Heart and Combined Heart–Kidney Transplantation in Patients With Concomitant Renal Insufficiency and End-Stage Heart Failure

Authors


Abstract

In patients with end-stage heart failure (ESHF) who are candidates for isolated heart transplant (HRT), dialysis dependence (DD) is considered an indication for combined heart–kidney transplantation (HKT). HKT remains controversial in ESHF transplant candidates with nondialysis-dependent renal insufficiency (NDDRI). Using United Network for Organ Sharing data, we examined the cumulative incidences of transplant and mortality in patients with DD and NDDRI waitlisted for HKT or HRT. In all groups, 3-month waitlist mortality was dismal: 31% and 21% for HRT- and HKT-listed patients with DD and 12% and 7% for HRT- and HKT-listed patients with NDDRI. Five-year posttransplant survival was improved in HKT recipients compared with HRT recipients for both patients with DD (73% vs. 51%, p < 0.001) and NDDRI (80% vs. 69%, p < 0.001). Likewise, multivariable analysis associated HKT with better outcomes than HRT in HKT-listed patients, although both improved survival. These data argue strongly for HKT in ESHF transplant candidates with DD. However, in patients with NDDRI, HKT must be weighed against the possibility of renal recovery with isolated HRT. Whether HRT (followed by a staged kidney transplant in patients who do not recover renal function after HRT), as opposed to HKT, maximizes organ benefit for patients with NDDRI and ESHF requires assessment. Nevertheless, given their dismal waitlist outcomes and excellent posttransplant results, we suggest that patients with DD and NDDRI with ESHF be considered for early listing and transplant.

Abbreviations
CI

confidence interval

DD

dialysis dependence

ESHF

end-stage heart failure

HKT

heart–kidney transplantation

HR

hazard ratio

HRT

heart transplant

NDDRI

nondialysis-dependent renal insufficiency

NYHA

New York Heart Association

UNOS

United Network for Organ Sharing

VAD

ventricular assist device

Introduction

Renal insufficiency is a comorbidity associated with early mortality in patients with end-stage heart failure (ESHF) who are waitlisted for isolated heart transplant (HRT) [1]. The magnitude of renal insufficiency affects these patients' waitlist outcomes: patients with dialysis dependence (DD) listed for HRT fare worse than patients with nondialysis-dependent renal insufficiency (NDDRI) [2]. Furthermore, patients with renal insufficiency have worse posttransplant outcomes, as do patients who develop renal dysfunction after HRT [3, 4]. In patients with post-HRT renal dysfunction, receiving a renal allograft greatly mitigates their increased risk of mortality [4]. Given these findings, combined heart–kidney transplantation (HKT) is increasingly being employed as a therapeutic strategy for the subgroup of patients with concomitant ESHF and renal insufficiency, for whom HRT may not be an optimal solution.

Early experience with HKT has shown that in well-selected patients, HKT recipients have equivalent survival to HRT recipients [5]. More recent analyses have suggested that the benefits of HKT are limited to patients with low-risk or DD, whereas patients with high-risk and NDDRI receive only limited benefit from HKT [6, 7]. However, these investigators did not match patients to account for treatment-selection bias, nor did they take into account differences in waitlist survival when making their recommendations [8]. For this reason, we reviewed the United Network for Organ Sharing (UNOS) Standard Transplant Analysis and Research database to compare waitlist and posttransplant outcomes in propensity-matched HKT and HRT waitlisted/transplanted patients with concomitant renal insufficiency.

Methods

Data collection and study population

We performed a retrospective review of de-identified data from the UNOS Standard Transplant Analysis and Research files. Our Institutional Research Ethics Board granted the study an exemption because no patient identifiers were included in the analysis. We identified 42 278 useable records of patients listed for HRT between January 1, 2000 and December 31, 2012. We excluded 6435 children (i.e. patients <18 years of age) listed for transplant. We also excluded an additional 260 patients listed for combined heart–liver transplant, 172 patients listed for heart–lung, 11 patients listed for heart–kidney–pancreas and 3 patients listed for heart–pancreas, leaving us with a total of 35 397 adults listed for HKT (n = 1356) or HRT (n = 34 041) over the period studied. We performed a separate analysis of posttransplant outcomes, using data from 28 562 patients who underwent HRT or HKT between January 1, 2000 and December 31, 2012. From this cohort, we excluded 4193 children (i.e. patients <18 years of age) and 112 heart–liver recipients, leaving us with a total of 24 257 adults who underwent HKT (n = 637) or HRT (n = 23 620).

Primary end points

The primary end point for the waitlisted patients was a composite of death on the waitlist and removal from the waitlist because of clinical deterioration (i.e. the patient became too sick to receive a transplant); we refer to this composite end point as “waitlist mortality” because nearly all patients who become too sick to receive a transplant die shortly thereafter. Patients who underwent transplantation and those who were removed from the list because of recovery or for other reasons were censored at the time of transplant or recovery; otherwise, patients were administratively censored at the last time of follow-up. The primary end point for the transplant recipients was posttransplant death; patients were administratively censored at the last time of follow-up. Clinical variables were defined at the time of listing for the waitlisted patients and at the time of transplant for the transplanted patients.

Statistical analysis

Demographics and clinical status at the time of listing and time of transplant were compared between HKT- and HRT-listed patients by using two-sample t-tests for continuous variables and the chi-square or Fisher exact test for categorical variables. Our analysis consisted of six comparisons between the patients with HKT and HRT, stratified by level of renal insufficiency (DD and NDDRI). Three of these comparisons examined the incidences of transplant and mortality in all waitlisted patients (unmatched), in waitlisted patients with DD (propensity-matched) and in waitlisted patients with NDDRI (propensity-matched). The other three comparisons examined posttransplant survival in all patients (unmatched), in patients with DD (propensity-matched) and in patients with NDDRI (propensity-matched).

Variables with missing values were imputed to avoid case deletion in our multivariable and propensity-matching analyses [9, 10]. Two separate multiple imputations involving all nonredundant variables were performed (one for our analyses from the time of listing and one for our analyses from the time of transplant) by using a regression switching (chained equations) approach with predictive mean matching [11, 12]. Twenty imputations were performed because of our large N and small amount of missing data (<25%) [13]. The complete sets of nonredundant observed values (at listing [Table 1] and at transplant [Table 2]) were used as covariates for prediction purposes [14].

