Kidney allograft survival: the long and short of it

The unfortunate fact that long-term renal allograft survival has not followed the gains made in short-term allograft survival has been documented repeatedly in the United States [1] and also worldwide [2]. Undoubtedly, large inclusive datasets are needed to document certain overarching trends. Some of the most accepted risk factors for poor outcomes after transplantation like prolonged pre-transplant dialysis time [3] and the use of expanded criteria donors [4] needed confirmatory analysis from large databases to quantify these effects. When in smaller datasets, risk factors that have been confirmed for larger populations are not evident, one of the two things might be happening. Either the local reality is different and the sampled population has different characteristics and treatments are possibly different than the overall population or there is inadequate power to adequately identify and quantify multiple risk factors at the same time. The authors of the paper and the staff of the Rogosin Institute are certainly to be congratulated for their excellent outcomes and the big progress their center has made [5]. Internal quality assessments are a key component to any program success. National trends might not be reflected in local populations and each center has to assess what works best in their local reality. Several centers have documented excellent local results despite the lack of national evidence of significant long-term progress. An important component of long-term survival is adequate care delivery to renal transplant patients. When centers are able to follow their patients very closely throughout their lifetime, this could make a substantial difference in outcomes. It is in fact very possible that much of the long-term attrition in renal transplant outcomes in the United States is driven by lack of medication availability and specialized medical care in general. The complicated socioeconomic realities of renal transplantation emphasize early transplant funding while later funding is clearly inadequate. Not only do Medicare beneficiaries lose their immunosuppressive drug coverage 3 years after transplantation, but clinic visits for renal transplant patients are a long-term financial liability for care providers because of low reimbursements. If these are in fact strong divers of long-term renal transplant attrition, then it is very possible that single institution might be able to overcome these difficulties. On the other hand, if established risk factors like ECD donors and prolonged dialysis time do not come up in the limited population as risk factors questions about the power of the analysis start surfacing. The authors conducted a single-center retrospective analysis of 1476 kidney transplant recipients from 1963 to 2006 stratified by immunosuppression protocol [5]. They report an increase in patient and kidney survival over time and as immunosuppression protocols changed. Their analysis of long-term survival includes an actual 5-year survival analysis, as well as a 1-year conditional analysis, thus evaluating a 4-year survival rate of only those grafts still functioning after 1 year. As in concordance with previous studies, patient and graft survival improved in the era of calcineurin inhibitor use beginning the mid-1980s (Protocols 2–5 in the study). Rejection rates have dramatically improved along with our ability to quell the impact of early rejection, thus leading to improved short-term survival. In their most-recent protocol, which involves induction therapy and steroid sparing, they report an acute rejection rate of 5.8% in 188 patients with an 86% 5-year survival rate and a 92% graft survival rate in those grafts functioning 1 year after transplant. While older recipient age, black race, diabetes, delayed graft function and presence of rejection in the first year emerged as significant risk factors for actual graft survival, only acute rejection was significant for conditional survival in this study. The methodology and era analysis in this paper illustrate some finer points to be considered when separating graft survival into short- and long-term rates. Improvements in acute rejection rates certainly have translated into very good first-year graft survival rates that now exceed 90%, leaving little room for improvement [6]. One-year conditional survival rates introduce a selection bias of those allografts that have already proven their robustness by surviving the critical period of higher risk of rejection and post-surgical complications. Another way to quantify long-term survival to patients in a clinically relevant manner is the concept of half-lives where both early and late effects are combined to give a clinically meaningful assessment of survival probability. The other scenario that can be helpful is death-censored data to quantify how long the kidney will last in a patient who stays alive. In the US renal transplant population as a whole, outcomes have unfortunately not mirrored the trend described in the Rogosin center report. In 1991, Gjertson reported that although survival took a dramatic rise upward between the years 1975 and 1990, starting at 44% survival and increasing to 81%, 2-year conditional allograft half-life remained stagnant at ∼7 years [7]. Subsequently, the report of kidney allograft half-lives assessed from 1995 to 2000 from UNOS registry data showed little improvement in the long term despite improved 1-year survival and rejection [8]. A recent report of the latest updated US data over the last two decades in press at AJT indicates that although continuing improvements have been made in short-term survival and attrition rates, the long-term attrition rates have stayed surprisingly constant [9]. In fact, there has been a steady increase in transplant half-lives in the United States, but this has been almost entirely driven by improvement in short-term survival. Deceased donor transplant half-lives were 6.6 years in 1989, 8.2 years in 2000 and 8.8 years in 2005 (Figure 1), yet attrition rates beyond the first year have stayed very constant between 5 and 7% year−1 in deceased donor transplantation (Figure 2). In living donor transplants that have better survival outcomes compared to deceased donor, half-lives have not improved at all going from 11.4 years in 1989 to 11.9 years in 2005 (Figure 1), with a somewhat lower but stable long-term attrition rate ∼4% [9]. This data shows impressively the dichotomy between short-term and long-term graft attrition, with dramatic improvements in the first year but negligible progress with longer-term survival. This very trend is also seen in the data presented in this paper from the Rogosin institute [5]. As illustrated in Figure 1 of their paper, the differences in the survival curves are mainly evident in the first year, while the subsequent attrition rates appear remarkably similar. Deceased (n = 164 480) and living (n = 88 430) donor kidney transplant half-lives in the US. Deceased donor kidney transplant attrition rates in the United States (n = 164 480). Survival data and half-lives allow us to determine the cumulative effects over time, but they do not account for the different risks of graft loss early after transplant versus later on. To explore these effects, attrition rates by time from transplant need to be explored in detail as illustrated in Figure 2. Given the observed dichotomy in the evolution of short versus long-term outcomes, it is in fact likely that different pathogenic processes are driving them. In the short term, the control of the immunological response seems to be more important, whereas the very same factors that protect the graft in the short term become a potentially important cause for graft attrition long term. A good example is antibody-depleting induction that can effectively protect patients from rejection and graft loss early after transplant but later might predispose to more polyoma virus nephropathy or PTLD. Serur et al. describe an increase in the use of ECD kidneys for their own center and that mirrors the national trend over time [5, 6]. In fact, this is a population where more substantial half-life improvements have been described in the United States [9] from 3 years in 1989 to 6.4 years in 2005. The relatively low half-lives in ECD transplants may be driven by patient death, as it is generally older recipients who receive ECD kidneys. In fact, death-censored half-lives are significantly higher and have improved from 4.3 years in 1989 to 9.2 years in 1999 and 10.1 years in 2005, closing the gap with SCD transplants. However, in this group also, long-term attrition rates for transplants in the 3–5 year time period have barely improved from 12.4 to 10% in 2004. This again illustrates that the improvement we see in graft survival is driven by improved short-term survival rates, while the long-term attrition rates are changing little. As it has only recently been recognized that long-term attrition rates are not substantially improving [1, 7–9], the pathogenesis of this phenomenon is still insufficiently explored. Traditional study end points in kidney transplantation have focused on the first year after transplantation. Later, graft attrition might be driven by different mechanisms such as CNI nephrotoxicity, metabolic problems and other toxicities induced by the very same medications that have allowed for short-term survival improvements. The individual optimal balance between over and under immunosuppression still eludes the understanding of the treating transplant physicians and patients. Conflict of interest statement. None declared.