Objectives: Increasing living-donor kidney transplant procedures via kidney paired donations could help combat organ shortages. We examined whether a higher volume of kidney paired donation transplants would lead to increased center performance.
Materials and Methods: Kidney paired donation, living-donor, and deceased-donor transplant data from 165 centers between 2012 and 2016 were obtained from the UNOS OPTN database. The fixed-effects model was used for panel analysis based on Durbin-Wu-Hausman tests (P < .001). Regression analyses tested asso-ciations between total transplant number and kidney paired donation-to-living donor kidney transplant ratio, incorporating up to 2-year lag terms. Regression analyses also tested associations between number of new wait list registrations and kidney paired donation transplant ratio.
Results: Mean and median number of kidney paired donor transplants equaled 3.59 and 1.2, respectively. Only 5 centers performed > 20 paired donation transplants annually. Mean and median ratios of paired donation transplants were 0.54 and 0.11. Total transplant number was not associated with paired donation transplant ratio of the same year (b= -.425, P = .662) or with that of 2 prior years (P = .830 and P = .629, respectively). Similarly, new wait list registrations were not correlated with paired donation transplant ratio of the same year (P= .501, P = .851) or that of 2 prior years (P = .792 and P = .816, respectively). When transplant centers were divided into those performing ≥ 10 paired donation trans-plants annually (18 transplant center, n = 90) versus those performing < 10 annually (147 transplant center, n = 735), no significant effects were shown (P > .10).
Conclusions: Kidney paired donation transplant does not appear to affect transplant center performance. This may be due to the small volume of these transplants currently performed by centers, thereby limiting overall growth in the number of transplants and new registrations.
Key words : Incompatible donors, Living-donor kidney transplantation, Paired exchange
Kidney transplantation is associated with better long-term survival rates than prolonged dialysis for patients with end-stage renal disease (ESRD).1 However, an inadequate supply of available deceased donor organs limits the number of transplants performed.2 To use as many available organs as possible, centers have implemented methods such as extended criteria donation, donation after cardiac death, and ABO-incompatible donation. However, these donated organs are often suboptimal and can carry a significant increase in costs, as well as result in worse long-term outcomes and early graft loss.3-5
For patients with ESRD who wait longer for an organ due to shortage of kidneys, long-term outcomes are poor.6 One method to combat this is to increase the number of living-donor kidney transplants (LDKT). Living-donor kidney transplant offers superior outcomes versus dialysis and deceased-donor kidney transplants (DDKT) and offers greater quality of life.7,8
One approach to increase living-donor trans-plants is through kidney paired donation (KPD).9 Kidney paired donation transplant pairs a living donor who is incompatible with the intended recipient but is compatible with another recipient whose donor is compatible with the first recipient. This process can be expanded to create chains of exchanges.10 It has been estimated that, of the nearly 6000 patients with willing but incompatible donors, participation in KPD could result in transplant in about half of them.11,12 Unlike extended criteria donations, donations after cardiac death, and ABO-incompatible donations, KPD is an excellent option to expand kidney transplant without the consequences stated above.
However, in the past decade, there has been a decrease in living kidney donations.8 Several reasons for this have been proposed. Rodrigue and associates suggested that changes in donor selection criteria, an increasing average age of the kidney transplant population, and the overall health status of the general population may contribute to fewer living donors and LDKT.8 Furthermore, the Centers for Medicare and Medicaid Services (CMS) recently published conditions that could potentially cost transplant centers their funding if their performance rates declined, which can be a disincentive for innovative programs like KPD that would increase the number of living donors.13 Despite these challenges, it is important for the long-term outcomes of ESRD patients that the transplant community increases the number of kidney transplants, specifically by increasing the volume of LDKT.
In this study, we aimed to investigate the relationship between KPD participation and LDKT transplant center performance. However, it was unclear how to appropriately gauge this. Traditionally, LDKT performance is calculated as a percentage of the total number of kidney transplants per center. However, this total includes DDKTs, which have no effect on the number of LDKTs performed and are therefore not relevant to the calculation. A second method proposed by Matar and associates suggested measuring LDKT rate “as a percentage of the number of new waitlist registrants.”14
The accurate measurement of the LDKT rate is crucial to the conditions of the CMS that are required for continued funding. The goal of a transplant program is to be efficient and productive, while adhering to CMS standards. Our objective was to determine the productivity of centers and whether they were underperforming or overperforming according to the LDKT rate. Specifically, we analyzed whether the rate of KPD participation affected the overall performance of the center as assessed by both the traditional method and the method described in Matar and colleagues.14 We aimed to assess whether an increase in KPD participation could increase the overall number of LDKTs as well as the efficiency of each transplant center.
