Age and the Associations of Living Donor and Expanded Criteria Donor Kidney With Kidney Transplant Outcomes

Background— Recent studies show a survival advantage with kidney transplant amongst elderly patients compared to those on dialysis. Study Design— In our present study we examined and compared the association of expanded donor criteria (ECD) kidney and living kidney donation with outcome of kidney transplant across different ages including elderly recipients. Setting and Participants— Using the Scientific Registry of Transplant Recipients, we identified 145,470 adult kidney transplanted patients. Mortality and death-censored graft failure risks were estimated by Cox proportional regression analyses over a follow-up period with a median of 3.9 years. Predictors— ECD kidney and living kidney donation and age compared to others. Outcomes— Mortality and death-censored graft failure risk. Results— Patients were 45±16 years old and included 40% women and 19% diabetics. Compared to transplanted patients 55-<65 years old, the fully adjusted death-censored graft failure risk was somewhat higher in patients ≥75 years old (HR, 1.30; 95% CI, 1.09–1.56), 35-<55 years (HR, 1.13; 95% CI, 1.08–1.17) and 18-<35 years (HR, 1.64; 95% CI, 1.57–1.71). Compared to non-ECD kidneys, ECD kidneys were significant predictors of mortality in non-elderly patients (18– <35 years: HR, 1.46 [95% CI, 1.19–1.77]; 35-<55 years: HR, 1.23 [95% CI, 1.14–1.32]; and 55-<65 years: HR, 1.26 [95% CI, 1.15–1.38]) and patients aged 65-<70 years (HR, 1.20; 9% CI, 1.05–1.36); but not in other groups of elderly patients (HRs of 1.12 [95% CI, 0.93–1.36] 70-<75 years and 1.04 [95% CI, 0.74–1.47] for ≥75 years). Similar results were found in risk of graft loss. Compared to deceased donor, living kidney was associated with better survival in all age groups and lower graft loss risk in patients aged <70 years. Limitations— Unmeasured confounders cannot be adjusted for. Conclusions— Among deceased donors, the ECD kidneys are not associated with either increased mortality or graft failure in recipients over 70 years. Among all types of donors, the persistent association between living donor kidneys and lower all-cause mortality across all ages suggests that, if possible, elderly patients gain longevity from living donor kidney transplant. span, along with a 61% reduction in the long-term risk of death. A more granular look at the data shows the projected increase life


