Long-Term Renal Allograft Survival in the United States: A Critical Reappraisal

Transcription

1 American Journal of Transplantation 2011; 11: Wiley Periodicals Inc. C 2010 The Authors Journal compilation C 2010 The American Society of Transplantation and the American Society of Transplant Surgeons doi: /j x Long-Term Renal Allograft Survival in the United States: A Critical Reappraisal K. E. Lamb, S. Lodhi and H.-U. Meier-Kriesche Division of Nephrology, Hypertension and Transplantation, University of Florida, Gainesville, FL *Corresponding author: Herwig-Ulf Meier-Kriesche, meierhu@medicine.ufl.edu Renal allograft survival has increased tremendously over past decades; this has been mostly attributed to improvements in first-year survival. This report describes the evolution of renal allograft survival in the United States where a total of patients received a single-organ kidney transplant between 1989 and Half-lives were obtained from the Kaplan Meier and Cox models. Graft half-life for deceaseddonor transplants was 6.6 years in 1989, increased to 8 years in 1995, then after the year 2000 further increased to 8.8 years by More significant improvements were made in higher risk transplants like ECD recipients where the half-lives increased from 3 years in 1989 to 6.4 years in In low-risk populations like living-donor-recipients half-life did not change with 11.4 years in 1989 and 11.9 years in First-year attrition rates show dramatic improvements across all subgroups; however, attrition rates beyond the first year show only small improvements and are somewhat more evident in black recipients. The significant progress that has occurred over the last two decades in renal transplantation is mostly driven by improvements in short-term graft survival but long-term attrition is slowly improving and could lead to bigger advances in the future. Key words: Graft half-life, graft survival, kidney transplantation Abbreviations: SRTR, Scientific Renal Transplant Registry; SCD, Standard criteria deceased donor; ECD, Expanded criteria deceased donor; Tx, Transplant; Rec/Don LE 45, Recipient and donor age less than or equal to 45; KM, Kaplan-Meier. Received 21 June 2010, revised 13 July 2010 and accepted for publication 04 August 2010 Introduction The specialty of kidney transplantation has made dramatic strides over the decades evolving from an experimental procedure to the standard of care in the treatment of patients with end-stage renal disease (1). Not only are the outcomes after kidney transplantation good enough to improve the quality of life (2) of our patients, but it has also been established as a life-saving procedure (3,4); yet the life-saving benefit of a kidney transplant lasts only as long as the transplanted kidney (4). Technical and pharmaceutical progress have helped to improve outcomes progressively even over the last decade when excellent outcomes were already considered standard of care. Now with graft survival rates in excess of 90% the question arises if any further improvements are possible or even necessary. In 2004 it became clear that the overall improvements in graft survival after kidney transplantation were really driven by improvements in first-year survival, whereas long-term graft attrition remained largely unchanged over decades (5,6). This highlighted a whole other area where improvements might be possible and necessary. Especially now when first-year survival rates are almost close to perfect it becomes clear that further improvements in long-term survival have to come through improvements in long graft maintenance. It is notoriously difficult to measure long-term survival, as lengthy follow-up is necessary to document it, yet periodic updates on the long-term trends can potentially yield important information especially when counseling patients in the pretransplant phase regarding expectations of future outcomes. The purpose of our present study was to reevaluate the evolution of short- and long-term renal allograft survival in the United States with the most recent data provided by the Scientific Renal Transplant Registry (SRTR). Materials and Methods Subjects We examined data from the national SRTR database for renal transplant recipients