Does Expanded Criteria Donor Status Modify the Outcomes of Kidney Transplantation From Donors After Cardiac Death?

Authors

  • S. K. Singh,

    1. Division of Nephrology, Department of Medicine, University of Toronto, Toronto, Ontario, Canada
    2. Institute of Health Policy, Management and Evaluation, University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author
  • S. J. Kim

    Corresponding author
    1. Division of Nephrology and the Kidney Transplant Program, Toronto General Hospital, University Health Network, Toronto, Ontario, Canada
    2. Division of Nephrology and the Renal Transplant Program, St. Michael's Hospital, Toronto, Ontario, Canada
    3. Institute of Health Policy, Management and Evaluation, University of Toronto, Toronto, Ontario, Canada
    • Division of Nephrology, Department of Medicine, University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author

Corresponding author: S. Joseph Kim, joseph.kim@uhn.ca

Abstract

The outcomes of kidney transplants that simultaneously exhibit donation after cardiac death (DCD) and expanded criteria donor (ECD) characteristics have not been well studied. We examined the outcomes of DCD versus non-DCD kidney transplants as a function of ECD status and the kidney donor risk index (KDRI). A cohort study of 67 816 deceased donor kidney transplant recipients (KTR), including 562 ECD/DCD KTR, from January 1, 2000 to December 31, 2009 was conducted using the Scientific Registry of Transplant Recipients. In a multivariable Cox proportional hazards model, the modestly increased risk of total graft failure in DCD versus non-DCD KTR was not significantly modified by ECD status (hazard ratio1.07 [95% CI: 1.01, 1.15] for non-ECD vs. 1.21 [95% CI: 1.04, 1.40] for ECD, p for interaction = 0.14).Moreover, the hazard ratios did not significantly vary by KDRI quintiles (p = 0.40). Similar trends were seen for death-censored graft failure and death with graft function. In conclusion, ECD status or higher KDRI score did not appreciably increase the relative hazard of adverse graft and patient outcomes in DCD KTR. These findings suggest that the judicious use of ECD/DCD donor kidneys may be an appropriate strategy to expand the donor pool.

Abbreviations
DCD

donation after cardiac death

DGF

delayed graft function

ECD

expanded criteria donor

ESRD

end-stage renal disease

HRSA

Health Resources and Services Administration

KDRI

kidney donor risk index

KTR

kidney transplant recipients

OPTN

Organ Procurement and Transplantation Network

PNF

primary nonfunction

SRTR

Scientific Registry of Transplant Recipients.

Introduction

Kidney transplantation is the renal replacement therapy of choice for patients with end-stage renal disease (ESRD) [1, 2]. However, there continues to be a significant gap between the demand and supply of organs for transplantation. The growth in the number of patients on the waiting list far exceeds the rate at which kidney transplantation is performed [3, 4]. In order to narrow this gap, several strategies have been used to expand the pool of deceased donors. In particular, kidney transplantation from expanded criteria donors (ECD) as well as donors after cardiac death (DCD) have been increasing in frequency [3, 5, 6].

Expanded criteria donors (ECD) are defined by widely accepted criteria, which includes donors older than 60 years of age or 50–59 years of age with two of the following characteristics: history of hypertension, cerebrovascular accident as the cause of death or terminal serum creatinine >1.5 mg/dL [7]. Although the outcomes of kidney transplants from these less than optimal donors are inferior to non-ECD kidney transplants, the former is superior to waiting on dialysis for certain subsets of the ESRD population [8, 9]. However, the ECD classification scheme poorly captures the spectrum of risk associated with these donor kidneys [10]. In order to overcome this limitation, Rao et al. derived the Kidney Donor Risk Index (KDRI), a continuous risk score that uses donor and transplant variables to provide more refined estimates of donor risk [11].

Donation after cardiac death (DCD), defined as cessation of cardiac function prior to organ recovery, has increased in frequency over the past decade. Although DCD kidneys generally have higher rates of delayed graft function (DGF) and primary nonfunction (PNF), it appears that recipients of DCD kidneys have similar long-term graft and patient outcomes as recipients of non-DCD kidneys [6, 12, 13].

