival curves with 95% Hall–Wellner bands. No significant differences in post-transplant Kaplan–Meier survival curves were found among groups stratified by LVEF. LVEF less than 40% blue, LVEF 40% to 59.9% red, LVEF 60% or greater green. EF, Ejection fraction.
      Distribution of transplants across donor heart function groups varied significantly among UNOS regions (Table 3); however, no trend emerged between regional volume and willingness to transplant hearts with LV dysfunction. In multivariate analysis, UNOS region did emerge as an independent predictor of mortality (Tables 4 and 5), with UNOS region 6 demonstrating best overall survival. Likewise, transplant era emerged as an independent predictor of survival (Tables 4 and 5), with the most recent era (2012-2016) having the best overall survival.
      Table 3Hearts transplanted in each left ventricular ejection fraction group by United Network for Organ Sharing region
      Group 1Group 2Group 3P value
      UNOS region (% of total transplants performed)<.001
       1 (3.8%)12 (1.0%)384 (32.2%)778 (65.3%)
       2 (13.3%)24 (0.57%)1348 (32.1%)2680 (63.7%)
       3 (12.1%)5 (0.13%)1105 (28.9%)2652 (69.4%)
       4 (10.7%)10 (0.29%)979 (28.8%)2392 (70.3%)
       5 (15.8%)16 (0.32%)1031 (20.6%)3923 (78.2%)
       6 (3.0%)0 (0.0%)193 (20.0%)761 (79.0%)
       7 (9.3%)11 (0.37%)797 (27.1%)1905 (64.8%)
       8 (5.5%)3 (0.17%)483 (27.5%)1256 (71.4%)
       9 (6.7%)14 (0.66%)739 (34.9%)1297 (61.2%)
       10 (8.7%)13 (0.47%)877 (31.9%)1791 (65.2%)
       11 (11.2%)16 (0.45%)1285 (36.0%)2213 (62.1%)
      UNOS, United Network for Organ Sharing.
      Table 4Cox regression hazard ratios and confidence intervals using left ventricular ejection fraction as a categoric variable
      ParameterParameter estimateStandard errorChi-squareP valueHR95% Wald confidence limits
      Group 2 (EF 40%-59.9%)0.0910.2950.095.7581.095
      HR compared with group 1 (EF >60%).
      0.615-1.951
      Group 1 (EF <40%)0.0310.0520.353.5531.031
      HR compared with group 1 (EF >60%).
      0.932-1.142
      Age recipient0.0150.00264.45<.0011.015
      HR reflects the effect of a 1-year change in age.
      1.011-1.019
      Age donor0.0030.0022.491.1151.003
      HR reflects the effect of a 1-year change in age.
      0.999-1.007
      Female gender0.0270.0530.250.6171.0270.925-1.140
      Race/ethnicity recipient
       Black0.0960.0552.996.0841.101
      HR compared with white ethnicity.
      0.987-1.227
       Hispanic0.0130.0950.019.8921.013
      HR compared with white ethnicity.
      0.841-1.220
       Asian0.3350.1863.256.0711.399
      HR compared with white ethnicity.
      0.971-2.013
       American Indian−0.0670.4170.026.8730.935
      HR compared with white ethnicity.
      0.413-2.117
       Native Hawaiian−0.4280.3441.543.2140.652
      HR compared with white ethnicity.
      0.332-1.280
       Multiracial0.0210.2950.005.9441.021
      HR compared with white ethnicity.
      0.573-1.818
      Race/ethnicity donor
       Black0.0610.0670.821.3651.063
      HR compared with white ethnicity.
      0.932-1.212
       Hispanic0.0960.0721.763.1841.101
      HR compared with white ethnicity.
      0.955-1.268
       Asian0.3670.2132.963.0851.443
      HR compared with white ethnicity.
      0.950-2.192
       American Indian−0.4470.3082.105.1470.640
      HR compared with white ethnicity.
      0.350-1.170
       Native Hawaiian−0.3340.7150.219.6400.716
      HR compared with white ethnicity.
      0.176-2.908
       Multiracial0.1640.3250.254.6141.178
      HR compared with white ethnicity.
      0.623-2.229
      BMI recipient0.0090.0053.375.0661.0090.999-1.018
      Days on wait list0.0000.0000.045.8331.0001.000-1.000
      UNOS region
       10.7590.2439.761.0022.135
      HR compared with UNOS region 6.