Table 1. Baseline characteristics in patients listed for heart transplant stratified by whether the patient was listed for a concomitant kidney transplant, in both unadjusted and propensity-matched analyses
 All patientsPropensity-matched patients with DDPropensity-matched patients with NDDRI
Nonmissing data (n = 35 397)Listed for HKT (n = 1356)Listed for HRT (n = 34 041)p*Listed for HKT (n = 146)Listed for HRT (n = 146)p*Listed for HKT (n = 525)Listed for HRT (n = 525)p*
  • Data are reported as number (%), mean ± standard deviation or median (interquartile range [IQR]).
  • CrCl, creatinine clearance; DD, dialysis dependence; ECMO, extracorporeal membrane oxygenation; HKT, heart–kidney transplantation; HRT, heart transplant; IABP, intra-aortic balloon pump; NDDRI, nondialysis-dependent renal insufficiency; NYHA, New York Heart Association; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure; VAD, ventricular assist device.
  • *p-Value based on chi-square or Student t-test analysis.
  • Boldface data denote statistical significance at the p < 0.05 level.
Baseline characteristics
Age at listing, years35 394 (99.9)51.4 ± 11.951.6 ± 12.40.6349.1 ± 12.848.6 ± 13.30.7253.7 ± 11.453.0 ± 12.80.35
Age >60 years at listing35 394 (99.9)337 (24.9)8947 (26.3)0.2432 (21.9)29 (19.9)0.67165 (31.4)182 (34.7)0.27
Male gender35 397 (100)1054 (77.7)25 621 (75.3)0.04109 (74.7)108 (740)0.89410 (78.1)409 (77.9)0.94
Caucasian race35 397 (100)813 (60.0)24 236 (71.5)<0.00193 (63.7)81 (55.5)0.15343 (65.3)344 (65.5)0.95
African-American race35 397 (100)379 (28.0)6095 (17.9)<0.00137 (25.3)37 (24.3)1.00125 (23.8)125 (23.8)1.00
Hispanic race35 397 (100)107 (7.9)2515 (7.4)0.4911 (7.5)18 (12.3)0.1733 (6.3)38 (7.2)0.54
College education28 357 (80.1)639 (57.5)13 906 (51.0)<0.00166 (55.9)64 (58.2)0.73239 (57.0)266 (63.5)0.07
Private insurance35 378 (99.9)669 (49.3)19 526 (57.4)<0.00182 (56.2)77 (52.7)0.56291 (55.4)293 (55.8)0.90
Year of listing, years35 397 (100)2008 (IQR: 2004–2010)2006 (IQR: 2003–2009)2007 (IQR: 2003–2010)2007 (IQR: 2003–2010)2007 (IQR: 2004–20102008 (IQR: 2004–2010)
Low-volume heart center (<250 listed)35 397 (100)285 (21.0)8258 (24.3)0.00627 (18.5)33 (22.6)0.39115 (21.9)95 (18.1)0.12
Low-volume heart–kidney center (<25 listed)1356 (100)608 (44.8)n/a 70 (48.0)n/a229 (43.6)n/a
BMI, kg/m235 285 (99.7)27.2 ± 5.627.3 ± 5.20.6427.3 ± 5.826.7 ± 5.10.3527.3 ± 5.126.7 ± 4.90.07
BMI >35 kg/m235 285 (99.7)99 (7.3)2372 (7.0)0.6515 (10.3)10 (6.9)0.3040 (7.6)28 (5.3)0.13
Diagnosis
Ischemic dilated cardiomyopathy35 397 (100)503 (37.1)14 491 (42.6)<0.00153 (36.3)52 (35.6)0.90217 (41.3)195 (37.1)0.16
Idiopathic dilated cardiomyopathy35 397 (100)388 (28.6)11 338 (33.3)<0.00134 (23.3)36 (24.7)0.78156 (29.7)162 (30.9)0.69
Heart retransplantation35 397 (100)221 (16.3)1191 (3.5)<0.00127 (18.5)31 (21.2)0.5648 (9.1)63 (12.0)0.13
Congenital heart disease35 397 (100)26 (1.9)1006 (3.0)0.034 (2.7)5 (3.4)0.7411 (2.1)13 (2.5)0.68
Functional status, life support
Status 1A35 397 (100)226 (16.7)7076 (20.8)<0.00169 (47.3)57 (39.0)0.1698 (18.7)100 (19.1)0.88
Status 1A or 1B35 397 (100)673 (49.6)18 217 (53.5)0.005107 (73.3)104 (71.2)0.70298 (56.8)295 (56.2)0.85
NYHA class IV28 535 (80.6)393 (33.7)8330 (30.4)0.0279 (68.7)63 (63.6)0.44157 (36.2)146 (32.2)0.21
NYHA class III–IV28 535 (80.6)953 (81.7)22 150 (80.9)0.50110 (95.7)92 (92.9)0.39357 (82.3)378 (83.3)0.69
Inotropes35 397 (100)451 (33.3)11 792 (34.6)0.2962 (42.5)65 (44.5)0.72200 (28.1)192 (36.7)0.61
Ventilator35 397 (100)60 (4.4)1501 (4.4)0.9831 (21.2)25 (17.1)0.3719 (3.6)23 (4.4)0.53
IABP35 397 (100)56 (4.1)1980 (5.8)0.00916 (11.0)18 (12.3)0.720 (0.0)0 (0.0)1.00
ECMO35 397 (100)15 (1.1)253 (0.7)0.139 (6.2)7 (4.8)0.615 (1.0)6 (1.1)0.76
Ventricular assist device35 397 (100)161 (11.9)4370 (12.8)0.3048 (32.9)37 (25.3)0.1666 (12.6)50 (9.5)0.12
Mechanical support (IABP/ventilator/VAD/ECMO)35 397 (100)247 (18.2)6813 (20.0)0.1078 (53.4)73 (50.0)0.56106 (20.2)92 (17.5)0.27
Renal function, diabetes
Creatinine clearance, mL/min34 273 (96.8)38.0 ± 24.883.1 ± 31.4<0.00140.4 ± 28.242.2 ± 26.40.5854.4 ± 25.755.3 ± 24.00.53
Creatinine clearance <50 mL/min34 273 (96.8)1005 (76.1)4624 (14.0)<0.00197 (69.3)105 (75.0)0.29266 (52.5)242 (46.8)0.10
Dialysis33 771 (95.4)504 (37.6)339 (1.2)<0.001146 (100)146 (100)1.000 (0.0)0 (0.0)1.00
Dialysis or CrCl <50 mL/min34 196 (96.6)1060 (79.8)4813 (14.6)<0.001146 (100)146 (100)1.00266 (52.5)242 (46.8)0.10
Diabetes35 056 (99.0)51 (38.9)8533 (25.3)<0.00160 (41.4)49 (34.3)0.21174 (33.66)181 (34.7)0.73
Hemodynamic parameters
Cardiac index, L/min/m231 539 (89.1)2.43 ± 0.822.15 ± 0.65<0.0012.39 ± 0.822.24 ± 0.760.152.22 ± 0.662.25 ± 0.580.58
Systolic PAP, mmHg33 012 (93.3)45.6 ± 13.643.9 ± 14.4<0.00146.6 ± 14.447.2 ± 15.80.7745.6 ± 13.944.7 ± 14.50.33
Mean PAP, mmHg31 767 (89.8)31.1 ± 9.429.8 ± 10.2<0.00132.3 ± 9.532.9 ± 11.00.6730.8 ± 9.930.2 ± 10.20.37
PCWP, mmHg30 990 (87.6)20.9 ± 8.220.2 ± 8.70.00722.9 ± 8.221.7 ± 7.80.2821.4 ± 8.520.5 ± 8.40.09
Table 2. Baseline characteristics at time of transplant in patients undergoing heart and heart–kidney transplantation, in both unadjusted and propensity-matched analyses
 Nonmissing data (n = 24 257)HKT (n = 637)HRT (n = 23 620)p*HKT, DD (n = 215)HRT, DD (n = 215)p*HKT, NDDRI (n = 284)HRT, NDDRI (n = 284)p*
  • Data are reported as number (%), mean ± standard deviation or median (interquartile range [IQR]).
