Donor age is a significant risk factor for graft loss after kidney transplantation. We investigated the question whether significant graft years were being lost through transplantation of younger donor kidneys into older recipients with potentially shorter lifespans than the organs they receive. We examined patient and graft survival for deceased donor kidney transplants performed in the United States between the years 1990 and 2002 by Kaplan-Meier plots. We categorized the distribution of deceased donor kidneys by donor and recipient age. Subsequently, we calculated the actual and projected graft survival of transplanted kidneys from younger donors with the patient survival of transplant recipients of varying ages. Over the study period, 16.4% (9250) transplants from donors aged 15–50 were transplanted to recipients over the age of 60. At the same time, 73.6% of donors above the age of 50 were allocated to recipients under the age of 60. The graft survival of grafts from younger donors significantly exceeded the patient survival of recipients over the age of 60. The overall projected improvement in graft survival, by excluding transplantation of younger kidneys to older recipients, was approximately 3 years per transplant. Avoiding the allocation of young donor kidneys to elderly recipients, could have significantly increased the overall graft life, by a total 27 500 graft years, between 1990 and 2002, with projected cost savings of about 1.5 billion dollars.
Kidney transplantation is now generally accepted as the optimal treatment for end stage renal disease (ESRD, Stage V chronic kidney disease). Recent work, pioneered by Wolfe and colleagues at the University of Michigan, indicates that the life expectancy of ESRD patients undergoing transplantation with deceased donor kidneys is almost twice as long as comparable patients remaining on chronic dialysis (1). Moreover, this survival benefit is evident across all age groups, even for the oldest transplant recipients (2,3). Subsequent studies have confirmed this robust finding, documenting significant survival benefit in obese patients, diabetics, African Americans, and those transplanted after prolonged periods of time on dialysis (4,5). The net impact of these data (along with improved outcomes, reduced morbidity and lifestyle advantages that have always been associated with renal transplantation), has been an unprecedented growth in the demand for transplantable kidneys (6). Despite increasing reliance on live donors, the waiting list for kidneys for deceased donors continues to grow, with more than 65 000 Americans now awaiting a suitable organ. With only approximately 10 000 deceased donor kidneys available each year, the shortage of organs is now the principal limiting factor to timely transplantation (7).
The best solution to the increasing disparity between suitable candidates and available organs is to increase the number (supply) of transplantable kidneys. Indeed, programmatic efforts such as HHS Secretary Tommy G. Thompson's Organ Donation Initiative, the United Network for Organ Sharing (UNOS) Expanded Criteria Donor (ECD) policy (8) and increased utilization of kidneys from donors after cardiac death (9) have resulted in recent incremental gains in organ availability. However, as noted in a recent analysis, if all potential donors in the United States became actual donors, substantial unmet needs would persist (10).
Alternatively, very little attention has been devoted to reducing demand for transplantation. Given the survival benefit documented across almost all demographic subgroups in the analyses noted above (11–14), UNOS allocation policies must remain consistent with the sometimes competing objectives of medical efficacy and justice. In the past, proposals to increase efficiency by allocating every kidney to the closest human leukocyte antigen (HLA)-matched candidate could not be implemented due to adverse impact on the justice component, particularly for minority candidates (15). In recent years, better immunosuppression has reduced the impact of HLA matching on graft survival. Now, nonimmunologic variables predictive of allograft quality have become recognized as more important predictors of outcome.
Perhaps, the most important determinants of allograft survival in the current era are donor and recipient age. As recognized in the policies defining the ECD program, kidneys from older deceased donors may offer more limited longevity than those from younger donors. Additionally, regardless of donor source, older recipients gain substantially fewer additional life years from transplantation than younger recipients, and are more likely to succumb to other medical illnesses despite a functioning allograft. Among younger recipients, a failing allograft is now the third most common reason for placing patients on the waiting list, and death with a functioning graft is relatively uncommon (16). Current policies allocating deceased donor kidneys in adults apply waiting time and HLA matching equally across all age groups. An unintended consequence of this policy is that the very youngest deceased donor organs, with the longest potential survival, can be transplanted into the oldest recipients with the shortest survival. Thus, predictable patient deaths with functioning allografts in older recipients might be said to reduce the overall potential life-years obtainable from the current allograft supply, and transplanting older kidneys into younger recipients might predispose these patients to graft failure, return to dialysis and relisting. The aim of our analysis was to quantify the impact a policy of matching the expected survival of the allograft to the expected survival of the transplant candidate might have on graft survival as a new approach to maximize utility of the donated kidneys.
