The current deceased donor kidney allocation policy has not changed substantially in >20 years.1 During this time, the gap between supply and demand has widened. Waiting time has become the dominant factor in allocation, and less emphasis has been placed on biologic criteria such as the degree of immune system sensitization or HLA matching. For minority candidates, such as African Americans, who have difficulty gaining access to the waiting list, delay in listing contributes to racial disparities in access to transplant.2–4 The current allocation system favors candidates who have waited the longest, but does not improve outcomes after transplant and discourages use of kidneys with a potentially shorter duration of functioning. These shortcomings have created inequities. Waiting times for blood type B candidates are much longer than waiting times for blood type A candidates.5,6 Kidneys with the potential to function longer may be allocated to candidates with shorter life expectancy; conversely, candidates with a longer estimated life span may be allocated kidneys with limited duration of functioning. These types of transplants result in high retransplant rates and increase the demand for donor kidneys. A new allocation policy was approved by the Organ Procurement and Transplantation Network (OPTN) on June 24, 2013.
The new allocation policy risk-stratifies deceased donors using the kidney donor profile index (KDPI).7,8 The KDPI takes into account donor age, height, weight, ethnicity, history of hypertension and diabetes, cause of death, serum creatinine level, hepatitis C status, and donation after circulatory death status (Table 1). Lower KDPI kidneys are associated with better post-transplant survival. Similarly, transplant candidates on the waiting list are risk-stratified based on estimated post-transplant survival (EPTS), which takes into account candidate age, dialysis duration, prior solid organ transplant, and diabetes status. Generally, older age, longer dialysis duration, prior solid organ transplant, and presence of diabetes are associated with higher EPTS scores and shorter expected post-transplant survival. The new allocation policy prioritizes candidates in the top 20th EPTS percentile to receive kidneys from donors with a KDPI≤20% (so-called “longevity matching”). It also prioritizes candidates with a calculated panel-reactive antibody (CPRA)≥98% and provides broader sharing for candidates with a CPRA≥99% (Table 1). The new policy maintains the current restriction in which kidneys from blood type B and O donors are allocated strictly to candidates with identical blood types (except for the zero-HLA mismatch category; Table 1).
As in the current system, points will be used to rank candidates in each category listed in Table 1, with more points leading to higher priority for receiving a kidney offer. One point will be awarded for each year spent waiting once the eGFR measurement is <20 ml/min per 1.73 m2 (Table 2). However, under the new allocation policy, candidates receiving dialysis at the time of listing will also receive waiting time credit from the first day of maintenance dialysis. These waiting time points will be awarded based on fractional years, by dividing the number of days waiting by 365 (Table 2). A new priority point scale of 0–202 will be awarded based on the CPRA (Tables 2 and 3). This scale was based inversely on the probability of receiving an organ offer. Other points will be awarded as in the current allocation policy (Table 2).1 Other features of the current and new kidney allocation policies are compared in Table 4.
In this study, we describe the final simulation models that were used to estimate how the current and new allocation systems allocate deceased donor kidneys. The results of these simulations were used to propose the new allocation system.
Candidates on the waiting list who are in the national top 20th percentile for EPTS constitute 15%–25% of all candidates on the waiting list for a donation service area (Figure 1). The percentage of kidney donors with KDPI in the top 20th percentile varied between 15% and 25% for most donation service areas (Figure 2). However, the candidate list is much larger than the number of donors available, with 122,669 candidates nationally and enough donors to allow <12,000 deceased donor kidney transplants annually (Table 5). The simulations resulted in 11,531 primary kidney and kidney-pancreas transplants (range, 11,463–11,586 across 10 iterations) under the current allocation policy and 11,599 (range, 11,538–11,681) under the new allocation policy. The 122,669 candidates represent the total number of kidney and kidney-pancreas candidates on the waiting list on the first day or added during the year, in simulations of current and new policy (Table 5).
