Individuals of Aboriginal ethnicity had a 3.26 times higher risk of graft failure (95% CI: 1.51–7.04) compared with whites. Deceased donor transplant was associated with worse graft survival compared with living donor transplant (adjusted HR 2.25 [95% CI: 1.32–3.83]). The risk of graft failure did not differ significantly by gender. Compared with observed experience at the age of younger than 13 years, the graft failure rate was 2.26 [95% CI: 1.04–4.93] times higher during experience at the age of 18 years or more, after adjustment for all other covariates, including time since transplant (which was the timescale of the model).
Transfer to adult care occurs during adolescence or young adulthood—a potentially turbulent time in which nonadherence to medications and decreased attendance at clinic visits may be more likely (13, 15, 18–20). Patients undergoing transfer must adapt to a new reality in adult- oriented care; adaptation during the vulnerable adolescent period may be particularly challenging. We demonstrate a progressively higher risk of graft failure with increasing age in the interval between birth and young adulthood, with a significantly higher risk associated with age older than 18 years compared with 0 to 13 years.
Even after accounting for age-related failure risk, we found that the adaptation period, defined as a 3-year interval beginning 6 months before the first-recorded visit to an adult center, was associated with a twofold increase in the risk of graft failure, compared with the period before adaptation for pediatric renal transplant patients in Canada. The magnitude of the relative hazard of graft failure associated with adaptation was even higher when we considered shorter adaptation intervals, suggesting that the risk is concentrated in the early posttransfer period. These observations call for implementation of additional support during the transition from pediatric to adult care, with most intense support immediately after transfer. Alternate models of care should also be considered; these include flexible age of transfer (to accommodate individual patient needs), a clinic devoted to transitional care with participation of both pediatric and adult care providers, and an interval during which care is shared between pediatric and adult care providers.
No prior studies considered the effects of both transfer of care and age on the risk of graft failure. A retrospective study from Ontario, Canada, found little difference in graft failure rates by age during adolescence (graft loss rates per 100 person-years were 5.06 between 14.1 and 17.9 years, 5.63 between 18.1 and 19.9 years, and 5.80 between 20.1 and 23.9 years) and concluded that there was no evidence for higher failure rates during transition from pediatric to adult-oriented care (21). In contrast, a U.S. Government Accountability Office report found that a higher percentage of transplant recipients who were younger than 18 years at the time of transplant and 18 years or older at the time of study experienced graft failure at 3, 5, and 7 years posttransplant than those who had not yet reached 18 years or who had received a transplant at the age of more than 18 years (22). However, neither this nor the Ontario study identified a transfer date; therefore, neither can make specific conclusions about the effect of transfer. Nonetheless, our results are consistent with the findings of the Government Accountability Office report suggesting that patients negotiating the transition from childhood to adulthood—and all that it entails, including transfer of care—are at high risk for graft failure.
Acute rejection rates were substantially and significantly lower after transfer than before (HR 0.10 [95% CI: 0.04–0.28] for Dutch born and 0.69 [95% CI: 0.33–1.40] for immigrant patients) in a recent study of pediatric renal transplant recipients (16). It is difficult to reconcile this large protective effect of transfer with the substantially increased risk of graft loss found in our study. It is possible that the lower posttransfer acute rejection rates were due to less-intense surveillance for rejection during the adaptation period. Adjustment for age as a continuous and non-time-dependent variable in the Dutch study may not have permitted complete control for the effects of age (16).
Although age at transfer of care may vary between centers, the median age at first adult visit in our study was 18 years, with a tight IQR, indicating that transfer occurs within a narrow age range in Canada. We are unable to account for differing transitional care practices across the country in this is registry-based study. Care patterns may vary across centers depending on regional health service delivery, pediatric and adult clinic staff availability, and patient volume. Furthermore, transitional care practices have likely changed over time. We acknowledge that there may be overlapping pediatric and adult care in some centers for part of the adaptation period.
An important strength of this study is the use of a longitudinal population-based cohort from the sole prospective Canadian organ failure registry within a universal access healthcare system. The cohort represents almost all pediatric renal transplants in Canada during the study interval, and as such it has greater generalizability. However, caution should be used in generalizing our results to other countries with different healthcare delivery systems.
This study has some limitations. First, because of the narrow age range at first visit to adult care, there was relatively little representation of the younger age intervals within the adaptation period and after adaptation. Similarly, there was little representation of the older age intervals before adaptation. As a result, it is possible that adjustment for age in the association between adaptation and graft failure risk was incomplete. Regardless of whether the age and adaptation effects are independent of each other, we can conclude that the interval around transition to adult-oriented care is a high-risk period. Second, as was performed by van den Heuvel et al. (16), we used the first visit to adult center to define the adaptation period. However, the first visit to adult center recorded in the registry may not have accurately reflected the first actual visit, that is some patients may have had visits to adult centers and had went unreported before a visit was recorded. This would result in a shift of the defined adaptation period to a later time in some individuals. Given that the adaptation period was quite broad—a 3-year time interval beginning 6 months before first adult ESRD center visit—the impact of small errors in transfer date is likely minimal. Because discrete rejection episodes are not captured within CORR, we could not evaluate the effect of transfer of care on rejection rates. Finally, we were unable to assess potential mechanisms for higher graft loss rates during the adaptation period, such as adherence to medications and clinic visits, because this information is not captured by CORR. Future retrospective studies with examination of primary medical charts, or prospective studies enrolling patients before the transition period, will be necessary to determine the reasons for graft loss during the transition period.