      1.327-3.437
       20.6270.2049.412.0021.873
      HR compared with UNOS region 6.
      1.254-2.796
       30.5570.1977.946.0051.745
      HR compared with UNOS region 6.
      1.185-2.569
       40.4290.1875.281.0221.536
      HR compared with UNOS region 6.
      1.065-2.215
       50.0380.1740.048.8271.039
      HR compared with UNOS region 6.
      1.739-1.461
       70.5720.2037.983.0051.772
      HR compared with UNOS region 6.
      1.192-2.636
       80.5090.1966.737.0091.664
      HR compared with UNOS region 6.
      1.133-2.445
       90.6810.2209.585.0021.976
      HR compared with UNOS region 6.
      1.284-3.041
       100.7800.22012.54<.0012.182
      HR compared with UNOS region 6.
      1.417-3.360
       111.0980.22523.81<.0012.998
      HR compared with UNOS region 6.
      1.929-4.660
      Era
       2000-20030.169090.0370620.814<.0011.184
      HR compared with era 2012-2016.
      1.101-1.273
       2004-20070.136190.0370513.508.00021.146
      HR compared with era 2012-2016.
      1.066-1.232
       2008-20110.039440.038181.0668.30171.040
      HR compared with era 2012-2016.
      0.965-1.121
      Propensity score0.1540.03419.97<.0011.166
      HR reflects 1-unit increase in the conditional probability of donor ejection fraction group given individual covariates.
      1.090-1.248
      Ischemic time (h)0.0680.0229.245.0021.0701.024-1.118
      Length of stay (d)−0.0010.0010.372.5420.9990.997-1.001
      HR, Hazard ratio; EF, ejection fraction; BMI, body mass index; UNOS, United Network for Organ Sharing.
      HR compared with group 1 (EF >60%).
      HR reflects the effect of a 1-year change in age.
      HR compared with white ethnicity.
      § HR compared with UNOS region 6.
      HR compared with era 2012-2016.
      HR reflects 1-unit increase in the conditional probability of donor ejection fraction group given individual covariates.
      Table 5Cox regression hazard ratios and confidence intervals using left ventricular ejection fraction as a continuous variable
      ParameterParameter estimateStandard errorChi-squareP valueHR95% Wald confidence limits
      Ejection fraction−0.0010.0030.171.6790.999
      HR reflects a 1% change in ejection fraction.
      0.993-1.005
      Age recipient0.0150.00264.48<.0011.015
      HR for age reflect the effect of a 1-year change in age.
      1.011-1.019
      Age donor0.0030.0022.491.1231.003
      HR for age reflect the effect of a 1-year change in age.
      0.999-1.007
      Female gender0.0270.0530.250.6201.0270.925-1.140
      Race/ethnicity recipient
       Black0.0960.0552.996.0841.101
      HR compared with white ethnicity.
      0.987-1.227
       Hispanic0.0130.0950.019.8861.013
      HR compared with white ethnicity.
      0.841-1.220
       Asian0.3350.1863.256.0751.399
      HR compared with white ethnicity.
      0.971-2.013
       American Indian−0.0670.4170.026.8750.935
      HR compared with white ethnicity.
      0.413-2.117
       Native Hawaiian−0.4280.3441.543.2140.652
      HR compared with white ethnicity.
      0.332-1.280
       Multiracial0.0210.2950.005.9461.021
      HR compared with white ethnicity.
      0.573-1.818
      Race/ethnicity donor
       Black0.0600.0670.821.3701.063
      HR compared with white ethnicity.
      0.932-1.212
       Hispanic0.0970.0721.763.1801.101
      HR compared with white ethnicity.
      0.955-1.268
       Asian0.3640.2132.963.0881.443
      HR compared with white ethnicity.
      0.950-2.192
       American Indian−0.4400.3082.105.1530.640
      HR compared with white ethnicity.
      0.350-1.170
       Native Hawaiian−0.3430.7150.219.6320.716
      HR compared with white ethnicity.
      0.176-2.908
       Multiracial0.1650.3250.254.6131.178
      HR compared with white ethnicity.
      0.623-2.229
      BMI recipient0.0090.0053.375.0671.0090.999-1.018
      Days on wait list0.0000.0000.045.8341.0001.000-1.000
      UNOS region
       10.7550.2439.761.0022.135
      HR compared with UNOS region 6.
      1.327-3.437
       20.6230.2049.412.0021.873
      HR compared with UNOS region 6.