Limitations-Unmeasured confounders cannot be adjusted for.
Conclusions-Among deceased donors, the ECD kidneys are not associated with either increased mortality or graft failure in recipients over 70 years. Among all types of donors, the persistent association between living donor kidneys and lower all-cause mortality across all ages suggests that, if possible, elderly patients gain longevity from living donor kidney transplant.
Keywords elderly; kidney transplantation; mortality; graft failure; living donor Individuals older than 65 years old, the so-called elderly, make up the fastest growing population group in United States and most developed countries. Based on data in the 2008 report of the United States census bureau, this sector of the population grew from 29.6 million in 1990 to 36.8 million in 2008, equivalent to a 20% incremental growth. As a result of the absolute increase of the elderly in the population distribution, the prevalence and incidence of chronic diseases such as hypertension, diabetes mellitus, kidney disease, coronary artery disease, and heart failure have also increased overall. (1)(2)(3)(4) The growth in the population of patients with end stage renal disease (ESRD) during the past 20 years has occurred principally in the elderly. (5) The adjusted incidence rate of ESRD for patients who are older than 75 years was 1744 per million population whereas it was only 127 per million population for those between 20 and 44 years old. (5) In 2006, 49% of the incident ESRD population was older than 65 years, and 26% were ≥75 years.(5) Based on these statistics, one of every 200 US adults older than 75 years is estimated to have ESRD. (5) Kidney transplant recipients are increasingly older at the time of transplant. In the 2008 Scientific Registry of Transplant Recipients (SRTR), for instance, recipients of deceased non-expanded criteria donor (non-ECD) kidneys in 2007 who were between ages 50 to 64 years and 65 years and older constituted 39 and 13 percent of the total pool, respectively.(6) By comparison, these age groups comprised only 29 and 5 percent of all such recipients in 1993, respectively. Currently, 16,496 ESRD patients older than 65 years are waiting for a kidney transplant; these individuals constitute 18.1% of all listed candidates(6) and need a special pre-transplant evaluation process. (7) The accuracy of previous studies comparing survival after kidney transplant with continued dialysis might have suffered methodological flaws such as selection bias. (8,9) Subsequent reports that have avoided these methodological problems show that, compared with those who are dialyzed, elderly individuals who receive transplants have a survival advantage, including transplants patients given ECD kidneys.(10-15) Wolfe et al (10) found that among patients aged 60-<75 years who received a primary deceased-donor transplant, the cumulative survival rate, in comparison to maintenance hemodialysis patients, improved after the first year posttransplant, with a projected 4-year increase in life span, along with a 61% reduction in the long-term risk of death. A more granular look at the data shows the projected increase life spans were 4.3 years, 2.8 years, and 1.0 year, for patients aged 60-64 years, 65-69 years, and 70-74 years, respectively.(10) Additionally, with present therapy, elderly patient survival at one, five, and ten years is approximately 80 to 90, 70, and 50 percent, respectively. (16)(17)(18)(19)(20)(21)(22)(23)(24) Recent studies show a survival advantage with kidney transplant among the elderly, including the recipients of expanded criteria donor kidneys, in comparison to those patients treated with dialysis. (10)(11)(12)(13)(14)(15) Very few studies have attempted to identify factors capable of predicting outcomes in elderly kidney transplant recipients. According to a 2009 study, acute rejection in the first 90 days and donor age ≥60 years were predictors of lower patient survival; by contrast, delayed graft function, donor age of 60 years or more, and HLA antibodies were associated with greater death-censored graft loss.(25) Time on dialysis prior to transplant was also a significant risk predictor in patients older than 70 years. (25) An article published in 2008 reported that elderly kidney transplant recipients who received organs from living donors aged >55 years had worse 3-year graft survival (86%) but nearly equivalent 3-year patient survival rates (88%) in comparison to counterparts receiving kidneys from living donors ≤55 years. (26) To our knowledge only a few studies have compared the association of expanded donor criteria (ECD) kidney and living kidney donation with kidney transplant outcomes across different age groups including in the elderly patients. In the present study, we examined the association of ECD kidney and living kidney donation with outcomes from kidney transplant amongst different age ranges with focus on the elderly recipients.

Patients
We extracted, refined, and examined data from all kidney transplant recipients listed in the SRTR up to December 2006. The SRTR data system includes data on all transplant donors, wait-listed candidates and transplant recipients in the US, which are submitted by members of the Organ Procurement and Transplantation Network. This study was approved by the Institutional Review Committees of Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center. Because of the large sample size, the anonymity of the patients studied and the non-intrusive nature of the research, the requirement for informed consent was waived.

Clinical, Demographic and Laboratory Measures
Demographic data and details of medical history were collected, including information on age, gender, race, ethnicity, type of insurance, presence of diabetes, and dialysis vintage. Dialysis vintage was defined as the duration of time between the first day of dialysis treatment and the day of kidney transplant. Information on the recipient's serum creatinine, serum albumin, weight and height (for calculation of body mass index (BMI), and information on six co-morbidities: coronary artery disease, chronic obstructive pulmonary disease, hypertension, peptic ulcer, peripheral vascular disease, cerebrovascular disease was also collected. We divided the entire population into six age groups of 18-<35, 35-<55, 55-<65, 65-<70, 70-<75 and ≥75 years. The last 3 groups (≥65 years) were considered the elderly.

