Matching Older Kidneys with Older Patients Does Not Improve Allograft Survival

Methods

Results

The numbers of first cadaver kidney transplantation procedures have increased only slightly from 1988 to 1998. However, an increasing proportion of these transplantations have been performed in patients ≥55 yr old. Similar trends have been observed in cadaver donor age (Figure 1). The percentages of the total number of recipients and donors ≥55 yr old have increased from 17.5 to 32.2% and from 6.5 to 12.0% from 1988 and 1998, respectively. During this time, giving older kidneys to older recipients was a common practice (Table 1). For example, recipients ≥55 yr old received kidneys that were ≥55 yr old 46.2% more often than would have been expected if differences in recipient and donor age had not been a factor in deciding to transplant a kidney (Table 1). Recipients ≥55 yr old received kidneys that were 18 to 29 yr old 33.6% less often than expected (Table 1). The differences between observed and expected numbers of transplantations in the different recipient and donor categories were similar in 1988 and 1993 compared with 1994 and 1998 (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Changes in the numbers of first cadaveric kidney transplant recipients and donors by age from 1988 to 1998.

Both recipient and donor age have important effects on graft survival, although the effects of donor age are much stronger than those of recipient age (Figure 2). Compared with recipients 18 to 29 yr old, recipients ≥55 old were 25% more likely to have graft failure over a median survival of 3.95 yr (25 to 75% interquartile range, 1.85 to 6.09 yr). Interestingly, there was a similar, increased risk of graft failure for recipients <18 yr old. On the other hand, donor kidneys ≥55 old were 78% more likely to have graft failure compared with kidneys 18 to 29 yr old. There was also a J-shaped curve effect for donor age because donor kidneys <18 yr old were 17% more likely to have graft failure compared with donor kidneys 18 to 29 yr old (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. The relative risk for graft failure attributable to specific donor and recipient age combinations (A) versus relative risks from donor age and recipient age per se (B) in Cox proportional hazard analyses. (A) Analysis includes only the 24 recipient and donor age categories shown (n = 74,297). The reference category (relative risk, 1.00) is for recipients and donors 18 to 29 yr old. Relative risks more than 1.00 indicate a proportionally greater risk of graft failure; relative risks <1.00 indicate a proportionally lower risk. (B) Analysis includes only the 4 recipient age categories (reference, 18 to 29 yr old) and 4 donor age categories (reference, 18 to 29 yr old); the relative risk in each of the 24 recipient and donor age categories was calculated on the basis of the additive effects of recipient and donor age. Differences in the relative risks between (A) and (B) should reflect the effects of combining donors and recipients in specific age categories. To the extent that (A) and (B) are similar, there is not likely to be an effect of combining donor and recipient ages in specific categories that are independent of the effects of recipient and donor age per se (B).

Giving older kidneys to older recipients does not appear to have a major effect on graft survival independent of the effects of recipient and donor age per se (Table 2). Indeed, there did not appear to be any consistent interactions of specific recipient-donor age combinations. Among the 16 recipient-donor age categories included in the multivariate Cox proportional hazard analysis, only 4 were statistically significant. However, all of these effects were only marginally significant (0.05 > P > 0.005), and some could have been the result of chance. The strongest of these effects were the 2 interactions for recipient and donor age for recipients <18 yr old receiving donor kidneys <18 yr old and for recipients <18 yr old receiving donor kidneys 30 to 41 yr old (Table 2). Both of these effects partly reduced the otherwise higher relative risk of graft failure for recipients <18 yr old. However, there was no reciprocal adverse effect of giving older donor kidneys to younger recipients, and in fact, there was a NS trend for a similar reduction in the risk of graft failure for giving older donor kidneys to younger recipients. In other words, there did not appear to be any consistent pattern across all of the possible recipient-donor age combinations that would suggest that giving younger or older kidneys to younger or older recipients altered the risk of graft failure.

