Outcome of Kidney Transplantation Using Expanded Criteria Donors and Donation After Cardiac Death Kidneys: Realities and Costs


  • This work was presented at 2007 American Transplant Congress, San Francisco, CA.

* Corresponding author: Dicken S. C. Ko, ko.dicken@mgh.harvard.edu


Expanded criteria donors (ECDs) and donation after cardiac death (DCD) provide more kidneys in the donor pool. However, the financial impact and the long-term benefits of these kidneys have been questioned. From 1998 to 2005, we performed 271 deceased donor kidney transplants into adult recipients. There were 163 (60.1%) SCDs, 44 (16.2%) ECDs, 53 (19.6%) DCDs and 11 (4.1%) ECD/DCDs. The mean follow-up was 50 months. ECD and DCD kidneys had a significantly higher incidence of delayed graft function, longer time to reach serum creatinine below 3 (mg/dL), longer length of stay and more readmissions compared to SCDs. The hospital charge was also higher for ECD, ECD/DCD and DCD kidneys compared to SCDs, primarily due to the longer length of stay and increased requirement for dialysis ($70 030, $72 438, $72 789 and $47 462, respectively, p < 0.001). Early graft survival rates were comparable among all groups. However, after a mean follow-up of 50 months, graft survival was significantly less in the ECD group compared to other groups. Although our observations support the utilization of ECD and DCD kidneys, these transplants are associated with increased costs and resource utilization. Revised reimbursement guidelines will be required for centers that utilize these organs.


The organ shortage crisis has challenged the transplant community to continually evaluate new approaches that can maximize and optimize the use of organs from all consented donors (1). According to the United Network for Organ Sharing (UNOS) data, more than 74 723 candidates were on the active waiting list for kidney transplantation in the United States, while 17 092 kidney transplants were performed in 2006 (2).

One strategy that is being increasingly utilized to address the critical shortage of kidneys for transplantation is the expansion of the deceased donor (DD) kidney pool to include those who might have been deemed unsuitable in earlier times, such as programs to use donation after cardiac death (DCD) donors. Efforts to increase DCD have been highly successful and this source now accounts for more than 5% of all deceased organ donors. Transplantation of kidneys from DCD donors yields 1-year graft and patient survival rates equivalent to kidneys from brain-dead donors (3,4). Expanded criteria donor (ECD) kidneys from donors greater than or equal to 60 years of age (or donors aged 50–59 years with certain co-morbidities) have also been shown to confer a survival benefit for end-stage renal disease (ESRD) patients compared to remaining on dialysis on the waiting list (5,6). However, the financial impact on the transplant centers and the long-term benefits of transplanting these kidneys have been questioned.

This study was designed to compare resource utilization and outcomes in adult recipients of ECD and DCD with standard criteria donor (SCD) kidney transplants at a single center during the period of increased utilization of these grafts.


We conducted a retrospective analysis of a prospectively collected database of all DD kidney transplants performed in adult recipients at Massachusetts General Hospital from January 1, 1998, to December 31, 2005. The only exclusions from consecutive kidney transplants performed during that period were pediatric (<20 years of age), simultaneous kidney-pancreas, other multi-organ and living donor kidney allograft recipients. A total of 271 DD kidney transplants met the entry criteria of which 163 (60.1%) were SCDs, 44 (16.2%) ECDs, 11 (4.1%) ECD/DCD and 53 (19.6%) DCDs.

ECDs were defined by the UNOS criteria as all DDs over age 60 years as well as DDs 50–59 years of age with at least two of the following comorbidities: (i) history of hypertension, (ii) cerebrovascular cause of brain death, or (iii) terminal serum creatinine (SCr) level >1.5 mg/dL. DCDs included all donors for whom resuscitative measures were discontinued before the patient progressed to brain death criteria. ECD/DCD met both criteria. Donor kidney biopsy was regularly used in the evaluation of preexisting and terminal parenchymal pathology in ECD donors. A biopsy showing >20% glomerulosclerosis or moderate to severe tubular, interstitial or vascular changes was a contraindication to kidney utilization. All ECD and, whenever possible, DCD kidneys were placed on a pulsatile perfusion apparatus to potentially minimize preservation injury (7). Although pump parameters were not exclusively used to discard kidneys, a flow rate >80 mL/min and a resistance <0.40 mmHg after a minimum of 6 h on the perfusion apparatus were considered reasonable thresholds for utilization.