Materials and Methods
Center-level data on the total numbers of LDKT, DDKT, and KPD transplant procedures performed between 2012 and 2016 were obtained from the United Network for Organ Sharing Organ Procurement and Transplantation Network (UNOS OPTN) database. Donor and recipient criteria and details on how patients were assessed were not included in the database but were not needed for the analysis. The records included information on 165 centers that performed at least one LDKT during the 5-year study period. All work was done in accordance with the Strengthening the Reporting of Cohort Studies in Surgery criteria.15
We used the fixed-effects model versus the random-effects model as the Durbin-Wu-Hausman tests indicated a bias due to center-level heterogeneity (P < .001). Regression analyses examined the asso-ciation between the total number of transplants and the KPD ratio (the total number of KPDs/the total number of LDKTs per year per transplant center). The regression model incorporated up to 2-year lag terms to examine whether an increase in the KPD ratio had any positive impact on the total number of transplants in 2 subsequent years. The association between the number of new registrants on the wait list and the KPD ratio was also examined in a similar fashion, specifically to find whether a higher relative volume of KPD transplants promoted registrations to wait lists at the center level.
The mean and median numbers of KPD transplants were 3.59 and 1.2, respectively. The distribution was fairly skewed, with only 5 transplant centers performing, on average, over 20 KPD transplants per year (standard deviation = 5.88, kurtosis = 16.61). The mean and the median KPD ratios were 0.54 and 0.11. These results are summarized in Table 1. The total number of transplants was not associated with the KPD ratio of the same year (b= -.425, P = .662) or that of 2 prior years (P = .830 and P = .629 for 1-year and 2-year lags, respectively; Table 2). The number of new wait list registrations was also not correlated with the KPD ratio of the same year (b= 0.501, P = .851; Table 3) or with that of 2 prior years (P = .792 and P = .816 for 1-year and 2-year lags, respectively).
When we examined the distribution of average KPDs per year per center, we detected a clear power law distribution (that is, most centers performed an extremely low number of KPDs per year, with only a small number of centers conducting large numbers of KPDs). Furthermore, the median number of KPDs did not show a significant increase over the study period. The distributions of average number of KPDs per year and the box plots showing the total number of KPDs by year are presented in Figure 1 and Figure 2.
To control for effects by year, fixed effects of year were explored. We observed a significant increase in the total number of transplants in 2015 and 2016 (b = 4.562, P = .011; b = 9.793, P < .001, respectively) compared with the reference year (2012). However, the regression results essentially remained the same after controlling for these fixed effects, leaving the coefficient of the KPD ratio insignificant (P = .575).
In contrast to the total number of transplants, the number of new wait list registrants did not vary significantly over time. All year fixed effects were statistically insignificant, with P values ranging from 0.103 in 2013 to 0.577 in 2016 (reference year = 2012).
To examine the potential size effect, we per-formed additional regression analyses by dividing the trans-plant centers into 2 groups: those performing 10 or more KPDs annually (18 transplant centers, n = 90) and those performing fewer than 10 KPDs annually (147 transplant centers, n = 735). All of the afore-mentioned conclusions remained the same, with no significant effects of KPD rate on the total number of kidney transplants of that and subsequent years (P > .10).
The theoretical potential for KPD to improve LDKT rates and therefore center performance is significant; however, the relationship between these 2 variables has been unclear. Traditional methods of calculating LDKT rates as a percentage of the total number of transplants occurring annually can be misleading.14 Centers that perform a significant number of LDKT transplants but also add a large volume of patients to their wait list each year are at risk of being overrated. In contrast, centers that perform a smaller number of LDKT transplants but are adding a smaller volume of patients to their wait list may be underrated. In an effort to acknowledge this discrepancy, we evaluated the effects of KPD participation on center performance by both the traditional calculation and the method proposed by Matar and colleagues.14
Our study of 165 transplant centers did not show a significant association between KPD participation and center performance by either method. We found no association between the total number of kidney transplants performed and the KPD ratio. Likewise, there was no association between the number of new registrants on the wait list and the KPD ratio. Additional analysis to determine whether centers performing more than 10 KPDs annually showed a significant association with performance also produced the same results by both calculations.