  • CrCl, creatinine clearance; DD, dialysis dependence; ECMO, extracorporeal membrane oxygenation; HKT, heart–kidney transplantation; HRT, heart transplant; IABP, intra-aortic balloon pump; NDDRI, nondialysis-dependent renal insufficiency; NYHA, New York Heart Association; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure; VAD, ventricular assist device.
  • *p-Value based on chi-square or Student t-test analysis.
  • Boldface data denote statistical significance at the p < 0.05 level.
Baseline characteristics
Age at listing, year24 257 (100)52.8 ± 11.252.1 ± 12.50.1651.9 ± 11.250.1 ± 12.20.1154.4 ± 12.155.4 ± 10.20.27
Age >60 years at transplant24 257 (100)171 (26.8)6633 (28.1)0.4953 (24.7)49 (22.8)0.6598 (34.5)101 (35.6)0.79
Male gender24 257 (100)505 (79.3)1815 (75.4)0.03166 (77.2)173 (80.5)0.41221 (77.8)226 (79.2)0.68
Caucasian race24 257 (100)411 (64.5)17 051 (72.3)<0.001141 (65.6)137 (63.7)0.69204 (71.8)200 (70.4)0.71
African-American race24 257 (100)153 (24.0)4015 (17.0)<0.00148 (22.3)48 (22.3)1.0052 (18.3)58 (20.4)0.52
Hispanic race24 257 (100)46 (7.2)1761 (7.5)0.8216 (7.4)18 (8.4)0.7218 (6.3)17 (6.0)0.86
College education19 325 (79.7)290 (56.0)9725 (51.7)0.0692 (52.6)89 (52.1)0.92137 (59.3)133 (59.6)0.94
Private insurance24 216 (99.8)310 (48.8)13 238 (56.1)<0.001106 (49.5)106 (49.8)0.96152 (53.7)154 (54.4)0.87
Year of transplant24 257 (100)2008 (IQR: 2004–2010)2006 (IQR: 2003–2009)2007 (IQR: 2004–2010)2007 (IQR: 2003–2010)2007 (IQR: 2004–2010)2008 (IQR: 2004–2010)
Low-volume heart center (<200 transplanted)24 257 (100)183 (28.7)7950 (33.7)0.00964 (29.8)65 (30.2)0.9283 (29.2)82 (28.9)0.93
Low-volume heart–kidney center (<10 transplanted)637 (100)201 (31.6)n/a73 (34.0)n/an/a78 (27.5)
BMI, kg/m224 240 (99.9)26.5 ± 4.826.7 ± 4.70.2826.8 ± 5.226.6 ± 4.90.7826.3 ± 4.426.6 ± 4.70.42
BMI >35 kg/m224 240 (99.9)33 (5.2)1099 (4.7)0.5416 (7.4)11 (5.1)0.3210 (3.5)14 (4.9)0.40
Diagnosis
Ischemic dilated cardiomyopathy24 257 (100)263 (41.3)10 079 (42.7)0.4981 (37.7)83 (38.6)0.84122 (43.0)126 (44.4)0.74
Idiopathic dilated cardiomyopathy24 257 (100)167 (26.2)8095 (34.3)<0.00164 (29.8)61 (2.4)0.7564 (22.5)66 (23.2)0.84
Heart retransplantation24 257 (100)101 (15.9)657 (2.8)<0.00132 (14.9)37 (17.2)0.5147 (16.6)40 (14.1)0.42
Congenital heart disease24 257 (100)10 (1.6)669 (2.8)0.062 (0.9)5 (2.3)0.259 (3.2)7 (2.5)0.61
Functional status, life support
Status 1A24 257 (100)261 (41.0)10 363 (43.9)0.15121 (56.3)118 (54.9)0.77103 (36.3)106 (37.3)0.79
Status 1A or 1B24 257 (100)521 (81.8)19 413 (82.2)0.80196 (91.2)192 (89.3)0.52231 (81.3)232 (81.7)0.91
NYHA class IV20 758 (85.6)249 (43.8)6927 (34.3)<0.00188 (49.4)86 (49.4)0.9984 (33.3)105 (40.4)0.10
NYHA class III–IV20 758 (85.6)457 (80.3)15 609 (77.3)0.09145 (81.5)147 (84.5)0.45196 (77.8)207 (79.6)0.61
Inotropes24 257 (100)286 (44.9)10 200 (43.2)0.3990 (41.9)91 (42.3)0.92121 (42.6)137 (48.2)0.18
Ventilator24 257 (100)18 (2.8)629 (2.7)0.8013 (6.1)13 (6.1)1.002 (0.7)5 (1.8)0.25
IABP24 257 (100)35 (5.5)1263 (5.4)0.8714 (6.5)15 (7.0)0.9214 (4.9)18 (6.3)0.47
ECMO24 257 (100)2 (0.3)142 (0.6)0.352 (0.9)6 (2.8)0.150 (0.0)0 (0.0)1.00
VAD24 257 (100)110 (17.3)5222 (22.1)0.00465 (30.2)66 (30.1)0.9241 (14.4)36 (12.7)0.54
Mechanical support (IABP/ventilator/VAD/ECMO)24 257 (100)152 (23.9)6642 (28.1)0.0285 (39.5)86 (40.0)0.9255 (19.4)55 (19.4)1.00
Renal function, diabetes, bilirubin
CrCl, mL/min24 032 (99.1)42.0 ± 28.581.4 ± 31.2<0.00149.3 ± 35.051.8 ± 27.90.4147.1 ± 22.446.9 ± 22.80.91
CrCl <50 mL/min24 032 (99.1)446 (70.4)3576 (15.3)<0.001133 (62.2)120 (56.1)0.20174 (61.3)183 (64.4)0.43
Dialysis24 205 (99.8)338 (53.1)681 (2.9)<0.001215 (100)215 (100)1.000 (0.0)0 (0.0)1.00
Dialysis or CrCl <50 mL/min24 024 (99.0)535 (84.1)4014 (17.2)<0.001215 (100)215 (100)1.00174 (61.3)183 (64.4)0.43
Diabetes24 012 (99.0)234 (37.0)5577 (23.9)<0.00176 (35.4)83 (38.8)0.46104 (37.4)104 (37.1)0.95
Serum bilirubin, mg/dL23 168 (95.5)1.14 ± 2.071.25 ± 2.280.251.51 ± 3.441.51 ± 2.280.991.01 ± 0.810.98 ± 0.760.63
Serum bilirubin ≥2 mg/dL23 168 (95.5)58 (9.5)2820 (12.5)0.0332 (15.9)40 (19.5)0.3421 (7.6)71 (7.6)1.00
Hemodynamic parameters
Cardiac index, L/min/m222 983 (94.7)2.48 ± 0.772.28 ± 0.73<0.0012.48 ± 0.812.42 ± 0.820.432.42 ± 0.782.39 ± 0.690.68
Systolic PAP, mmHg23 497 (96.9)24.5 ± 13.542.0 ± 14.20.00744.9 ± 13.343.3 ± 14.80.2543.1 ± 14.443.1 ± 13.80.96
Mean PAP, mmHg23 038 (95.0)29.7 ± 9.528.5 ± 10.20.00431.0 ± 9.430.2 ± 10.60.4229.3 ± 10.029.1 ± 9.60.82
PCWP, mmHg22 717 (93.7)19.9 ± 8.719.0 ± 8.80.0121.2 ± 9.020.6 ± 8.90.4920.3 ± 8.819.6 ± 8.20.35
Donor/operative characteristics
Bicaval anastomosis24 211 (99.8)407 (64.1)14 063 (59.7)0.02132 (61.7)135 (63.4)0.72155 (54.8)172 (60.8)0.15
Organ ischemic time, h22 947 (94.6)3.17 ± 1.073.21 ± 1.050.373.29 ± 1.313.29 ± 0.990.993.17 ± 1.103.18 ± 0.950.98
Distance organ transported, min24 233 (99.9)112 ± 175167 ± 208<0.001130 ± 216154 ± 1880.2293 ± 132104 ± 1550.37
Donor age, years24 257 (100)31.7 ± 12.631.6 ± 12.20.8232.0 ± 12.532.9 ± 12.50.4630.5 ± 11.431.5 ± 12.40.30
Gender match24 257 (100)466 (73.2)17 105 (72.4)0.68157 (73.0)160 (74.4)0.74201 (70.8)211 (74.3)0.35
Race match24 253 (99.9)319 (50.1)13 332 (56.5)0.001114 (53.0)109 (50.7)0.63158 (55.6)145 (51.1)0.27
Gender/race match24 253 (99.9)234 (36.7)9708 (41.1)0.0385 (39.5)83 (38.6)0.84118 (41.6)108 (38.0)0.39