Despite the increasing use of ECD and DCD kidneys for transplantation, the outcomes of kidneys from deceased donors with both ECD and DCD characteristics (i.e. combined ECD/DCD) has received little attention in the literature. There may be a general reluctance to use these kidneys since the combination of insults may cause unacceptably poor outcomes. However, data to support this assertion are generally lacking. Thus, the purpose of this study is to evaluate the outcomes of combined ECD/DCD kidney transplantation versus other ECD and DCD categories. In particular, we evaluate the outcomes of DCD versus non-DCD kidney transplantation as a function of ECD status. To better appreciate the role of donor organ quality, we also explored the impact of DCD versus non-DCD kidney transplantation on graft outcomes as a function of the KDRI.

Materials and Methods

Study design and participants

This study used data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donor, wait-listed candidates and transplant recipients in the United States, submitted by the members of the Organ Procurement and Transplantation Network (OPTN), and has been described elsewhere. The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services provides oversight to the activities of the OPTN and SRTR contractors.

Using the SRTR Standard Analysis File, we conducted a retrospective cohort study of patients receiving a deceased donor kidney transplant from January 1, 2000 to December 31, 2009. Exclusion criteria included (1) age less than 18 years at the time of transplantation, (2) multiorgan transplant recipients (including simultaneous kidney–pancreas recipients), (3) living donor kidney transplant recipients and (4) patients with missing data to calculate the KDRI. Patients with PNF were excluded from the DGF and time-to-event analyses. Patient follow-up time started on the day of transplant and continued until the conclusion of the observation period, death, graft failure or loss to follow-up.

Exposure and outcome classification and assessment

Based on ECD and DCD status, patients were classified into four subgroups: (1) non-ECD and non-DCD (non-ECD/non-DCD), (2) non-ECD and DCD (non-ECD/DCD), (3) ECD and non-DCD (ECD/non-DCD) and (4) ECD and DCD (ECD/DCD). The KDRI was also calculated for each recipient's deceased donor kidney and categorized into quintiles. The KDRI requires the input of 15 distinct variables that are determined at the time a deceased donor kidney becomes available [11].

The primary outcome of interest was total graft failure, which was defined as the need for chronic dialysis, preemptive retransplantation or death with graft function. Secondary outcomes included death-censored graft failure, death with graft function, DGF and PNF. Death-censored graft failure was defined as the need for chronic dialysis or preemptive retransplantation. Death with graft function was defined as death from any cause with a functioning transplant. DGF was defined as the need for at least one dialysis session within the first week after kidney transplantation. PNF was defined as a graft never achieving sufficient function to allow discontinuation of dialysis. The latter was operationalized as the date of graft failure equal to the date of transplant.

Potential confounders

A range of recipient, donor and transplant characteristics were included as potential confounders in multivariable statistical models. Recipient factors included age, sex, race, cause of ESRD, peak PRA, time on dialysis and body mass index. Donor factors included age, sex, race, terminal serum creatinine, history of hypertension, cause of death and body mass index. Transplant factors included use of induction therapy, cold ischemia time, HLA mismatches and year of transplant.

Sensitivity analyses

Additional analyses were performed to evaluate the robustness of the primary results. These analyses included (1) reparameterizing KDRI into deciles or a continuous variable, (2) excluding the DCD component of the KDRI, (3) adjustment for clustering of outcomes by transplant center, (4) restricting the cohort to a lower risk subpopulation (i.e. recipients less than 60 years, nondiabetic, waiting time less than 3 years, peak PRA zero), (5) restriction to patients receiving kidneys that underwent pulsatile perfusion after recovery, (6) including PNF in time-to-event analyses, (7) adjusting for warm ischemia time in the subcohort of patients without missing values and (8) only examining transplant centers performing at least one ECD/DCD kidney transplant over the study period.

Statistical analysis

Baseline characteristics by ECD/DCD status were compared using the one-way analysis of variance for continuous variables and the chi-square test for categorical variables. The distributions of KDRI across ECD/DCD categories were visually explored using box plots. Logistic regression models were fitted to examine the relation between ECD/DCD status (or KDRI quintile) and the risk of DGF or PNF while adjusting for potential confounders.

The Kaplan–Meier product limit method was used to assess time to total graft failure, death-censored graft failure or death with graft function across ECD/DCD categories. The log-rank test was used to evaluate differences between survival curves. Cox proportional hazards models were fitted to examine the independent association between ECD/DCD status (or KDRI quintile) and each of the time-to-event outcomes. The proportionality assumption was examined using Schoenfeld residuals and log(cumulative hazard) curves. No important departures from proportionality were observed. Of note, an interaction term for DCD and ECD (or DCD and KDRI quintile) was included in all multivariable models. Likelihood ratio testing was performed to assess the statistical significance of the interaction term(s).

