Journal Logo

Clinical and Translational Research

The Influence of Deceased Donor Age and Old-for-Old Allocation on Kidney Transplant Outcome

Moers, Cyril; Kornmann, Nirvana S.S.; Leuvenink, Henri G.D.; Ploeg, Rutger J.

Author Information
doi: 10.1097/TP.0b013e3181b0fa8b

Abstract

With an increasing number of patients on the waiting list, transplantation of kidneys from suboptimal donors has gained wide acceptance in most countries. Older donors, expanded criteria donation (ECD), and donation after cardiac death (DCD) have nowadays become important sources of kidney grafts (1–4). Various studies indicate that organs derived from such elevated-risk donors can be used successfully, provided that careful selection criteria are used. Nevertheless, in most large registry analyses advanced donor age remains one of the most important risk factors for inferior posttransplant outcome (5–10). Although a number of previous studies have superficially assessed the impact of donor age, detailed data on the extent to which a few years increase in donor age will influence early graft function and graft survival (GS) are scarce.

Within Eurotransplant, an international organ exchange organization in Europe, a well-established old-for-old allocation program exists since 1999 (11). In this kidney exchange program, 65+ deceased donor grafts are allocated to nonimmunized recipients of 65 years and older, using only ABO blood group matching and a policy to keep preservation times short. Results of this program and other senior-recipient organ exchange programs are encouraging, with a higher utilization rate for older donor kidneys, shorter waiting times for older patients, and reduction of the number of older patients on the waiting list. Overall long-term outcomes after transplantation seem not to be negatively affected by this policy (12).

Our current analysis focuses on the influence of deceased donor age on renal transplantation in the United States and addresses the question whether old-for-old allocation is safe. Aims of the study were to obtain regression models that show in detail the effect of donor age on short- and long-term outcome and to simulate kidney GS rates if an old-for-old kidney allocation program were implemented in the United States.

METHODS

Study Population

A January 10, 2007 extract of the Organ Procurement and Transplantation Network (OPTN) database was used. The study population consisted of deceased donor single- kidney recipients who were transplanted from January 1, 1994, to December 31, 2006. Only transplants from donors aged more than or equal to 11 years were included in the analysis. We chose 1994 as lower boundary of this cohort because several important variables were not collected before this year and also because postoperative care before this year would be too different from today's regimen. The upper limit was 2006 as database completeness for transplants performed thereafter was still too low at the time of analysis.

Endpoints

Endpoints for short-term outcome after kidney transplantation were delayed graft function (DGF) and primary nonfunction. DGF was defined as any dialysis requirement in the first week after transplantation. As reliable data on primary nonfunction could not be easily derived from the OPTN database, graft loss within 3 months posttransplant was used as a surrogate. GS up to 10 years posttransplant served as an endpoint for long-term outcome.

Statistical Method

Donor, transplant, and recipient demographics were calculated for the study cohort and plotted in graphs showing causes of death and the number of transplants per year. For each year between 1994 and 2006, kidney discard rate was visualized as the percentage of kidneys actually transplanted from all deceased donors recovered in various donor age categories. The correlation between donor age and recipient age was calculated by Pearson's method. A binary logistic regression model was used to identify independent donor-, preservation-, and recipient-related risk factors for DGF and for graft loss within 3 months posttransplant. Cox regression models examined which factors significantly contributed to the risk of graft failure and death with a functioning graft up to 10 years posttransplant (13). We used the Kaplan-Meier method to analyze death-censored GS in recipients. Univariate linear regression models were constructed with DGF, or 1-, 5-, and 10-year death-censored GS as dependent variable and donor age as independent variable. In the model for DGF, the data were split into patients who received a DCD kidney and those who received a graft derived from donation after brain death (DBD), because DCD has a well-documented independent effect on the incidence of DGF (14, 15).