We analyzed the Scientific Registry of Transplant Recipient (SRTR) database of 74 998 primary solitary deceased donor kidney transplants performed between 1990 and 2002, in order to describe the allocation of kidney grafts, as a function of donor and recipient age. Recipient age was categorized into the following groupings: 0–17, 18–29, 30–39, 40–49, 50–59, 60–64, 65–69 and 70+. We examined the frequency of kidney transplants in which younger organs were allocated to old recipients and older organs to young recipients. For the purpose of this analysis, we created a binary variable of younger (ages 15–50) and older donors (>50). The reasons for our selection of this particular cutoff point were twofold: we observed a more rapid increase in the risk for graft loss coinciding with this donor age cutoff (as evidenced in the expanded criteria donor designation). In addition, we noted that there was a sufficient volume of organs from younger donors transplanted to older recipients below this cutoff, allowing for a potentially equal exchange of these donations for older organs, such that overall waiting time for a transplant would not be affected.
We analyzed univariate patient and graft survival with Kaplan-Meier product-limit models. The overall patient survival rates were analyzed for the categorized recipient age groups and separately as a function of donor age strata. To estimate the graft survival potential of kidneys from young deceased donors, we examined death censored graft survival rates of these grafts in younger recipients (aged 20–40). We compared this potential graft survival, with patient survival in older recipients (above age 60) who had received younger donations. In order to quantify the differences in the survival, we estimated the average graft and patient years over the study period by calculating the area under the respective survival curves, and assessed the difference in the life expectancy of older recipients to the potential life expectancy of a young kidney. In order to assess the total impact of this difference over the entire lifetime of the graft, we projected graft and patient survival beyond the study period by extending the survival curves until they reached 0%. For the projection of long-term graft and patient survival, we utilized data beyond 2 years post-transplant until last follow-up available. This was done in order to base the projections on a relatively stable long-term attrition rate. We used the same projections to estimate patient survival differences, which would result from a hypothetical reallocation of deceased donor kidneys, so that younger patients would preferentially get kidneys from younger donors.
Over the study period of 1990–2002, the distribution of deceased donor organs to the recipient groups by donor age groups are displayed in Figures 1 and 2. Historically, for donors aged 15–50, approximately 16.4% of transplants were allocated to recipients over the age of 60 (9250 cases). Conversely, over the same period, 73.6% of kidneys from deceased donors over the age of 50 have been allocated to recipients under the age of 60 (134 34 cases). Considering these numbers, we assumed that 9250 total cases would be available for a potential exchange, limiting younger donations to younger recipients, while not affecting the overall likelihood of transplantation.
Figure 3 displays patient survival by recipient age category for those patients receiving a young deceased donor organ (donor age 15–50). There was a strong association between patient age and survival, and more importantly there was a notable distinction between the age groups chosen and their relative prognosis for survival, even with younger donors, which are typically low-risk, are included in the analysis. In addition, Figure 3 shows the death censored graft survival for recipients between the ages of 20 and 40 who received a younger donation (15–50), depicting the full survival potential of a deceased donor kidney from a young donor. Recipients over the age of 60 had progressively inferior survival expectancy, as compared to the potential survival of a graft from a young deceased donor.
The average number of graft years for young grafts (aged 15–50) transplanted into young recipients below age 60, as calculated from the area under the 10-year survival curve was 7.4 years (potential graft survival). In comparison, the average patient survival for patients 60–64 over the same time span was 7.0 years. Average patient survival for 65–69 year old recipients was 6.4 years, and for recipients 70+, 5.9 years. Accordingly, the differences in graft years that, on average, were not utilized by transplant recipients with inferior patient survival were 0.5 years (for recipients 60–64), 1.0 years (for recipients 65–69) and 1.5 years (recipients 70+). The graphical model of this relationship for the oldest recipient cohort is displayed in Figure 4. By calculating the number of cases in which these transplants actually occurred, we also estimated the total number of graft years theoretically lost for each transplant performed. For the 60–64, 65–69 and 70+ recipient age groups, respectively, there were 2009, 3003 and 1838 total graft years lost to patient death with functioning graft on average within the 10-year follow-up relative to the potential graft survival were the organs allocated to younger recipients. To mirror these results, yearly death with functioning graft rates were 1.2, 2.0, 2.9, 4.4, 5.7, 6.6 and 7.5% in 18–29, 30–39, 40–49, 50–59, 60–64, 65–69 and over 70-year old recipients, respectively.