Confirming the results of the multivariate Cox analysis was the fact that the unadjusted relative risks of several different combinations of recipient and donor age appeared to be qualitatively similar to what would be expected in these categories from the independent effects of recipient and donor age alone—that is, without interactions from recipient and donor age mismatches (Figure 2). For example, for donor kidneys ≥55 yr old, the risk of graft failure was not higher or lower for recipients of different ages in a multivariate Cox proportional hazard analysis (Table 2). Similar results were obtained if the Cox analysis was limited to only patients that survived at least 1 yr or for 3 yr (data not shown).

As was the case for graft survival, giving older kidneys to older recipients did not appear to have any major effects on patient survival, or death-censored graft survival, independent of the effects of recipient and donor age per se (Figures 3 and 4). The Cox proportional hazard analyses confirmed the qualitative comparisons of the effects of donor-recipient age combinations with the intrinsic risk from donor and recipient age for patient survival, and death-censored graft survival (data not shown). However, the effects of recipient age on death-censored graft failure and death with function were very different from each other (Figures 3 and 4). Specifically, older recipients were much more likely to die than their younger counterparts, whereas older recipients were less likely to return to dialysis (or to undergo retransplantation) compared with younger recipients. In contrast, the effects of donor age on death-censored graft failure and death with function were similar (Figures 3 and 4). In Cox analysis adjusted for the same risk factors shown in Table 2, recipients of kidneys ≥55 yr old were 72% (95% confidence interval, 52 to 94%) more likely to return to dialysis (or to require retransplantation) than recipients of kidneys 18 to 29 yr old (full model not shown). Similarly, recipients of kidneys ≥55 yr old were 80% (30 to 250%) more likely to die than recipients of kidneys 18 to 29 yr old (full model not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. The relative risk for death-censored graft failure attributable to specific donor and recipient age combinations (A) versus relative risks from donor age and recipient age in any combination (B). For a full explanation, see Figure 2 caption.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. The relative risk for death with function attributable to specific donor and recipient age combinations (A) versus relative risks from donor age and recipient age in any combination (B). For a full explanation, see Figure 2 caption.

Discussion

The results of this study imply that transplantation centers are more likely to accept older donor kidneys for older recipients (and younger kidneys for younger recipients) than would be expected if decisions to accept or reject a kidney were not influenced by differences between the age of the recipient and donor. One reason for this might be a belief that overall graft survival is improved by giving older kidneys to older recipients. However, the results of this analysis suggest that giving older kidneys to older recipients has no effect on overall graft survival.

There is a widely held belief that placing older donor kidneys in older recipients reduces the physiologic stress on the allograft and thereby prolongs allograft survival (7,12–16). Most of the studies that have provided data to support this hypothesis have been single-center investigations of relatively small numbers of patients (12–14). However, two single-center studies failed to find an independent effect of donor and recipient age differences on graft survival (10,17). Studies from large registries have shown that older donor age is associated with reduced graft survival, but these studies have not systematically examined the effects of interactions between donor and recipient age, independently from the effects of donor and recipient age per se. For example, Busson and Benoit (16) showed that when the donor was more than 10 yr older than the recipient, graft survival was lower, but they did not report whether differences in donor and recipient age predicted graft survival better than (or independently of) recipient age and donor age alone.

In comparing separate analyses of transplantations grouped by donor age, Hariharan et al. (18) found that older recipient age negatively affected graft survival in recipients of older donor kidneys, but not in recipients of younger donor kidneys. However, they did not report the effects of specific donor and recipient age combinations (interactions) in multivariate analysis that independently adjusted for the intrinsic effects of recipient and donor age (18). Thus, it is unclear from these studies whether giving older kidneys to older recipients improves graft survival. The results of our analysis suggest that it does not.

It is theoretically possible that the current practice of giving older kidneys to older recipients has made it impossible to detect an effect of this age matching on graft survival. However, this seems unlikely for two reasons. First, the number of patients in the analysis we performed is large, making it improbable that even small effects would have been missed. Second, giving older kidneys to older recipients was not completely efficient; there were large numbers of younger recipients who received older donor kidneys and vice versa (Table 1).