A two-sided p value of <0.05 was considered statistically significant. All analyses were performed using Stata/MP4, version 12.0 (StataCorp, College Station, TX, USA). The study was approved by the Research Ethics Board at Toronto General Hospital, University Health Network.

Results

A total of 157 381 kidney transplant recipients were eligible for inclusion over the study period. The following patients were excluded: 62 296 living donor recipients, 4213 recipients under 18 years of age at the time of transplantation, 11 768 prior kidney transplant recipients, 1664 multiorgan transplant recipients, 8 466 patients with missing data to calculate the KDRI and 1158 patients with other missing covariate data. A total of 67816 patients were included in the final cohort. There were 562 (0.8%) patients in the combined ECD/DCD subgroup (i.e. donors fulfilling criteria for both ECD and DCD).

Recipient, donor and transplant characteristics across ECD/DCD subgroups are shown in Table 1. Baseline recipient characteristics were similar across all subgroups, with the exception of age and cause of ESRD. Recipients were older and more likely to have diabetes mellitus as the cause of ESRD in the ECD subgroups when compared to the non-ECD subgroups. However, these characteristics were similar between the ECD/non-DCD and ECD/DCD subgroups. As expected, donors were older, more likely to be hypertensive and to have died of a cerebrovascular accident in the ECD subgroups. Mean terminal serum creatinine as well as baseline transplant characteristics were similar across ECD and DCD subgroups.

Table 1. Distribution of baseline characteristics by ECD and DCD status
Study characteristicsNon-ECD/Non-DCD (N = 50,242)Non-ECD/DCD (N = 4,840)ECD/Non-DCD (N = 12,172)ECD/DCD (N = 562)p-Value
  1. SD= standard deviation; IQR= interquartile range; ESRD= end-stage renal disease; HLA= human leukocyte antigen.

Recipient factors
Age (years), mean (SD)50.5 (12.9)52.1 (12.5)59.2 (10.8)60.6 (10.2)<0.001
Sex (%)     
 Male30 203 (60.1)2 994 (61.9)7 652 (62.9)361 (64.2)<0.001
 Female20 039 (39.9)1 846 (38.1)4 520 (37.1)201 (35.8) 
Race (%)     
 White23 333 (46.4)2 254 (46.5)5 732 (47.1)274 (48.7) 
 Black15 789 (31.4)1 720 (35.5)3 845 (31.6)186 (33.1)<0.001
 Hispanic7 392 (14.7)497 (10.3)1 527 (12.5)51 (9.1) 
 Other3 728 (7.4)369 (7.6)1 068 (8.8)51 (9.1) 
Cause of ESRD (%)     
 Glomerulonephritis11 339 (22.6)1 023 (21.1)1 810 (14.9)67 (11.9) 
 Diabetes mellitus12 554 (25.0)1 234 (25.5)4 037 (33.2)193 (34.3) 
 Polycystic kidney disease4 618 (9.2)457 (9.4)1 043 (8.6)49 (8.7)<0.001
 Hypertension12 572 (25.0)1 309 (27.1)3 289 (27.0)154 (27.4) 
 Other9 159 (18.2)817 (16.9)1 993 (16.4)99 (17.6) 
Peak panel reactive antibody (%), median (IQR)2 (16)2 (15)0 (11)1 (10)<0.001
Time on dialysis (years), median (IQR)2.9 (3.5)3.1 (3.4)2.9 (3.1)2.7 (3.2)0.002
Body mass index (kg/m2), median (IQR)27.0 (7.5)27.6 (7.9)27.3 (7.1)28.2 (7.3)<0.001
Donor factors
Age (years), mean (SD)33.6 (14.7)36.0 (14.4)60.4 (6.4)59.2 (5.2)<0.001
Sex (%)     
 Male30 866 (61.4)3 226 (66.7)5 899 (48.5)320 (56.9)<0.001
 Female19 376 (38.6)1 614 (33.3)6 273 (51.5)242 (43.1) 
Race (%)     
 White34 436 (68.5)4 175 (86.3)8 860 (72.8)503 (89.5) 
 Black6 705 (13.3)306 (6.3)1 512 (12.4)19 (3.4)<0.001
 Hispanic7 496 (14.9)268 (5.5)1 359 (11.2)21 (3.7) 
 Other1 604 (3.2)91 (1.9)441 (3.6)19 (3.4) 
Serum creatinine (mg/dl), mean (SD)1.1 (0.9)1.0 (0.7)1.2 (1.0)1.0 (0.5)<0.001
History of hypertension (%)     
 Yes8 066 (16.1)822 (17.0)8 788 (72.2)427 (76.0)<0.001
 No42 176 (83.9)4 018 (83.0)3 384 (27.8)135 (24.0) 
Cause of death (%)     
 Anoxia8 273 (16.5)1 632 (33.7)674 (5.5)121 (21.5) 
 Cerebrovascular accident15 512 (30.9)797 (16.5)10 150 (83.4)350 (62.3)<0.001
 Head trauma24 840 (49.5)2 107 (43.5)1 184 (9.7)66 (11.7) 
 Other1 617 (3.2)304 (6.3)164 (1.3)25 (4.5) 
Body mass index (kg/m2), median (IQR)25.1 (7.2)25.8 (7.8)27.2 (7.2)28.8 (7.9)<0.001
Transplant factors
Induction therapy (%)     
 Yes37 172 (74.0)3 900 (80.6)9 239 (75.9)441 (78.5)<0.001
 No13 070 (26.0)940 (19.4)2 933 (24.1)121 (21.5) 
Cold ischemia time (hours), median (IQR)17.4 (11.2)18 (10.4)18 (11.2)18.3 (10.0)<0.001
HLA mismatches, median (IQR)4 (2)4 (2)5 (1)5 (1)<0.001
Year of transplant     
 2000–200318 097 (36.0)680 (14.1)3 747 (30.8)78 (13.9) 
 2004–200615 566 (31.0)1 504 (31.1)3 921 (32.2)189 (33.6)<0.001
 2007–200916 579 (33.0)2 656 (54.9)4 504 (37.0)295 (52.5) 