Statistical Methods
For survival analysis we used Cox proportional hazard regression models separately across the different age groups. Two predefined outcomes were analyzed: all-cause death and death censored graft failure. Graft failure was defined as re-initiation of dialysis treatment or repeat transplant. In our graft failure analysis, the patients were followed until graft failure or censoring (death or the end of the follow-up period) whichever happened first. Based on the detected association between age groups and death censored graft failure, which appeared to be U-shaped, we tested non-linearity by adding the quadric term of age to the models which already had the linear term. For each analysis, three models were examined based on the level of multivariate adjustment: I. An unadjusted model that included mortality or graft failure data and time postkidney transplant, and entry calendar quarter (year and quarter of kidney transplant date); II. Case-mix adjusted models that included age, gender, race-ethnicity (African Americans and other self-categorized African-Americans, Non-Hispanic Whites, Asians, Hispanics and others), diabetes mellitus, dialysis vintage, six previously mentioned co-morbidities and three most recent laboratory measures before transplant: serum creatinine, serum albumin, and BMI.
III. fully adjusted models which included all of the covariates in the case-mix model as well as the following transplant data: (1) donor type (deceased or living), (2) donor age, (3) panel reactive antibody (PRA) titer (last value prior to transplant), (4) number of HLA mismatches, (5) cold ischemia time, (6) expanded donor criteria (EDC) using standard definition (donor history of hypertension and/or donor serum creatinine > 1.5 mg/dL and/or cause of death in donor is cerebrovascular event) and (7) history of acute rejection. EDC analysis was restricted to deceased donors.
All analyses were carried out with SAS version 9.1, SAS Institute Inc., www.sas.com.

Results
Using the SRTR database we identified 145,470 adult kidney transplanted patients. There were 25,616 deaths (18%) and 22,876 graft failures (16%). The median follow-up time was 1449 (25th-75th percentile, 636-2497) days. Figure 1 shows the age distribution of the 145,470 kidney transplanted patients. Only 11% of patients were older than 65 years. Table  1 shows the clinical, demographic, laboratory, and transplant data of the 145,470 kidney transplanted patients across six different age groups including 3 group younger and 3 older than 65 years. The crude mortality rate was incrementally higher across older groups, whereas the graft failure rate was lower in older groups. Older patients included less women and more non-Hispanic white recipients. The percentage of deceased donors, usage of expanded donor criteria kidneys and older donors was incrementally higher in older groups.
As shown in Figure 2 we examined the association between age groups and all cause mortality using 18 (Table 2), although subtle differences were observed across age groups (age-donor interaction p-value <0.001).
Additional analyses using the subgroup of non-ECD of the deceased donors as the reference confirmed the above findings, although the power was somewhat mitigated (see Table 2).
As shown in Table 3 we also examined predictors of death-censored graft failure within the age groups. Compared to the non-ECD kidneys, the ECD kidneys were significant predictors of graft loss in younger patients, i.e., 40%, 31%, 38% and 36% increased risk of graft loss in 18-<35, 35-<55, 55-<65 and 65-<70 years old patients, respectively, but not significantly so in older patients, although similar trends were observed in the elderly as well (see Table 3). Age was an effect modifier of the association between ECD kidney and graft loss (p for interaction <0.001). Compared to the deceased kidney, living kidney donation was significantly associated with better graft survival in non-elderly patient, i.e., 22%, 29%, 32% and 49% lower graft failure in 18-<35, 35-<55, 55-<65 and 65-<70 years old patients. The effect modifying role of age was consistent with a significant statistical interaction (p<0.001). Similar results were found when the subgroup of non-ECD of the deceased donors was used as the reference (see Table 3).