Differences in characteristics between patients with HKT and HRT were controlled for by the use of propensity-score matching [15]. Multivariable logistic regression models that included all available prelisting or pretransplant variables were employed to develop a propensity score for all patients in each of our four propensity-matched comparisons; to handle missing data, propensity scores were calculated across all imputed data sets (n = 20) [10] using the “Across” approach described by Mitra and Reiter [16] and combined according to Rubin's rules [17]. We next carried out a 1:1 nearest-neighbor matching algorithm without replacement (using a caliper of 0.25 of the standard deviation of the linear propensity score); balance was achieved in our model by using the standardized differences approach [10, 18-20]. A propensity-matched subgroup was thereby generated for our four previously described comparisons [21].

Semi-parametric estimation of cumulative incidence functions (competing outcomes analysis) was performed to assess the incidences of transplant and mortality for patients on the waitlist [22]. The Pepe and Mori test statistic was used to compare differences between cumulative incidence functions [23]. Posttransplant survival distributions were estimated with the Kaplan–Meier method [24] and compared by using the log-rank test [25]. Comparisons of waitlist and posttransplant outcomes were performed first in our two unmatched comparisons and were then repeated in our four propensity-matched sub-analyses of patients with DD and NDDRI.

Univariable and multivariable time-varying Cox proportional hazards regression analyses assessed the effect of baseline characteristics in HKT-listed patients, as well as the time-varying variables of HKT and HRT, on mortality from the time of listing for HKT (notably, 166 patients listed for HKT underwent HRT instead) [26, 27]. A multivariable risk model was constructed by using variables previously shown to predict long-term survival in posttransplant patients. Potential interactions between covariates were tested. Nonsignificant variables were still included in our final multivariable models as long as they were plausible predictors of survival, unless they overly decreased the variance of our predictive values (i.e. had an untoward bias-variance tradeoff). The final model included the following variables: age, gender, race, insurance status, listing year, listing center volume, diagnosis, BMI, listing status, New York Heart Association (NYHA) status, mechanical life support requirement, DD or a creatinine clearance less than 50 mL/min, diabetes, mean pulmonary artery pressure, pulmonary capillary wedge pressure and the time-varying variables of HKT and HRT. A second Cox proportional hazards regression analysis (not time-varying) was constructed in the same manner to assess the effect of pretransplant characteristics on posttransplant mortality. This additional model included the following variables: age, gender, race, insurance status, year of transplant, diagnosis, BMI, listing status at transplant, NYHA functional status, mechanical life support requirement, DD or creatinine clearance less than 50 mL/min, diabetes, serum bilirubin, mean pulmonary artery pressure, pulmonary capillary wedge pressure, donor age, and a combined gender/race match.

Means are presented with standard deviations and hazard ratios (HRs) are presented with 95% confidence intervals (CIs). For all analyses, p-values are two-sided, and a p < 0.05 was considered significant. Because of the exploratory nature of this study, no adjustments were made for multiple comparisons [28]. Analyses were conducted with STATA software (version 12; StataCorp LP, College Station, TX).