from 1989 to November 1, Analyses were conducted on adult transplant recipients 18 years or older. Multiorgan transplants were excluded from the analysis. Data were analyzed separately for living and deceased-donor transplants, black recipients, nonblack recipients, first transplants and repeat transplants, for standard criteria donor (SCD) kidneys and expanded criteria donor (ECD) kidneys and for recipients with donor and recipient ages below 45 years. Outcome measures We analyzed graft, patient and death-censored graft survival by estimating survival half-lives and we analyzed attrition rates all stratified by year of transplant. 450

2 Long-Term Renal Allograft Survival Half-lives; Univariate half-lives were calculated as median half-lives, i.e. the intersection point of the Kaplan Meier curve with the 50% survival threshold. We differentiated between actual half-lives for those instances where all patients had reached the 50% mark, actuarial half-lives for those instances when only a proportion of patients had reached the 50% mark and projected half-lives when none of the patients reached the 50% mark. In the tables and figures actual and actuarial half-life were grouped together but projected half-lives are shown separately. Projections were obtained by forecasting the Kaplan Meier curves from a point of stable attrition, which was fairly consistently located between 3 and 8 years of survival yielding a period of 5 years from which the forecasts were based. Forecasted projections were carried out using ordinary least squares point estimates. Multivariate half-lives were obtained in the same fashion from the Cox proportional hazard models. Attrition rates; Attrition rates were calculated by first acquiring actual 1-year, 3-year, 5-year and 10-year survival rates. The total number of patients failed during the time period was subtracted from the number of patients originally entering the cohort and divided then by the original number entering the cohort to obtain an absolute failure percentage. The percentage of absolute failures was then divided by the total number of years in the follow-up interval to obtain a yearly failure rate. Independent variables Covariates used to calculate the adjusted half-lives from the Cox model included recipient s transplant age (reference group 18 34), pretransplant diagnosis of diabetes (reference group diabetic), candidate race (reference group Caucasian) and candidate gender (reference group males) summarized for transplant year 1999 to yield the most up-to-date case mix for complete 10-years follow-up (7). half-lives were utilized in both the univariate Kaplan Meier and the multivariate Cox Regression model for allograft failure and only in the univariate Kaplan Meier for death-censored allograft failure. Multivariate models were corrected for the same variables uses in the SRTR annual data report as described above. Proportional hazard assumptions were tested by visually assessing log log survival curves. The Exact method was used to handle tied outcome occurrences. All analyses were conducted using SAS (v.9.2, Cary, NC) and a type-one error probability of 0.05 was utilized as an indication of statistical significance. Results Patients We analyzed a total of adult kidney recipients transplanted between 1989 and 2005 excluding multiorgan transplants. Of these, were deceased-donor transplants and living-donor transplants. Of deceased-donor transplants were ECD transplants. Of deceased standard criteria donor recipients, were first transplant recipients and were repeat transplants. Statistical models Graft survival Figure 1(A) shows overall graft survival for standard criteria deceased-donor transplants between 1989 and 2005 and the respective median half-lives based on where the survival curve crosses the 50% survival line. Outcomes were measured by the Kaplan Meier models and the Cox multivariate proportional hazard models. Half-lives were calculated based on actual and projected follow-up where applicable. Half-lives based on actual versus projected follow-up are displayed distinctly in the results. Projected Figure 1(B) shows death-censored graft survival for standard criteria deceased-donor transplants and the respective half-lives. Figure 1: (A) Kaplan Meier cumulative graft failure and (B) deathcensored graft failure, by year of first deceased SCD transplants from transplant year American Journal of Transplantation 2011; 11:

3 Lamb et al. Half-lives Table 1 displays the overall both actual or actuarial halflives and the projected half-lives marked as in the second shaded line by transplant year. The overlap between the actuarial half-lives and projected half-lives represents instances where still reasonable conclusions can be drawn from the actuarial data but forecast were generated in parallel. This gives a sense also about how well the forecasts might be working. When evaluating all deceased-donor transplants jointly, the half-life was 6.6 years in 1989, increased to close to 8 years in 1995, stayed around 8 years until transplant year 2000 and then further increased to 8.8 years in When looking only at first-donor transplants (N = ) excluding retransplants, the half-life was 6.8 years in 1989, increased to 8 years in 1995 and increased to 9 years in When limiting the analysis to just standard criteria deceased-donor transplants but including retransplants (N = ) the half-lives improved from 6.7 years in 1989 to 9.5 years in Slightly (but not dramatically) better half-lives were achieved in first standard criteria kidney graft recipients (N = ). When donor and recipient age was limited to less than 45 (N = ), graft half-lives improved from 7.7 years in 1989 to 11 years in First ECD transplant (N = ) half-lives were dramatically lower with 3 years half-life for first deceased-donor ECD transplants in 1989 that increased to 6.4 years in Living-donor transplant (N = ) half-lives were substantially higher than deceased-donor half-lives but there was no appreciable change in living-donor half-life over the years. Whether repeat transplants were included or not half-lives were 11.4 years in 1989 and years in When dividing the populations into black versus nonblack recipients the half-lives display the well-known outcomes difference between blacks and nonblacks. There has been a similar absolute increase in graft half-lives when comparing nonblack and black recipients, with a modest improvement in black ECD kidney half-life when comparing an increase in 3.9 years from to only 2.9 years in nonblack ECD half-life over the same time period. Black standard criteria deceased-donor (N = ) transplant half-life increased from 4.1 years in 1989 to 7.4 years in 2005; however, this was still substantially lower than nonblack (N = ) half-life in the same year (10.9 years). Living-donor transplant half-lives in 1989 were 6.3 years in black recipients and 12.3 years in nonblack recipients. In 2005 living-donor recipient half-lives were 7.5 years in black versus 13.5 years in nonblack patients. Figure 2(A) contrasts the overall half-lives between first living-donor transplants and first standard criteria deceased-donor transplants. Figure 2(B) contrasts first standard criteria deceased-donor graft survival between black and nonblack recipients. Table 2 displays the death-censored graft half-lives by transplant year. Death-censored half-lives for all deceased-donor kidneys were 10.2 years in 1989 and increased to 14.3 years in Death-censored graft half-lives for standard criteria deceased-donor transplants have increased from 10.6 years in 1989 to 15.5 years in When deceaseddonor transplants were limited to first transplants and both donor and recipient age of less than 45, death-censored graft half-life was 10.1 years in 1989 and 12.4 years in First ECD transplant death-censored half-lives increased from 4.3 years in 1989 to 10.1 years in Living-donor death-censored half-life was 16.5 years in 1989 and 16.6 years in Standard criteria deceased-donor half-life was 5.2 years in black recipients in 1989 versus 12.8 years in nonblack and 10 years in 2005 in black versus 21.2 years in nonblack. In contrast to cumulative graft half-life, death-censored half-lives were notably more improved from 1989 to 2005 in nonblack recipients as compared to black recipients. Death-censored standard criteria kidney half-life was 12.8 years and forecast to be 21.2 years in 2005, resulting in an increase by 8.4 