Abstract
Certain patient groups are predicted to derive significant survival benefit from transplantation with expanded criteria donor (ECD) kidneys. An algorithm published in 2005 by Merion and colleagues characterizes this group: older adults, diabetics and registrants at centers with long waiting times. Our goal was to evaluate ECD listing practice patterns in the United States in terms of these characteristics. We reviewed 142 907 first-time deceased donor kidney registrants reported to United Network for Organ Sharing (UNOS) between 2003 and 2008. Of registrants predicted to benefit from ECD transplantation according to the Merion algorithm (’ECD-benefit’), 49.8% were listed for ECD offers (’ECD-willing’), with proportions ranging from 0% to 100% by transplant center. In contrast, 67.6% of adults over the age of 65 years were ECD-willing, also ranging from 0% to 100% by center. In multivariate models, neither diabetes nor center waiting time was significantly associated with ECD-willingness in any subgroup. From the time of initial registration, irrespective of eventual transplantation, ECD-willingness was associated with a significant adjusted survival advantage in the ECD-benefit group (HR for death 0.88, p < 0.001) and in older adults (HR 0.89, p < 0.001), but an increased mortality in non-ECD-benefit registrants (HR 1.11, p < 0.001). In conclusion, ECD listing practices are widely varied and not consistent with published recommendations, a pattern that may disenfranchise certain transplant registrants.
Keywords: Deceased donor transplantation, expanded criteria donor, older adults, organ utilization
Background
Organ shortages remain a significant problem for those with end-stage renal disease (ESRD) seeking kidney transplantation. Despite this chronic disparity between organ supply and demand, many recovered kidneys are still discarded (1–4). Historically, organs with the highest rates of discard are those with an increased relative risk of graft loss (’marginal’ donor kidneys) (1,5,6). These organs, while inferior to standard criteria donor (SCD) kidneys, still provide a survival advantage over maintenance dialysis (7). In an effort to reduce discard rates, the Organ Procurement and Transplantation Network (OPTN) implemented in 2002 an expanded criteria donor (ECD) program (8).
Under the ECD program, transplant registrants indicate at the time of listing their willingness to consider an ECD kidney. By design, these organs have a 1.7-fold higher risk of graft loss compared with SCD organs (6,9). Despite this increased risk, certain groups of registrants are predicted to derive a long-term survival benefit from transplantation with an ECD organ compared with waiting for an SCD one, due to high waiting list mortality rates and shorter waiting times for ECD organs (7,10). On the other hand, given the increased risk of graft loss and mortality, some patients may fare worse with an ECD organ (11,12). In order to efficiently allocate organs, therefore, it is critical to deliver ECD kidneys to those who would benefit from them (11,13,14).
The estimation of which registrants might benefit from transplantation with an ECD organ typically falls to clinicians and their patients. To aid in this decision, several models for evidence-based ECD listing behavior have been proposed (10,12,15,16). Perhaps the most publicized and straightforward was an algorithm introduced by Merion and colleagues in the Journal of the American Medical Association in 2005 (10). They demonstrated that patients over 40 years of age, with either diabetes as the primary cause of ESRD or non-Hispanic race and listing at centers with median waiting times greater than 1350 days, are predicted statistically to have long-term survival benefit from accepting ECD kidney transplantation (’ECD-benefit’) compared with waiting on the list for the next available non-ECD kidney.
Intuitively, another group of patients who might derive a survival advantage from ECD transplantation are older adults (aged 65 or older). In Europe, this intuition has translated into clinical practice through the Eurotransplant Seniors Program (ESP), where kidneys from older donors are expressly allocated to older patients on the waiting list (17–20).Older adults represent over one-half of dialysis patients in the United States and a growing proportion of the national waiting list (21,22). In addition, registrants aged 65 or older have a stepwise reduction in likelihood of transplantation (and a stepwise increase in death while waiting) as waiting time increases, significantly greater than those seen in younger age groups (16). For all age groups, the realized survival advantage after transplantation is delayed because of an increased mortality in the perioperative period; however, with a waiting list annual mortality rate of 10%, older adults reach a survival benefit much earlier than their younger counterparts (7,12). If ECD organs confer a shorter waiting time, older recipients would be expected to derive a significant gain in survival when compared to waiting for an SCD organ. Schnitzler and colleagues modeled this ’break-even point’, or minimum waiting time difference between ECD and SCD organs that would have to exist in order for ECD organs to confer a survival benefit, as 11 months for patients over the age of 60 years (12).