The characteristics of recipients and donors were compared for the following: (1) actual transplants in 2010, (2) 10 simulations using current allocation policy, and (3) 10 simulations using the new allocation policy (Table 6). The characteristics of actual transplant recipients were similar to the characteristics of recipients under simulations of the current allocation policy. The new allocation policy resulted in more transplants for candidates aged 18–49 years and fewer transplants for candidates aged ≥50 years. As expected, the new allocation policy also resulted in more transplants for candidates with blood type B and fewer transplants for candidates with blood type A compared with current allocation policy. Because of the change in prioritization points, the number of transplants in candidates with a CPRA>20% increased with the new allocation policy compared with the current policy (Figure 3). Median life spans and graft-year survival were longer for simulated transplant recipients under the new policy compared with recipients under the current policy (Table 5). The new policy results in transplants with more median extra life-years for recipients versus wait-list candidates compared with current policy, without increasing mortality on the waiting list (Table 5).
The new allocation policy uses KDPI and EPTS to rank order kidneys and candidates, respectively. We found that the number of candidates in the top 20th percentile for survival constitutes 15%–25% of all candidates on the waiting list, depending on the donation service area. Similarly, the percentage of kidney donors with a KDPI<20% varied between 15% and 25% for most donation service areas. Because candidates far outnumber donors, not all candidates in the top 20th percentile for survival will undergo transplant within 1 year. The new allocation policy results in increases in average projected median lifespan after transplantation and increases in time with a functional allograft. The distribution of transplants did not change substantially by candidate race or HLA mismatches. The number of kidneys that were shared beyond the local area increased only modestly due to broader sharing for candidates with a CPRA≥99% and regional sharing of kidneys from donors with a KDPI>85%. The new allocation policy results in more transplants for highly sensitized candidates, particularly those with a CPRA≥99%, and increases the number of transplants for candidates with blood type B.
The new allocation policy is projected to result in improved patient and allograft survival (Table 5). The simulations predict an average 7.0% increase in median patient life-years per transplant and an average 2.8% increase in median allograft years of life under the new allocation policy compared with the current policy (Table 5). Assuming 11,000 transplants, this could lead to a gain of 9130 life-years of patient survival and 2750 years of allograft survival. This increase is likely due to a greater number of younger candidates (aged 18–49 years) undergoing transplant than older candidates (aged≥50 years), and to better matching of patient and graft expected longevity. Adolescent and young adult age is a risk factor for poor adherence to immunosuppressive medications,9 and this increases the theoretical risk of overall reduced allograft years of life under the new policy. However, the simulations also predict improved graft survival.
Despite transplants in some higher-risk candidates, the simulations show that the net benefit of the new policy is to improve patient and allograft survival (Table 5). Giving priority to high-CPRA candidates could result in worse outcomes, because high CPRA is a risk factor for rejections and poor outcomes. However, finding immunologically compatible donors will remain challenging for candidates with a CPRA>80% (Figure 3).
The new allocation policy maintains many features of the current policy, including the local, regional, and national categories, and organ offers would be made first to all local candidates before regional candidates in the top 20th percentile of survival. Likewise, organ offers would be made to all regional candidates before national candidates in the top 20th percentile of survival. Pediatric candidates in general maintain the same priority over adult candidates as in current policy, but they are prioritized to receive local offers from donors with a KDPI<35% instead of from donors aged<35 years. The candidates with offers of zero HLA-A, HLA-B, and HLA-DR mismatched organs are prioritized over candidates with one or more HLA mismatches at the A, B, and DR loci. This priority for zero HLA mismatched organs could continue to incentivize candidates to list even before eGFR falls to <20 ml/min per 1.73 m2, despite the waiting time being back-dated to dialysis initiation for candidates who are listed after starting dialysis. The new policy does not change current policy allowing candidates to list even before eGFR falls to <20 ml/min per 1.73 m2, and candidates not on dialysis can start accruing waiting time as soon as eGFR falls to <20 ml/min per 1.73 m2. Currently, only three donation service areas, namely One Legacy in California, Iowa Donor Network, and Gift of Life Michigan, have an approved variance to policy allowing calculation of waiting time from the start of dialysis, even if this occurred before listing.
Despite these simulations, the potential effect of the new policy remains to be determined. For example, the simulations project a slight increase in the number of transplants, from 11,531 to 11,599 (Table 5). It is unclear whether this increase will be borne out in reality, or whether it results from the organ acceptance criteria in the kidney-pancreas simulated allocation model (KPSAM) being based on current organ acceptance behavior. The new allocation policy would also remove the kidney allocation variances (deviations in organ allocation policy approved by the OPTN) that are currently in place. The KPSAM does not take these variances into account when it simulates the current allocation policy. Once the policy is enacted, the OPTN Kidney Committee will evaluate its intended and unintended consequences.