For purposes of this study, any DD not meeting the above ECD or DCD criteria was defined as an SCD. Delayed graft function (DGF) was defined as the need for dialysis in the first week posttransplant. Renal allograft loss was defined as death with a functioning graft, allograft nephrectomy, resumption of dialysis or return to the pre-transplant SCr level in recipients who had undergone pre-emptive transplantation.

All transplant candidates were evaluated with comprehensive pre-transplant medical and psychosocial evaluations with emphasis on the cardiovascular system and any other nonrenal organ failure to determine operative risks. All patients who were offered an ECD kidney underwent a standardized informed consent process. Listing patients for an ECD kidney neither mandated nor restricted them to receiving an ECD kidney, as the decision to accept any kidney, whether from an ECD or SCD, is reaffirmed at the time of the offer with informed consent.

At the time of transplantation, patients were selected on the basis of blood type compatibility, waiting time, HLA-matching, a negative cross match and special listing for ECD (when applicable) in accordance with UNOS guidelines.

All DD kidney transplant patients received depleting antibody induction using rabbit antithymocyte globulin at a dose of 1–1.5 mg/kg/day based on actual body weight. The first infusion was begun prior to allograft reperfusion, intraoperatively, and subsequent infusions were administered at postoperative day 1 and 2. Depending on initial graft function, one or two more doses were administered to enable a delay in initiating tacrolimus (TAC). Maintenance immunosuppression consisted of TAC, mycophenolate mofetil (MMF) and tapering doses of steroids. The administration of TAC was delayed until the patient had exhibited a diuresis and a declining SCr level to <5.0 mg/dL. Target 12-h TAC trough levels were based on donor quality and recipient immunologic risk but typically ranged from 8 to 12 ng/mL.

All patients received surgical site prophylaxis with a first-generation cephalosporin for 24 h, antifungal prophylaxis with clotrimazole for 5 days and anti-Pneumocystis prophylaxis with sulfamethoxazole/trimethoprim for 6 months. The prophylaxis regimen was tailored if there was a history of allergy to standard protocol drugs. Antiviral prophylaxis consisted of oral valganciclovir for 4–6 months if either donor or recipient cytomegalovirus (CMV) serologic status was positive, while oral famciclovir was used if both donor and recipient serologies were negative. Before valganciclovir become commercially available (June 2001), oral ganciclovir 1000 mg pot id was used.

Univariate analysis was performed by the unpaired t-test for continuous variables, the chi-square test for categorical variables and the Fisher's exact test when data were sparse. Unadjusted actual patient and graft survival rates were reported. Patient and graft survival curves were also computed using the Kaplan–Meier method and compared using the log-rank test. Cox regression analysis was performed to test the impact of different factors on allograft survival. Categorical data were summarized as proportions and percentages, and continuous data were summarized as means and standard deviations. A p-value of <0.05 was considered to be significant.


From January 1, 1998, through December 31, 2005, we performed a total of 271 DD kidney transplants into adult recipients including 163 (60.1%) SCDs, 44 (16.2%) ECDs, 11 (4.1%) ECD/DCD and 53 (19.6%) DCDs.

Donor, recipient and allograft characteristics are summarized in Table 1. Due to the defining criteria by UNOS, ECDs were older (61.2 vs. 36.1), had a higher incidence of preexisting hypertension (70.2% vs. 16.8%) and a higher SCr (1.1 vs. 0.8 mg/dL) compared with SCDs (p < 0.05). ECD kidney transplant recipients were also older (Table 1) compared with SCD and DCD kidney recipients, 57.1, 49.2 and 48.6, respectively (p < 0.01). As shown in Table 1, there were no significant differences between donor and recipient characteristics in the SCD and DCD groups. There were no significant differences between any of the groups based on donor body mass index or cold ischemic time.