Four of the 16 recipient-donor age combinations examined appeared to have statistically significant effects on graft survival over and above the effects of recipient and donor age per se (Table 2). However, there did not appear to be any consistent pattern to these effects, raising the possibility that these were the result of chance. For example, the apparent reduction in the risk for recipients <18 yr old that received younger donor kidneys, although statistically significant, was similar in magnitude to the reduction in risk for recipients <18 yr old that received older donor kidneys (Table 2). Similarly, the apparent reduction in risk for recipients 42 to 54 yr old that received younger donor kidneys was comparable to the risk reduction for recipients 42 to 54 yr old that received older donor kidneys. Importantly, none of the interactions testing the hypothesis that giving older kidneys to older recipients improves graft survival was statistically significant. It is also apparent from the qualitative comparison of relative risks that most of the risk associated with different recipient-donor age combinations (Figure 2A) is attributable to the intrinsic effects of recipient and donor age (Figure 2B).

There may be ethical reasons to give older kidneys to older recipients, even if outcomes are not improved. Society may decide that it is justifiable to give “good” kidneys to younger recipients and “marginal kidneys” to older recipients. To a large extent, donor age determines which kidneys are good and which are marginal. However, our analysis suggests that using differences in recipient and donor age to help decide whether to accept or reject a cadaver donor kidney does little to enhance overall graft survival. In other words, the total number of years that the population of kidney transplant recipients enjoys graft function is not affected by current differences between the ages of the recipients and donors.

As we examined these data, we were interested in testing the hypothesis that the age of the host could modulate the rate of senescence of the transplanted kidney. To the extent that mechanisms of normal aging might be both intrinsic and extrinsic to the kidney, it is theoretically possible that the age of the kidney and the age of the host could interact to accelerate or slow the rate of graft failure attributable to senescence (19). Our results suggest that this is not the case in kidney transplant recipients. However, these results certainly cannot distinguish how much of the effects of aging are intrinsic or extrinsic to the kidney. It is possible that either intrinsic or extrinsic factors are alone responsible for the effects of aging on the kidney. It is also possible that both intrinsic and extrinsic factors are important and that both continue at a fixed, linear rate over time. Either of these pathophysiologic scenarios for renal senescence would be entirely consistent with the lack of any interaction between recipient and donor age on graft survival seen in this study. Of course, the use of nephrotoxic immunosuppressive medications, allograft rejection, and partial denervation of the kidney could all affect the natural, physiologic aging of the kidney (19).

The effects of recipient age on death with function compared with death-censored graft failure were markedly different (Figures 3 and 4). Recipients ≥55 yr old were much more likely to die with a functioning allograft compared with younger recipients. However, what is even more remarkable is the fact that recipients ≥55 yr old were less likely than younger recipients to return to dialysis or to undergo retransplantation. Others have reported that older transplant recipients are more likely to die of infectious complications and less likely to have acute rejection compared with younger recipients (20). Taken together, these observations suggest that strong consideration should be given to reducing immunosuppression in older transplant recipients.

Finally, like others, we noted a J-shaped curve relationship between both donor and recipient age and graft survival. Specifically, donors and recipients <18 yr of age had an increased risk of graft failure, compared with donors and recipients 18 to 29 yr old. Decreased graft survival for kidneys taken from children is well known (21–23) and is likely the result of technical complications associated with transplanting small kidneys. It is also known that children have decreased graft survival, which is attributable to both a heightened immune response and problems with medication compliance that lead to a higher incidence of acute rejection.

In summary, giving older cadaver kidneys to older recipients is a common practice. Our analysis showed that although both recipient and donor age affect graft survival, giving older kidneys to older recipients does not. Transplantation centers and doctors should examine their policies and the current practice of giving older kidneys to older recipients. In particular, who decides what recipient and donor age combinations are appropriate, and what role the transplantation center and the patient should play in this decision, should be openly discussed.