The distributions of KDRI scores across ECD/DCD categories are graphically depicted in Figure 1. Median (interquartile range) KDRI scores were generally higher among ECD versus non-ECD kidneys (1.95 [1.00], 2.24 [1.14], 3.66 [1.57] and 3.94 [1.59] for non-ECD/non-DCD, non-ECD/DCD, ECD/non-DCD and ECD/DCD, respectively). Although the difference in KDRI scores between DCD and non-DCD kidneys within a given ECD group was statistically significant (p < 0.0001), there was a considerable overlap in their distributions. In fact, the distribution of KDRI scores for DCD kidneys was completely nested within the distribution of non-DCD kidneys for each ECD group.

Figure 1.

Distribution of kidney donor risk index scores across ECD/DCD subgroups.

The incidence of PNF was 0.7%, 0.9%, 1.5% and 2.9% in recipients of non-ECD/non-DCD, non-ECD/DCD, ECD/non-DCD and ECD/DCD kidneys, respectively. Using the non-ECD/non-DCD group as the referent category, the adjusted odds ratios (aOR) for PNF in the non-ECD/DCD, ECD/non-DCD and ECD/DCD groups were 1.41 (95% CI, 1.02–1.94), 1.58 (95% CI, 1.22–2.14) and 3.19 (95% CI, 1.85–5.52), respectively. The aOR for the association of DCD status and PNF among non-ECD and ECD kidney recipients is shown in Table 2 (see Supporting Table S1 for full model results). Despite a quantitative difference in the point estimates, the p value for interaction was not statistically significant (p = 0.24). A similar nonsignificant interaction was observed for aOR across KDRI quintiles (p = 0.50).

Table 2. Odds ratios for DCD versus non-DCD kidney transplant recipients by ECD status for primary nonfunction and delayed graft function
OutcomeECD statusOdds ratio for DCD vs. non-DCD (95% CI)p-Value for interaction
  1. 95% CI = 95% confidence interval.

Primary non-Non-ECD1.41 (1.02–1.94)0.24
 functionECD2.02 (1.20–3.42) 
Delayed graftNon-ECD2.36 (2.21–2.53)0.17
 functionECD2.70 (2.25–3.24) 

The incidence of DGF was 21.3%, 39.6%, 30.5% and 53.3% in recipients of non-ECD/non-DCD, non-ECD/DCD, ECD/non-DCD and ECD/DCD kidneys, respectively. Using the non-ECD/non-DCD group as the referent category, the adjusted odds ratios (aOR) for DGF in the non-ECD/DCD, ECD/non-DCD and ECD/DCD groups were 2.36 (95% CI, 2.21–2.53), 1.19 (95% CI, 1.12–1.28) and 3.23 (95% CI, 2.68–3.88), respectively. However, the aOR for the association of DCD status and DGF was 2.36 (95% CI, 2.21–2.53) among non-ECD kidney recipients and 2.70 (95% CI, 2.25–3.24) among ECD kidney recipients (Table 2 and Supporting Table S1). The p value for interaction was not statistically significant (p = 0.17). A similar nonsignificant interaction was observed for aOR across KDRI quintiles (p = 0.61).