Table 1.  Donors and recipient characteristics*
 SCD (n = 163)ECD (n = 44)ECD/DCD (n = 11)DCD (n = 53)
  1. BMI = body mass index; CIT = cold ischemic time; Cr = creatinine; HTN = hypertension.

  2. *Mean ± standard deviation.

  3. **p < 0.001 (comparing ECD and ECD/DCD groups with SCD and DCDs).

Donor age36.1 ± 6.461.2 ± 7.2**56.1 ± 8.6**36.8 ± 9.8 
Donor BMI27.4 ± 7.426.3 ± 6.8     25.5 ± 6.7  28.2 ± 5.9 
Donor HTN16.8%70.2%27.3%15.1%
Terminal serum Cr   0.8 ± 0.29 1.1 ± 0.4** 1.08 ± 0.42** 0.8 ± 0.25
CIT(h)15.9 ± 7.416.2 ± 6.6   16.5 ± 6.8  15.4 ± 7.9 
Recipient age49.2 ± 6.757.1 ± 6.4**55.6 ± 7.2**48.6 ± 6.8 

Overall, 3 of the 271 recipients (1.1%) died during the immediate postoperative period (30 days), due to cardiovascular events. The incidence of DGF graft function was significantly higher in ECD, DCD and ECD/DCD recipients compared to SCDs, 35.6%, 33.3%, 38.2% and 15.1% (p < 0.001), respectively (Table 2). In addition, time to reach SCr below 3 mg/dL was significantly longer in all non-SCD groups compared to SCDs. The length of initial hospitalization was also longer for non-SCD kidney recipients compared to SCDs, 9.8–11 in other groups compared to 6.1 days in SCDs (p < 0.05). Readmissions (within 90 days) were also more common in non-SCD groups compared to SCDs. The incidence of acute rejection and BK nephropathy were comparable among all groups.

Table 2.  Outcomes*
 SCD (n = 163)ECD (n = 44)ECD/DCD (n = 11)DCD (n = 53)
  1. Cr = creatinine; DGF = delay graft function.

  2. *Mean ± standard deviation.

  3. **p < 0.001 (comparing SCDs with other groups).

  4. 1rejection at 1 year.

Perioperative mortality n = 3 (1.1%) 1.2%1.8%00
Time to Cr < 3 (mg/dL) 4 ± 2.910.6 ± 4.5**10.8 ± 4.9**9.7 ± 3.9**
Length of stay6.1 ± 2.910.2 ± 5.2**   11 ± 6.2**9.8 ± 5.6**
Acute rejection114.7%13.6%14.2%13.9%
BK nephropathy3%2%2%2%
Hospital charge ($)47 462 ± 108970 030 ± 2849**72 438 ± 3012**72 789 ± 3259**

The data clearly show that the utilization of ECD, ECD/DCD and DCD kidneys was associated with higher initial hospitalization charges (Table 2). These kidneys had a higher incidence of DGF that required dialysis and also longer length of stay (LOS). This difference in initial hospitalization charges for ECD, ECD/DCD and DCD kidney recipients compared to SCDs, $70 030, $72 438, $72 789 and $47 462, respectively (p < 0.001), continued to escalate as more re-admissions were required during the peri-transplant period.

Mean follow-up in all groups was 50 months except for the ECD/DCD group, which was 27 months. Patient and graft survival are shown in Figures 1 and 2. Early (12 and 24 months) graft survival rates were comparable between all groups. However, after a mean follow-up of 50 months, graft survival was significantly less in the ECD group compared to DCD and SCDs, 65%, 79% and 80%, respectively (p = 0.0116). As ECD kidney recipients had a higher incidence of DGF, we analyzed the effects of DGF on graft survival. As shown in Table 3, DGF had not a negative impact on the graft survival.