This study demonstrates a significantly higher risk of renal graft loss during the adaptation period between pediatric and adult care in pediatric renal transplant patients in a universal access healthcare system. The findings presented here support the development and implementation of programs to support young transplant recipients after transfer from pediatric to adult care. Future studies will identify determinants of risk during the adaptation period, permitting the development of targeted interventions with the goal of improving long-term graft survival among pediatric renal transplant recipients.
MATERIALS AND METHODS
Study Design and Population
After obtaining institutional ethics approval from the University of Calgary, we performed a retrospective cohort study of all patients in 9 of the 10 Canadian provinces and three territories who initiated renal replacement therapy (dialysis or transplant) at younger than 18 years of age between January 1, 1992, and December 31, 2007, who received their first transplant in a pediatric center, and who had at least 3 months of graft function (Fig. 1). Patients from the province of Quebec were not included as their data were not available for release to investigators. Patients were followed up from date of transplant until outcome (graft failure), death, loss to follow-up, or end of observation (December 31, 2007).
The CORR is the sole national organ failure registry in Canada and is maintained by the Canadian Institute for Health Information. All pediatric and adult dialysis and transplant centers in Canada submit data every 6 months to CORR using standardized forms. CORR data elements include demographics, ESRD treatment modalities, changes in treatment modalities, renal transplant details, comorbidities, and outcomes. The registry data have been used extensively in multiple studies in patients with kidney failure (23–27). Almost all renal transplants in Canada (98.5%) are captured by CORR; demographic data are coded with high reliability (28).
We excluded pediatric patients for whom a clear pattern of care (i.e., consistent care at a pediatric center followed by consistent care at an adult center) could not be ascertained. Multiorgan transplant recipients were also excluded.
Graft failure is reliably and consistently reported in the registry. The primary exposure of interest was the period of adaptation associated with transfer to an adult care environment. We defined adaptation as a 3-year time period beginning 6 months before the first visit to an adult healthcare center (adaptation period is 0.5 years before transfer date to 2.5 years after transfer date). The date of first visit to an adult center is recorded in CORR. This date will not capture missed appointments; therefore, we included a 6-month time interval before first adult clinic visit to allow capture of missed appointments and to minimize the impact of any delays in reporting first visit at an adult center. A 3-year adaptation period is consistent with what has been considered in prior studies and should be sufficiently long to allow adaptation to a new care environment (16). We also examined alternate definitions of the adaptation period (0.5 years before to 0.5 years after transfer and 0.5 years before to 1.5 years after transfer) and the results are presented. Potential confounders in the relationship between graft failure and the adaptation period include age, donor source (living or deceased), gender, ethnic origin (white, Aboriginal, black, Asian/Indian/other), primary renal disease, and socioeconomic status.
The primary outcome was graft failure.
Demographic and baseline clinical characteristics were described with medians (IQR) or proportions, as appropriate. A significance level α equal to 0.05 was used in all statistical tests.
Crude age-specific graft failure rates were calculated and plotted for each 3-year age interval from 0 to less than 22 years of age by adding all graft failures observed within a 3-year age interval and dividing by the total person-time observed within that age interval. Therefore, patients could contribute observation time to multiple age intervals. Patients were censored at death or loss to follow-up.
We used Cox proportional hazards models, with time from transplant as the timescale, to estimate the excess graft failure risk associated with being in the adaptation period, compared with the periods before and after adaptation; a three-level, time-dependent adaptation variable was included in the model. All patients started observation in “before adaptation.” Six months before the first adult care visit, they switched to “within adaptation,” where they remained until 2.5 years later (in the primary analysis), when they entered the “after adaptation” period. Additional models in which the adaptation period ended 0.5 years, and 1.5 years after the first adult care visit were also fitted. We adjusted all models for the potentially confounding effect of current age by entering age in the model as a time-dependent variable: younger than 13 years, 13 years to younger than 15 years, 15 years to younger than 18 years and 18 years or older. The models were also adjusted for gender, donor source, ethnicity, primary renal disease, and socioeconomic status. Socioeconomic status, estimated from median neighborhood income, was classified by quintile within the 2001 Statistics Canada census data by linking residence postal codes to census data. Retention of covariates in the final models was determined by statistical significance and minimizing of Akaike's (29) Information Criterion in nested models. Using this model-selection process, socioeconomic status and causes of primary renal disease were not included in the final models. The proportional hazards assumption in all models was tested based on the scaled Schoenfeld residuals (30). All analyses were performed using R (The R Foundation for Statistical Computing, http://cran.r-project.org).