Discussion
In 145,470 kidney transplant recipients, older age was associated with an incrementally higher risk of death. The association between age and death-censored graft failure was Ushaped, indicating the highest risk of graft failure in both young adults (18-<35 years) and very old patients (≥75 years) and the lowest risk in patients aged 55-<75 years. In the patients older than 70 years, the ECD kidney was not a significant predictor of death or graft loss; however, in patients younger than 70 years ECD was a strong predictor of poor outcomes. Living donor kidney was associated with better patient survival in all age groups, while patients younger than 70 years also exhibited the advantage of lower risk of graft loss.
Compared to patients 18-<35 years old, elderly patients had higher risk of mortality. These results are not surprising as age is an important predictor of death in both the ESRD and non-ESRD populations. (27) The association between age groups and graft failure was also U-shaped. Compared to patients aged 18-<35 years (reference), patient 65-<70 years, 70-<75 years and ≥75 years had a 37%, 36% and 20%, respectively, lower risk of graft failure, whereas the risk was similar in patients aged 55-<75 years. Survival in the elderly kidney transplant recipient is currently excellent. With present therapy, patient survival at one, five, and ten years is approximately 85, 70, and 50 percent, respectively. (16)(17)(18)(19)(20)(21)(22)(23)(24) Although these data are promising, elderly patients may not live long enough to suffer graft failure. In elderly transplanted recipients, death with functioning graft due to cardiovascular disease, (21) infection, or malignancy(28) was the most common cause of graft loss.(29) Moreover, kidney function predicts non-cardiovascular mortality from multiple causes in the elderly. (30) Another potential explanation for greater graft survival in patients ≥75 years is that activation of the immune system in old patients is lower than in younger counterparts. For example, acute rejection is less common in older transplant recipients, presumably on account of an immune system that is less active. (24). The most common cause of graft failure after the first year post transplant is chronic renal allograft nephropathy. (31) However multiple factors (alloantigen-dependent and alloantigen-independent) seem to be involved in the pathogenesis of chronic graft dysfunction, (32). Data from both experimental models and humans indicate an important role for many elements of the immune system in the pathogenesis of chronic renal allograft nephropathy. (33)(34)(35)(36) Decreased activation of the alloantigen-dependent factors of chronic renal allograft nephropathy might contribute to longer graft life in elderly patients.
In contrast to patients aged <70 years, ECD kidneys did not correlate with all-cause mortality in patients older than 70 years. Similar associations were found with graft loss. This is an important consideration in the selection of a suitable kidney for elderly patients. Our study is notable for its large sample size and for the several important transplant covariates that were accounted for in the multivariate analyses. As with all registry-based observational studies, these results are subject to certain limitations. Like all observational studies, our study also cannot prove causality. Immunosuppressive and other regimens, which have potential impacts on patient and graft survival, were not available in the database. Nonetheless, in the fully adjusted model, we controlled for a number of important variables.
In conclusion, similar to the general population, among the kidney transplant recipients older age is associated with higher risk of death. The association between age and (death censored) graft failure is somewhat U-shaped. In the elderly patients aged over 70 years, EDC kidney is not predictor of death or graft loss; however, in patients younger than 70 years, EDC was a strong predictor of both increased death and graft loss rates. Living donor kidney appears associated with greater survival across all age groups including the elderly, although the significantly lower graft loss rate is observed mainly among those younger than 70 years. Hence, our study suggests that the elderly ESRD patients gain years of life if they receive a kidney transplant, in particular from a living donor. Given other data indicating that elderly transplant recipients have a 41% lower overall risk of death compared with waitlisted candidates, (14) elderly ESRD patients should be transplanted with a living donor kidney if possible. However, among those elderly individuals over 70 years who expect deceased kidneys, receiving ECD kidneys exhibit virtually the same survival and graft failure outcomes. These findings need to be verified in additional studies.  Hazard ratios (95% confidence intervals) of all-cause mortality using Cox regression analyses in 145,470 kidney transplant patients (reference category: 18-<35 years) Hazard ratios (95% confidence intervals) of death censored graft failure using Cox regression analyses in 145,470 kidney transplant patients (reference category: 18-<35 years)  (23) 10,101 (7) 4,271 (3) 1,295 (1) <0.001 No. of Deaths * 3050 [10] 10258 [15] 7259 [22] 2759 [27] 1178 [28] 397 [31] <0  (23) 3409 (15) 1393 (11) 365 (9) 180 (9) 59 (9) <0.001 Graft thrombosis 55 (7) 121 (7) 52 (6) 16 (5) 3 (2) 2 (4)     ECD: Expanded criteria donor * case mix model was adjusted for : age, gender, race-ethnicity, diabetes mellitus, dialysis vintage, six co-morbidities and three laboratory measures: serum creatinine, serum albumin, and BMI. ** fully adjusted models was adjusted for all of the covariates in the case-mix model as well as donor type (deceased or living), donor age, panel reactive antibody (PRA) titer (last value prior to transplant), number of HLA mismatches, cold ischemia time, expanded donor criteria (EDC) and history of acute rejection.