Results

From January 1, 2000 to December 31, 2012, 35 397 adults were listed for either HKT (n = 1356) or HRT (n = 34 041) and 24 257 adults underwent either HKT (n = 637) or HRT (n = 23 620). A histogram of patients listed for and receiving HKT by year is shown in Figure 1, revealing a doubling in the yearly incidence of HKT waitlisting and transplantation over the past decade.

Figure 1.

Number of patients who were listed for or underwent heart–kidney transplant, stratified by year of listing/transplant.

Patient characteristics at time of listing

Thirty-six variables detailing patient characteristics at the time of listing are described in Table 1, stratified by whether the patient was listed for HKT or HRT. The amount of missing data was significant (i.e. >15%) for only two variables: level of education and NYHA functional status. Statistically significant differences between patients listed for HRT and HKT were noted in 23 of the 36 variables assessed (Table 1). Our propensity-matching algorithm matched 146 patients with DD listed for HKT with 146 patients with DD listed for HRT; baseline characteristics were not significantly different between matched DD HKT- and HRT-listed patients in any of the variables assessed. We also matched 525 patients with NDDRI listed for HKT with 525 patients with NDDRI listed for HRT; baseline characteristics were not significantly different between NDDRI HKT- and HRT-listed patients in any of the 36 variables assessed.

Waitlist incidence of transplant and mortality

Figure 2 is a competing outcomes depiction of the cumulative incidence of mortality and transplant for patients on the waitlist for HKT and HRT. Below, we provide the p-value for each set of curves compared and the respective 3-month and 1-year cumulative incidences for each group. The HKT-listed patients had a higher cumulative incidence of death (p < 0.001; 10% vs. 7% and 22% vs. 12%) and a lower cumulative incidence of transplant (p < 0.001; 26% vs. 36% and 47% vs. 58%) than did their HRT-listed counterparts (Figure 2A). Propensity-matched patients with DD listed for HKT (n = 146) and HRT (n = 146) had similar cumulative incidences of death on the waitlist (p = 0.14; 21% vs. 31% and 31% vs. 39%) and of transplant (p = 0.32; 38% vs. 36% and 57% vs. 47%) (Figure 2B). Propensity-matched patients with NDDRI listed for HKT (n = 525) and HRT (n = 525) had a similar cumulative incidence of death on the waitlist (p = 0.27; 7% vs. 12% and 19% vs. 17%), but HKT-listed patients had a lower cumulative incidence of transplant than HRT-listed patients (p < 0.001; 26% vs. 35% and 47% vs. 60%) (Figure 2C).

Figure 2.

Competing outcomes depiction of the cumulative incidence of transplant and mortality for patients listed for transplant, stratified by organ listed (heart transplant vs. heart–kidney transplant). Graphs are shown for the full study cohort (A), the propensity-matched dialysis-dependent patients (B) and the propensity-matched nondialysis-dependent renal insufficient patients (C). The sum of the percentages of the three heart transplant curves at any time point equals 100%, as does the sum of the three heart–kidney transplant curves. The cumulative incidence of death in patients listed for heart and heart–kidney transplant (shown in red), as well as the cumulative incidence of transplant in heart- and heart–kidney-listed patients (shown in blue), are compared by using the Pepe and Mori [23] cumulative incidence test. Curves depicting the number of patients still at risk on the waitlist are shown in black.

Patient characteristics at time of transplant

Forty-five variables detailing pretransplant and operative characteristics are described in Table 2, stratified by whether the patient underwent HKT or HRT. The amount of missing data was significant (i.e. >15%) for only one variable, level of education. Statistically significant differences between patients receiving HKT and HRT were noted in 24 of the 42 variables assessed. Our propensity-matching algorithm matched 215 patients with DD who underwent HKT with 215 patients with DD who underwent HRT; all pretransplant characteristics were similar between matched DD HKT- and HRT-listed patients. We also matched 284 patients with NDDRI who underwent HKT with 284 patients with NDDRI who underwent HRT; pretransplant characteristics were not significantly different between propensity-matched NDDRI HKT- and HRT-listed patients in any of the variables assessed.

Posttransplant survival

A Kaplan–Meier graph of posttransplant survival in HKT and HRT recipients is shown in Figure 3. The p-value is given for each set of curves compared, and the 1- and 5-year survival rates are provided for each group. Our unadjusted comparison of posttransplant survival between the 637 HKT recipients and the 23 260 HRT recipients showed no difference in survival (p = 0.84; 87% vs. 88% and 75% vs. 75%) (Figure 3A). The patients with DD who underwent HKT had better posttransplant survival than propensity-matched HRT recipients (p < 0.001; 84% vs. 69% and 73% vs. 51%) (Figure 3B). Propensity-matched patients with NDDRI who underwent HKT also had better posttransplant survival than HRT recipients (p = 0.004; 91% vs. 81% and 80% vs. 69%) (Figure 3C), although the survival difference was smaller than it was in patients with DD.

Figure 3.

Kaplan–Meier estimates of posttransplant survival in patients who underwent isolated heart or heart–kidney transplant. (A) Unadjusted Kaplan–Meier estimate of posttransplant survival after either isolated heart or heart–kidney transplant. (B) Kaplan–Meier estimate of posttransplant survival after either isolated heart or heart–kidney transplant in propensity-matched subgroups of patients with dialysis dependence. (C) Kaplan–Meier estimate of posttransplant survival after either isolated heart or heart–kidney transplant in propensity-matched subgroups of nondialysis-dependent renal insufficient patients.

Time-varying Cox proportional hazards analysis of survival from time of listing

As described in the Methods section, 17 variables were included in our multivariable time-varying model of predictors of survival from waitlisting (Table 3). Both HKT and HRT were significant predictors of survival from the time of listing. However, HKT had a greater protective effect (HR: 0.25, 95% CI: 0.20–0.32, p < 0.001) than HRT (HR: 0.67, CI: 0.50–0.90, p = 0.008). Other significant predictors of survival included having private insurance, whereas significant predictors of mortality included being listed at a low-volume transplant center, being listed for retransplantation, 1A listing status, NYHA class IV functional status, DD or a creatinine clearance less than 50 mL/min, and an elevated pulmonary capillary wedge pressure.