We followed the approach outlined in Figure 4(a), and described in more detail online (see Supplemental Digital Content 1,https://links.lww.com/TP/A83), to simulate GS, as an old-for-old allocation program had been used in the time period studied. Old-for-old matching was performed following the Eurotransplant Senior Program (ESP) allocation rules: donor and recipient age more than or equal to 65 years, only recipients with no prior transplants, recipient panel-reactive antibodies less than or equal to 5%, no human leukocyte antigen (HLA) matching, and a policy to keep cold ischemic time (CIT) relatively short. For our old-for-old simulation, CITs of 65+ grafts were artificially reduced by a factor 12/19, thus, mimicking the effect observed in the ESP (16). For each existing or newly matched donor kidney+recipient combination, a theoretical GS time was calculated. Based on the shape of the actual baseline survival data points underlying a Cox model for graft failure in our dataset, we estimated that the baseline survival function would follow an exponential course

FIGURE 4.
FIGURE 4.:
(a) Schematic overview of old-for-old simulation methodology; (b) 10-year non-death-censored graft survival (GS) after simulation; (c) 10-year death-censored GS after simulation; (d) impression of model simulation accuracy: overall 10-year death-censored GS for the whole study cohort, based on real GS data, and on simulated values for the same cohort of recipients (n=99,860). The real, nonsimulated old-for-old survival curves in (b) and (c) should be regarded as reference only and cannot be compared with the simulated old-for-old curves. The reason is that in those 1011 real transplants human leukocyte antigen matching was performed, whereas in the simulation kidneys were not matched for human leukocyte antigen and cold ischemic times were artificially reduced by a factor 12/19. *=Newly matched donor graft+ recipient combinations.
FIGURE 4.
FIGURE 4.:
(Continued).

where t is time posttransplant. Values for a and c were derived by means of a least square fit to the baseline survival data points derived from this Cox model. Next, a survival function was obtained for each existing or newly matched combination

where bi is the ith regression coefficient, and xi is the value of the ith factor in the Cox model (13). From Eqs. (1) and (2), an equation for GS time (time-to-failure) of any donor kidney+recipient combination was derived

where s is a random number between 0 and 1 generated for each recipient, and T is the simulated time-to-failure for the graft.

Statistical analyses were conducted using SPSS and SigmaPlot software. Two-sided P values less than 0.05 were considered to indicate statistical significance.

RESULTS

Demographics

Between January 1, 1994, and December 31, 2006, 99,860 deceased donor single-kidney transplants from donors aged more than or equal to 11 years were performed in the United States. Table 1 shows basic demographic statistics for the study population. Figure 1(a) shows that in young donors, the leading cause of death was trauma, whereas in older donors death after a cerebrovascular accident was predominant. Between 1994 and 2006, the total number of kidney transplants per year from deceased donors increased by 39.5% (Fig. 1b). This increase came primarily from donors of age 35 years and older, and therefore the relative share of older donor kidney transplants has risen during these 13 years. There was no relevant correlation between donor and recipient age in this dataset (R2=0.05).

TABLE 1
TABLE 1:
Donor, recipient, and transplant demographics for the whole study cohort (n=99,860 deceased donor kidney transplants between 1994 and 2006) and for all kidney transplants performed from deceased donors aged 65 years and older in this same cohort (n=1011)
FIGURE 1.
FIGURE 1.:
(a) Distribution of deceased donor kidney transplants (cohort 1994–2006) per donor age, stratified into the two major causes of death, and (b) total number of deceased donor kidney transplants per year, distributed over five different donor age categories.

Kidney Discard

In Figure 2(a) relative deceased donor kidney usage is plotted as the percentage of kidneys that were actually transplanted from donors recovered each year. Thirty-five+ donor kidney usage decreased between 1994 and 2006, especially in kidneys derived from 50+ donors. Discard rates were higher for each subsequent donor age category, up to 36.9% and 66.9% for 50 to 64-year old, and 65+ donor kidneys, respectively, in 2006.

FIGURE 2.
FIGURE 2.:
(a) Relative deceased donor kidney usage: percentage of kidneys actually transplanted from all donors recovered each year, per donor age category, and (b) Kaplan-Meier plots of 10-year death-censored graft survival in recipients of deceased donor kidneys, stratified into five different donor age categories.

Risk Factors for DGF and Early Graft Loss

In a binary logistic regression model, all included factors were significant independent determinants of the risk for DGF, except for ECD donor versus non-ECD donor, cerebrovascular accident as cause of death in the donor and donor history of diabetes mellitus (Table 2). Donor age increased DGF risk with an odds ratio (OR) of 1.02 (P<0.0005), indicating that for each year increase in donor age, the relative risk for DGF in the recipient increased by 2%. Each additional year of donor age significantly increased the risk of early graft loss with an OR of 1.01 (P<0.0005).