Variations in ECD listing practices across the United States have been reported (23,24). To our knowledge, however, these listing variations have not been evaluated in terms of published recommendations or evidence-based predicted benefit from ECD listing. Given that registering as willing to accept an ECD kidney (’ECD-willing’) increases the likelihood of future transplantation (23), evidence-based listing behavior is likely to impart a survival advantage (25). The goal of this study, therefore, was to evaluate the incorporation of evidence into ECD listing practices in a national cohort of kidney transplant candidates. We analyzed listing practices among (a) all registrants, (b) patients predicted to benefit from ECD transplantation according to the Merion algorithm and (c) older registrants (aged 65 or older). We addressed both patient-level and center-level factors associated with listing practices, as well as the implications of ECD listing on survival.
Methods
Source data and study population
This was a secondary analysis of prospectively collected transplant and waiting list data as reported to the United Network for Organ Sharing (UNOS). Death information was augmented by linkage to the Social Security Death Master File. In order to create a cohort consistent with the Merion study (i.e. to which the ECD-benefit algorithm was applicable), we adopted inclusion and exclusion criteria similar to that used in their survival benefit model (10). Of the 257 035 new registrants for kidney-only deceased donor transplants between 2000 and 2008, we excluded 9207 who had previously received a nonkidney organ transplant and 34 887 who had previously received a kidney transplant. We also excluded 14 729 registrants who received a live donor transplant within 6 months of initial registration, as these patients likely had identified a live donor at initial registration and thus had little incentive to list as ECD-willing. The remaining 198 212 registrations were used to calculate median deceased donor transplant waiting time in 262 transplant centers by the Kaplan-Meier method, censoring for death, removal from the waiting list or end of study. For the remainder of the analysis, the cohort was limited to the years 2003 through 2008. Using different time periods was an attempt to more accurately capture historical median wait times that transplant centers and registrants might have used in deciding on ECD listing status.
Exposure and outcome definitions
Our primary outcome was ECD listing status, as reported to UNOS. Those who were listed as willing to accept an ECD kidney were labeled ’ECD-willing’. We also studied a secondary outcome of death from the time of initial registration to compare ECD-willing and non-ECD willing patients, using survival models for the event of death and censored for loss to follow-up or end-of-study (but not censored for transplantation, so as to become an intention-to-treat type of design). An important exposure of interest was whether a person was predicted to benefit from ECD transplantation as directed by the recommendations of Merion et al. (’ECD-benefit’) (10). Using their algorithm, we defined ECD-benefit registrants as those with age greater or equal to 40 years and with either (1) diabetes as a cause of ESRD, or (2) non-Hispanic race and listing at a transplant center with a median wait time of 1350 days or longer. Other calculated exposures included median center waiting time, which is described in the previous section, and transplant center volume, which was defined as total number of deceased donor renal transplants performed at a center over the study period.
Analytical methods
All regression models included the following covariates: age category, gender, ethnicity, body mass index, primary cause of ESRD, panel reactive antibody (PRA), primary insurance, median transplant center waiting time and center transplant volume over the study period. All variables were tested for collinearity by examining variance inflation factors. Associations between covariates and ECD listing status were estimated using a generalized linear model with a Poisson family, log link and robust variance estimates to determine the relative rate for each variable, as previously described (26). In our outcome analyses, time to death was estimated using the Kaplan–Meier method, and adjusted comparisons by ECD listing status were made using Cox proportional hazards models. To account for center-level differences in listing practice, organ availability and latent registrant factors, all regression models were adjusted for center-level clustering. Analyses were performed using Stata 11.0/MP (StataCorp, College Station, TX).
Results
ECD-willing registrants
Of 142 907 first-time kidney transplant registrants between 2003 and 2008, 44.9% were ECD-willing (Table 1, columns 2 and 3). Those classified as ECD-benefit (by the Merion algorithm) were split almost equally between ECD-willing (49.8%) and not-ECD willing (50.2%). In contrast, 40.2% of ECD-willing registrants were not predicted to benefit from ECD transplantation. ECD-willing registrants were older (39.8% over 60 years vs. 18.2%), more likely to be male (61.0% vs. 58.7%) and more likely to have diabetes as the primary cause of their kidney failure (38.4% vs. 31.8%). Age was correlated in a stepwise manner with ECD-willingness, with 71.7% of adults over 70 years and 62.2% of adults between 60 and 69 years listed for ECD kidneys, as compared with 27.1% of those between 18 and 39 years and only 6.3% of those under 18 years. In addition, while unadjusted for age, more ECD-willing registrants used Medicare as primary insurance (50.2% vs. 39.2%) and more had education beyond a college degree (7.0% vs. 5.7%). Finally, on average, ECD-willing registrants came from transplant centers with shorter waiting times than non-ECD willing registrants (median 1512 vs. 1675 days for any deceased donor kidney; 1780 vs. 1887 days for an SCD kidney).
Table 1.