Kaplan–Meier curves for total graft failure, stratified by ECD/DCD status, are shown in Figure 2. The cumulative probability of total graft failure among DCD kidney transplant recipients was slightly greater than non-DCD recipients, irrespective of ECD status, at 1, 3, 5 and 9 years posttransplant. Similar patterns were noted for the outcomes of death-censored graft failure and death with graft function.

Figure 2.

Kaplan–Meier curves for total graft failure, death-censored graft failure and death with graft function by ECD/DCD subgroups. Log-rank p < 0.0001 for the overall comparison of ECD/DCD subgroups for each of the three study outcomes. See Supporting Table S2 for log-rank p values from pairwise comparisons of ECD/DCD subgroups for the three study outcomes.

A multivariable Cox proportional hazards models for total graft failure demonstrated adjusted hazard ratios (aHR)of 1.07 (95% CI, 1.01–1.15), 1.19 (95% CI, 1.13–1.25) and 1.43 (95% CI, 1.23–1.66) for recipients of non-ECD/DCD, ECD/non-DCD and ECD/DCD kidneys, respectively, when compared to non-ECD/non-DCD recipients. A similar monotonic increase across ECD/DCD groups was observed for death-censored graft failure (1.12 [95% CI, 1.03–1.23], 1.33 [95% CI, 1.24–1.42] and 1.65 [95% CI, 1.35–2.01]) and death with graft function (1.01 [95% CI, 0.92–1.12], 1.05 [95% CI, 0.97–1.14] and 1.23 [95% CI, 0.98–1.54]).

However, the modest increase in the risk of total graft failure for DCD versus non-DCD patients was not significantly modified by ECD status (Table 3). Furthermore, Table 4 shows that the aHR for total graft failure among DCD versus non-DCD patients were not significantly different across KDRI quintiles (likelihood ratio test p value = 0.40). A comparable low degree of heterogeneity was observed in the aHR across KDRI quintiles for death-censored graft failure (likelihood ratio test p value = 0.59) and death with graft function (likelihood ratio test p value = 0.65).

Table 3. Hazard ratios for DCD versus non-DCD kidney transplant recipients across ECD subgroups for total graft failure, death-censored graft failure and death with graft function
OutcomeECD statusHazard ratio for DCD vs. Non- DCD (95% CI)p-Value for interaction
  1. 95% CI = 95% confidence interval.

Total graft failureNon-ECD1.07 (1.01–1.15)0.14
 ECD1.21 (1.04–1.40) 
Death-censoredNon-ECD1.12 (1.03–1.23)0.34
 graft failureECD1.24 (1.04–1.54) 
Death with graftNon-ECD1.01 (0.92–1.12)0.26
 functionECD1.17 (0.93–1.45) 
Table 4. Hazard ratios for DCD versus non-DCD kidney transplant recipients across KDRI quintiles for total graft failure, death-censored graft failure and death with graft function
KDRI quintileNumber of patientsNumber of DCDTotal graft failure hazard ratios (95% CI)*Death-censored graft failure hazard ratios (95% CI)aDeath with graft function hazard ratios (95% CI)b
  1. *p-Value for likelihood ratio test of interaction terms = 0.40.

  2. a

    p-Value for likelihood ratio test of interaction terms = 0.59.

  3. b

    p-Value for likelihood ratio test of interaction terms = 0.65.