years for nonblack recipients. Deathcensored black standard criteria half-life is forecast to only increase by 4.8 years by Similar trends are seen in death-censored ECD, recipient and donor age less than 45 and living-donor half-lives. Figure 3(A) contrasts standard criteria deceased-donor first transplants with first living-donor transplants and Figure 3(B) contrasts first standard criteria deceased-donor transplants between black versus nonblack patients. Table 3 displays the adjusted half-lives derived from the Cox proportional hazard models. The adjusted half-lives show similar patterns to the unadjusted half-lives. The adjusted standard criteria deceased-donor half-life in 1989 was 6.6 years and 9.9 years in For living-donor transplants the adjusted half-lives were 11.4 years in 1989 and 12.2 years in Graft attrition rates Table 4 displays the graft attrition rates by transplant year, where both graft loss and patient death are counted as an event. As also shown in Figure 4, for all categories the American Journal of Transplantation 2011; 11:

4 Long-Term Renal Allograft Survival Table 1: Kaplan Meier estimates of cumulative graft half-lives by transplant year Transplant years All races ECD/SCD (N = ) ECD/SCD st Tx ECD/SCD (N = ) 1st Tx ECD/SCD SCD (N = ) SCD st Tx SCD (N = ) 1st Tx SCD st Tx Rec/Don LE (N = 40529) 1st Tx Rec/Don LE 45 1st Tx ECD (N = 21523) st Tx ECD Living donor (N = 88430) Living donor Living donor 1st Tx (N = 76579) Living donor 1st Tx Nonblack/black 1 ECD/SCD (N = /46896) /4 7.5/ /5 7.9/5 7.9/ / / / / /6.5 /6.4 /6.3 /6.9 ECD/SCD 9.1/ 9.2/ 9/ / /7 9.6/ / /7.1 SCD (101072/39828) 7.8/ / / / / / /6 9.5/6.7 10/ /7 10/7 /6.7 /7.2 SCD 9.8/7 10/ /6.8 10/ / / / /7.4 1st Tx ECD (14956/6567) 3.7/ / / / /2.3 5/ / / /4.4 6/ / / /5 6.1/4.9 /5 1st Tx ECD 6.3/4.3 6/ / /5.2 6/ /5.7 1st Tx Rec/Don LE / / / / / / /6 12/ /7.6 /7.2 /7.7 /7.9 /7.7 (N = 27253/13276) 1st Tx Rec/Don LE 45 Living donor (N = 75640/12790) 12.3/ /7 12.6/ / / /7 12/ /7.6 /8.7 / / / / / / / / /8 Living donor 12.9/ 12.3/ 12.2/ /9.2 13/ / / / / /7.5 1 Nonblack-recipient half-lives listed before hash mark followed by black recipients after the hash mark. Where no half-life could be obtained respective years are blank. Shaded rows are projections. American Journal of Transplantation 2011; 11:

5 Lamb et al. A B Figure 2: Actuarial (diamond marker, solid line) and projected (round marker, dotted line) cumulative half-lives. (A) First SCD deceased versus first living donor and (B) first SCD deceased black versus nonblack recipients. year attrition rate improved dramatically and progressively since The 1 3, 3 5 and 5 10 year attrition rates show also small but consistent improvements. Figure 4 shows graft attrition rates for (A) standard deceased-donor kidney recipients, (B) living-donor kidney recipients, (C) nonblack standard deceased-donor recipients and D) black standard deceased-donor recipients. Compared to nonblack SCD attrition rates, black SCD attrition rates showed a modest improvement in the attrition rates for years 1 3, 3 5 and Table 5 displays the death-censored graft attrition rates counting only graft failure as an event and censoring in case of death. For donor and recipient age below 45, the death-censored graft attrition rates have changed very little. Three to five years graft attrition was 5.0% per year in 1989 and 4.7% in Comparing black to nonblack patients there is a substantially higher long-term attrition rate in black recipients. In 1989 the 3 5 year attrition rate in blacks was 8.7% versus 3.2% in nonblacks and 4.9% in blacks in 2004 versus 3% in nonblacks. Table 6 displays graft and death-censored graft attrition rates in living-donor transplants comparing black to nonblack recipients. Again the dramatic improvement in 0 1 year attrition rates is evident also in living-donor transplants. The first year attrition rate in black recipients in 1989 was 14.7% compared to 7.8% in nonblack patients and 4.6% in black patients compared to 4.5% in nonblack patients in Similarly for the 3 5 year attrition rate there has been a more striking improvement in black recipients with an attrition rate of 10.3% per year in black recipients in 1989 compared to 5.1% in For deathcensored