Baseline characteristics among all registrants during the study period (columns 2–3), those predicted to benefit from ECD kidneys according to the Merion algorithm (columns 4–5) and registrants aged 65 or over (columns 6–7), stratified by willingness to consider ECD offers
| All registrants | ECD-Benefit | Age ≥ 65 | ||||
|---|---|---|---|---|---|---|
| Willingness to consider ECD offers | No | Yes | No | Yes | No | Yes |
| N | 78 730 | 64 177 | 38 244 | 37 981 | 6 938 | 14 498 |
| % | 55.1% | 44.9% | 50.2% | 49.8% | 32.4% | 67.6% |
| ECD-benefit (by Merion criteria), % | 48.6 | 59.2 | 100.0 | 100.0 | 72.3 | 67.3 |
| Age (mean) | 46.1 | 54.8 | 54.2 | 58.3 | 68.7 | 69.3 |
| Male (%) | 58.7 | 61.0 | 60.1 | 62.8 | 60.5 | 64.3 |
| Race (%): | ||||||
| White | 43.6 | 46.0 | 46.7 | 46.4 | 53.2 | 58.9 |
| Black | 29.1 | 31.1 | 28.8 | 31.4 | 20.7 | 21.5 |
| Hispanic | 17.6 | 15.6 | 11.9 | 13.2 | 14.2 | 12.1 |
| Asian | 7.4 | 5.3 | 9.8 | 6.5 | 10.2 | 5.7 |
| BMI (mean) | 27.7 | 28.3 | 28.7 | 28.8 | 27.9 | 27.7 |
| Primary cause of ESRD (%): | ||||||
| Glomerulonephritis | 20.6 | 13.8 | 10.9 | 7.0 | 10.5 | 9.7 |
| Diabetic Nephropathy | 31.8 | 38.4 | 54.2 | 60.4 | 40.9 | 37.9 |
| Polycystic kidney disease | 7.5 | 7.1 | 6.4 | 4.5 | 5.7 | 5.9 |
| Hypertensive nephrosclerosis | 19.4 | 21.8 | 15.6 | 15.1 | 24.6 | 25.4 |
| Diabetes (%) | 39.3 | 49.2 | 61.9 | 69.3 | 52.8 | 50.5 |
| PRA (%): | ||||||
| 0–19% | 79.8 | 77.7 | 80.5 | 79.3 | 82.6 | 80.6 |
| 20–79% | 13.1 | 14.1 | 12.7 | 13.4 | 11.6 | 12.9 |
| 80–100% | 7.1 | 8.1 | 6.8 | 7.4 | 5.8 | 6.6 |
| Primary insurance (%): | ||||||
| Private | 48.0 | 40.9 | 52.5 | 41.6 | 33.4 | 22.8 |
| Medicare | 39.0 | 50.2 | 37.8 | 49.8 | 61.4 | 72.0 |
| Medicaid | 10.0 | 6.3 | 7.0 | 5.8 | 3.5 | 3.0 |
| Education (%): | ||||||
| High school or less | 56.1 | 55.9 | 52.7 | 54.5 | 56.9 | 56.4 |
| College | 38.2 | 37.1 | 40.4 | 37.8 | 33.5 | 33.6 |
| Beyond college | 5.7 | 7.0 | 6.8 | 7.7 | 9.6 | 10.0 |
| Median waiting time (days) | 1675 | 1512 | 1968 | 1737 | 1798 | 1502 |
| Median waiting time for SCD (days) | 1887 | 1780 | 2203 | 2048 | 1998 | 1783 |
Within the ECD-benefit cohort, there were 76 225 first-time registrants (Table 1, columns 4 and 5). As in the total cohort, those listed as ECD-willing were older, more often male, more likely to have diabetes as a cause of their renal failure, more likely to have Medicare as a primary insurer and more likely to have postcollege education. Again, ECD-willing, ECD-benefit registrants were more commonly from centers with shorter waiting times, with 27.1% vs. 18.5% from centers with waiting times under 1350 days (the cutoff used in the Merion algorithm). The median center waiting time, whether calculated as time to any deceased donor kidney transplantation or time to SCD kidney transplantation, was shorter for ECD-willing registrants.
Finally, in the cohort of 21 436 registrants aged 65 years or older, a much larger proportion (67.6%) was listed as ECD-willing (Table 1, columns 6 and 7). Characteristics associated with ECD-willingness were very similar to those in the two other cohorts. As expected, the proportion predicted as ECD-benefit was higher in this cohort. However, when categorized by ECD-willing status, 67.3% of those listed as ECD-willing were ECD-benefit and 72.3% of those who were non-ECD willing were ECD-benefit. Here, too, center-level median waiting time was paradoxically lower in those who were ECD-willing, regardless of whether time to transplantation included ECD transplants in the calculation.