TABLE 2
TABLE 2:
Multivariate risk analysis for delayed graft function, early graft loss, graft failure, and death with a functioning graft
TABLE 2
TABLE 2:
Continued

Risk Factors for Graft Failure and Death With a Functioning Graft

All factors included in the Cox model, except traumatic cause of death of the donor and the number of years the recipient spent on the waiting list, significantly influenced the risk of graft failure. Donor age increased the risk of graft failure with a hazard ratio of 1.01 (P<0.0005) for each subsequent year of age. Each year increase in donor age was also associated with a significantly higher risk of recipient death with a functioning graft (OR 1.004, P<0.0005).

Kaplan-Meier GS Analysis

Figure 2(b) shows that for each subsequent donor age category more than 11 to 34 years, GS up to 10 years was significantly lower (log-rank test, P<0.0005). GS was as low as 39% at 10 years posttransplant for 65+ donor grafts versus 70% for kidneys derived from donors aged 11 to 34 years.

DGF Risk as a Function of Donor Age

In Figure 3(a) the incidence of DGF is plotted with donor age as an independent variable, stratified into DBD and DCD donor kidneys. These univariate regression analyses show that the absolute risk of DGF increased by 0.35% and 0.37% with each year increase in donor age for DBD and DCD grafts, respectively (P<0.0005). DGF risk in DCD kidneys derived from donors aged 11 to 65 years was 17% to 18% thigher than in DBD grafts. Both effects were present for the whole spectrum of donor ages between 11 and 75 years (R2=0.90 for DBD recipients and R2=0.30 for DCD recipients).

FIGURE 3.
FIGURE 3.:
(a) The incidence of delayed graft function (DGF) per year of donor age, with fitted regression lines predicting DGF risk. Linear regression equations obtained from these data are as follows: (%DGF in DBD recipients)=0.35×(donor age)+10; (%DGF in DCD recipients)=0.37×(donor age)+27. In the univariate regression analysis for DGF, a quadratic (or even higher order) fit did not improve the R 2 to such an extent that it would be relevant to incorporate this more complex approximation. (b) Graft survival rates at 1, 5, and 10 years posttransplant per year of donor age, with fitted quadratic regression curves predicting graft survival. The equations obtained for these curves are as follows: (1-year GS)=−0.0039×(donor age)2+0.15×(donor age)+93 (R 2= 0.81, P<0.0005); (5-year GS)=−0.0061×(donor age)2+ 0.13×(donor age)+82 (R 2=0.86, P<0.0005); (10-year GS)= −0.0065×(donor age)2+0.0079×(donor age)+69 (R 2=0.91, P<0.0005). For this analysis, a quadratic fit yielded much better R 2 values than a linear fit.

GS as a Function of Donor Age

Figure 3(b) shows 1-, 5-, and 10-year death-censored GS rates as a function of donor age, for three different cohorts within the study population. Quadratic univariate regression functions were fitted to these data points, which yielded three curves predicting death-censored GS. For all three follow-up intervals, GS rates decreased when donor age increased. This effect was present for the whole range of donor ages between 11 and 75 years, and most pronounced with advanced donor age.

Simulation of Old-for-Old Allocation

To give an impression of the simulation accuracy of our model, Figure 4(d) shows 10-year death-censored GS for the entire study population. One Kaplan-Meier curve is based on real survival data, whereas the other curve was simulated for the same cohort of recipients. Although the model failed to accurately predict GS in the first year, differences between the real and the simulated Kaplan-Meier curve were only minimal from 1 year posttransplant onwards. At 10 years after transplantation both curves virtually overlapped.

In our analysis, donor age was a significant independent risk factor for death with a functioning graft in a Cox model for this outcome (Table 2), but there was no significant interaction between donor age and recipient age in the model (P=0.5), implying that an old patient receiving an old kidney is not more at risk for death with a functioning graft than a younger patient who receives an old kidney. Figure 4(b and c) show the results of the old-for-old allocation simulation outlined in Figure 4(a). Sixty-five+ grafts that are normally predominantly allocated to recipients aged less than 65 years (old-to-young) showed a drop in 10-year GS when artificially reallocated to recipients aged more than or equal to 65 years (old-to-old: 20.9%→12.9%, P<0.0005). When cases were censored on death with a functioning graft (Fig. 4c), this difference disappeared (39.7%→38.9%, P=0.9). Donor grafts younger than 65 years of age, which were previously allocated to recipients of 65 years and older (yound-to-old) may have a better 10-year GS if these kidneys are allocated to recipients less than 65 years (young-to-young), although the difference observed did not reach statistical significance (19.4%→26.2%, P=0.4). When censored on death with a functioning graft, no such improvement could be observed anymore (56.1%→53.5%, P=0.05). Recipients aged 65 years and older, who had previously received a graft derived from a donor less than the age of 65 years had a significantly worse 10-year GS in the hypothetical new situation where they would have received a 65+ graft (non-death censored: 19.4%→12.9%, P<0.0005; death censored: 56.1%→38.9%, P<0.0005). Conversely, recipients less than the age of 65 years who had originally received a 65+graft would now receive a kidney with a significantly improved 10-year GS (nondeath censored: 20.9%→26.2%, P=0.001; death censored: 39.7%→53.5%, P<0.0005). As the curves in Figure 4(b and c) demonstrate, the disadvantage for the former recipients would approximately equal the benefit in the latter group. For reference, Figure 4(b and c) also show a real, nonsimulated survival curve of those 1011 65+ donor kidneys that were actually transplanted to nonimmunized 65+ recipients in the study cohort.

