A rapid growth in kidney transplantation (KTX) over the last five decades and the shortage of donor kidneys relative to the demand necessitated the development of deceased donor (DD) allocation criteria (1). Donor age has long been recognized as a key determinant of allograft outcome (2, 3). Higher donor age is associated with greater degrees of glomerular sclerosis, interstitial fibrosis, and reduced functioning nephron number. Thus, higher donor age is independently associated with worse allograft outcomes. Conversely, younger-age donor kidneys have better allograft outcomes. Nevertheless, the shortage of donor organs has led to the development of extended donor criteria to use older donors (4).
Children have been given long-standing societal preference for DD KTX, such that the Organ Procurement and Transplant Network (OPTN) in the United States had special allocation preferences built for pediatric waitlisted patients. However, the prior allocation criteria did not necessarily lead to the desired results of accomplishing DD KTX in children within 6 months after listing. Also, pediatric programs were often waiting for the better donor kidneys. Therefore, on October 2005, the OPTN introduced a new allocation policy that has become known as Share-35 (S35) (5). Under this policy, DD kidneys from donors ages <35 years are preferentially allocated to pediatric (ages <18 years) recipients after zero mismatch, multiorgan recipients, highly sensitized recipients, and local candidates who have been prior living donors (LD). The intent of this new policy (6, 7) was threefold: (a) to shorten the wait times for children to less than 6 months after listing, (b) to reduce the allocation of older donor kidneys to children who need to maximize the life of the allograft, and (c) to reduce the delay in the assignment of donor organs to adult recipients because many pediatric transplant programs routinely refuse to accept donor organs from older patients deemed as being of lower quality.
The subsequent results from the OPTN have revealed that waiting times for children did indeed shorten and that children received predominantly younger donor kidneys (6–8). Because pediatric transplants still form a small minority of all KTX, a substantial number of S35-KTX still went to adult recipients. We took this opportunity to evaluate the outcome from these S35 kidneys on graft survival in pediatric and adult recipients, further stratified by age, to assess which age group derives the best benefit from these young donor and presumably ideal kidneys.
A total of 18,461 S35-KTX were performed from October 2005 to November 2010. Of these, 12.54% (2315) were allocated to children and the remaining 87.46% (16,146) were allocated to adult recipients. Data on the study group demographics are presented in Table 1.
Before S35 implementation, 6.3% (2674 of 43,106) transplants from DD ages <35 years were allocated to pediatric recipients. After October 2005, the number of KTX performed from DD ages <35 years doubled. The number of DD ages <35 years per year was stable between 3500 and 4000 per year between 1994 and 2009, so the absolute number of DD ages <35 years did not increase, and only the proportion going to pediatric recipients changed.
Effect of Share-35 Kidney Transplants on Cumulative Graft Survival
The cumulative graft survival (CGS) from S35-KTX was compared between adult and pediatric recipients. As shown in Figure 1A, Kaplan-Meier (K-M) univariate estimates showed that S35-KTX to pediatric (ages <17 years) recipients had significantly worse CGS than S35-KTX to adult recipients (P=0.0027). At 4 years after transplantation, the pediatric recipient group had approximately 7% lower CGS than S35-KTX to adult recipients. Among adults, CGS was noted to be superior in middle-aged (ages 31–45 years) adults (Fig. 1B).
To analyze this result further, we then subdivided the pediatric age group into young children (ages <12 years) and adolescents (ages 13–17 years) because the adolescent age group has been associated with worse graft outcomes. Adolescents received 57.54% of all S35-KTX among pediatric recipients and indeed had worst CGS (Fig. 1B). In comparison with the middle-aged adults with the best CGS, adolescents had 12% lower CGS at 4 years after transplantation. In contrast, the K-M curve for young children was statistically not different from the CGS curve for adult recipients ages 31–45 years. Thus, pediatric recipients are not a single homogeneous entity but are actually a mix of two disparate populations (young children and adolescents) with divergent results.
We then fitted multivariate Cox models that analyzed for the effect of recipient age on CGS in association with other relevant variables. As shown in Table 2, S35-KTX to adolescents (ages 13–17 years) and young adults (ages 18–30 years) had significantly higher risk for cumulative graft failure than S35-KTX to adult recipients ages 31–45 years. In contrast, S35-KTX to children ages <12 years and older adults ages >45 years did not have significantly higher risk for cumulative graft failure than S35-KTX to adult recipients ages 31–45 years. These models took into account the degree of human leukocyte antigen (HLA) mismatch, which tends to be higher for S35-KTX than other DD KTX.
Because Black recipient race is a known risk factor for worse CGS, we then analyzed the proportion of Black recipients among different age groups. Blacks comprised 20.7% of the adolescent population, 16.7% of young children, 23.0% of recipients ages 18–30 years, 30.0% of recipients ages 31–45 years, and 23.7% of recipients ages ≥46 years. Thus, Black recipients were not disproportionately higher in representation in the age groups with worst CGS (ages 13–17 and 18–30 years).
We then stratified recipient age group for pediatric and adult recipients into smaller subgroups in the Cox multivariate models within S35-KTX to see if changes in risk could be localized further. We depict all the resultant adjusted hazard ratios (AHR) in Figure 2A. CGS in young children (ages <12 years) and adults (ages 26–30 and 46–55 years) was comparable (P=0.43, 0.76, and 0.84, respectively) with adult recipients ages 31–45 years with the best CGS, whereas CGS was poor among adolescents (ages 13–17 years; P=0.0013), young adults (ages 18–25 years; P<0.0001), and older adults (ages 56–65 and >65 years; P=0.02 and <0.0001).
Effect of Share-35 Kidney Transplants on Death-Censored Graft Survival
Because death is uncommon in children and young adults compared with older adults, we analyzed the effect of S35-KTX on death-censored graft survival (DCGS) to estimate the effect of death on graft failure in different recipient age groups. K-M univariate estimates showed that S35-KTX in children (ages ≤17 years) had significantly worse DCGS compared with S35-KTX in adults ages 31–45 years (Fig. 1C; P<0.01). At 4 years after transplantation, the S35-KTX in children had a 10% lower DCGS than adults ages 31–45 years.
We then further analyzed the data by splitting the pediatric population into younger children and adolescents, as we had done for CGS. When compared with adult recipients ages 31–45 years, the K-M estimate for DCGS was 15% lower for the ages 13–17 years recipient age group at 4 years after transplantation (Fig. 1D; P<0.001). In comparison with young children (ages ≤12 years), adolescents (ages 13–17 years) had 12% lower 4-year DCGS (P=0.0126). In contrast to our CGS results, the DCGS curve for young children was 3% lower than for adults ages 31–45 years; this difference reached statistical significance (P=0.0022).
We then fitted multivariate Cox models that analyzed for the effect of recipient age on DCGS in association with other relevant variables, the results of which are depicted in Table 3. In these adjusted multivariate Cox models, compared with S35-KTX to adult recipients ages 31–45 years, S35-KTX had the highest risk for death-censored graft failure in recipients ages 13–17 and 18–30 years. In this Cox model, children ages ≤12 years also had a significantly higher risk for death-censored graft failure. In contrast, the recipient ages ≥45 years had significantly lower risk for death-censored graft failure. These data pointed to the fact that children, adolescents, and young adults have reasons for graft failure unrelated to death.
We then stratified the recipient age group further into smaller intervals within the S35-KTX group in our multivariate models for DCGS, as we had done for CGS. DCGS was poor among adolescents (ages 13–17 years) and young adults (ages 18–25 years) and somewhat lower in young children compared with those ages 31–45 years. Among adults, DCGS improved significantly with each advancing recipient age group (ages 46–55, 56–65, and >65 years) as death risk rose in these age groups as shown in Figure 2B.
Our results highlighted a key and provocative finding that the CGS in recipients of kidneys from DD ages <35 years (presumably ideal kidneys) is affected by recipient age. CGS is actually worst in certain young age groups that included adolescents (ages 13–17 years) and young adults (ages 18–25 years) and is better in children (ages <12 years) and middle-aged adults (ages 31–45 years). These results were sustained in multivariate models. The detriment in allograft survival in adolescents and young adults was not accounted for by traditional (donor, recipient, and immunosuppression-related) factors within our model. The effect of acute rejection (AR) as a cause of graft failure was not studied because data on AR are not complete in the OPTN database after 1 year of transplantation. However, DCGS showed poor graft survival in adolescents and young (ages 18–25 years) adults, pointing to the cause of allograft failure to factors unrelated to death. CGS in older adults (ages >55 years) was lowered due to death. Recent allocation concepts (9, 10) in adults assume that the maximal utility is obtained from donor and recipient age matching, which may not be the case as indicated in our study. These concepts do not take into account the “at-risk” age adults ages between 18 and 25 years.
Much of the detriment in children seems to occur within the adolescent age group, an age group well characterized for higher rates of medication nonadherence (11, 12). Nonadherence to posttransplantation care is a prevalent problem in adolescents particularly of an older age and greater years after transplantation. AR is identified as a significant consequence of medication nonadherence and subsequent graft failure, except in a subgroup with presumed graft tolerance (12). Our study shows that, among pediatric age group, adolescents continue to do poorly despite receiving ideal kidneys and additional strategies are needed to improve graft survival.
Not much attention has been paid in the past to young adults ages between 18 and 25 years, which is a novel finding. Our data suggest that adults ages 18–25 years experience the same problems as adolescents. It is not until at ages 26–30 years that improvements in CGS are seen. Poor CGS in young adults may be related to lack of maturity and responsibility, as recent brain maturation data have shown that young adult brains are still undergoing maturation that may be responsible for nonadherence in them just like adolescents (13).
The S35 policy was meant to provide children (ages <17 years) with better access to young DD (presumably ideal kidneys) to give them advantage of maximum graft years. Our date and recent data confirm (8) the increase in access to transplantation in pediatric recipients, a desired outcome of the S35 policy. However, adolescents (ages 13–17 years) due to their unique characteristics are not benefiting adequately from this policy. Levine et al. also found lower graft survival in adolescents from ideal kidneys (14). The question of whether the needs of adolescents could be equally served by allocation of ages 35–50 years kidneys is worth investigating. In an analysis of UNOS data, Dale-Shall et al. (15) showed that 5-year graft survival in children was somewhat lower at 64.9% from DD ages 35–50 years compared with 68.9% from those ages 18–34 years. In their study, they did not differentiate between children and adolescents, a question that is worth exploring.
Furthermore, a decrease in LD was observed for children of all ages and races with implementation of S35 policy (8). In a prior analysis of the UNOS database, Hardy et al. (16) analyzed the long-term graft survival in children from LD versus DD restricted to ages <35 years. Their date was obtained from 1994 to December 2006, which is largely before the implementation of S35. In their analysis, adjusted graft survival with LD KTX was clearly superior to DD KTX ages <35 years (AHR, 1.22; P<0.001), translating into a 12% better 5-year survival. These data suggested that, in the long-term, ideal DD KTX ages <35 years were not equivalent to LD. Future studies should compare the long-term graft survival of all pediatric recipients (including both LD and DD) before and after S35 implementation policy to assess the full impact of S35 policy on allograft survival.
In summary, this large-scale national registry analysis, 4 years after implementation of the S35 policy, demonstrates that the best CGS from a young donor ages <35 years is seen in middle-aged adults and in children ages <12 years even after accounting for more HLA mismatches and other factors. CGS was lower in adolescents and young adults due to higher graft loss (presumably due to AR and other factors) and in older recipients due to higher death rates. From a pure utilitarian point of view, the allocation of S35-KTX preferentially to children ages between 13 and 17 years may not lead to the most optimum allograft usage. We are speculating that this age group and young adults may do equally well with less ideal kidneys, which needs to be confirmed in future studies. Physicians responsible for the care of pediatric patients with end-stage renal disease should continue to encourage and pursue living donation, although it is easier to get DD for them with the S35 policy. Similarly, everyone caring for adolescents and young adults (ages 18–25 years) should work hard at improving adherence in them for maximizing graft survival, as these are very vulnerable ages for graft loss.
The limitations of our study are the limitations of all transplant databases reported, including the retrospective nature of the data collection and the relative drop in completeness of reporting at later posttransplantation time points. This latter limitation precluded an accurate assessment of the contributions of AR to the results. We also could not assess the function of these kidneys through estimated glomerular filtration rate assessments, which may also impact long-term graft survival. In this study, the time period after transplantation is still relatively short and may not reflect longer-term results, although the time period is longer than in prior analyses of S35 results and a reversal of the CGS slopes is extremely unlikely.
MATERIALS AND METHODS
First, we extracted data on the DD KTX that were from donors ages <35 years reported to the OPTN national transplant database in the United States from the master database going back to 1987 to assess if more of these organs were being allocated to pediatric recipients after the implementation of S35 policy compared with the years prior.
Thereafter, we analyzed the effect of S35 kidneys on allograft survival using data reported to the OPTN national transplant database in the United States from October 2005 to November 2010. CGS, DCGS, and patient survival were calculated using K-M estimates. Differences between survival curves were tested for significance by log-rank method after the confirmation of proportional hazards assumptions. Cox proportional hazards models were fitted to determine the AHRs for outcomes from S35-KTX in pediatric and adult recipients further stratified by recipient age. The S35-KTX were limited to DD ages ≤35 years, single-donor/single-organ transplants, and non–donation after cardiac death organs. The other variables fitted into the Cox models included donor demographics (sex, race, and cause of death), recipient demographics (sex and race), transplant factors (dialysis in the first week; pump use or not; HLA match at A, B, and DR loci; and percent panel reactive antibody), and initial immunosuppression. The level of significance was set at P<0.05, with Bonferroni corrections as appropriate for simultaneous multiple comparisons of different recipient age groups. This study was conducted after obtaining an approval from the institutional review board of the University of Florida under an exempt category approval of deidentified Scientific Registry of Transplant Recipients data analyses.
1. Hauptman PJ, O’Connor KJ. Procurement and allocation of solid organs for transplantation. N Engl J Med
1997; 336: 422.
2. Schold JD, Kaplan B, Baliga RS, et al. The broad spectrum of quality in deceased donor kidneys. Am J Transplant
2005; 5: 757.
3. Schold JD, Kaplan B, Chumbler NR, et al. Access to quality: evaluation of the allocation of deceased donor kidneys for transplantation. J Am Soc Nephrol
2005; 10: 3121.
4. Port FK, Bragg-Gresham JL, Metzger RA, et al. Donor characteristics associated with reduced graft survival: an approach to expanding the pool of kidney donors. Transplantation
2002; 74: 1281.
5. Allocation of deceased kidneys. 2010 [cited 2011 January 12]; Available from: http://optn.transplant.hrsa.gov/PoliciesandBylaws2/policies/pdfs/policy_7.pdf
6. Agarwal S, Oak N, Siddique J, et al. Changes in pediatric renal transplantation after implementation of the revised deceased donor kidney allocation policy. Am J Transplant
2009; 5: 1237.
7. Gritsch HA, Veale JL, Leichtman AB, et al. Should pediatric patients wait for HLA-DR-matched renal transplants? Am J Transplant
2008; 10: 2056.
8. Amaral S, Patzer RE, Kutner N, et al. Racial disparities in access to pediatric kidney transplantation since Share 35. J Am Soc Nephrol
2012; 6: 1069.
9. Hippen BE, Thisthlethwaite R, Ross LF. Risk, prognosis, and unintended consequences in kidney allocation. N Engl J Med
2011; 364: 1285.
10. Leichtman AB, McCullough KP. Improving the allocation system for deceased-donor kidneys. N Engl J Med
2011; 364: 1287.
11. Hsu DT. Biological and psychological differences in the child and adolescent transplant recipient. Pediatr Transplant
2005; 3: 416.
12. Berquist RK, Berquist WE, Esquivel CO, et al. Non-adherence to post-transplant care: prevalence, risk factors and outcomes in adolescent liver transplant recipients. Pediatr Transplant
2008; 2: 194.
13. Sowell ER, Thompson PM, Holmes CJ, et al. In vivo evidence for post-adolescent brain maturation in frontal and striatal regions. Nat Neurosci
1999; 10: 859.
15. Levine MH, Reese PP, Wood A, et al. Inferior allograft outcomes in adolescent recipients of renal transplants from ideal deceased donors. Ann Surg
2012; 255: 556.
15. Dale-Shall AW, Smith JM, McBride MA, et al. The relationship of donor source and age on short- and long-term allograft survival in pediatric renal transplantation. Pediatr Transplant
2009; 6: 711.
16. Hardy BE, Shah T, Cicciarelli J, et al. Kidney transplantation in children and adolescents: an analysis of United Network for Organ Sharing Database. Transplant Proc
2009; 5: 1533.