Figure 1.

Kaplan–Meier analysis of kidney allograft survival by donor type. Survival is comparable among groups until 2 year posttransplant, after which a significantly increase rate of graft loss is noted in the ECD kidney recipients (p = 0.0116).

Figure 2.

Kaplan–Meier analysis of patient survival by donor type. Long-term recipient survival is comparable among all groups.

Table 3.  Impact of DGF on risk of graft failure
Variable HRp-Value
  1. HR = hazard ratio.

DGFYes vs. no1.2560.44

The long-term outcome of the ECD/DCD group cannot be determined yet since there were only 11 patients in this group with relatively short follow-up (mean = 27 months). However, these kidneys did have comparable short- and intermediate-term survival to that observed in SCD, ECD and DCD recipients. Both intermediate- and long-term patient survival were comparable among all groups.


Kidney transplantation is unquestionably the preferred therapy for most patients with ESRD since both survival and quality of life are superior in allograft recipients compared to similar patients who remain on dialysis (1). As outcomes of renal transplantation have improved, the number of renal transplant candidates listed for DD kidney transplantation has increased dramatically over the years. One of the main strategies to address the discrepancy between supply and demand in organ transplantation is expansion of the DD kidney pool utilizing ECD and DCD donors (1–3).This has been a major focus of the US Department of Health and Human Services organ donation breakthrough collaborative which was initiated in 2003, with the objective of increasing access to transplantable organs.

Before the organ donation collaborative, the term ECD was used to classify subsets of DD that are aged 60 years or older and those aged 50–59 years with at least two of the following characteristics: history of hypertension, SCr level greater than 1.5 mg/dL (132.6 μmol/L) and cerebrovascular cause of death (7). These criteria define a donor population from whom the risk of graft failure after transplantation was anticipated to be 70% higher than after a non-ECD transplant (8–10). Despite this expected higher rate of graft failure compared to SCD kidneys, multiple studies have subsequently shown that kidney transplantation using ECDs is still associated with a substantial reduction in morbidity and improvement in life expectancy when compared with suitable transplant candidates who remained on maintenance dialysis treatment (8–10). If we consider an increased risk of complications and death for those who have to wait, kidney transplants clearly save more lives and cost less as a treatment than does dialysis.

Ojo and colleagues showed that on average, recipients of ECD kidney transplants lived 5 years longer than transplant candidates who remained on dialysis, whereas SCD transplant recipients had a 13-year survival benefit (6). Accordingly, ECD kidney transplants have continued to expand and comprise 15% of national DD activity (1,2).

Another approach to the organ shortage has been the utilization of donors after cardiac death. The Institute of Medicine has studied the issues surrounding the use of DCDs, reaching the conclusion that ‘the recovery of organs from nonheart beating donors is an important, medically effective and ethically acceptable approach to reducing the gap that exists now and will continue to exist in future between the demand for and available supply of organs for transplantation’ (8). A number of investigators have reported excellent short-term outcomes using these donors, and several different organ procurement organizations have demonstrated 10–15% growth in organ donation as a result of the use of DCD donors. Multiple studies have shown that the overall results of DCD (without ECD characteristics) and SCD kidney transplants are comparable (9–14).

Our observations regarding long-term DCD kidney survival are consistent with the earlier reports (11–13) Doshi et al. reviewed the UNOS database (from 1998 to 2004) and showed that the 5-year allograft and patient survival rates of 66.9% and 81.3% in recipients of DCD kidneys was comparable with 66.5% and 80.8% graft and patient survival in the SCD group, respectively (14). This most recent review also confirmed a higher incidence of DGF in the DCD group compared to SCDs (41% vs. 24%) which led to a longer LOS (10.2 vs. 9 days). In addition, the incidence of rejection was similar in both groups (9.4% vs. 10%). Doshi's review did not include the fate of ECD kidneys in the long-term analysis. Our study at the Massachusetts General Hospital, similarly, showed no differences in the long-term graft and patient survival, a higher incidence of DGF, longer LOS and similar incidence of rejection in recipients of DCD versus SCD kidneys. Our long term follow-up, however, reveals a more rapid attrition in survival of the ECD allografts (65% survival at 50 months vs. 79–80% for SCDs).