DISCUSSION

In this retrospective OPTN database analysis, we have shown that donor age has an important impact on short- and long-term outcome after deceased donor kidney transplantation. This is true not only for selected categories but also for the whole range of donors aged 11 years and older. From 1994 to 2006, the relative share of older donor grafts in the kidney transplant pool has increased. Older donor kidney discard rates in the United States are known to be higher than in other parts of the world, and there is much debate why this discrepancy exists (17). Donor biopsies may have contributed to these statistics: Cecka et al. (17) showed that in the United States the percentage of donor kidneys biopsied increases with donor age, thus giving recipient centers an additional diagnostic tool to select kidney grafts derived from older donors. It is plausible that transplant centers are hesitant to accept kidneys with a documented high percentage of glomerulosclerosis, even if other organ quality measures are favorable and the centers would have accepted the kidney without a biopsy. Apart from biopsy interpretation, donor age itself may also be an important motivation for not accepting kidneys from older donors (18). Irrespective of the reason for kidney discard, the impact of advanced donor age on outcome may be underestimated because of the fact that more stringent qualitative selection criteria are applied to older donor grafts compared with that of younger donor kidneys. It should be kept in mind that our results represent the effect of donor age in the context of current clinical transplantation practice in the United States, and in the presence of a high kidney discard rate for donors of advanced age.

An important limitation of our study is that we had to use early graft loss as a surrogate marker for primary nonfunction and that DGF was the only available indicator for early renal function. Serum creatinine or creatinine clearance values would have offered a more detailed tool to assess graft function, but these were not available in the database at standardized time points after transplantation.

In the context of relevant covariates, donor age proved to be an independent predictor of DGF risk, early graft loss, and graft failure within 10 years posttransplant. Although the odds and hazard ratios seem rather low at first sight, it should be noted that donor age was included in the model as a continuous variable, in contrast to most analyses where donor age is a categorical variable with only a few possible alternatives. The Kaplan-Meier analyses show that for each subsequent donor age category more than 11–34 years, death-censored GS up to 10 years was significantly lower. The quadratic regression models present a more detailed analysis, which reveals that there is a weak negative effect of donor age on GS for donor ages less than or equal to 34 years. This effect becomes increasingly stronger from this age onwards. Not surprisingly, the incidence of DGF was higher in DCD kidney recipients. Overall, the 17% to 18% higher rate versus DBD graft recipients is well in line with findings of other studies (14, 19–22). The higher variance in the DCD data can be explained by the smaller number of available cases per single year of donor age (∼60 DCD recipients vs. ∼1500 DBD recipients). These univariate data show that donor age accurately predicts DGF risk not only in the upper extremes but also for the whole range of deceased donor ages between 11 and 75 years.

Advanced recipient age does not seem to be a relevant risk factor for an inferior short- or long-term posttransplant graft outcome, when death with a functioning graft is not considered a failure. The multivariate Cox model even points at a GS benefit when recipient age increases. However, this observation is most likely to be an artifact associated with censoring cases on death with a functioning graft in combination with the long survival period studied by the model.