      1.254-2.796
       30.5520.1977.946.0051.745
      HR compared with UNOS region 6.
      1.185-2.569
       40.4260.1875.281.0231.536
      HR compared with UNOS region 6.
      1.065-2.215
       50.0330.1740.048.8471.039
      HR compared with UNOS region 6.
      1.739-1.461
       70.5580.2037.983.0051.772
      HR compared with UNOS region 6.
      1.192-2.636
       80.5060.1966.737.0101.664
      HR compared with UNOS region 6.
      1.133-2.445
       90.6770.2209.585.0021.976
      HR compared with UNOS region 6.
      1.284-3.041
       100.7780.22012.54<.0012.182
      HR compared with UNOS region 6.
      1.417-3.360
       111.0950.22523.81<.0012.998
      HR compared with UNOS region 6.
      1.929-4.660
      Era
       2000-20030.169090.0370620.814<.0011.184
      HR compared with era 2012-2016.
      1.101-1.273
       2004-20070.136190.0370513.508.00021.146
      HR compared with era 2012-2016.
      1.066-1.232
       2008-20110.039440.038181.0668.30171.040
      HR compared with era 2012-2016.
      0.965-1.121
      Propensity score0.1530.03419.85<.0011.166
      HR reflects 1-unit increase in the conditional probability of donor ejection fraction group given individual covariates.
      1.090-1.248
      Ischemic time (h)0.0680.0229.245.0021.0701.025-1.118
      Length of stay (d)−0.0010.0010.373.5410.9990.997-1.001
      HR, Hazard ratio; BMI, body mass index; UNOS, United Network for Organ Sharing.
      HR reflects a 1% change in ejection fraction.
      HR for age reflect the effect of a 1-year change in age.
      HR compared with white ethnicity.
      § HR compared with UNOS region 6.
      HR compared with era 2012-2016.
      HR reflects 1-unit increase in the conditional probability of donor ejection fraction group given individual covariates.
      Cox regression analysis using EF as a categoric variable (Table 4) and as a continuous variable (Table 5) showed that after adjustment for propensity variation, transplant year, and transplant region, EF had no statistically significant impact on mortality.

      Discussion

      Our study provides modern, robust, long-term data on survival of recipients of donor hearts with suboptimal LV systolic function and suggests a pathway to explore expansion of the donor pool by reclaiming a portion of the donor pool unused because of poor function. Limitations of our work include those inherent in all retrospective, observational studies and in those dependent on manual data entry. In addition, during the period of study, digitization of echocardiograms became more common, and therefore there was increased ability for transplant centers to review copies of the echocardiogram performed and interpreted at the procurement center. Therefore, there may have been disparities between the EF recorded at the procurement center and the transplant center's cardiologist's or surgeon's interpretation of the same echocardiogram. Furthermore, because only 1 EF is recorded in the UNOS database, it is also possible that subsequent echocardiograms were performed demonstrating a change in function but not recorded in the UNOS database.
      Despite these drawbacks, several useful conclusions emerge. Using a large, contemporary database and a balancing methodology that accounts for baseline differences in donor and recipient characteristics, and looking at EF as both a categoric variable guided by spline analysis and a continuous variable, we showed that after adjustment for transplant year and transplant region, EF had no statistically significant impact on mortality. At the very least, this should provide reassurance about using donor hearts with LVEF 50% to 54.9% and hopefully discourage the surprisingly common practice of nonuse of these hearts based exclusively on the perception that their function is compromised.
      • Tryon D.
      • Hasaniya N.W.
      • Jabo B.
      • Razzouk A.J.
      • Bailey L.L.
      • Rabkin D.G.
      Effect of left ventricular dysfunction on utilization of donor hearts.
      Perhaps more compelling than the survival data was the demonstration that ventricular function across all donor heart groups normalized at approximately 1 year (Table 2 and Video 1). This supports prior experimental studies demonstrating functional recovery of hearts damaged by the pathophysiology of brain death after the organ is removed or protected from the hostile environment produced by brain death
      • Galinañes M.
      • Hearse D.J.
      Brain death-induced impairment of cardiac contractile performance can be reversed by explantation and may not preclude the use of the hearts for transplantation.
      • Mikhova K.M.
      • Don C.W.
      • Laflamme M.
      • Kellum J.A.
      • Mulligan M.S.
      • Verrier E.D.
      • et al.