Era effects
To determine whether practice patterns changed after the 2005 publication of Merion’s algorithm, we calculated the percentage of ECD-benefit registrants who were listed as ECD-willing in 2003 through 2004, which was prepublication, and compared that to the proportion in 2006 through 2007, which was postpublication. There was no difference in ECD listing of ECD-benefit patients, with 49.5% listed in the earlier time period, compared to 49.7% in the later time period (p = 0.7). When stratified by center, the median change at a given center in proportion of ECD-benefit patients who were ECD-willing pre- to postpublication was 0% (IQR −6.6% to +7.5%).
Transplant center characteristics
Of transplant centers listing new registrants during our study period, 40% had median waiting times over 1350 days. These centers represented 80 772 listings, or 56.5% of our cohort. When calculated using waiting times for SCD organs only, 48.8% of centers (representing 66.7% of our cohort) had waiting times over 1350 days. Listing practices differed significantly by center (Figure 1). Among ECD-benefit registrants, there was large by-center variation in the proportion listed as ECD-willing, which was not associated with center median waiting time. For example, there were 41 centers that listed over 90% of the ECD-benefit patients as ECD-willing, yet there were 17 that listed fewer than 10% as such. Similarly, among older patients, 57 centers listed more than 90% as ECD-willing, while 8 listed fewer than 10%. There was more center-level consistency in ECD listing practices for older registrants than there was in listing practices for registrants identified by the Merion algorithm as predicted to benefit from ECD transplants.
Figure 1. Variation in transplant center listing in (A) registrants predicted to benefit from ECD kidneys according to the Merion algorithm and (B) registrants aged 65 or over.
Shown are distributions, with the X-axis indicating a proportion from 0 to 100% and the Y-axis indicating how many transplant centers listed that proportion of their ECD-benefit or older registrants for ECD kidneys. Bars on the Y-axis are further stratified by center waiting time.
Predictors of ECD-willingness
In a multivariate generalized linear model of demographics and center-level data on ECD listing, older age groups, male sex, black race, glomerulonephritis as a cause of renal disease, higher PRA levels and Medicare primary insurance were significantly associated with ECD listing in the total cohort (Table 2, columns 2 and 3). Among ECD-benefit registrants, similar associations were seen, with older age, male sex, black race, glomerulonephritis, polycystic kidney disease, and Medicare as primary insurance significantly associated with ECD listing (Table 2, columns 4 and 5). Finally, among older registrants, older age, male sex, Asian race, glomerulonephritis as a cause of renal failure, and Medicare primary insurance were associated with ECD listing (Table 2, columns 6 and 7). Diabetes as a primary cause of renal failure (part of the Merion algorithm), median center waiting time and center volume were not associated with ECD listing.
Table 2.
Association of registration characteristics with willingness to consider ECD offers, among all registrants during the study period (columns 2–3), those predicted to benefit from ECD kidneys according to the Merion algorithm (columns 4–5) and registrants aged 65 or over (columns 6–7). Relative rates (RR) and p-values from generalized linear models are shown.
| All registrants | ECD-Benefit | Age ≥ 65 | ||||