Table 3. Predictors of mortality from time of listing for heart–kidney transplant (HKT) and from time of the HKT procedure according to two separate Cox proportional hazards regression analyses
 Predictors of Survival from Time of Listing for HKTPredictors of Survival after HKT
Multivariable hazard ratio (95% CI)p*Univariate hazard ratio (95% CI)pMultivariable hazard ratio (95% CI)p**Univariate hazard ratio (95% CI)p
  • CI, confidence interval; CrCl, creatinine clearance; ECMO, extracorporeal membrane oxygenation; IABP, intra-aortic balloon pump; NYHA, New York Heart Association; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure.
  • *p-Value obtained from multivariable time-varying Cox proportional hazards regression, as described in the Methods section.
  • **p-Value obtained from multivariable Cox proportional hazards regression (not time-varying), as described in the Methods section.
  • Boldface data denote statistical significance at the p < 0.05 level.
Baseline characteristics
Age, years1.002 (0.994–1.009)0.640.998 (0.91–1.005)0.540.998 (0.982–1.014)0.801.001 (0.987–1.015)0.89
Age >60 years1.05 (0.87–1.26)0.610.93 (0.65–1.35)0.71
Male gender0.92 (0.76–1.12)0.400.90 (0.75–1.08)0.250.89 (0.59–1.35)0.590.88 (0.61–1.28)0.51
Caucasian race0.98 (0.83–1.15)0.781.00 (0.71–1.40)0.99
African-American race0.90 (0.74–1.08)0.261.02 (0.85–1.21)0.850.90 (0.59–1.37)0.631.10 (0.76–1.60)0.62
Hispanic race0.95 (0.70–1.30)0.761.03 (0.54–1.96)0.92
College education1.01 (0.85–1.20)0.890.91 (0.64–1.29)0.59
Private insurance0.81 (0.69–0.96)0.010.87 (0.74–1.02)0.080.76 (0.54–1.07)0.110.79 (0.58–1.09)0.15
Listing/transplant year1.00 (0.97–1.03)0.971.01 (0.98–1.03)0.560.98 (0.92–1.04)0.471.01 (0.72–1.42)0.96
Low-volume heart center (<250 listed)1.21 (1.00–1.46)0.051.19 (0.99–1.43)0.070.90 (0.63–1.31)0.591.01 (0.95–1.06)0.81
Low-volume heart–kidney center (<25 listed)1.10 (0.94–1.28)0.250.87 (0.61–1.25)0.47
Ischemic dilated cardiomyopathy0.96 (0.82–1.13)0.611.05 (0.76–1.44)0.78
Idiopathic dilated cardiomyopathy0.93 (0.78–1.11)0.430.82 (0.55–1.22)0.320.88 (0.61–1.26)0.48
Heart retransplantation1.33 (1.06–1.67)0.011.23 (1.00–1.51)0.050.98 (0.64–1.51)0.94
Congenital heart disease1.14 (0.66–1.98)0.640.84 (0.21–3.39)0.81
BMI, kg/m21.000 (0.987–1.013)0.991.029 (0.997–1.063)0.08
BMI >35 kg/m21.08 (0.80–1.46)0.601.18 (0.89–1.57)0.261.77 (1.01–3.24)0.041.87 (1.08–3.25)0.03
Status 1A1.32 (1.02–1.72)0.041.44 (1.17–1.76)<0.0011.06 (0.70–1.58)0.791.32 (0.96–1.82)0.09
Status 1A or 1B1.26 (1.08–1.47)0.0040.96 (0.65–1.39)0.81
NYHA class IV1.59 (1.30–1.94)<0.0011.55 (1.30–1.84)<0.0011.02 (0.67–1.57)0.911.18 (0.83–1.69)0.36
NYHA class III–IV1.43 (1.11–1.84)0.0061.44 (0.88–2.36)0.14
Life support
Inotropes1.15 (0.98–1.36)0.091.06 (0.78–1.46)0.70
Ventilator2.71 (1.99–3.68)<0.0013.90 (2.15–7.08)<0.001
IABP1.32 (0.91–1.90)0.141.22 (0.64–2.32)0.54
ECMO4.19 (2.30–7.62)<0.001Perfectly predicts outcome
Ventricular assist device1.23 (0.96–1.59)0.101.34 (0.86–2.10)0.19
Mechanical life support1.23 (0.96–1.58)0.101.45 (1.19–1.76)<0.0011.66 (1.09–2.53)0.021.68 (1.17–2.40)0.005
Renal function, diabetes, bilirubin
CrCl, mL/min1.002 (0.998–1.005)0.330.994 (0.988–1.000)0.07
CrCl <50 mL/min0.99 (0.82–1.19)0.891.27 (0.89–1.82)0.19
Dialysis1.33 (1.13–1.56)<0.0011.24 (0.90–1.72)0.19
Dialysis or CrCl <50 mL/min1.28 (1.04–1.57)0.011.09 (0.90–1.33)0.381.02 (0.64–1.63)0.931.12 (0.71–1.76)0.63
Diabetes1.13 (0.95–1.34)0.161.16 (0.98–1.36)0.081.10 (0.77–1.58)0.601.20 (0.87–1.67)0.27
Bilirubin, mg/dLn/a1.03 (0.98–1.08)
Bilirubin ≥2 mg/dLn/a1.69 (1.06–2.71)0.031.59 (1.00–2.54)0.05
Hemodynamic parameters
Cardiac index, L/min/m20.95 (0.85–1.05)0.330.89 (0.72–1.10)0.29
Systolic PAP, mmHg1.006 (1.000–1.012)0.041.009 (0.997–1.021)0.11
Mean PAP, mmHg0.994 (0.979–1.009)0.431.010 (1.001–1.019)0.021.027 (0.996–1.058)0.081.012 (0.996–1.029)0.15
PCWP, mmHg1.020 (1.002–1.037)0.031.016 (1.005–1.025)0.0030.977 (0.946–1.009)0.151.005 (0.987–1.023)0.60
Donor/operative characteristics
Bicaval anastomosis0.94 (0.68–1.30)0.72
Organ ischemic time, h1.09 (0.95–1.25)0.22
Distance organ transported, min1.000 (0.999–1.001)0.50
Donor age, years1.010 (0.997–1.025)0.141.012 (0.999–1.025)0.06
Gender match0.70 (0.48–0.99)0.050.65 (0.47–0.91)0.01
Race match0.96 (0.70–1.31)0.79
Combined gender/race match0.64 (0.45–0.90)0.01
Intervention
Heart–kidney transplant0.25 (0.20–0.32)<0.0010.34 (0.28–0.42)<0.001
Heart-only transplant0.67 (0.50–0.90)0.0081.37 (1.06–1.78)0.02

Cox proportional hazards analysis of survival from time of transplant

Eighteen variables were included in our multivariable model of predictors of survival from transplant as described in the Methods section (Table 3). Patient characteristics that predicted post-HKT mortality included a BMI >35 kg/m2, a preoperative mechanical life support requirement, and a serum bilirubin level above 2.0 mg/dL. Pretransplant DD or a creatinine clearance less than 50 mL/min did not predict outcomes in HKT recipients. Finally, we noted a significant protective effect of a gender match between donor and recipient (HR: 0.70, 95% CI: 0.48–0.99, p = 0.05).