graft failure the gap between the outcomes in black compared to nonblack persists until the most recent data even when limiting to donors and recipients below age 45. Discussion Long-term renal allograft survival in the United States has made small but measurable progress over the years. The majority of the gain has been documented in the first year after transplantation. This is important progress as with improved 1-year survival the graft attrition starts at a higher intercept with most likely better long-term outcomes. Yet the attrition rate has to at least stay constant to achieve this. In fact the data suggests that the long-term attrition rates have improved slightly despite of arguably more highrisk patients now at least reaching the 1-year mark. Both progresses combined have led to an improvement in renal allograft half-lives for deceased donor from 6.6 years in 1989 to 8.2 years in 2000 and a continued improvement to a projected 8.8 years in These numbers corroborate previous publications that have shown similar data (5,6). This is important as the herein presented data has significantly more follow-up relative to the previous report and now allows us to compare the previous actuarial half-life projections (6) with the actual half-lives until the year 2000, which in fact does confirm the previous projections; the actuarial half-life projections beyond the year 2000 point toward possible small further improvements in half-lives. As half-lives seem to be the best way to give the patient a good clinical understanding of how long their 454 American Journal of Transplantation 2011; 11:

6 Long-Term Renal Allograft Survival Table 2: Kaplan Meier estimates of death-censored graft failure half-lives by transplant year Transplant years Recipient population ECD/SCD ECD/SCD st Tx ECD/SCD st Tx ECD/SCD SCD SCD st Tx SCD st Tx SCD st Tx Rec/Don LE st Tx Rec/Don LE st Tx ECD st Tx ECD Living Donor Living donor Living donor 1st Tx Living donor 1st Tx Nonblack/black 1 SCD 12.8/ /6.4 14/ /7 13.8/8 15/6.3 /8.2 /9.3 /9.3 /9.8 /9.7 /9 SCD 14.4/ 15.1/ 14.2/ 16.4/ 15.6/ / / / / / / /10 1st Tx ECD 5.2/2 6.6/ / / / /5.1 9/ / / /6.2 /5.9 /6.4 /7.1 1st Tx ECD 10.7/7 11.3/ / / /7.5 1st Tx Rec/Don LE / /6 13.6/ / /7 13.9/7.8 /7.1 /9.5 /8.7 /8 /9.2 /8.7 1stTx Rec/DonLE45 14/ 13.9/ 13.5/ 13.7/ 14.1/ /8 16.6/9 21.2/ / / / /8 ECD/SCD 12.5/ / / / / /7.8 /7.9 /8.6 /8.7 /9.2 /9.2 /8.9 ECD/SCD 13.2/ 13.7/ 12.9/ 15/ 14.4/ 15.7/8.5 16/8.8 18/ / / / /9.3 Living donor 17.4/ / /10.2 /9.2 /9.4 /8.9 /9.7 /10.3 /12.0 Living donor 17.3/ 16.5/ 17.4/ 16.9/ 17.6/ / / / / / / / / / /8.4 1 Nonblack-recipient half-lives listed before hash mark followed by black Recipients after the hash mark. No KM half-life reported for respective year where blank. Shaded rows are projections. American Journal of Transplantation 2011; 11:

7 Lamb et al. A B Figure 3: Actuarial (diamond marker, solid line) and projected (round marker, dotted line) deathcensored half-lives. (A) First SCD deceased versus first living donor and (B) first SCD deceased black versus nonblack recipients. future kidney transplants might actually last it is important to break down the populations especially where substantial differences in outcomes are expected. Another way to generate clinically useful data is to analyze death-censored half-lives to give the patients an idea of how long a kidney might last independent of the patient s survival. This is particularly important in young patients where the likelihood of death is remote. In fact, there are several interesting points to be made when looking at this data in detail. For example it is striking how in the low-risk population where both donors and recipients were below age 45 there has been a very limited improvement in death-censored graft survival. The cumulative gain in death-censored graft survival in this population is about 2.5 years. When looking at the same population for overall graft survival there was a 3.3-year gain in halflife over the 16 years between 1989 and That means that there were small increments in graft survival and even smaller increments in patients survival. This is important to realize as in the past it has often been hypothesized that the lack of long-term