|---|---|---|---|---|---|---|
| RR (95% CI) | p-Value | RR (95% CI) | p-Value | RR (95% CI) | p-Value | |
| Age category: | ||||||
| 18– 39 years | Reference | |||||
| 40–49 years | 1.41 (1.30–1.53) | <0.001 | Reference | |||
| 50–59 years | 1.87 (1.62–2.15) | <0.001 | 1.34 (1.22–1.47) | <0.001 | ||
| 60–69 years | 2.37 (1.98–2.83) | <0.001 | 1.70 (1.47–1.96) | <0.001 | Reference | |
| 70–99 years | 2.67 (2.18–3.27) | <0.001 | 1.89 (1.58–2.26) | <0.001 | 1.07 (1.03–1.12) | <0.001 |
| Sex: | ||||||
| Male | 1.05 (1.02–1.07) | <0.001 | 1.06 (1.03–1.09) | <0.001 | 1.06 (1.03–1.09) | <0.001 |
| Race: | ||||||
| White | Reference | Reference | Reference | |||
| Black | 1.10 (1.02–1.18) | 0.012 | 1.09 (1.02–1.17) | 0.011 | 1.01 (0.96–1.08) | 0.6 |
| Hispanic | 1.02 (0.86–1.20) | 0.8 | 1.03 (0.86–1.23) | 0.7 | 0.98 (0.88–1.09) | 0.7 |
| Asian | 0.88 (0.75–1.05) | 0.2 | 0.85 (0.71–1.03) | 0.1 | 0.85 (0.74–0.97) | 0.02 |
| Body Mass Index | ||||||
| >35 | 1.00 (0.95–1.05) | 0.9 | 0.98 (0.93–1.03) | 0.4 | 0.96 (0.92–1.00) | 0.1 |
| Diagnosis: | ||||||
| GN | 0.90 (0.83–0.97) | 0.009 | 0.83 (0.74–0.94) | 0.003 | 0.93 (0.88–1.00) | 0.042 |
| PKD | 0.95 (0.89–1.01) | 0.1 | 0.88 (0.80–0.97) | 0.011 | 0.96 (0.91–1.01) | 0.1 |
| DM | 1.02 (0.94–1.10) | 0.7 | 0.95 (0.84–1.08) | 0.4 | 0.96 (0.90–1.03) | 0.2 |
| HTN | 0.98 (0.90–1.06) | 0.6 | 0.93 (0.84–1.04) | 0.2 | 0.97 (0.92–1.01) | 0.2 |
| PRA: | ||||||
| 0–19% | Reference | Reference | Reference | |||
| 20–79% | 1.06 (0.99–1.12) | 0.1 | 1.05 (0.98–1.12) | 0.2 | 1.05 (1.00–1.10) | 0.1 |
| 80–100% | 1.11 (1.03–1.19) | 0.004 | 1.07 (0.99–1.15) | 0.1 | 1.06 (1.00–1.12) | 0.1 |
| Insurance: | ||||||
| Private | Reference | Reference | Reference | |||
| Medicare | 1.13 (1.05–1.22) | 0.001 | 1.15 (1.06–1.26) | 0.001 | 1.15 (1.01–1.30) | 0.032 |
| Medicaid | 0.98 (0.88–1.10) | 0.8 | 1.05 (0.94–1.17) | 0.4 | 1.10 (0.99–1.22) | 0.1 |
| Center waiting time: | ||||||
| 0–749 days | Reference | Reference | Reference | |||
| 750–1349 | 0.93 (0.71–1.21) | 0.6 | 0.92 (0.75–1.12) | 0.4 | 0.96 (0.84–1.11) | 0.6 |
| 1350–1999 | 0.83 (0.61–1.12) | 0.2 | 0.85 (0.67–1.07) | 0.2 | 0.96 (0.83–1.12) | 0.6 |
| 2000–9999 | 0.76 (0.51–1.11) | 0.2 | 0.74 (0.53–1.04) | 0.1 | 0.78 (0.57–1.07) | 0.1 |
| Center volume: | ||||||
| 0–299 | Reference | Reference | Reference | |||
| 300–499 | 1.06 (0.84–1.35) | 0.6 | 0.89 (0.71–1.13) | 0.3 | 0.94 (0.82–1.08) | 0.4 |
| 500–999 | 1.28 (1.02–1.60) | 0.031 | 1.18 (0.96–1.44) | 0.1 | 1.05 (0.92–1.20) | 0.5 |
| Over 1000 | 1.07 (0.81–1.41) | 0.6 | 0.96 (0.72–1.26) | 0.7 | 0.98 (0.82–1.18) | 0.9 |
Outcomes by ECD listing
A greater proportion of ECD-willing registrants eventually received deceased donor transplants when compared with non-ECD willing registrants in all three cohorts (27.0% vs. 21.1% for all registrants, 22.2% vs. 14.4% for ECD-benefit registrants, and 28.2% vs. 16.7% for older registrants, p < 0.001 for all comparisons) (Table 3A). Not surprisingly, more ECD-willing registrants also received ECD transplants. Non-ECD benefit patients (i.e. those less than 40, or without diabetes and a long projected waiting time) had the highest rates of LD or DD transplantation of any group.
Table 3.
(A) Proportion of registrants eventually undergoing transplantation. Deceased donor (DD) includes both SCD and ECD kidneys. (B) Relative risk (shown as Hazard Ratio) of death from time of listing, comparing those who were ECD-willing with those who were not. Note that, since this is an intention-to-treat design, some from each group might have received ECD kidneys, SCD kidneys, LD kidneys, or no kidneys at all.