Discussion

The current study examined waitlist and posttransplant outcomes in patients with ESHF listed for HKT and HRT with various levels of renal insufficiency (DD or NDDRI). We found that the cumulative incidence of waitlist mortality in patients with renal insufficiency listed for HRT or HKT was remarkably high—at least as high as that of status 1A HRT-listed patients in patients with DD, and at least as high as that of status 1B HRT candidates in patients with NDDRI [2]. Posttransplant outcomes were identical in unmatched HKT and HRT recipients, implying that HKT addresses renal insufficiency and its associated effect on mortality without conferring significant procedural risk. We also found that DD HKT recipients had a striking improvement in posttransplant outcomes compared to matched DD HRT recipients, whereas NDDRI HKT recipients had a less dramatic but still significant survival improvement over matched NDDRI HRT recipients. We identified multiple predictors of mortality from the time of listing for HKT, listing center volume, being listed for retransplantation, 1A listing status, NYHA class IV functional status, DD or a creatinine clearance less than 50 mL/min, and having diabetes. Furthermore, we identified multiple predictors of survival from HKT-listing, including having private insurance status and undergoing HRT or HKT. Importantly, on multivariable analysis, HKT was associated with a greater survival benefit than HRT in HKT candidates. Finally, we assessed predictors of mortality from time of transplant in HKT recipients, finding that a BMI ≥35 kg/m2, a mechanical life support requirement, and a serum bilirubin level ≥2 mg/dL predicted worse outcomes, whereas a donor/recipient gender match predicted survival.

Waitlist outcomes in patients listed for HKT and HRT

In previous analyses of waitlist outcomes in HRT-listed patients, the 3-month incidences of mortality and transplant have been 13–16% and 58–59% in status 1A patients, 7–9% and 48–51% in status 1B patients and 4% and 23–30% in status 2 patients [1, 2]. The 3-month waitlist outcomes for matched patients with DD listed for HKT and HRT were significantly worse than the reported outcomes for status 1A patients: the incidence of death was 21% and 31% and the incidence of transplant was 38% and 36%, respectively. Despite high waitlist mortality in these patients, more than half of patients with DD were listed as status 1B or 2. Three-month waitlist outcomes in patients with NDDRI were also grim: the incidence of death for matched HKT- and HRT-listed patients was 7% and 12%, and the incidence of transplant was 26% and 35%, respectively. Waitlist outcomes in patients with NDDRI were comparable to those of status 1B listed patients, yet nearly half of the patients with NDDRI were listed as status 2. Clearly, the hemodynamics-focused organ allocation algorithm used for the HRT waitlist does not account for the waitlist mortality associated with renal insufficiency [29]. Because the guiding principles of organ allocation include necessity, equitability and utility, we suggest that these data argue in favor of status 1A priority exception for ESHF patients with DD and status 1B priority exception for NDDRI ESHF patients. Furthermore, we believe these data support combined HKT for patients with DD. However, the benefit of combined HKT over HRT is less dramatic in NDDRI patients with ESHF than in DD patients. In the interest of maximizing organ distribution, unless patient characteristics indicate a low likelihood of renal recovery after HRT, we suggest that the majority of NDDRI patients with ESHF should be listed for HRT and not HKT. A subsequent kidney transplant in patients who did not have renal recovery after HRT could then considered. Again, we suggest that although most patients with NDDRI should be listed for HRT, they should be listed as status 1B because of their high waitlist mortality. The utility of a staged renal transplant in ESHF patients with NDDRI who do not have renal recovery after HRT requires assessment, and outcomes in these patients must be compared with outcomes in combined HKT recipients to determine which strategy more appropriately allocates organs and promotes patient survival.

Posttransplant survival in HRT and HKT recipients

Our findings confirm the initial report by Narula et al [5] that patients who undergo HKT have survival comparable to that of HRT recipients; the postoperative risk of renal insufficiency noted in HRT recipients appears to be entirely mitigated by HKT without additional operative risk [30]. Moreover, our propensity-matched comparison of DD HKT and HRT recipients shows a clear survival advantage for DD HKT recipients. These data confirm previous reports that in patients with DD, HRT has inferior outcomes, whereas HKT is associated with survival rates similar to those of non-DD HRT recipients [3, 4].

HRT versus HKT in NDDRI patients

Outside of the present study, outcomes for NDDRI HRT and HKT recipients remain poorly described: we could only find one reported study of NDDRI transplant recipients, which compared NDDRI HKT recipients with HRT recipients (who ostensibly had normal renal function) and found no difference in posttransplant outcomes [7]. Importantly, no adjustment for baseline differences (including preoperative creatinine clearance) between groups was made, making the analysis susceptible to treatment-selection bias [8]. Our analysis comparing matched NDDRI HKT and HRT recipients tries to fill this gap. Unfortunately, our findings are not clear cut: although some survival advantage was noted in our matched comparison between HKT and HRT recipients, the benefit was less pronounced than for patients with DD, and many NDDRI HRT recipients had excellent posttransplant survival.

We suggest that the majority of patients with NDDRI with ESHF listed for transplant have a cardiorenal syndrome (i.e. their ESHF underlies their NDDRI), and that alleviating their ESHF through HRT alone should improve their cardiac-related NDDRI [31]. Unfortunately, we were unable to identify variables in the UNOS database that could predict which patients with NDDRI would fare well after HRT and which would require HKT. To make this decision, our institution uses renal ultrasonography; biopsy findings; patterns in the serum creatinine, creatinine clearance and electrolyte abnormalities; and multidisciplinary meetings among transplant nephrologists, cardiologists and cardiac surgeons to ascertain which patients with NDDRI are likely to recover renal function after HRT alone.