Although we observed an increased incidence of DGF using DCD and ECD kidney allografts, we found no difference in the long-term graft and patient survival between DCD and SCD recipients. This contrasts with other studies which have shown a negative impact of DGF on kidney allograft survival from SCD donors (11–16). DGF is a multifactorial phenomena resulting from warm and cold ischemia, poor donor quality and/or early rejection (16). Since there was no correlation between DGF and graft survival in our series it might be postulated that our protocol of routine induction therapy with antithymocyte globulin and delayed tracrolimus introduction limited the likelihood of early subclinical rejection. This could provide an explanation for the lack of a detrimental impact of DGF on long-term graft survival observed in our series. Of course, these observations require further studies to confirm the hypothesis.

Nationwide, a significant number of DCD and ECD kidneys continue to be discarded. However, as already noted, despite reports demonstrating that these kidneys have a higher rate of DGF and a greater susceptibility to preservation injury as well as drug toxicity, the long-term outcome is quite satisfactory (17–25). Thus, the major obstacle to more widespread utilization of these organs seems to be economic. We anticipated that the transplantation of ECD and DCD kidneys would result in higher costs when we embarked on utilizing these organs for our patients.

In this review, we have documented the more frequent need for hemodialysis, longer LOS, more hospital readmissions due to poor or late onset graft function and more CMV infections in recipients of ECD and DCD kidneys. In our institution, this resulted in a $20 000–25 000 higher cost for their initial medical care. Whiting et al. have compared the economic costs of ECD kidney transplantation to those on hemodialysis (17). They found the break-even point at which transplantation became less costly than ongoing hemodialysis ranged from 4.4 years for SCDs to as long as 13 years if an ECD kidney was transplanted into a high-risk recipient. The 5-year present value of payments for ECDs was significantly higher compared to SCDs: $143 329 versus $121 698, respectively. The average 1-year cost of hemodialysis for this cohort was $28 666. They concluded that transplantation was always less expensive than hemodialysis. Although, in a strict cost analysis, ECD kidney recipients take longer to reach the break-even point and this point might not be reached consistently if allograft survival proved to be dramatically reduced in some patient cohorts, in general, utilization of DCD and ECD kidneys is still a cost-saving treatment strategy when compared to hemodialysis (26–28).

In summary, review of our experience has demonstrated that the cost for utilizing marginal donors increases due to several factors discussed. Nevertheless, reimbursement for the transplantation services is based upon a single diagnostic-related group (DRG) which assumes that all renal transplants require comparable resource utilization. In reality, the increased resource needs demonstrated by our study emphasize that centers who pursue the use of ECD and DCD kidneys are penalized financially for attempting to serve the largest possible number of ESRD patients. As the transplant community seeks to utilize as many organs as possible in order to benefit increasing number of patients on the waiting list, we must take into consideration the true economic impact of such undertakings. This may not be addressed based on current remuneration schemes for services rendered. When costs increase significantly, as we have observed, transplant centers must absorb the deficit spending for each of these organs utilized. As a result, the current policies for reimbursement impose a significant burden on those transplant centers that accept these organs and jeopardize their financial viability despite the medical, social and ethical benefits of maximizing organ utilization, We advise, therefore, that a separate DRG or modifier should be considered for ECD and DCD kidney recipients so that reimbursement is adjusted based on the donor and recipients risk. Such action can sustain the organ breakthrough collaboration efforts for increasing organ utilization. Otherwise, we risk ongoing underutilization of ECD and DCD allografts despite the documented benefits provided to ESRD recipients by these organs.


We appreciate Dr. David Schoenfeld's assistance in the statistical analysis.