| (A) | |||
|---|---|---|---|
| Cohort | % LD transplant | % DD transplant | % ECD transplant |
| All registrants | |||
| Non-ECD willing | 8.15 | 21.13 | 0.23 |
| ECD-willing | 5.80 | 27.04 | 9.51 |
| ECD-benefit | |||
| Non-ECD willing | 6.87 | 14.44 | 0.22 |
| ECD-willing | 5.00 | 22.17 | 9.23 |
| Non-ECD-benefit | |||
| Non-ECD willing | 9.36 | 27.45 | 0.24 |
| ECD-willing | 6.95 | 34.10 | 9.91 |
| Age ≥ 65 | |||
| Non-ECD willing | 5.22 | 16.71 | 0.52 |
| ECD-willing | 3.72 | 28.18 | 13.56 |
| (B) | ||
|---|---|---|
| Subgroup | HR (95% CI) for ECD-willing | p-Value |
| ECD-benefit | 0.88 (0.85–0.91) | <0.001 |
| Non-ECD benefit | 1.11 (1.05–1.18) | <0.001 |
| Age ≥ 65 | 0.89 (0.83–0.94) | <0.001 |
Survival differed by listing status. In a multivariate ’intention-to-treat’ analysis, estimating survival from time of listing until death or end-of-study, there was indeed a significant survival benefit for being ECD-willing among ECD-benefit registrants (HR for death from the time of listing, regardless of whether or not the patient was transplanted, was 0.88, 95% CI 0.85–0.91, p < 0.001) (Table 3B). We note that all members of the cohort, including those transplanted with LD, SCD, or ECD organs, remained part of the cohort and part of the analysis until either death or end-of-study. This survival advantage was also true among older registrants (HR 0.89, 95% CI 0.83–0.94, p < 0.001). In contrast, among those who were not predicted to derive benefit from ECD transplantation under Merion’s algorithm, ECD-willing status was associated with a higher risk of death (HR 1.13, 95% CI 1.07–1.20, p < 0.001).
Discussion
Despite convincing evidence that ECD transplantation can impart a long-term survival benefit in carefully selected patients, our study shows that half the patients who are predicted to benefit from an ECD transplant are not listed with UNOS as willing to consider one. Compared with previous reports, our study finds fewer pediatric patients listed as ECD-willing (a group that is generally discouraged from accepting ECD organs); also, longer median waiting time, previously found to have a negative association with ECD-willingness, was simply not significantly associated in our sample (23,24). Although these observations appear as nominal improvements in listing behavior, it is clear that a large proportion of ECD listing decisions are not consistent with published recommendations. In addition, we find that there has been no significant change in ECD-benefit listing practice over time or following the high-profile publication of an evidence-based selection algorithm.
Characteristics associated with ECD listing were fairly consistent across cohorts and with previous studies (16,23). Men, African Americans, older patients and those with Medicare as their primary insurance were more likely to be ECD-willing. Surprisingly, diabetes as a cause of renal failure, a characteristic demonstrated to confer proportional benefit from ECD transplantation in selected patients, was not associated with ECD-willingness in any of the cohorts, contrary to previous studies. Addressing these differences in ECD listing practices may attenuate some of the known inequalities in access to transplantation. For instance, older women typically have lower rates of kidney transplantation than their male counterparts and, where appropriate, increased listing as ECD-willing could attenuate that disparity (26).
Some of the variation in ECD listing practice clearly exists at the level of the transplant center (23,24). We found an enormous range of listing practice patterns (the distribution depicted in Figure 1 nearly resembles a uniform random distribution). As noted previously, perhaps the least logical finding is that those who are listed at centers with long waiting times are no more likely (and perhaps even slightly less likely) to list as ECD-willing. One possible explanation for this might be reverse causation; in other words, centers with more ECD-willing registrants and ECD transplants might have shorter waiting times, and rather than waiting times driving ECD listing practices, it could be that practices drive waiting times. To examine this possibility, we repeated the analysis using median waiting time for SCD organs only; using this metric, it would be unlikely for a center to lower their SCD waiting times just by listing more patients for ECDs, so the possibility of reverse causation is eliminated. This sensitivity analysis did not change the direction of the association in any way, suggesting that our findings are indeed robust. Centers with longer waiting times, which should reasonably be listing more patients for ECD, do not necessarily do so.
Our study provides strong support for the utilization of the Merion algorithm in ECD listing decisions. As in previous studies, we find that patients who listed as ECD-willing were more likely to eventually receive a transplant (23). However, more importantly and not previously shown, we have demonstrated that ECD listing was associated with a survival benefit for registrants predicted to benefit from ECD transplantation. We also found that those registrants not predicted to benefit from ECD transplantation actually had higher rates of death when listed for ECD kidneys. The implications of this are unclear; while one conclusion might be that appropriate ECD listing (i.e. listing ECD-benefit patients for ECD kidneys) is a proxy for better medical care, an alternate inference might be that patients listed for ECD kidneys ’outside of the algorithm’ are sick in ways not captured by the algorithm and, as such, would be more likely to die.