1 (0.63–1.49)13 4478811.19 (1.02–1.38)1.28 (1.04–1.57)1.10 (0.88–1.37)
2 (1.50–1.92)13 4468091.04 (0.89–1.22)1.07 (0.87–1.33)1.02 (0.80–1.29)
3 (1.93–2.38)13 44611491.13 (0.99–1.29)1.12 (0.93–1.35)1.14 (0.94–1.38)
4 (2.39–3.14)13 44613651.00 (0.88–1.13)1.03 (0.87–1.21)0.95 (0.79–1.15)
5 (3.15–9.75)13 44611371.04 (0.93–1.17)1.08 (0.93–1.25)0.97 (0.81–1.17)

Sensitivity Analyses

Repeating the analysis for total graft failure after categorizing KDRI into deciles or modeling KDRI as a continuous variable did not appreciably change our results. In addition, the primary results were robust to exclusion of the DCD variable in calculating the KDRI, adjustment for clustering of outcomes by transplant center, cohort restriction to a lower risk subpopulation and restricting to patients receiving kidneys that underwent pulsatile perfusion after recovery. The findings of the time-to-event analyses did not change if 0.5 days or a random number from 0 to 1 were imputed as survival times for patients with PNF. Moreover, adjustment for warm ischemia time in the subcohort of patients not missing these values (n = 6047) did not significantly alter the main results. Finally, restricting the analysis to transplant centers that performed at least one ECD/DCD kidney transplant over the study period (111 of 260 centers) had little impact on the overall results.

Discussion

The results of our study confirm existing knowledge that, despite the increase in DGF and PNF in recipients of DCD kidneys, their long-term unadjusted cumulative probability of graft failure is quite similar to that of recipients of standard kidneys (i.e. non-ECD/non-DCD). It also confirms that the graft failure rate of ECD kidney transplant recipients is higher than non-ECD recipients. However, the most notable result is that the relatively modest independent association between DCD status and graft/patient outcomes is not significantly modified by ECD status.

Upon examination of the characteristics of ECD/DCD kidneys chosen for transplantation, it is clear that the donors are somewhat younger, more likely hypertensive and less likely to have died of a cerebrovascular accident. When compared to donors of ECD/non-DCD kidneys. This suggests careful selection by transplant clinicians in the process of evaluating seemingly ‘higher risk’ donor kidneys for transplantation. Additionally, when the distribution of KDRI was plotted across ECD and DCD subgroups, the scores from ECD/DCD kidneys showed lower variability than the ECD/non-DCD subgroups. If fact, the ECD/DCD subgroup scores were entirely nested within the ECD/non-DCD subgroup scores. Multiple donor and transplant variables, in addition to the classic ECD variables of age, hypertension, death by CVA and serum creatinine, are included in the KDRI. Therefore, given the overlapping KDRI scores, we hypothesize that kidneys from more ‘favorable’ ECD donors were selected in order to offset the potential risk of a concomitant DCD-associated insult.

Our multivariable Cox proportional hazards model demonstrated a stepwise increase in the risk of graft failure across ECD and DCD subgroups, using non-ECD/non-DCD as the referent group. However, we believe that the most relevant clinical question is as follows: given the availability of an ECD kidney, does a concomitant DCD insult increase the risk of poor outcomes above and beyond that which is expected for an ECD kidney alone? Our study demonstrates that although DCD kidneys have a slightly greater adjusted risk of graft failure when compared to non-DCD kidneys, this increased risk is not significantly greater for ECD versus non-ECD kidneys. Given the potential limitations of using a dichotomous classification scheme to evaluate donor risk, we also used the KDRI as a more granular measure of donor risk and found no significant increase in the risk of graft failure for DCD versus non-DCD kidney transplantation across quintiles of the KDRI.

Several single center retrospective studies have described outcomes of combined ECD/DCD kidney transplantation with conflicting results. Saidi et al. showed similar graft survival at 2 years posttransplant when comparing ECD, SCD and DCD kidneys, although the mean follow-up for the ECD/DCD subgroup was limited to 27 months [14]. Tojimbara et al. evaluated short- and long-term outcomes of 256 DCD KTR, of which 29% were combined ECD/DCD kidneys. In this study, there was no difference in overall graft survival of ECD/DCD versus non-ECD/DCD. Additionally, when comparing those with DGF versus immediate graft function, there was no significant difference in the proportion of ECD kidneys with DGF [15]. However, other studies have shown inferior graft survival in recipients of ECD/DCD kidneys when compared to non-ECD/DCD kidneys [16, 17]. These results should be interpreted with caution given the small numbers of combined ECD/DCD kidney transplants in all of these studies, the relatively short follow-up time, and the presence of confounders that were not accounted for in the analyses.