Our study has several limitations. The KPSAM cannot account for changes in organ acceptance behaviors.10 Therefore, if the new policy results in dramatic changes in organ acceptance behavior, the estimates of number of transplants from the simulations will differ from reality. However, given the limited supply of kidney allografts and the large number of candidates on the waiting list, the overall number of transplants is unlikely to decrease. The KPSAM simulates transplants, discards, and removals and deaths on the waiting list for a 1-year period.10 Therefore, it is not possible to determine whether the new allocation policy could lead to changes in numbers of transplants in subsequent years of policy implementation. The KPSAM was not designed to predict outcomes at the level of a center or a donation service area because it assumes similar organ acceptance behavior across the country. The majority of the data shown in this study were used by the OPTN Kidney Transplantation Committee to submit the policy for public comment. Based on feedback from the public and other OPTN committees and regions, the new allocation policy underwent minor changes that were further incorporated. These include the following: (1) zero HLA mismatched candidates will receive priority in the categories of candidates with a CPRA≥98%; (2) pediatric candidates will be excluded from EPTS, thereby allowing more adults (approximately 700, data not shown) to be in the top 20th percentile of survival; (3) pediatric candidates will be offered kidneys with a KDPI>85% only for zero HLA mismatched kidneys; (4) EPTS scores will be updated daily or whenever EPTS factors change; and (5) each transplant center will be required to establish center-specific criteria in order to receive offers for blood type A2 and A2/B kidneys for candidates with blood type B. These changes were not accounted for by the simulations, but would affect only a small number of candidates and kidneys, thereby not invalidating the findings of this study.
Development of the new kidney allocation policy represents a 9-year effort by the OPTN committee. The simulations predict that the new policy could potentially improve kidney transplant outcomes. After implementation of the new policy, expected at the end of 2014, OPTN and the Scientific Registry of Transplant Recipients (SRTR) will evaluate the policy’s intended and unintended consequences on an ongoing basis.
SRTR data were used. The SRTR data system includes data on all donors, wait-listed candidates, and transplant recipients in the United States, submitted by the members of OPTN, and has been described elsewhere.11 The Health Resources and Services Administration of the US Department of Health and Human Services provides oversight of the activities of the OPTN and SRTR contractors. All kidney transplant candidates on the kidney, kidney-pancreas, and pancreas waiting lists from January 1, 2010, to December 31, 2010, and any kidney or pancreas donors whose organs were offered for transplant during this period were included. The EPTS was calculated for each candidate at listing or using the candidate’s data on January 1, 2010, whichever was later. EPTS thresholds for determining whether a candidate is in the top 20th percentile of survival were based on the national EPTS distribution from candidates for each blood type on the waiting list between January 1, 2007, and December 31, 2009. The KDPI for each kidney allograft was calculated using the distribution of kidney donor risk index of donors between January 1, 2007, and December 31, 2009, as the reference donors. We assumed that 20% of blood type A or AB white, African American, or Hispanic donors had a blood type A2 or A2B, respectively. We also assumed that 70% of the candidates with blood type B had low anti-A antibody titers and thus could accept kidneys from donors with an A2 or A2B blood type.
This study conducted simulations using the KPSAM, which is a program that has been used routinely by the OPTN committees to assess policy proposals.10 The KPSAM simulates the arrival of donated organs and new candidates on the waiting list over a 1-year period, checks compatibility of organs with candidates on the waiting list at the time an organ becomes available, creates ordered lists of compatible candidates (candidates with more points have priority for receiving the organ over candidates with fewer points in each ordered list), simulates candidate acceptance or refusal of organ offers using a logistic regression model based on organ acceptance behavior in 2010, calculates the number of transplants and number of organs discarded, and uses linear approximations to Cox proportional hazard models12 to project outcomes such as median allograft and patient survival for each transplant. Allograft failure was defined as the need for dialysis or retransplant. The KPSAM repeated this process 10 times each for the current and new allocation policies (Tables 1–4), each time randomly permuting the order of donor arrivals and generating new random numbers to determine organ offer acceptance. Because the same donors and candidates are used in each of the simulations, and they are the actual donors and candidates from calendar year 2010 and not independent samples, statistical tests of comparisons are not possible. Instead, the average and the minimum to maximum range of results for the 10 iterations are described for the current and the new allocation policies. Of note, this range reflects variability of the simulation modeling, not variability in actual organ allocations.