The authors thank the director and personnel of the Canadian Organ Replacement Register at the Canadian Institute for Health Information (Dr. John Gill, Lilyanna Trpeski, Yingbo Na, Robert Williams) for their assistance in providing the dataset used in this study.
1. Freed GL, Hudson EJ. Transitioning children with chronic diseases to adult care: Current knowledge, practices, and directions. J Pediatr
2006; 148: 824.
2. Frank M. Factors associated with non-compliance with a medical follow-up regimen after discharge from a pediatric
diabetes clinic. Can J Diabetes Care
1996; 20: 13.
3. Rosen DS, Blum RW, Britto M, et al. Transition to adult health care for adolescents and young adults with chronic conditions: Position paper of the Society for Adolescent Medicine. J Adolesc Health
2003; 33: 309.
4. Smith JM, McDonald RA. Renal transplantation in adolescents. Adolesc Med Clin
2005; 16: 201.
5. 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.
6. McDonald R, Ho PL, Stablein DM, et al. Rejection profile of recent pediatric renal transplant
recipients compared with historical controls: A report of the North American Pediatric Renal Transplant
Cooperative Study (NAPRTCS). Am J Transplant
2001; 1: 55.
7. Magee JC, Krishnan SM, Benfield MR, et al. Pediatric
transplantation in the United States, 1997–2006. Am J Transplant
2008; 8: 935.
8. Foster B. The impact of transfer from a pediatric
to an adult care centre on graft survival. J Am Soc Nephrol
2008; 19: 229A.
9. Ettenger RB, Rosenthal JT, Marik JL, et al. Improved cadaveric renal transplant
outcome in children. Pediatr Nephrol
1991; 5: 137.
10. Penkower L, Dew MA, Ellis D, et al. Psychological distress and adherence to the medical regimen among adolescent renal transplant
recipients. Am J Transplant
2003; 3: 1418.
11. Dobbels F, Van Damme-Lombaert R, Vanhaecke J, et al. Growing pains: Non-adherence with the immunosuppressive regimen in adolescent transplant recipients. Pediatr Transplant
2005; 9: 381.
12. Ferris ME, Gipson DS, Kimmel PL, et al. Trends in treatment and outcomes of survival of adolescents initiating end-stage renal disease care in the United States of America. Pediatr Nephrol
2006; 21: 1020.
13. Shaw RJ, Palmer L, Blasey C, et al. A typology of non-adherence in pediatric renal transplant
recipients. Pediatr Transplant
2003; 7: 489.
14. Watson AR. Non-compliance and transfer from paediatric to adult transplant unit. Pediatr Nephrol
2000; 14: 469.
15. Watson AR. Problems and pitfalls of transition from paediatric to adult renal care. Pediatr Nephrol
2005; 20: 113.
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. Kramer A, Stel VS, Tizard J, et al. Characteristics and survival of young adults who started renal replacement therapy during childhood. Nephrol Dial Transplant
2009; 24: 926.
18. Magee JC, Bucuvalas JC, Farmer DG, et al. Pediatric
transplantation. Am J Transplant
2004; 4(suppl 9): 54.
19. Morgenstern BZ, Murphy M, Dayton J, et al. Noncompliance in a pediatric renal transplant
population. Transplant Proc
1994; 26: 129.
20. Meyers KE, Thomson PD, Weiland H. Noncompliance in children and adolescents after renal transplantation. Transplantation
1996; 62: 186.
21. Koshy SM, Hebert D, Lam K, et al. Renal allograft loss during transition to adult healthcare services among pediatric renal transplant
2009; 87: 1733.
22. Office USGA. End-stage renal disease characteristics of kidney transplant recipients, frequency of transplant failures, and cost to medicare. Office USGA 2007. Available at: http://www.gao.gov/
23. Tonelli M, Hemmelgarn B, Culleton B, et al. Mortality of Canadians treated by peritoneal dialysis in remote locations. Kidney Int
2007; 72: 1023.
24. Tonelli M, Hemmelgarn B, Gill JS, et al. Patient and allograft survival of Indo Asian and East Asian dialysis patients treated in Canada. Kidney Int
2007; 72: 499.
25. Tonelli M, Hemmelgarn B, Kim AK, et al. Association between residence location and likelihood of kidney transplantation in Aboriginal patients treated with dialysis in Canada. Kidney Int
2006; 70: 924.
26. Tonelli M, Hemmelgarn B, Manns B, et al. Death and renal transplantation among Aboriginal people undergoing dialysis. CMAJ
2004; 171: 577.
27. Tonelli M, Klarenbach S, Manns B, et al. Residence location and likelihood of kidney transplantation. CMAJ
2006; 175: 478.
28. Canadian Institute for Health Information. Health services databases. Available at: http://www.cihi.ca
, 2010. Accessed December 17, 2010.
29. Akaike H. Modern development of statistical methods. New York, Pergamon Press 1981.
30. Grambsch P, Therneau T. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika
1994; 81: 515.
Keywords:© 2011 Lippincott Williams & Wilkins, Inc.
Pediatric; Renal transplant; Transfer to adult care; Allograft failure