If advanced recipient age does not negatively contribute to death-censored posttransplant graft outcome, old-for-old allocation could become interesting. Between 1994 and 2006, some organ procurement organizations in the United States have practised old-for-old kidney allocation (23). When we correlated donor age with recipient age, however, we found that in the time period studied such policies did not exist on a large scale. An important rationale for senior recipient exchange policies is that a transplanted kidney which outlives its recipient can be considered a success. Theoretically, it would be expected that the disadvantage of receiving an older kidney is less severe in aged recipients, because they have a higher chance of dying before the relatively short lifespan of their graft is over. Conversely, kidney grafts derived from younger donors would be expected to have a higher yield in younger recipients, as these patients would make more use of the longer lifespan that a young graft has to offer. Although both effects mentioned earlier were observed in our simulation, the net effect of implementing an old-for-old allocation program in terms of total functional graft time gained or lost would most likely be close to zero. Our analysis did show that 65+ grafts will function equally well in non-HLA-matched older recipients compared with their performance in HLA-matched younger patients. The simulated data suggest that it could be safe to sacrifice HLA matching for obtaining shorter CITs and—perhaps even more important—reducing waiting times of selected senior transplant candidates. Results of the ESP show that the system also significantly reduced discard rates of aged donor kidneys (24), but it is difficult to predict whether the same would happen in the United States, because the exact reasons for current high kidney discard rates remain elusive. As a side effect, allocation of kidney grafts derived from relatively young donors to relatively young recipients may even further improve outcomes for this group. We found that retransplantation is associated with a marked increase in the risk of graft failure. This may be a good reason to give longer-lasting kidneys to longer-living patients, to keep the number of retransplants as low as possible for recipients in this group. Our calculations also showed that advanced donor age puts a recipient more at risk for death with a functioning graft. Although the magnitude of this effect does not differ between old and young recipients, an old-for-old policy might imply that the average patient survival of older recipients decreases, whereas patient survival will increase with the same amount in younger recipients. Such potential shifting of life years from one group to another is likely to cause serious ethical dilemmas for policy makers. It is difficult to predict whether a shorter time on dialysis for older patients will balance or even outweigh this survival disadvantage, but reports from existing old-for-old programs suggest that the algorithm can be implemented without compromising graft and patient survival (12, 16, 23–28). Nevertheless, only 3% of all kidney transplants in our study's cohort came from donors aged 65 years and older. If a senior recipient allocation program is to make any difference on overall renal transplant outcome, older kidney utilization rates in the United States would probably have to become substantially higher. Presently, introducing such a new algorithm seems to be much effort for only a small percentage of the transplant population, with a net benefit which is likely to be nearly zero.

A possible limitation of our old-for-old analysis is that part of it is inevitably based on theoretical assumptions about the survival conduct of a graft in its recipient. However, all calculations are also based on real donor kidneys and real recipients in the OPTN database, with many of their relevant characteristics taken into account by means of a Cox model. Formal mathematical sensitivity analyses are beyond the scope of this article, but our model did replicate the overall real GS curve with a high accuracy. This suggests that the simulated survival curves are likely to reliably predict clinical reality. Another potential limitation is our assumption that old-for-old allocation would decrease CIT for 65+ kidneys in the United States with the same amount as it did in Eurotransplant. It remains to be seen whether this is true, because the current American ECD program already accounts for a relatively short CIT when donor age is 60 years and older.

In summary, our results show that deceased donor age is a strong predictor of inferior short- and long-term outcome after kidney transplantation for the whole range of donor ages from 11 years and older. When it is taken into account that older donor grafts are subjected to an exceptionally strict quality selection in the United States (17), the real biological adverse effect of donor age on outcome may even be more pronounced. As a result of this selection, kidney graft utilization rates from old donors are low, with discard rates up to 66.9% for 65 years and older donors. Even with such high kidney discard in older donors, the risk of DGF, early and late graft failure increases for each subsequent year of donor age. Provided that the average CIT of 65+ kidneys would decrease with a few hours by abandoning HLA matching, broad implementation of old-for-old kidney allocation in the United States is likely to be safe and could be a tool to reduce waiting time for older patients. However, our simulation suggests that it will not necessarily improve the overall outcome after kidney transplantation.