To account for results beyond the follow-up period of available data, we also repeated calculations based on projected patient and graft survival over the respective lifetimes of the cohort. By aggregating all the recipients aged 60 and over, the projected patient survival of the entire older cohort with a relatively ideal graft from a younger deceased donor, was 15.5 years following transplantation. On the other hand, the projected potential graft survival of a graft from a young donor in ideal situations (in younger recipients) was projected to be 21.9 years (displayed on Figure 5). The average difference between potential graft survival and patient survival over the lifetime following transplant in recipients aged 60 and over was 3.0 years (represented by the differences in the area under the respective survival curves). This difference equates to 27 500 cumulative graft years for the 9250 transplants we identified for the potential exchange.
To estimate the impact on patient survival, we estimated the effect of reallocating younger organs into younger recipients and older organs into older recipients for the number of cases that could have been exchanged without affecting the likelihood of a transplant (i.e. all older recipients that received a younger organ now received an older organ in exchange for each case of a younger recipient who received an older organ). This reallocation simulation resulted in a projected patient survival increase in the younger recipient group of 27.1 months, and a reduced projected survival of 18.4 months in older recipients. Cumulatively, for the 9250 recipients over this time period that we identified that could have been exchanged, there was an average 9 months improved projected lifespan per patient.
Currently in the US, kidneys from young donors with the best survival expectancy are transplanted into older patients with a limited life expectancy at a significant rate. Between 1990 and 2002, if the 9250 kidneys from younger (<50 years of age) deceased donors transplanted into older (>60 years of age) recipients were allocated instead to recipients less than 60 years of age, our analysis indicates that an additional 27 750 graft years would have been gained. This benefit would have accrued by simply matching kidneys with longer lifespans to recipients projected to have longer lifespans, in turn reducing the loss of well-functioning allografts to patient death, and without impacting the number of kidneys available for recipients older than 60 years of age. Therefore, such a policy would not necessarily reduce the likelihood of older patients receiving a transplant, or increase waiting time.
Since allograft function and survival correlate with patient survival, there may be additional benefits of allocating younger kidneys to the younger ESRD population. Statistically, we estimated an average life expectancy gain of 27.1 months per patient in the younger population, while there would be a correspondent net loss in the life expectancy of the older population affected by this exchange, by about 18.8 months, with an estimated net gain of 9 months per patient for the overall population.
The stipulations of equal opportunity and equal access have been fundamental principles of the organ allocation process in the United States since its inception. This presents a difficult challenge when considering a policy change that might affect the overall population positively, but could impact negatively on segments of the population. It is difficult to decide whether equal opportunity is represented by the same absolute benefit or by the same relative benefit of transplantation. Intrinsic to the process, there are significant disparities the benefits kidney transplantation confers. The relative benefit is about equal in younger and older patients, with a death risk reduction by about 50% and approximate doubling in life expectancies. On the other hand, there is a 30-year gain in life expectancy in the youngest versus a 4-year gain in life expectancy in the very oldest segment of the population. Clearly, the assessment of the greatest benefit to the patient is very different from the assessment of the benefit to the population.
When considering the question of equity, and whether allocating younger kidneys to younger recipients would breach this principle by increasing the survival of the younger population at expense of the older population, one should also consider the opposite side of this argument. Death with a functioning graft is considered a success for the patient, because the patient could live without returning to dialysis (while obviously a loss of resource). The older patient population benefits by the current allocation process from a greatly increased rate of this type of ‘success’, with annual rates of death with functioning graft over twice as high as younger recipients (6.2% vs. 2.8%). By reallocating younger kidneys to younger patients, younger patients would benefit from higher death with functioning graft rates, similar to the older population, suggesting that the younger population would enjoy more of their lifespan with a functioning graft similar to the older patients.