There are several additional limitations to our study. First, we must rely on data reported to UNOS, and we are therefore susceptible to the presence of reporting bias, but only if reporting errors occurred in a systematic manner. Second, we have no information on an individual’s decision to list as ECD-willing. The survival benefit shown by Merion and colleagues was a long-term benefit, defined as 3-year post-transplant, which was determined by statistical significance of relative risk and not effect size. Indeed, ECD transplantation, like all transplantation, is associated with an increased risk of death in the immediate postoperative period; the survival benefit exists only if waiting time for an ECD organ is significantly shorter than that for an SCD organ. An individual’s valuation of the short-term risk versus long-term benefit (and their perceived value of being off of dialysis, even with equal or shorter survival) will inevitably play into their decision whether to list as ECD willing. This is only one of the innumerable unmeasured variables taken into consideration by a provider and transplant candidate when deciding on listing status. In fact, a candidate may not even have been aware of the ECD option. We see only the final choice. However, given that we aim only to observe practice patterns throughout the United States and the clinical outcomes in these patients, the factors that went into the listing decision may be irrelevant. Furthermore, given the wide disparities in practice, with some centers listing all of their ECD-benefit patients for ECD organs and some listing none, it is unlikely that our findings are driven only by subtle latent factors specific to outstanding circumstances.
In conclusion, this study identifies a striking discordance between kidney transplant registrants predicted to benefit from ECD transplantation and those who are willing to consider ECD organ offers. It also identifies large variation in practice patterns by transplant center. Coupled with the fact that those predicted to benefit from ECD transplants experienced a survival benefit from the time of listing for ECD offers and those listed for ECD organs were more likely to be transplanted, these findings indicate the possible need for focused education for patients and their transplant providers about evidence-based ECD listing decisions. Improved ECD listing practice not only may help bridge the gap in access to transplantation, but also may improve registrant survival for subgroups predicted to benefit from, or suffer harm from, ECD transplantation.
Acknowledgments
We report an analysis of data collected by the Organ Procurement and Transplantation Network (OPTN). The OPTN is supported by Health Resources and Services Administration contract 234–2005-370011C. The analyses described here are the responsibility of the authors alone and do not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
References
- 1.Cecka JM, Cohen B, Rosendale J, Smith M. Could more effective use of kidneys recovered from older deceased donors result in more kidney transplants for older patients? Transplantation. 2006;81:966–970. doi: 10.1097/01.tp.0000216284.81604.d4. [DOI] [PubMed] [Google Scholar]
- 2.Merion RM. What can we learn from discarded deceased donor kidneys? Transplantation. 2006;81:973. doi: 10.1097/01.tp.0000216278.51480.d8. [DOI] [PubMed] [Google Scholar]
- 3.Sung RS, Galloway J, Tuttle-Newhall JE, et al. Organ donation and utilization in the United States, 1997–2006. Am J Transplant. 2008;8(4 Pt 2):922–934. doi: 10.1111/j.1600-6143.2008.02171.x. [DOI] [PubMed] [Google Scholar]
- 4.Organ and Transplantation Network: National Data. [Accessed on October 1, 2009]; Available at: http://optn.transplant.hrsa.gov/latestData/rptStrat.asp.
- 5.Sung RS, Christensen LL, Leichtman AB, et al. Determinants of discard of expanded criteria donor kidneys: impact of biopsy and machine perfusion. Am J Transplant. 2008;8:783–792. doi: 10.1111/j.1600-6143.2008.02157.x. [DOI] [PubMed] [Google Scholar]
- 6.Port FK, Bragg-Gresham JL, Metzger RA, et al. Donor characteristics associated with reduced graft survival: an approach to expanding the pool of kidney donors. Transplantation. 2002;74:1281–1286. doi: 10.1097/00007890-200211150-00014. [DOI] [PubMed] [Google Scholar]
- 7.Ojo AO, Hanson JA, Meier-Kriesche H, et al. Survival in recipients of marginal cadaveric donor kidneys compared with other recipients and wait-listed transplant candidates. J Am Soc Nephrol. 2001;12:589–597. doi: 10.1681/ASN.V123589. [DOI] [PubMed] [Google Scholar]
- 8.Rosengard BR, Feng S, Alfrey EJ, et al. Report of the Crystal City meeting to maximize the use of organs recovered from the cadaver donor. Am J Transplant. 2002;2:701–711. doi: 10.1034/j.1600-6143.2002.20804.x. [DOI] [PubMed] [Google Scholar]
- 9.Organ Procurement and Transplantation Network: Policies. [Accessed October 18, 2009]; Available at: http://optn.transplant.hrsa.gov/PoliciesandBylaws2/policies/pdfs/policy_7.pdf.