      Effect of cytokine hemoadsorption on brain death-induced ventricular dysfunction in a porcine model.
      and suggests that studies of LV function in the setting of brain death should be interpreted cautiously.
      Figure thumbnail fx2
      Video 1Dr David G. Rabkin and the recipient of a donor heart with low LVEF discuss the patient's experiences. Video available at: https://www.jtcvs.org/article/S0022-5223(18)32826-5/fulltext.
      Our study demonstrates that reduced LV systolic function alone should not be used as a basis for nonuse. This is a common occurrence. Our group has demonstrated that isolated LV systolic dysfunction is used as the sole explanation for nonuse in 19% of potential donors whose hearts were declined for transplantation, amounting to approximately 1300 hearts per year.
      • Tryon D.
      • Hasaniya N.W.
      • Jabo B.
      • Razzouk A.J.
      • Bailey L.L.
      • Rabkin D.G.
      Effect of left ventricular dysfunction on utilization of donor hearts.
      Of the donor hearts with recorded LVEF included in that study, 60.41% had LVEF 40% or greater, included in group 2 or 3. This represents approximately 785 hearts per year.
      After our article had been accepted for presentation, Dr Chen and colleagues
      • Chen C.W.
      • Sprys M.H.
      • Gaffey A.C.
      • Chung J.J.
      • Margulies K.B.
      • Acker M.A.
      • et al.
      Low ejection fraction in donor hearts is not directly associated with increased recipient mortality.
      published a study with somewhat different methodology but demonstrating similar results. They categorized patients into 3 cohorts by donor EF: less than 40% (reduced EF), 40% to 50% (borderline EF), and 50% or more (normal EF). Propensity score matching was performed using 1-to-1 matching without replacement to separately compare reduced EF cases with normal EF cases, and borderline EF cases with normal EF cases. Baseline variables with statistically significant differences among the cohorts were chosen for propensity score matching. Using the UNOS database, Chen and colleagues
      • Chen C.W.
      • Sprys M.H.
      • Gaffey A.C.
      • Chung J.J.
      • Margulies K.B.
      • Acker M.A.
      • et al.
      Low ejection fraction in donor hearts is not directly associated with increased recipient mortality.
      showed that after matching baseline donor and recipient characteristics, patients who received hearts with “reduced” and “borderline” function had equivalent 1-year survival compared with patients who received hearts with LVEF greater than 50%. Our study supports and extends their results to demonstrate equivalent long-term survival of donor hearts with diminished ventricular function.

      Conclusions

      Alternative therapies for patients on the heart transplant waiting list include continued waiting for primary transplant with modern waiting list mortality (or removal from transplant list because of deterioration) for status 1a of 15% at 180 days,
      • Singh T.P.
      • Almond C.S.
      • Taylor D.O.
      • Graham D.A.
      Decline in heart transplant wait list mortality in the United States following broader regional sharing of donor hearts.
      bridge to transplant using an LV assist device with equivalent short-term survival but higher incidence of pretransplant complications (gastrointestinal bleeding, pump thrombosis/failure, LV assist device-related infection, cerebrovascular accident, and development of human leukocyte antigen antibodies) with unclear impact on long-term outcomes and no guarantee of receiving a heart with normal LV function,
      • Shah H.
      • Kahanda M.
      • Schilling J.
      • LaRue S.
      • Shuster J.
      • Sitner S.
      The impact of LVAD complications on short-term outcomes after heart transplantation.
      • Drakos S.G.
      • Kfoury A.G.
      • Kotter J.R.
      • Reid B.B.
      • Clayson S.E.
      • Selzman C.H.
      • et al.
      Prior human leukocyte antigen-allosensitization and left ventricular assist device type affect degree of post-implantation human leukocyte antigen-allosensitization.
      • Arnaoutakis G.J.
      • George T.J.
      • Kilic A.
      • Weiss E.S.
      • Russell S.D.
      • Conte J.V.
      • et al.
      Effect of sensitization in US heart transplant recipients bridged with a ventricular assist device: update in a modern cohort.
      • Bull D.A.
      • Reid B.B.
      • Selzman C.H.
      • Mesley R.
      • Drakos S.
      • Clayson S.
      • et al.
      The impact of bridge-to-transplant ventricular assist device support on survival after cardiac transplantation.
      high-risk conventional surgery for selected patients,
      • Kawajiri J.
      • Manlhiot C.
      • Ross H.
      • Delgado D.
      • Billia F.
      • McDonald M.
      • et al.