We propose a few reasons as to why this was observed. First, it could largely be due to the fact that the median number of KPD transplants performed by single centers per year was fewer than 2, which may not be an adequate number to show significant associations. To this end, Massie and colleagues16 demonstrated that more than 1000 additional kidney transplants would occur per year if all centers in the United States performed at least 2 KPD transplants per 100 LDKT-eligible patients. They further examined centers with the highest KPD rates (3.66 KPD per 100 LDKT-eligible) and determined that, if all centers were to participate in KPD at these rates, the number of KPD transplants would increase per year by a factor of 3.2.16
In a single-center study of one of the highest performing KPD programs in the world, Kute and associates observed that LDKT increased by 25% in the year of 2015 as a result of KPD.17 However, this center performed 77 KPD transplants in 1 year. The centers we designated as “high performing” in our final analysis still only completed more than 10 KPDs annually. It may therefore be necessary to increase the volume of KPD transplant procedures before beneficial effects on performance and LDKT become apparent.
The problem of the low number of KPD trans-plants despite its potential to increase LDKT is not new. Several efforts have been made to address this issue.18-20 Analyses of current matching algorithms yielded improvements that optimize pairing in KPD databases.12,21 Other strategies, such as increasing compatible pair participation in KPD pools, have also been proposed.22 As these strategies are imple-mented and further developed, the effects of KPD on LDKT may be able to be better assessed.
Second, due to the overall decrease in living donors in the past 10 years, it is possible that KPD actually stabilized the rate of LDKT. If fewer traditional LDKTs occurred because of a lack of living donors, the LDKT rates and center performance would have been expected to drop. Even at a low volume, KPD may provide enough LDKTs that would not have happened through the traditional system to maintain these rates. If true, continued efforts to sustain or even increase center KPD participation would be strongly indicated.
This study was the first to analyze the impact of KPD on performance by both the traditional method and that proposed by Matar and colleagues.14 Per our assessment, neither showed a significant impact on center performance by KPD. Despite this, we believe that efforts among transplant centers to increase KPD are necessary, as KPD provides a means to increase organ donation without compromise in outcomes. We recommend the creation of one uniform, nationwide, or even worldwide service that would encompass all living donors and incompatible recipients. This would increase chances of finding pairs. Additionally, we suggest that participation in KPD should be a quality initiative that all centers are compelled to comply with in order to be accredited. With a national or international database and an increased commitment of centers to perform KPDs, we believe the effects could be highly significant.
Finally, a true effort to increase organ donation must involve methods to improve not just KPD participation but organ donation among donors of all types. One approach to this is aggressive edu-cation in relevant populations about donor types. Understanding why certain groups of people have lower rates of donation and addressing those issues is also critical to reducing the shortage of organs. Additionally, other models such as presumed consent and donor vehicles should be explored and considered as viable means to increase donation.
Limitations include the previously discussed low volume of KPD transplants, even at high performing centers, and the inability to precisely determine the impact of KPD on the already declining LDKT rate. Further studies are needed to examine these effects and to more accurately describe the role of KPD in transplant center performance. The database used also did not provide detailed information about donors and recipients or how patients were assessed. However, this information is largely unnecessary to answer the question of whether more KPD transplants would result in more LDKTs.
Kidney paired donation is a viable and promising strategy to increase kidney transplant in the United States. We believe that its beneficial effects on LDKT will be better observed as the number of KPD transplants continues to increase. Current objectives for transplant centers should include thoughtful strategies for increasing the volume of transplants through KPD, which will likely clarify the ability of KPD to affect center performance.
DOI : 10.6002/ect.2018.0266
From the 1Department of Surgery, University of Toledo Medical Center, Toledo,
Ohio; and the 2School of Policy and Government, George Mason University,
Fairfax, Virginia, USA
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare.
Corresponding author: Alexandra J. Wenig, Department of Surgery, University of Toledo Medical Center, 1505 Bernath Parkway Apt B, Toledo, OH 43615, USA
Phone: +1 937 638 4666
Table 1. Summary Statistics for Types of Transplants and Number of New Wait List Registrations Based on the Median Number of Each Type of Donation Per Center Per Year
Table 2. Association Between Total Number of Transplants and Kidney Paired Donation Ratio of the Same Year
Table 3. Relationship Between Number of New Wait List Registrants and Kidney Paired Donation Ratio
Figure 1. Average Kidney Paired Donations Transplants per Transplant Center Per Year
Figure 2. Total Number of Transplants per Year