REFERENCES

1.Koffman G, Gambaro G. Renal transplantation from non-heart- beating donors: A review of the European experience. J Nephrol 2003; 16: 334.
2.Cohen B, D'Amaro J, De Meester J, et al. Changing patterns in organ donation in Eurotransplant, 1990–1994. Transpl Int 1997; 10: 1.
3.Steinbrook R. Organ donation after cardiac death. N Engl J Med 2007; 357: 209.
4.Ruggenenti P, Perico N, Remuzzi G. Ways to boost kidney transplant viability: A real need for the best use of older donors. Am J Transplant 2006; 6: 2543.
5.Quiroga I, McShane P, Koo DD, et al. Major effects of delayed graft function and cold ischaemia time on renal allograft survival. Nephrol Dial Transplant 2006; 21: 1689.
6.Woo YM, Gill JS, Johnson N, et al. The advanced age deceased kidney donor: Current outcomes and future opportunities. Kidney Int 2005; 67: 2407.
7.Giblin L, O'Kelly P, Little D, et al. A comparison of long-term graft survival rates between the first and second donor kidney transplanted—The effect of a longer cold ischaemic time for the second kidney. Am J Transplant 2005; 5: 1071.
8.Keith DS, DeMattos A, Golconda M, et al. Factors associated with improvement in deceased donor renal allograft function in the 1990s. J Am Soc Nephrol 2005; 16: 1512.
9.Schold JD, Kaplan B, Baliga RS, et al. The broad spectrum of quality in deceased donor kidneys. Am J Transplant 2005; 5: 757.
10.Ploeg RJ, Visser MJ, Stijnen T, et al. Impact of donor age and quality of donor kidneys on graft survival. Transplant Proc 1987; 19: 1532.
11.Cohen B, Smits JM, Haase B, et al. Expanding the donor pool to increase renal transplantation. Nephrol Dial Transplant 2005; 20: 34.
12.Fritsche L, Horstrup J, Budde K, et al. Old-for-old kidney allocation allows successful expansion of the donor and recipient pool. Am J Transplant 2003; 3: 1434.
13.Cox DR. Regression models and life-tables. J R Stat Soc 1972; 34: 187.
14.Moers C, Leuvenink HGD, Ploeg RJ. Non heart beating organ donation: Overview and future perspectives. Transpl Int 2007; 20: 567.
15.Moers C, Smits JM, Maathuis MH, et al. Machine perfusion or cold storage in deceased-donor kidney transplantation. N Engl J Med 2009; 360: 7.
16.Smits JM, Persijn GG, van Houwelingen HC, et al. Evaluation of the Eurotransplant Senior Program. The results of the first year. Am J Transplant 2002; 2: 664.
17.Cecka JM, Cohen B, Rosendale J, et al. Could more effective use of kidneys recovered from older deceased donors result in more kidney transplants for older patients? Transplantation 2006; 81: 966.
18.Matas AJ, Delmonico FL. Transplant kidneys sooner: Discard fewer kidneys. Am J Transplant 2001; 1: 301.
19.Keizer KM, de Fijter JW, Haase-Kromwijk BJ, et al. Non-heart-beating donor kidneys in the Netherlands: Allocation and outcome of transplantation. Transplantation 2005; 79: 1195.
20.Brook NR, Waller JR, Richardson AC, et al. A report on the activity and clinical outcomes of renal non-heart beating donor transplantation in the United Kingdom. Clin Transplant 2004; 18: 627.
21.Hattori R, Ono Y, Yoshimura N, et al. Long-term outcome of kidney transplant using non-heart-beating donor: Multicenter analysis of factors affecting graft survival. Clin Transplant 2003; 17: 518.
22.Koffman G, Gambaro G. Renal transplantation from non-heart-beating donors: A review of the European experience. J Nephrol 2003; 16: 334.
23.Delmonico FL, Burdick JF. Maximizing the success of transplantation with kidneys from older donors. N Engl J Med 2006; 354: 411.
24.Frei U, Noeldeke J, Machold-Fabrizii V, et al. Prospective age-matching in elderly kidney transplant recipients—A 5-year analysis of the eurotransplant senior program. Am J Transplant 2008; 8: 50.
25.Opelz G, Dohler B. Effect of human leukocyte antigen compatibility on kidney graft survival: Comparative analysis of two decades. Transplantation 2007; 84: 137.
26.Pascual J, Marcen R, Ortuno J. Clinical experience with everolimus (Certican) in elderly recipients: The “old-for-old” concept. Transplantation 2005; 79: S85.
27.Segoloni GP, Messina M, Squiccimarro G, et al. Preferential allocation of marginal kidney allografts to elderly recipients combined with modified immunosuppression gives good results. Transplantation 2005; 80: 953.
28.Arns W, Citterio F, Campistol JM. ‘Old-for-old’—New strategies for renal transplantation. Nephrol Dial Transplant 2007; 22: 336.
Keywords:

Donor age; Kidney transplantation; Regression models; Old-for-old; Simulation

Supplemental Digital Content

© 2009 Lippincott Williams & Wilkins, Inc.