      High-risk cardiac surgery as an alternative to transplant or mechanical support in patients with end-stage heart failure.
      or forgoing these therapies in favor of end-of-life care.
      • Whellan D.J.
      • Goodlin S.J.
      • Dickinson M.G.
      • Heidenreich P.A.
      • Jaenicke C.
      • Stough W.G.
      • et al.
      End-of-life care in patients with heart failure.
      Although even if a large number of donor hearts with reduced ventricular function were used, there would still be a significant disparity between available and needed organs. Given these alternatives, nonuse of donor hearts with diminished LVEF on the basis of depressed function alone may represent a waste of valuable organs for transplantation.

       Conflict of Interest Statement

      Authors have nothing to disclose with regard to commercial support.

      Supplementary Data

      References

        • Lund L.H.
        • Edwards L.B.
        • Kucheryavaya A.Y.
        • Benden C.
        • Christie J.D.
        • Dipchand A.I.
        • et al.
        The registry of the International Society for Heart and Lung Transplantation: thirty-first official adult heart transplant report–2014.
        J Heart Lung Transplant. 2014; 33: 996-1008
        • Khush K.K.
        • Zaroff J.G.
        • Nguyen J.
        • Menza R.
        • Goldstein B.A.
        National decline in donor heart utilization with regional variability: 1995-2010.
        Am J Transplant. 2015; 15: 642-649
        • Galeone A.
        • Varnous S.
        • Lebreton G.
        • Barreda E.
        • Hariri S.
        • Pavie A.
        • et al.
        Impact of cardiac arrest resuscitated donors on heart transplant recipients' outcome.
        J Thorac Cardiovasc Surg. 2017; 153: 622-630
        • Trivedi J.R.
        • Gallo M.
        • Slaughter M.S.
        Use of resuscitated donor hearts for transplantation: same results either side of the Atlantic Ocean.
        J Thorac Cardiovasc Surg. 2017; 154: 539
        • Prieto D.
        • Correlia P.
        • Baptista M.
        • Antunes M.J.
        Outcome after heart transplantation from older donor age: expanding the donor pool.
        Eur J Cardiothorac Surg. 2015; 47: 672-678
        • Tryon D.
        • Hasaniya N.W.
        • Jabo B.
        • Razzouk A.J.
        • Bailey L.L.
        • Rabkin D.G.
        Effect of left ventricular dysfunction on utilization of donor hearts.
        J Heart Lung Transplant. 2018; 37: 349-357
        • Smits J.M.
        • De Paum M.
        • de Vries E.
        • Rahmel A.
        • Meiser B.
        • Laufer G.
        • et al.
        Donor scoring system for heart transplantation and the impact on patient survival.
        J Heart Lung Transplant. 2012; 31: 387-397
        • Khush K.K.
        • Menza R.
        • Nguyen J.
        • Zaroff J.G.
        • Goldstein B.A.
        Donor predictors of allograft utilization and recipient outcomes after heart transplantation.
        Circ Heart Fail. 2013; 6: 300-309
        • Rao S.D.
        • Watson A.
        • Lo P.
        • Jabbour A.
        • Hayward C.S.
        • Keogh A.M.
        • et al.
        Successful orthotopic heart transplantation using donors with left ventricular systolic dysfunction: Evidence for brain death-induced Takotsubo cardiomyopathy.
        J Heart Lung Transplant. 2015; 34: S71-S72
        • Galinañes M.
        • Hearse D.J.
        Brain death-induced impairment of cardiac contractile performance can be reversed by explantation and may not preclude the use of the hearts for transplantation.
        Circ Res. 1992; 71: 1213-1219
        • Mikhova K.M.
        • Don C.W.
        • Laflamme M.
        • Kellum J.A.
        • Mulligan M.S.
        • Verrier E.D.
        • et al.
        Effect of cytokine hemoadsorption on brain death-induced ventricular dysfunction in a porcine model.
        J Thorac Cardiovasc Surg. 2013; 145: 215-224
        • Wilhelm M.J.
        • Pratschke J.
        • Beato F.
        • Taal M.
        • Kusaka M.
        • Hancock W.W.
        • et al.
        Activation of the heart by donor brain death accelerates acute rejection after transplantation.
        Circulation. 2000; 102: 2426-2433
        • Pratschke J.
        • Wilhelm M.J.
        • Laskowski I.
        • Kusaka M.
        • Beato F.
        • Tullius S.G.
        • et al.