Beyond the implications for graft and patient survival, the possibility of extending the average life of an allograft would have important implications for the waiting list and the economics of ESRD care. Currently, a failed transplant is the third most common cited reason for transplant wait listing in the US, with 17% of the patients listed in 2002 having previously lost an allograft (16). It is likely that extending graft longevity would reduce the number of new wait listings secondary to transplant failure, limiting the seemingly inexorable growth of the waiting list. In addition, given the cost efficiency of transplantation relative to dialysis ($13 749 vs. $67 506 annually per patient), saving 27 500 graft years would imply a cost saving of approximately 1.5 billion dollars for the transplants exchanged over this 13-year period (or $160 000 per transplant).
Unfortunately, the gains for younger patients are at least partially offset by an 18.4 months reduction in life expectancy for older recipients. It must be stressed that such transplants would continue to offer significant survival benefits compared to maintenance dialysis. In addition, the potential negative impact on older recipients could in part be mitigated by more expedited and efficient allocation of the older donor kidneys. Other interventions to improve outcomes in this segment of the population might evolve via prospective studies. The Eurotransplant Senior Program, currently underway in six European nations, preferentially allocates kidneys from donors greater than 65 years old in an expedited fashion to older recipients, and has shown encouraging preliminary results. Nonetheless, caution is warranted when considering such an approach, not only regarding long-term outcomes, but also for the potential negative impact of short-term complications (e.g. delayed graft function) that can impact may be to a greater extend on morbidity and mortality in older renal transplant recipients.
In this analysis, we estimated the number of graft years that were potentially lost retrospectively by virtue of allocating younger deceased donor kidneys to older recipients. As the allocation of younger donations to older recipients was not an infrequent event, we were able to estimate the implications of this practice. We reported the graft years saved based on survival estimates within the study period and found a relatively clear recipient age threshold for which graft years were lost due to patient mortality. Moreover, to understand the full impact of this practice, we projected the effects over the lifetime of transplant recipients. The projections were important to demonstrate that this impact would likely apply more dramatically in the long-term. However, only prospective application of this policy will fully allow for analysis of the actual graft, patient and cost savings with such an amendment.
The present analysis is not intended as a comprehensive solution to the broad organ shortage, and represents only one of many approaches to this subject. Obviously, donor age is not the only factor defining the quality of the organ and recipient age is not the only variable affecting patient outcomes; more sophisticated systems to match quality of donors and recipients could be devised. Donor and recipient age are the strongest risk factors for both graft and patient loss, respectively, and in fact explained the highest amount of variability in both models for patient death and graft loss. On the other hand, most certainly other risk factors can be used to get a better estimation of graft and patient survival. A more radical approach could match survival expectancies of grafts and patients based on a wide variety of patient and graft demographics and comorbid conditions. As that would possibly the most efficient way of allocating organs, it would be also the most exclusive, by really only allocating organs of poor quality to high-risk recipients. By using just donor and recipient age, still a little variation allows for a chance of getting a relatively good kidney even in a high-risk recipient group. In addition, age is a fairly easy to understand concept, and it might make sense to a patient to get a kidney from a donor age group similar to his own age, as opposed to just getting the overall worst kidney.
Certainly, incorporating an age variable into any allocation algorithm would be accompanied by substantial ethical and societal debate regarding implications and propriety. However, UNOS policies already favor children and adolescents, increasing chances for transplantation for candidates less than 18 years of age three-fold. On the surface, these discussions would be no more suspect than others currently underway involving novel approaches to remedy the organ shortage, such as financial incentives for donors.
We have to emphasize that our research is meant only to provide some of the data needed for an adequate discussion of the potential for age matching in deceased donor kidney transplantation. By no means is this article intended to advocate for or against such a policy, and ethical, financial and logistical implications are far more complex than what can be discussed within the limits and scope of a scientific publication.
In summary, by not including the impact of donor and recipient age in allocation of kidneys from deceased donors, valuable potential graft years are lost through patient death with a functioning graft. Between 1990 and 2002, 27 500 years of dialysis could have been saved by allocating younger deceased donor kidneys to younger recipients. These data indicate that proposals incorporating age matching into allocation algorithms should be considered as the transplant community continues to struggle with the implications of the donor shortage.
The data reported here have been supplied by the University Renal Research and Education Association (URREA) as the contractor for the Scientific Registry of Transplant Recipients (SRTR). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of or interpretation by the SRTR or the U.S. Government. IRB approval or exemption determination is the responsibility of the authors as well. We would like to express our appreciation to Suzanne C. Johnson who has helped in editing and review of the article.