The authors thank SRTR colleagues Delaney Berrini for manuscript preparation, and Nan Booth, MPH, ELS, for manuscript editing.
This work was conducted under the auspices of the Minneapolis Medical Research Foundation, contractor for the SRTR, as a deliverable under US Department of Health and Human Services, Health Resources and Services Administration, Healthcare Systems Bureau, Division of Transplantation Contract HHSH250201000018C. As a US Government–sponsored work, there are no restrictions on its use. The views expressed herein are those of the authors and not necessarily those of the US Government.
Published online ahead of print. Publication date available at www.jasn.org.
See related editorial, “Simulating the New Kidney Allocation Policy in the United States: Modest Gains and Many Unknowns,” on pages 1617–1619.
1. Smith JM, Biggins SW, Haselby DG, Kim WR, Wedd J, Lamb K, Thompson B, Segev DL, Gustafson S, Kandaswamy R, Stock PG, Matas AJ, Samana CJ, Sleeman EF, Stewart D, Harper A, Edwards E, Snyder JJ, Kasiske BL, Israni AK: Kidney, pancreas and liver allocation and distribution in the United States. Am J Transplant 12: 3191–3212, 2012
2. Kasiske BL, London W, Ellison MD: Race and socioeconomic factors influencing early placement on the kidney transplant waiting list. J Am Soc Nephrol 9: 2142–2147, 1998
3. Kasiske BL, Snyder JJ, Matas AJ, Ellison MD, Gill JS, Kausz AT: Preemptive kidney transplantation: The advantage and the advantaged. J Am Soc Nephrol 13: 1358–1364, 2002
4. Alexander GC, Sehgal AR: Barriers to cadaveric renal transplantation among blacks, women, and the poor. JAMA 280: 1148–1152, 1998
5. Matas AJ, Smith JM, Skeans MA, Lamb KE, Gustafson SK, Samana CJ, Stewart DE, Snyder JJ, Israni AK, Kasiske BL: OPTN/SRTR 2011 annual data report: kidney. Am J Transplant 13(s1): 11–46, 2013
6. Hurst FP, Sajjad I, Elster EA, Falta EM, Patel P, Abbott KC, Agodoa LY, Jindal RM: Transplantation of A2 kidneys into B and O recipients leads to reduction in waiting time: USRDS experience. Transplantation 89: 1396–1402, 2010
7. Rao PS, Schaubel DE, Guidinger MK, Andreoni KA, Wolfe RA, Merion RM, Port FK, Sung RS: A comprehensive risk quantification score for deceased donor kidneys: The kidney donor risk index. Transplantation 88: 231–236, 2009
8. Organ Procurement and Transplantation Network (OPTN): Kidney Donor Profile Index Calculator. Available at: http://optn.transplant.hrsa.gov/resources/allocationcalculators.asp?index=81
. 2013. Accessed December 19, 2013
9. Shaw RJ, Palmer L, Blasey C, Sarwal M: A typology of non-adherence in pediatric renal transplant recipients. Pediatr Transplant 7: 489–493, 2003
10. Scientific Registry of Transplant Recipients: Kidney-Pancreas Simulation Allocation Model (KPSAM) User Guide, Version 5. Available at: http://www.srtr.org/sam/KPSAM.pdf
. 2012. Accessed December 19, 2013
11. Leppke S, Leighton T, Zaun D, Chen SC, Skeans M, Israni AK, Snyder JJ, Kasiske BL: Scientific Registry of Transplant Recipients: Collecting, analyzing, and reporting data on transplantation in the United States. Transplant Rev (Orlando) 27: 50–56, 2013
12. Wolfe RA, McCullough KP, Schaubel DE, Kalbfleisch JD, Murray S, Stegall MD, Leichtman AB: Calculating life years from transplant (LYFT): Methods for kidney and kidney-pancreas candidates. Am J Transplant 8: 997–1011, 2008