improvement after kidney transplantation might be due to more risky donors and recipients. However, in actuality there is the least improvement in the lowest risk populations. This is further corroborated by the living-donor data where half-lives have been oscillating between 11 and 12 years without any appreciable era effect since Half-life assessments are difficult especially when halflives are long because lengthy follow-up is needed to generate meaningful data. In order not to have to wait for the full length of time to understand outcomes data half-lives can be estimated based on incomplete follow-up or even projected. Actual half-lives are derived from data when all patients have follow-up for at least the same time as the half-life itself. For example if in one population the half-life is 8 years, a minimum follow-up of 8 years for all patients is needed to establish an actual half-life. For an actuarial half-life, follow-up derived from a Kaplan Meier estimated survival would rely on a minimum of one patient with at least as much follow-up as the half-life. In a scenario where not even one patient has enough follow-up to visually forecast the half-life, projections are necessary. The projection is based on the proportion of the Kaplan Meier curve where somewhat stable attrition rates have been established. The projections may be based totally or partially on actual follow-up and to a variable extent on actuarial follow-up. It is important to realize that the reliability of the projections does decrease with the length of the projection. The higher the half-life, the longer the projections are in order to estimate a half-life. Good examples are the death-censored projected half-lives displayed in Table 2. With death-censored half-lives around 20 years for living-donor kidney patients transplanted in 1999 there is only 10 years of actual follow-up, thus the remaining 10 years of the displayed half-lives are based on projections. For patients transplanted in 2004 the same projected halflife is only based on 5 years of actual follow-up. An indirect measure of the reliability of the forecasts is the year-to-year variation, if the half-lives projected vary widely this is probably an indication that the forecasts are not reliable. Clearly the longer the half-lives are the more imprecise the projections will be. In fact the death-censored survival data have a higher variability in the predictions because the deathcensored half-lives are longer than in the uncensored data. 456 American Journal of Transplantation 2011; 11:

8 Long-Term Renal Allograft Survival Table 3: Cox adjusted estimates of cumulative graft failure half-lives by transplant year Transplant years All races SCD SCD st Tx SCD st Tx SCD st Tx SCD Rec/Don LE 45 1st Tx SCD Rec/Don LE Living donor Living donor Living donor 1st Tx Living donor 1st Tx Nonblack/black 1 SCD 7.8/ / / / / / /6 9.6/6.7 10/6.8 10/7 10/7 /6.8 /7.2 SCD 9.8/7 10.1/ / / / /7.7 11/ /7.5 1st Tx SCD 8.2/4.3 8/ / / / /6 9.6/ / /6.9 10/7.1 10/7 /6.9 /7.4 1st Tx SCD 10/7 10.1/ / / / / / /7.5 1st Tx SCD Rec/Don LE 45 1st Tx SCD Rec/Don LE / / / / / / /6 12.2/ /7.6 /7.2 /7.5 /7.9 / / / / / / / / /7.2 1 Nonblack-recipient half-lives listed before hash mark followed by black Recipients after the hash mark. No KM half-life reported for respective year where blank. Shaded rows are projections. American Journal of Transplantation 2011; 11:

9 Lamb et al. A B C D Figure 4: Cumulative graft failure yearly attrition rates of first kidney transplants (A) Deceased SCD donor, (B) living donor, (C) Nonblack deceased SCD donor and (D) black deceased SCD donor. An interesting area in this data is the comparison between black and nonblack recipients. In general the data highlights the profound differences in outcomes between black and nonblack patients, which have been described previously (8). The small improvements in half-lives have been nearly parallel between black and nonblack patients as shown in Figure 2(B). It is interesting to note though that deathcensored half-lives have been somewhat more stagnant in black patients compared to nonblack patients (Figure 3B) with the caution again that the forecasts in nonblack recipients death-censored