- 10.Merion RM, Ashby VB, Wolfe RA, et al. Deceased-donor characteristics and the survival benefit of kidney transplantation. JAMA. 2005;294:2726–2733. doi: 10.1001/jama.294.21.2726. [DOI] [PubMed] [Google Scholar]
- 11.Ojo AO. Expanded criteria donors: process and outcomes. Semin Dial. 2005;18:463–468. doi: 10.1111/j.1525-139X.2005.00090.x. [DOI] [PubMed] [Google Scholar]
- 12.Schnitzler MA, Whiting JF, Brennan DC, et al. The expanded criteria donor dilemma in cadaveric renal transplantation. Transplantation. 2003;75:1940–1945. doi: 10.1097/01.TP.0000076381.16276.1B. [DOI] [PubMed] [Google Scholar]
- 13.Danovitch GM, Bunnapradist S. Allocating deceased donor kidneys: maximizing years of life. Am J Kidney Dis. 2007;49:180–182. doi: 10.1053/j.ajkd.2006.11.046. [DOI] [PubMed] [Google Scholar]
- 14.Meier-Kriesche HU, Schold JD, Gaston RS, Wadstrom J, Kaplan B. Kidneys from deceased donors: maximizing the value of a scarce resource. Am J Transplant. 2005;5:1725–1730. doi: 10.1111/j.1600-6143.2005.00923.x. [DOI] [PubMed] [Google Scholar]
- 15.Pascual J, Zamora J, Pirsch JD. A systematic review of kidney transplantation from expanded criteria donors. Am J Kidney Dis. 2008;52:553–586. doi: 10.1053/j.ajkd.2008.06.005. [DOI] [PubMed] [Google Scholar]
- 16.Schold JD, Meier-Kriesche HU. Which renal transplant candidates should accept marginal kidneys in exchange for a shorter waiting time on dialysis? Clin J Am Soc Nephrol. 2006;1:532–538. doi: 10.2215/CJN.01130905. [DOI] [PubMed] [Google Scholar]
- 17.Frei U, Noeldeke J, Machold-Fabrizii V, et al. Prospective age-matching in elderly kidney transplant recipients–a 5-year analysis of the Eurotransplant Senior Program. Am J Transplant. 2008;8:50–57. doi: 10.1111/j.1600-6143.2007.02014.x. [DOI] [PubMed] [Google Scholar]
- 18.Cohen B, Smits JM, Haase B, Persijn G, Vanrenterghem Y, Frei U. Expanding the donor pool to increase renal transplantation. Nephrol Dial Transplant. 2005;20:34–41. doi: 10.1093/ndt/gfh506. [DOI] [PubMed] [Google Scholar]
- 19.Persijn GG, Smits JM, Frei U. Three-year experience with the new Eurotransplant Kidney Allocation System. Nephrol Dial Transplant. 2001;16 Suppl 6:144–146. doi: 10.1093/ndt/16.suppl_6.144. [DOI] [PubMed] [Google Scholar]
- 20.Segev DL. Evaluating options for utility-based kidney allocation. Am J Transplant. 2009;9:1513–1518. doi: 10.1111/j.1600-6143.2009.02667.x. [DOI] [PubMed] [Google Scholar]
- 21.Andreoni KA, Brayman KL, Guidinger MK, Sommers CM, Sung RS. Kidney and pancreas transplantation in the United States, 1996–2005. Am J Transplant. 2007;7(5 Pt 2):1359–1375. doi: 10.1111/j.1600-6143.2006.01781.x. [DOI] [PubMed] [Google Scholar]
- 22.United States Renal Data System. 2009 Annual Data Report. [Accessed October 18, 2009]; Available at: http://www.usrds.org/2009/pdf/V2_07_09.PDF.
- 23.Sung RS, Guidinger MK, Leichtman AB, et al. Impact of the expanded criteria donor allocation system on candidates for and recipients of expanded criteria donor kidneys. Transplantation. 2007;84:1138–1144. doi: 10.1097/01.tp.0000287118.76725.c1. [DOI] [PubMed] [Google Scholar]
- 24.Schold JD, Howard RJ, Scicchitano MJ, Meier-Kriesche HU. The expanded criteria donor policy: an evaluation of program objectives and indirect ramifications. Am J Transplant. 2006;6:1689–1695. doi: 10.1111/j.1600-6143.2006.01390.x. [DOI] [PubMed] [Google Scholar]
- 25.Wolfe RA, Ashby VB, Milford EL, et al. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med. 1999;341:1725–1730. doi: 10.1056/NEJM199912023412303. [DOI] [PubMed] [Google Scholar]
- 26.Segev DL, Kucirka LM, Oberai PC, et al. Age and comorbidities are effect modifiers of gender disparities in renal transplantation. J Am Soc Nephrol. 2009;20:621–628. doi: 10.1681/ASN.2008060591. [DOI] [PMC free article] [PubMed] [Google Scholar]