Thus, although our results confirm the intuition that HKT is associated with modestly better outcomes than HRT in patients with NDDRI with ESHF, this finding does not compel us to recommend HKT for most patients with NDDRI. Instead, we think it reasonable to limit HKT to patients whose renal function is unlikely to recover and, thus, to avoid transplanting kidneys to patients with NDDRI ESHF who would otherwise recover renal function after HRT. Contingency plans to perform an urgent kidney transplant (perhaps by identifying potential living-related donors) should be made in anticipation of post-HRT renal failure because the prognosis of HRT recipients with renal failure is poor [4]. Moreover, expeditious and high-priority listing for kidney transplant in NDDRI HRT recipients who do not recover renal function is prudent, given the dismal outcomes of such patients who do not undergo kidney transplant. This option of urgently listing organ recipients when renal recovery fails was recently suggested at a summit where the guidelines for liver–kidney transplant were reassessed [4, 32]. As well, the summit on simultaneous liver–kidney transplant noted that we still lack knowledge as to which tools best diagnose unrecoverable renal disease, whether the particular cause of a patient's acute or chronic kidney disease affects the odds of recovery without kidney transplant, and the utility of delayed kidney transplant in patients with multi-organ failure requiring concomitant transplant [32]. Answers to these questions are needed in order to better define which patients should be listed for kidney-after-HRT and which require HKT. It bears repeating, however, that we present here strong evidence that regardless of which procedure (HRT or HKT) patients with NDDRI with ESHF are listed for, their dismal outcomes on the waitlist indicate that a status exception (1A for DD and 1B for patients with NDDRI) would be prudent.

As an alternative therapy, it has been suggested that supporting patients with ESHF with a cardiorenal syndrome by using a ventricular assist device (VAD) may serve to sustain renal function [33, 34]. However, these studies are preliminary, and the utility of VAD support in these patients has yet to be rigorously assessed in a multi-institutional cohort [35]. Notably, we performed a sensitivity analysis of patients with DD and NDDRI supported with a VAD, and we found no difference in outcomes between supported patients and the rest of our cohort—a finding supported by our univariate analysis in which VAD support did not predict improved waitlist outcomes. More data regarding this potential therapy are needed before it can be recommended as an alternative to transplant in patients with DD and NDDRI with ESHF.

Further complicating matters, HKT removes two organs from the transplant donor pool; thus, to determine whether the procedure meets the organ allocation goals of justice and utility, the benefits of HKT must be weighed against the opportunity cost of transplanting a heart and kidney to two separate patients. Analyses of waitlist outcomes in HRT recipients who are subsequently listed for a kidney transplant have shown that their hazard of waitlist death is nearly twice (HR: 1.92) that of isolated-kidney-transplant candidates (3-year incidence of mortality, >40% vs. 14–18%, respectively). Our findings in DD and NDDRI HKT candidates are similar: a 3-year incidence of mortality of >40% and >30%, respectively [4, 36]. As well, Cassuto et al [4] reported that among patients who received a kidney transplant (compared to those who remained waitlisted), patients who had undergone a previous HRT had a greater reduction in their risk of mortality than patients who had not undergone a prior HRT (HR: 0.38, CI 0.29–0.50 vs. HR: 0.52, CI: 0.51–0.53). These results underscore the significant hazard of death in patients with combined heart and renal failure and in patients who have undergone HRT but who continue to have renal failure, suggesting that in appropriate patients, HKT represents a reasonable allocation of organs in terms of both justice and organ utility.

Predictors of outcomes at time of listing and time of transplant

In our multivariable time-varying Cox proportional hazards model of outcomes from the time of listing, although HKT was associated with better outcomes than HRT, both HKT and HRT were associated with a significant reduction in patient mortality (Table 3). Our model also found private insurance to be a predictor of survival. In contrast, being listed at a low-volume cardiac center, being listed for heart retransplantation, 1A listing status, NHYA class IV functional status, DD or a creatinine clearance less than 50 mL/min, and an elevated pulmonary capillary wedge pressure all predicted worse waitlist outcomes, a result that is consistent with those of previous studies of patients waitlisted for HRT [1, 2].

Interestingly, only four variables were found to predict post-HKT outcomes in our multivariable Cox proportional hazards model for posttransplant outcomes. A BMI ≥35 kg/m2, a preoperative mechanical life support requirement and a serum bilirubin level above 2 mg/dL all predicted worse outcomes in HKT recipients, as previous studies have also shown [6, 7]. Meanwhile, we found that a donor/recipient gender match conferred a survival advantage; notably, a gender match has been previously shown to predict improved outcomes in HRT [37]. Importantly, DD or a creatinine clearance less than 50 mL/min did not predict post-HKT outcomes, suggesting that kidney transplantation mitigated the known negative effect of poor renal function on post-HRT outcomes.

Limitations

Our study is limited by its retrospective nature, its susceptibility to selection bias, and the unavailability of complete data for all patients in the UNOS database. We attempted to address these limitations by using multiple imputations to account for missing variables, propensity matching to account for treatment-selection bias and multivariable analysis to account for confounding variables. Nevertheless, unmeasured variables and the observational nature of this study leave ample room for residual bias to confound our results. Furthermore, evaluating the benefits of transplant in waitlisted patients involves complex statistical methods, none of which allow for definitive conclusions; our findings of improved outcomes with patients with HKT and HRT should thus be considered with appropriate circumspection [38]. We were also unable to assess staged procedures, in which a patient receives a heart or kidney in one operation and subsequently receives the other organ in a separate procedure; staged procedures remain an option for patients with concomitant ESHF and renal insufficiency. Finally, a broader regional sharing algorithm in the heart allocation policy was implemented on July 12, 2006, and our cohort includes patients listed before and after its implementation. Others have found that HRT waitlist outcomes improved after this policy was implemented [2], although year of listing and year of transplant did not appear to significantly affect waitlist outcomes in our multivariable analyses of HKT candidates and recipients.

Conclusions

The waitlist survival of DD HKT and HRT candidates is at least as bad as status 1A HRT candidates, and early listing and transplantation are crucial to improving survival in these patients. Moreover, HKT appears to provide a significant survival advantage over HRT in patients with DD with ESHF, and strong consideration should be given to listing these patients for HKT. Waitlist survival of patients with NDDRI is at least as poor as that of status 1B HRT-listed patients, and early listing and transplantation is crucial to improving survival in these patients. However, the utility of HKT is lower in patients with NDDRI ESHF than in patients with DD ESHF, probably because many of these patients' renal function improves after HRT. Determining which patients with NDDRI should be listed for HKT remains a complex undertaking involving renal ultrasonography, a multidisciplinary discussion of the risks and benefits and, possibly, biopsy. We suggest that a staged procedure (HRT with possible staged kidney transplant) be strongly considered, particularly when HKT-listing would prolong time on the waitlist, or when a living-related kidney donor is identified. An analysis is surely needed that compares patients with NDDRI ESHF who undergo HKT with patients who undergo HRT and subsequent kidney transplant if their post-HRT renal function is poor. Finally, the benefits of a donor/recipient gender match in HKT recipients were pronounced, and matching recipients on gender should be considered in stable patients.

Acknowledgments

Stephen N. Palmer, PhD, ELS, contributed to the editing of the manuscript. No external sources of funding were used.

Disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

Ancillary