Locke et al. conducted a large retrospective cohort study of over 78 174 deceased donor KTR, including over 2500 DCD kidney transplants, using data from the UNOS registry from 1993 to 2005. Graft survival was explored across donor subgroups (non-ECD, ECD and DCD). Similar to the previous studies, 5-year graft survival was only slightly inferior in the DCD subgroup versus the non-ECD subgroup (79.9% vs. 77.9%, p value<0.001), despite a high incidence of DGF among DCD KTR. Consistent with existing data, each component of the ECD classification predicted an increased risk of graft loss among non-DCD KTR. However, within the DCD subgroup, the only predictor of graft loss was age. DCD kidneys from donors >50 years of age had similar graft survival as the ECD subgroup, and DCD kidneys from donors <50 years of age similar to the non-ECD subgroup [18].

Our study differs from existing studies as it evaluates the outcomes of DCD kidney transplantation across ECD subgroups as well as KDRI scores, which takes into account additional risk factors not included in the classical ECD definition. In effect, we explored the role of ECD status or KDRI as an effect modifier of the relation between DCD status and outcomes [19].

Our main findings are perhaps counterintuitive, as one may speculate that the DCD insult should potentiate the impact of other ‘marginal’ donor characteristics on kidney transplant outcomes. Although these kidneys qualified as ECD/DCD, their characteristics may be more comparable to ECD/non-DCD kidneys based on the significant overlap in KDRI scores. The relatively small number of ECD/DCD kidneys transplanted over the study period also likely reflects careful organ selection by transplant clinicians and thus generalizing these findings to all combined ECD/DCD kidneys may not be appropriate. However, combined ECD/DCD kidneys with KDRI scores in the range observed in this study may warrant consideration for transplantation. The proportion of discarded ECD/DCD kidneys that have KDRI scores within the range observed in our study is not currently known, but they may represent an important untapped source of kidneys to safely expand the donor pool.

To our knowledge, this study is the largest to date exploring short- and long-term outcomes of combined ECD/DCD kidneys. Moreover, we evaluated the risk of graft failure risk in a more granular way using the KDRI along with the classical ECD definition. The strengths of this study include its large sample size, the use of a contemporary cohort reflective of the national experience in the United States and multivariable statistical modeling with adjustment for an extensive set of covariates.

Despite its strengths, our study has some limitations. First, the overall sample size is large but the subset of ECD/DCD kidney transplants is relatively small. As a result, we may have missed small but true differences in the risks of PNF, DGF, graft failure or death in DCD versus non-DCD kidney transplants across ECD groups. Larger numbers of ECD/DCD kidney transplants, along with more outcomes, will be needed to determine if the differences observed in our study are clinically and statistically meaningful. Second, the selection of ECD/DCD kidneys may involve factors other than those captured in the KDRI thus limiting generalizability of these findings to clinical practice. However, most of the key donor and transplant variables available to clinicians at the time of kidney offer are incorporated into the KDRI calculation. This makes it less likely that important factors were missed. Finally, residual confounding due to unmeasured or incompletely measured variables (e.g. warm ischemia time) may have influenced the study results. This is a perennial problem in observational research but our comprehensive multivariable modeling strategy and the conduct of sensitivity analyses likely enhances the validity and robustness of our primary analysis.

In summary, this study suggests that combined ECD/DCD kidneys, when carefully selected, may be an appropriate strategy to increase the deceased donor pool. Although the unadjusted outcomes of ECD/DCD kidney transplants are slightly inferior to non-ECD/DCD kidney transplants, they appear to be acceptable. Furthermore, the adjusted analyses suggest that the modifying effect of ECD status on the association of DCD status and outcome is likely modest at best. As the use of DCD and ECD kidneys increases in frequency over time, the availability of combined ECD/DCD kidneys for transplantation may also increase. Future studies should quantify the expected increase in the donor pool with the wider adoption of combined ECD/DCD kidneys and examine the factors that influence clinical decision making around the use of kidneys deemed to be at higher risk for graft failure. Furthermore, data from a Dutch cohort of wait-listed kidney transplant candidates suggest that both DCD and combined ECD/DCD kidney transplants provide a survival benefit when compared to dialysis [20]. This warrants further investigation to determine the potential life-years saved by using ECD/DCD kidney transplants to expand the deceased donor pool.

Acknowledgments

The data reported here have been supplied by the Minneapolis Medical Research Foundation (MMRF) as the contractor for the Scientific Registry of Transplant Recipients (SRTR). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of or interpretation by the SRTR or the U.S. Government.

Disclosure

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

Ancillary