        Influence of donor brain death on chronic rejection of renal transplants in rats.
        J Am Soc Nephrol. 2001; 12: 2474-2481
        • Halejcio-Delophont P.
        • Siaghy E.M.
        • Devaux Y.
        • Richoux J.P.
        • Bischoff N.
        • Carteaux J.P.
        • et al.
        Consequences of brain death on coronary blood flow and myocardial metabolism.
        Transplant Proc. 1998; 30: 2840-2841
        • White M.
        • Wieschmann R.J.
        • Roden R.L.
        • Hagan M.B.
        • Wollmering M.M.
        • Port J.D.
        • et al.
        Cardiac ß-adrenergic neuroeffector systems in acute myocardial dysfunction related to brain injury; evidence for catecholamine-mediated myocardial damage.
        Circulation. 1995; 92: 2183-2189
        • D'Amico T.A.
        • Meyers C.H.
        • Koutlas T.C.
        • Peterseim D.S.
        • Sabiston D.C.
        • Van Trigt P.
        • et al.
        Desensitization of myocardial ß-adrenergic receptors and deterioration of left ventricular function after brain death.
        J Thorac Cardiovasc Surg. 1995; 110: 746-751
        • Dujardin K.S.
        • McCully R.B.
        • Wijdicks E.F.M.
        • Tazelaar H.D.
        • Seward J.B.
        • McGregor C.G.A.
        • et al.
        Myocardial dysfunction associated with brain death: clinical, echocardiographic, and pathologic features.
        J Heart Lung Transplant. 2001; 20: 350-357
        • Mehra M.R.
        • Ventura H.O.
        Brain injury and ventricular dysfunction: insights into reversible heart failure.
        Congest Heart Fail. 2005; 11: 99-101
        • Rosenbaum P.R.
        • Rubin D.B.
        The central role of the propensity score in observational studies for causal effects.
        Biometrika. 1983; 70: 41-55
        • Chen C.W.
        • Sprys M.H.
        • Gaffey A.C.
        • Chung J.J.
        • Margulies K.B.
        • Acker M.A.
        • et al.
        Low ejection fraction in donor hearts is not directly associated with increased recipient mortality.
        J Heart Lung Transplant. 2017; 36: 611-615
        • Singh T.P.
        • Almond C.S.
        • Taylor D.O.
        • Graham D.A.
        Decline in heart transplant wait list mortality in the United States following broader regional sharing of donor hearts.
        Circ Heart Fail. 2012; 5: 249-258
        • Shah H.
        • Kahanda M.
        • Schilling J.
        • LaRue S.
        • Shuster J.
        • Sitner S.
        The impact of LVAD complications on short-term outcomes after heart transplantation.
        J Heart Lung Transplant. 2016; 35: S264-S265
        • Drakos S.G.
        • Kfoury A.G.
        • Kotter J.R.
        • Reid B.B.
        • Clayson S.E.
        • Selzman C.H.
        • et al.
        Prior human leukocyte antigen-allosensitization and left ventricular assist device type affect degree of post-implantation human leukocyte antigen-allosensitization.
        J Heart Lung Transplant. 2009; 28: 838-842
        • Arnaoutakis G.J.
        • George T.J.
        • Kilic A.
        • Weiss E.S.
        • Russell S.D.
        • Conte J.V.
        • et al.
        Effect of sensitization in US heart transplant recipients bridged with a ventricular assist device: update in a modern cohort.
        J Thorac Cardiovasc Surg. 2011; 142: 1236-1245
        • Bull D.A.
        • Reid B.B.
        • Selzman C.H.
        • Mesley R.
        • Drakos S.
        • Clayson S.
        • et al.
        The impact of bridge-to-transplant ventricular assist device support on survival after cardiac transplantation.
        J Thorac Cardiovasc Surg. 2010; 140: 169-173
        • Kawajiri J.
        • Manlhiot C.
        • Ross H.
        • Delgado D.
        • Billia F.
        • McDonald M.
        • et al.
        High-risk cardiac surgery as an alternative to transplant or mechanical support in patients with end-stage heart failure.
        J Thorac Cardiovasc Surg. 2017; 154: 517-525
        • Whellan D.J.
        • Goodlin S.J.
        • Dickinson M.G.
        • Heidenreich P.A.
        • Jaenicke C.
        • Stough W.G.
        • et al.
        End-of-life care in patients with heart failure.
        J Card Fail. 2014; 20: 121-134