data might not be very solid because the half-lives are so long. On the other hand looking at the graft attrition data in black patients who clearly have the highest graft attrition there seems to be a more perceptible trend of improvement over the years as can be seen in Figure 4(D). In fact even though the absolute increase in half-life was only about 4 years, the half-life almost doubled from 1989 (4 years) to 2005 (7.4 years). We specifically analyzed attrition rates because half-lives assess only the cumulative effect of graft loss overtime, but there is a clear difference in the early rate of graft loss versus late and the underlying pathogenetic processes might even be somewhat distinct. Clearly first-year attrition rates are significantly higher for all subpopulations and have improved significantly overtime. The subsequent graft attrition rates are fairly constant. When comparing the 1 3, 3 5 and 5 10-year attrition rate there do not seem to be dramatic differences with increasing time posttransplant but overall the 5 10-year attrition rate is for the most part somewhat higher. This seems to be most evident in living-donor transplants where the 1 3, 3 5 and 5 10 year yearly attrition rates increase by about 1% for each subsequent episode in both black and nonblack patients. This phenomenon is neither evident in the death-censored data nor in low-risk deceased-donor transplants namely with both donor and recipient age below 45 years. This means that the mild increase in attrition rates is probably driven by death rather than graft loss. In fact when looking at the death-censored attrition rates there is no increase in the attrition with increasing time posttransplant. When comparing attrition rates between populations the most striking difference is between living- and deceaseddonor recipients. Most of this difference is based on the well-known better short-term outcomes in living-donor transplants. Long-term attrition rates are also lower in living-donor transplants but only in nonblack patients. In black patients there seems to be a higher long-term attrition rate in living-donor transplants compared to nonblack 458 American Journal of Transplantation 2011; 11:

10 Long-Term Renal Allograft Survival Table 4: Kaplan Meier estimates of cumulative graft attrition rate by transplant year for first-transplant deceased donor Transplant years All recipients Year Years Years Years Recipient and donor age below Year Years Years Years Nonblack/black Year 18.3/ / / / / / / / / / / /13 8.3/ /10.4 8/ / / /9 6.4/ / Years 11.8/ / / /9.9 5/ / / /8 4.5/ /7 4.1/ / / / / / / / Years 5.7/ / / / / / /10 6/ / /9.8 6/ / / / / / Years 6.8/10.7 7/ / / /9 6.7/ / / / / /7.7 Recipient and donor age below Year 17.3/ / / / /16 12/ / / /9 6.2/ / /9.7 6/ / / / / / / / Years 6/ / /10.5 5/9.9 4/ / / / / / /7.9 4/ /7.1 4/ /8.1 4/ / / Years 5.1/ / / / / / / / / /9 4.3/ / / / / / Years 6.1/ /9 6.3/8.4 6/ / / /8.7 5/ / /7.8 5/6.3 1 Nonblack recipient half-lives listed before hash mark followed by black recipients after the hash mark. American Journal of Transplantation 2011; 11:

11 Lamb et al. Table 5: Kaplan Meier estimates of death-censored graft attrition rate by transplant year for first-transplant deceased donor Transplant years All recipients 0 1 Year Years Years Years Recipient and donor age below Year Years Years Years Nonblack/black 0 1 Year 13.8/ / / / / / /9.6 7/9 6.6/ /7.9 6/ / / / / / / / / / Years 4.2/ / / / / / / / / /4.9 2/ / / / / /5.2 2/4.8 2/ Years 3.2/ / / / / / / / / / / / / / /4.8 3/ Years 3.7/ / / / / / / /6 3.2/ / /5 Recipient and donor age below Year 14.4/ / / / / / / / / / / / /6.1 5/6.4 5/ / / /5 3.5/ / Years 4.7/ / / / / / / /6.9 3/ / /7.2 3/7 2.9/ / / / / /6 3 5 Years 3.6/ / / / /7.1 4/ / / /8 4.2/ /8 3.5/7 2.7/ / / / Years 4.1/ / / /7.5 4/ / / / / / /5.3 Nonblack recipient half-lives listed before hash-mark followed by Black Recipients after the hash mark. 460 American Journal of Transplantation 2011; 11:

12 Long-Term Renal Allograft Survival Table 6: Kaplan Meier estimates of cumulative graft attrition and death-censored graft attrition rate by transplant year for first transplant living-donor nonblack versus black Transplant years All recipients graft failure 0 1 year year year year Nonblack/black 0 1 year 7.8/ / / / / /10 7.1/ / / / / / / / / / / /5 3/4.1 3/6 1 3 year 3.4/ / / / / / / / / / / / /6 3.1/ / / / year 3.9/ / / / / / / /8 4.3/ / / / / / / / year 5.1/ / / / / /9 5.4/ /7.8 5/ /6.6 5/6.7 All recipients death censored graft failure 0 1 year year year year Nonblack/black 0 1 year 6.2/ / / /10.8 5/ / / / / / / / /4.6 3/ / / / / /3 2/ year 2.2/ /5.3 2/ / / / / / / / / / / / / / / / year 2.7/ / / /6 2.7/7 3.4/ /6 2.9/ / / / / /5.5 2/4.7 2/ / year 3.2/ / / / / / / / /5.7 3/ /5 Nonblack recipient half-lives listed before hash mark followed by black recipients after the hash mark. American Journal of Transplantation 2011; 11:

13 Lamb et al. deceased-donor transplants and similar to black deceaseddonor transplants. Significant progress has occurred over the decades in renal transplantation and is mostly driven by improvements in short-term graft and patient survival. Further improvement, which now has to come mainly from long-term survival improvements, has been more difficult to achieve. The multifactorial nature of chronic renal allograft loss (9) makes specific interventions for populations difficult. Increased immunosuppression has decreased acute rejection rates but led to more graft loss driven by opportunistic infections or over-immunosuppression (6), thus keeping long-term graft loss a constant phenomenon not only in the United States but also in other countries (10). Immunosuppression minimization strategies as a population approach to deal with toxicities are likely to incur the same problems (11). Despite these obvious difficulties and in the absence of any good tools to individualize immunosuppression to each patient, modest but measurable progress has occurred in long-term graft attrition resulting ultimately in longer kidney allograft half-lives in the United States. Acknowledgment We would like to extend our appreciation to the Central Florida Kidney Center, Inc. for supporting this work through the endowment of the Eminent Scholar Chair in Nephrology and Hypertension. We would also like to thank Melissa Smiles for editing and proof reading the manuscript. The authors do not have any conflicts of interest or disclosures with regards to the data presented in this manuscript. Disclosure The authors of this manuscript have no conflicts of interest to disclosure as described by the American Journal of Transplantation. References 1. Murray JE, Merrill JP, Harrison JH, Wilson RE, Dammin GJ. Prolonged survival of human-kidney homografts by immunosuppressive drug therapy. N Engl J Med 1963; 268: Hricik DE, Halbert RJ, Barr ML et al. Life satisfaction in renal transplant recipients: preliminary results from the transplant learning center. Am J Kidney Dis 2001; 38: Wolfe RA, Ashby VB, Milford EL et al. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med 1999; 341: Kaplan B, Meier-Kriesche HU. Death after graft loss: An important late study endpoint in kidney transplantation. Am J Transplant 2002; 2: Meier-Kriesche HU, Schold JD, Kaplan B. Long-term renal allograft survival: Have we made significant progress or is it time to rethink our analytic and therapeutic strategies? Am J Transplant 2004; 4: Meier-Kriesche HU, Schold JD, Srinivas TR, Kaplan B. Lack of improvement in renal allograft survival despite a marked decrease in acute rejection rates over the most recent era. Am J Transplant 2004; 4: OPTN/SRTR Annual Report HHS/HRSA/HSB/ DOT; UNOS; Arbor Research Collaborative for Health. Available at: Accessed June 15, Eckhoff DE, Young CJ, Gaston RS et al. Racial disparities in renal allograft survival: a public health issue? J Am Coll Surg 2007; 204: Pascual M, Theruvath T, Kawai T, Tolkoff-Rubin N, Cosimi AB. Strategies to improve long-term outcomes after renal transplantation. N Engl J Med 2002; 346: McDonald S, Russ G, Campbell S, Chadban S. Kidney transplant rejection in Australia and New Zealand: Relationships between rejection and graft outcome. Am J Transplant 2007; 7: Srinivas TR, Meier-Kriesche HU. Minimizing immunosuppression, an alternative approach to reducing side effects: Objectives and interim result. Clin J Am Soc Nephrol 2008; 3(Suppl 2): S101 S American Journal of Transplantation 2011; 11: