Adolescent age at transplant has long been recognized as a risk factor for graft failure. In 1997, analyses of the United Network for Organ Sharing data showed poorer 5-year graft survival among 13- to 21-year-old kidney transplant recipients compared with those younger at transplantation (1). This elevated risk for adolescent recipients was also observed in an analysis of the North American Pediatric Renal Trials and Collaborative Studies data (2); numerous independent analyses of United Network for Organ Sharing data have also shown that adolescent and young adult age at transplant is associated with a higher risk for graft failure than younger or older age (3–5). Investigators and clinicians speculate that nonadherence to immunosuppressive therapy—more common in adolescents than other age groups (6–8)—may be the most important factor contributing to poor graft survival in this age group. All previous analyses considering the association between age and graft survival focused on age at transplantation. However, all pediatric transplant recipients will eventually have to navigate adolescence and young adulthood. If poor adherence during adolescence and young adulthood is indeed the link between adolescent age and elevated graft failure risk, then it is likely that graft failure risk is elevated during this age interval for all kidney transplant recipients, regardless of age at transplant.
We hypothesized that graft failure rates would be higher in the adolescent and young adult years than at younger or older ages, irrespective of age at transplant. The aims were to determine age-specific death-censored graft failure rates among young (<40 years) recipients of a first renal transplant and to compare these failure rates across age categories, controlling for time since transplant and other potential confounders. Graft failure was defined as loss of graft function requiring return to dialysis (death censored). In additional analyses, we considered the outcome of death or loss of graft function.
In a retrospective cohort study, we evaluated graft outcomes in 90,689 individuals recorded in the United States Renal Data System (USRDS) database who received a first renal transplant in the United States when younger than 40 years, between January 1988 and October 2009; recipients were followed up until October 2009. Of these, 18,310 were younger than 21 years at transplant. We were most interested in those younger than 21 years at transplant; individuals between 21 and 39 years at transplant were included to allow comparisons between younger and older individuals with the same time since transplant. Individuals with graft failure on the same date as the transplant (n=404) were excluded. The median interval of observation was 5.9 years (interquartile range [IQR], 2.5–10.5 years; maximum 21.8 years). Demographic and disease characteristics are presented in Table 1. There were 31,857 graft failures, including 8861 deaths during 631,977 person-years of observation, of which 6555 graft failures (20.6%) and 1808 deaths (20.4%) occurred in the first year after the transplant. Failures within the first year of transplant occurred at a median of 54 days (IQR, 11–178 days) posttransplant. Among those younger than 21 years at transplant, the median age at graft failure occurring at any time posttransplant was 20 years (IQR, 16–23 years).
Crude Death-Censored Graft Failure Rates by Age
Figure 1(A) shows crude death-censored graft failure rates for all recipients, unadjusted for time since transplant, in 1-year age intervals between 0 and 40+ years. The high failure rates among children younger than 3 years reflect the fact that a large proportion of the observation time in this age interval is from the first posttransplant year, when surgical complications occur; there was very little observation time in children younger than 3 years beyond the first posttransplant year. Failure rates were fairly stable at approximately 2 to 3 per 100 person-years between 3 and 11 years of age, but were progressively higher with each 1-year increment in age after 11 years, peaking at 6.6 per 100 person-years among 19 year olds; failure rates then fell gradually with increasing age.
Because the association between age and failure risk may differ in the early posttransplant period compared with later (2), we also examined the association between age and failure rate among those with at least 1 year of graft function. Crude death-censored graft failure rates for recipients with at least 1 year of graft function, in 1-year age intervals, starting from the end of the first posttransplant year, are presented in Figure 1(B). Failure rates were fairly stable at about 1.4 per 100 person-years until 10 years of age, when the rates began to increase. Failure rates reached a maximum of 6.3 per 100 person-years at 19 years, after which there was a gradual decline.
Crude Death-Censored Graft Failure Rates by Time Since Transplant
Figure 2(A,B) shows crude graft failure rates by time since transplant, stratified by age at transplant. These plots demonstrate the association between age and graft failure risk, independent of age at transplant. Individuals younger than 5 years at transplant showed the highest failure rates 13 to 17 years posttransplant (when they were ∼17–21 years), whereas higher failure rates were observed in the years immediately after transplant among those transplanted during adolescence.
Adjusted Comparison of Failure Rates by Age
Table 2 shows the results of two time-dependent Cox models with time-varying covariates. The first includes all first graft recipients, observed since the transplant date, and estimates the relative hazards of death-censored graft failure within the following age categories compared with age 25 to 29 years: 0 to 4 years, 5 to 9 years, 10 to 12 years, 13 to 16 years, 17 to 20 years, 21 to 24 years, 30 to 34 years, 35 to 39 years, and 40 years and older. The second model includes only those with at least 1 year of graft function. Age category was entered as a time-dependent variable. The potential confounders listed in Table 2 were included in the model.
Compared with individuals observed at 25 to 29 years, death-censored graft failure rates were 20% higher for those with the same time since transplant observed between 17 and 20 years and between 21 and 24 years of age (P<0.0001). Failure rates were significantly lower among 5 to 16 year olds than 25 to 29 year olds: failure risk was 40% lower in 5 to 9 year olds (P<0.0001), 44% lower in 10 to 12 year olds (P<0.0001), and 9% lower in 13 to 16 year olds (P=0.01). Failure rates were also significantly lower among those 30 years and older than among 25 to 29 year olds. There was no significant difference in failure rates between 25 and 29 year olds and those younger than 5 years of age (P=0.5). When we used death or graft failure as the outcome, the hazard ratios (HR) for all age groups between 5 and 34 years were unchanged. Compared with 25 to 29 year olds, the HR for death or graft failure associated with age 0 to 4 years was 1.36 [1.17, 1.57] (P<0.0001), for age 35 to 39 years was 0.80 [0.77, 0.83] (P<0.0001), and for 40 years and older was 0.79 [0.76, 0.82] (P<0.0001).
Recipients With More Than or Equal to 1 Year of Graft Function
The results of this model were very similar to those for the model including the first year. Under these conditions, compared with individuals observed at 25 to 29 years of age, graft failure rates were 19% higher for those with the same time since transplant observed between 17 and 20 years and 22% higher for those observed between 21 and 24 years (both P<0.0001). Graft failure risk was substantially and significantly lower in all other age intervals than in 25 to 29 year olds (Table 2). Again, the results of a model with death or graft failure as the outcome were unchanged for age groups between 5 and 34 years. Compared with 25 to 29 year olds, the HR for death or graft failure associated with age 0 to 4 years was 0.41 [0.25, 0.66] (P=0.0002), for age 35 to 39 years was 0.68 [0.65, 0.72] (P<0.0001), and for 40 years and older was 0.62 [0.59, 0.65] (P<0.0001).
Other Independent Correlates of Graft Failure
As noted in previous studies (5,9), males were at significantly lower risk for graft failure than females. Also as previously noted, compared with whites, African Americans had a significantly higher risk of graft failure (3,4,9), and individuals of other races had a significantly lower risk of failure. Younger donor age was associated with a significantly lower risk of graft loss: for every 10 years younger the donor, the graft failure rate was 12% lower. Consistent with previous reports (10,11), failure rates were 26% to 29% lower in living donor versus deceased donor grafts (P<0.0001). Focal segmental glomerulosclerosis was associated with significantly higher failure risk, and “unknown” renal disease with a significantly lower risk than congenital anomalies of the kidneys or urinary tract. Also consistent with previous reports (12,13), compared with the lowest socioeconomic status (SES) quartile, each progressively higher SES quartile was associated with a progressively and significantly lower graft failure risk. Compared with those with 0 human leukocyte antigen (HLA) mismatch, all other levels of HLA mismatch were associated with significantly higher graft failure risks. The risk of graft failure was significantly and progressively higher in more remote eras compared with the most recent era (all P<0.0001).
Numerous studies have shown higher graft failure rates in individuals in their adolescent and young adult years at the time of kidney transplant than in individuals who were younger or older at the time of transplant (1–5). In this study, we show that the risk of graft failure increases with increasing age beyond approximately 11 years, peaking at 17 to 24 years of age, irrespective of age at transplant.
This study builds on the results of previous studies that examined graft survival by age at transplant (1–5). Our findings suggest that those in the adolescent and young adult age group at the time of transplant immediately enter a high-risk interval, explaining their poorer graft survival compared with those transplanted during a lower risk period (younger or older age). However, those transplanted before the high-risk adolescent and young adult age period must eventually traverse this interval—during which the risk of graft failure seems to be heightened.
There were some differences in the results of analyses including all recipients compared with the analyses including only those with at least 1 year of graft survival. Whereas there was no significant difference in graft failure risk among 0 to 4 year-olds compared with 25 to 29 year olds in the analysis considering all recipients, when we evaluated only those with at least 1 year of graft function, the failure risk among 0 to 4 year olds was significantly lower than that of 25 to 29 year olds. This difference is likely explained by the known higher risk of failure due to surgical complications in the immediate posttransplant period in very young recipients (1,2).
Our findings are consistent with the results of a US Government Accountability Office report, which showed that individuals negotiating the transition between childhood and adulthood (defined as those who were younger than 18 years of age at the time of transplant and 18 years and older a minimum of 4 years later) had poorer 3-, 5-, and 7-year graft survival than both younger and older individuals who had not negotiated this age interval (14). In contrast, a study of pediatric kidney transplant recipients in Ontario, Canada, found no significant difference in graft failure rates by age (15). However, there was no adjustment for time since transplant in this relatively small study.
We can only speculate on the reasons behind the observed age-related differences in graft failure risk. Poor immunosuppressive medication adherence and risk-taking behavior among adolescents and young adults have been proposed as mediators in the association between age and graft failure (1,2). Previous studies support this idea: the relationship between age and immunosuppressive medication nonadherence mirrors the relationship that we observed between age and graft failure risk (6–8). Two patterns of graft failure are possible: rapid failure due to irreversible acute rejection, or slow progression of graft dysfunction due to chronic rejection, numerous acute rejections, or other factors. There may be a lag between poor adherence and graft failure, such that poor adherence beginning in early adolescence results in progressive graft damage leading to graft failure in late adolescence and early adulthood. Although failure rates were slightly lower in 13 to 16 year olds than in 25 to 29 year olds, the increasing failure rates observed at ages more than 11 years (Fig. 1) suggest that the risk likely begins in early adolescence. In addition, some failures experienced by 25 to 29 year olds may be the result of accumulated damage during adolescence and early young adulthood.
It has been hypothesized that difficulties experienced by young people in adapting to the new care team following transfer from pediatric to adult-oriented care, may contribute to poor medication adherence, and therefore to higher graft failure rates in the late adolescent and young adult age interval (15–17). However, this is controversial and is difficult to prove or disprove. A Canadian study suggested that transition to adult care and age are independently associated with graft failure risk (18). Neither type of care facility (pediatric vs. adult) nor timing of transfer to adult-oriented care is recorded in the USRDS; therefore, the effect of transfer of care could not be evaluated in our analysis. Gaps in health insurance coverage, common among adolescents and young adults (19), may also play a role in the association between the adolescent and young adult age interval and higher graft failure rates (13,20). The USRDS records insurance status at the time of transplant, but does not capture changes in status over time. Therefore, we were unable to consider the potential effect of change in insurance status on graft failure rates.
We must also consider the possibility that age-related changes in the immune system may play a role in the higher risk of graft failure observed in adolescence and young adulthood. Exposure to a variety of viral infections over the course of childhood and adolescence may result in expanded alloreactivity, increasing the likelihood of rejection (21,22).
It is important to recognize that the HR for graft failure presented here are not completely independent of age at transplant. Individuals of different ages with the same time since transplant must have also been of different ages at transplant. For example, recipients who are 17 to 20 years 10 years posttransplant were 7 to 10 years at transplant, whereas recipients who are 25 to 29 years 10 years posttransplant were 15 to 19 years at transplant. Age at transplant partly explains the observed relationship between age and graft failure rates, because individuals may undergo transplantation during a high-risk period (i.e., adolescence). Our analysis accounts for the possibility that risks may change as age changes, instead of considering the static association between age at transplant and graft survival.
The results of this study should be generalized to kidney transplant recipients outside the United States with caution. A similar effect of age was observed in Canadian pediatric kidney transplant recipients (18). However, cultural differences, differences in transitional care practices, and differences in health insurance systems may result in differing associations between age and graft failure risk in other countries. This study highlights the substantial increase in graft failure risk during late adolescence and early adulthood, and emphasizes the importance of accounting for observed age (and not only age at transplant) in studies examining graft outcomes among young kidney transplant recipients. Furthermore, our findings suggest that adolescents and young adults may warrant more intense vigilance than recipients in other age groups. Protocol biopsies and more frequent clinic visits and surveillance blood work may be important ways of detecting poor adherence and evidence of graft rejection or dysfunction before accumulation of irreversible graft damage. Consideration should be given to intensified surveillance in this high-risk age interval.
MATERIALS AND METHODS
This was a retrospective cohort study of individuals recorded in the USRDS database who received a first renal transplant in the United States when younger than 40 years, between January 1988 and October 2009. The primary exposure variable was age, which was treated as a time-dependent variable (i.e., a variable that changes over time). No patient identifiers are included in the datasets.
Crude Graft Failure Rates by Age
For each 1-year age interval, failure rates, with 95% confidence intervals, were calculated (graft losses per 100 person-years of observation within that age interval). Age-specific failure rates were calculated first for all recipients, and then for recipients with at least 1 year of graft function. Individuals could contribute person-time to multiple 1-year age intervals.
Association Between Age and Graft Failure Rate
Time-dependent Cox models with time-varying covariates (23), matched on time since transplant, were used to estimate the additional graft failure risk associated with each of the age categories compared with the risk at 25 to 29 years, adjusted for potential confounders. The age interval 25 to 29 years was chosen as the reference to ensure that the reference age interval would include observation time from a wide range of times since transplant (i.e., at one extreme, individuals may have been transplanted at 1 year of age, reaching the reference age interval 24 years posttransplant, or, at the other extreme may have been transplanted at 26 years of age, within the reference interval).
Two distinct models were constructed. The first considered all recipients and all observation time after the first transplant, and the second considered only the observation time after the end of the first posttransplant year among those with at least 1 year of graft survival.
Multivariable analyses considered potential confounders with a known association with graft survival, including the following: sex, race, primary renal disease, era of transplant, SES, donor source, duration of dialysis before transplant, HLA mismatch, and donor age. Transplant era categories were based on changes in immunosuppression practices over time (24). SES, estimated using median household income by zip code, was classified by quartile within the U.S. Census data (1999) (13).
Data analysis was performed using SAS version 9.2 (SAS Institute, Cary, NC) and R (version 2.10.1); a P value less than 0.05 was considered statistically significant. The study received a waiver from the Montreal Children's Hospital Research Ethics Board due to the deidentified nature of the data.
1. Cecka JM, Gjertson DW, Terasaki PI. Pediatric
renal transplantation: A review of the UNOS data. United Network for Organ Sharing. Pediatr Transplant
1997; 1: 55.
2. Smith JM, Ho PL, McDonald RA. Renal transplant outcomes in adolescents: A report of the North American Pediatric
Renal Transplant Cooperative Study. Pediatr Transplant
2002; 6: 493.
3. Gjertson DW, Cecka JM. Determinants of long-term survival of pediatric
kidney grafts reported to the United Network for Organ Sharing kidney transplant registry. Pediatr Transplant
2001; 5: 5.
4. Hwang AH, Cho YW, Cicciarelli J, et al. Risk factors for short- and long-term survival of primary cadaveric renal allografts in pediatric
recipients: A UNOS analysis. Transplantation
2005; 80: 466.
5. Keith DS, Cantarovich M, Paraskevas S, et al. Recipient age and risk of chronic allograft nephropathy in primary deceased donor kidney transplant. Transpl Int
2006; 19: 649.
6. Dew MA, Dabbs AD, Myaskovsky L, et al. Meta-analysis of medical regimen adherence outcomes in pediatric
solid organ transplantation. Transplantation
2009; 88: 736.
7. Dobbels F, Ruppar T, De Geest S, et al. Adherence to the immunosuppressive regimen in pediatric
kidney transplant recipients: A systematic review. Pediatr Transplant
2010; 14: 603.
8. Pinsky BW, Takemoto SK, Lentine KL, et al. Transplant outcomes and economic costs associated with patient noncompliance to immunosuppression. Am J Transplant
2009; 9: 2597.
9. Omoloja A, Mitsnefes M, Talley L, et al. Racial differences in graft survival: A report from the North American Pediatric
Renal Trials and Collaborative Studies (NAPRTCS). Clin J Am Soc Nephrol
2007; 2: 524.
10. 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; 13: 711.
11. Hariharan S, Johnson CP, Bresnahan BA, et al. Improved graft survival after renal transplantation in the United States, 1988 to 1996. N Engl J Med
2000; 342: 605.
12. Goldfarb-Rumyantzev AS, Koford JK, et al. Role of socioeconomic status in kidney transplant outcome. Clin J Am Soc Nephrol
2006; 1: 313.
13. Woodward RS, Page TF, Soares R, et al. Income-related disparities in kidney transplant graft failures are eliminated by Medicare's immunosuppression coverage. Am J Transplant
2008; 8: 2636.
14. Office USGA. End-Stage Renal Disease Characteristics of Kidney Transplant Recipients, Frequency of Transplant Failures, and Cost to Medicare: Report to Congressional Requesters. 2007.
15. Koshy SM, Hebert D, Lam K, et al. Renal allograft loss during transition
to adult healthcare services among pediatric
renal transplant patients. Transplantation
2009; 87: 1733.
16. van den Heuvel ME, van der Lee JH, Cornelissen EA, et al. Transition
to the adult nephrologist does not induce acute renal transplant rejection. Nephrol Dial Transplant
2010; 25: 1662.
17. Watson AR. Non-compliance and transfer from paediatric to adult transplant unit. Pediatr Nephrol
2000; 14: 469.
18. Samuel SM, Nettel-Aguirre A, Hemmelgarn BR, et al. Graft failure and adaptation period to adult healthcare centers in pediatric
renal transplant patients. Transplantation
2011; 91: 1380.
19. Callahan ST, Cooper WO. Continuity of health insurance coverage among young adults with disabilities. Pediatrics
2007; 119: 1175.
20. Woodward RS, Schnitzler MA, Lowell JA, et al. Effect of extended coverage of immunosuppressive medications by medicare on the survival of cadaveric renal transplants. Am J Transplant
2001; 1: 69.
21. Dhanireddy KK, Maniscalco J, Kirk AD. Is tolerance induction the answer to adolescent non-adherence? Pediatr Transplant
2005; 9: 357.
22. Ford ML, Larsen CP. Overcoming the memory barrier in tolerance induction: Molecular mimicry and functional heterogeneity among pathogen-specific T-cell populations. Curr Opin Organ Transplant
2010; 15: 405.
23. Therneau TM, Grambsch PM. Modeling survival data: Extending the Cox model. New York: Springer 2000, pp 130.
24. Gulati A, Sarwal MM. Pediatric
renal transplantation: An overview and update. Curr Opin Pediatr
2010; 22: 189.