The article by Baskin-Bey et al1 in this issue provides support for the growing pressure to change the rules that govern the allocation of deceased donor kidneys in the United States. The changes that are being considered represent a major shift in the philosophy of kidney allocation and have critical consequences for the more than 70,000 patients2 who are facing a prolonged wait. However, before we embark on change, it is incumbent on us to understand the nature of the major deficiencies in our current allocation system.
The philosophy behind the allocation rules has traditionally, if somewhat simplistically, been viewed as balance between the seemingly competing factors of equity (fairness in access for all those in need), medical utility (optimal exploitation of the function of available organs), and efficiency (including time, cost, organ wastage, and geography). However, determinants of these factors have not remained constant, and the changes that have taken place in the allocation rules during the past two decades reflect an ongoing reassessment of the balance between them. For example, close HLA matching improves long-term graft outcome, yet the role of HLA matching in the algorithm has been steadily deemphasized during the years, not because matching is unimportant or irrelevant, but because its emphasis led to unforeseen inequity and inefficiency.3
In the current allocation algorithm for adults awaiting a standard criteria donor (SCD) kidney, one point is given for each year of waiting time and one point each is given for matching at each of the two DR loci: zero-mismatched kidneys are given priority and are shared nationally (∼17% of patients receive them, usually within two years of being placed on the list).3 All patients on the list, irrespective of age, are eligible to receive SCD kidneys, although children (<18 years) are given priority for kidneys from donors younger than 35 years. Nonstandard so-called extended criteria donor kidneys (from donors > 60 years or those > 50 years with additional risk factors) account for approximately 15% of all transplanted kidneys and are allocated on the basis of waiting time alone, typically to older patients.4 When we explain the algorithm to patients, we like to use a simple analogy. The waiting time component is like waiting in line for a bus; frustrating perhaps, yet ultimately predictable. The matching component can be likened to a game of bingo or a lottery; someone playing (on the waiting list) will get lucky, but the process essentially is random. Because waiting times are long and zero-mismatched kidneys are uncommon, the list currently is most like the bus line: when the bus comes, the next person gets on. The system seems fair, it is intuitive and easy to explain, so what is the problem?
The problem is that, other than the small component given for HLA matching and the priority given to children, the algorithm for the 85% of all kidneys that come from SCDs makes no attempt to match the features of the donor kidney to those of the recipient. However, patients vary in their life expectancy and donor kidneys vary in their quality. In an ideal world, all patients, no matter what their life expectancy, would get high-quality organs with a long functional life without a prolonged wait. However, we do not live in such a world! At a time of shortage, does it make sense to give a kidney that could be expected to function for decades to patients who will probably live for no more than a few years?5 To put it more concretely, in the current algorithm, an “ideal” SCD kidney from a 20-year-old victim of a motor vehicle accident could just as well be allocated to a 25-year-old with immunoglobulin A nephropathy without significant comorbidity as to a 65-year-old patient with diabetes with a history of coronary heart disease. It is considerably more likely that the kidney would be allocated to the older recipient because as the waiting list ages, older patients are receiving more of the available organs.5 There thus is “wastage” of the potential functioning years of deceased donor kidneys, particularly those from younger donors. Meier-Kriesche et al6 estimated that between 1990 and 2002, a total of 27,500 graft years were wasted as a result of placement of kidneys from young donors into older recipients. As stewards of the organs that are bequeathed to society by the altruism of the deceased and their families, we have a choice: we can accept that wastage as an inevitable and reasonable consequence of equitable allocation; we can allocate kidneys in a solely utilitarian fashion, concerned only with keeping them functioning as long as possible and paying only lip service to equity; or we can strive, as we always have, to find a new balance.
The study by Baskin-Bey et al1 shows that matching the features of the donated kidney with those of its recipient has the potential to improve the effective donor kidney supply. Others have reached a similar conclusion.6 However, it is one thing to show that matching of donor and recipient features can increase the utility of the donor kidney supply and it is another to design an algorithm that does so in an equitable and efficient manner. The Kidney Allocation Review Subcommittee of the United Network for Organ Donation is in the process of designing such an algorithm.7 The proposal currently under discussion and evaluation differs in 1 major respect from that proposed by Baskin-Bey et al.1 Rather than using years of renal allograft function (renal-years) as their marker of effectiveness, the Kidney Allocation Review Subcommittee proposed a more nuanced approach. They introduced a concept named “net life survival benefit” (NLSB) that provides an estimate of a recipient’s lifespan after receiving a transplant, but then subtracts the estimated lifespan of that patient if he or she were to remain on the waiting list. The NLSB score then can be modified further to take into account that the quality of life of patients waiting for a kidney is likely to be less than that of patients who receive one. The purpose is to maximize the quality-adjusted person-years of life for the entire wait-list population.
It is one thing to make survival and quality-of-life estimates for groups of patients and another to make these estimates for individual patients. After all, we perform transplantation on individual patients, not groups. Some of the stimulus for change in the algorithm for allocation of kidneys has come from effective changes made in recent years in the allocation of livers and lungs. For both these vital organs, the allocation algorithm now direct the organs to recipients who are deemed most likely to benefit from them, and the waiting-time component of allocation has been eliminated. In the case of liver allocation, the likelihood of dying is estimated in a verifiable and objective fashion by using the so-called Model of End-stage Liver Disease (MELD) score (calculated from prothrombin time, bilirubin level, and creatinine level).8 Any seasoned clinician can evaluate a patient with advanced liver disease and relate their MELD score to their clinical condition and risk of early death. However, for kidney transplantation, which is more properly regarded as not so much a lifesaving procedure, but rather a life-prolonging and life-enhancing one, estimation of life expectancy is much more difficult. Baskin-Bey et al1 used age, diabetes mellitus, time on dialysis therapy, and a “history of angina” to predict survival by using an arbitrary scoring system. To estimate NLSB, a predictive modeling system is used based on extensive survival data available at the Scientific Registry of Transplant Recipients. Not surprisingly, age is by far the most important component in both systems.5 However good both predictive formulas are in estimating the lifespan or survival benefit of patient cohorts with certain characteristics in common, we suspect that clinicians, no matter how seasoned, might find it difficult to relate them confidently to the prognosis of an individual patient.
The bottom line is that application of a formulaic estimate of life expectancy based to a large extent on age and some other essentially immutable biological factors may be the only way by which we can maximize the “value” of deceased donor kidneys and reduce the wastage of life-years that is inherent in the current kidney allocation algorithm. In the United Kingdom, which faces a kidney shortage similar to that in the United States, an allocation algorithm was introduced in 2006 that includes points for a limited difference between the age of the donor and recipient.9 It is not surprising that some commentators are concerned by the anticipated changes.10 Any change made in the current allocation algorithm will advantage some patients, but inevitably disadvantage others. Allocation algorithms, no matter how sophisticated and well designed, do not increase the supply of donor organs. Concerted efforts are being made in this regard, and in the United States, deceased donation has increased significantly during the last few years.11 However, for the foreseeable future, there will be a shortage of kidneys. We cannot escape our obligation to strive to allocate them in the best interests of both our individual patients and society at large. All those caring for actual and potential transplant candidates are strongly encouraged to actively engage in the public policy discussions that will accompany the development and application of new allocation rules.
FULL TEXT