SEARCH

SEARCH BY CITATION

Keywords:

  • Donor pool;
  • donor/recipient matching;
  • donor risk;
  • donor selection;
  • extended donor criteria;
  • interaction;
  • kidney allocation;
  • kidney graft survival;
  • kidney transplantation;
  • recipient outcome;
  • risk

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. Reference

As of May 2012, over 92 000 patients were awaiting a solitary kidney transplant in the United States and new waitlist registrations have been rising for over a decade. The decreasing availability of donor organs makes it imperative that organ allocation be as efficient and effective as possible. We performed a retrospective cohort study of adult recipients in the United States (n = 109 392) using Scientific Registry of Transplant Recipients data. The primary aim was to evaluate the interaction of donor risk with recipient characteristics on posttransplant outcomes. Donor quality (based on kidney donor risk index [KDRI]) had significant interactions by race, primary diagnosis and age. The hazard of KDRI on overall graft loss in non-African Americans was 2.16 (95%CI 2.08-2.25) versus 1.85 (95%CI 1.75-1.95) in African Americans (p < 0.0001), 2.16 (95%CI 2.08-2.24) in nondiabetics versus 1.84 (95%CI 1.74-1.94) in diabetics (p < 0.0001), and 2.22 (95%CI 2.13-2.32) in recipients <60 years versus 1.83 (95%CI 1.74-1.92) in recipients ≥60 (p < 0.0001). The relative hazard for diabetics at KDRI = 0.5 was 1.49 but at KDRI = 2.0 the hazard was significantly attenuated to 1.17; among African Americans the respective risks were 1.50 and 1.17 and among recipients 60 and over, it was between 1.64 and 1.22. These findings are critical considerations for informed decision-making for transplant candidates.


Abbreviations
AA

African American

DCD

death after cardiac donation

ECD

expanded criteria donor

KDRI

kidney donor risk index

OPTN

Organ Procurement and Transplantation Network

PKD

polycystic kidney disease

PRA

serum panel reactive antibody

SCD

standard criteria donor

SRTR

Scientific Registry of Transplant Recipients

UNOS

United Network for Organ Sharing

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. Reference

As of May 2012, over 92 000 patients were awaiting a solitary kidney transplant in the United States [1]. The number of new waitlist registrations for a solitary kidney transplant has been steadily increasing from 22 285 in 2000 to 34 091 in 2009 [2]. The rate of donation has not increased sufficiently to meet the growing need for transplantation with a slow decline in the number donations from just over 13 600 in 2009 to 13 201 in 2011 [3]. According to a report from 2001, the population of end-stage renal disease patients was increasing worldwide at a rate of 7% per year [4]. The majority of patients listed for a transplant have initiated dialysis and waiting time on dialysis is a strong risk factor for decreased graft and patient survival [5]. Life expectancy for waitlisted dialysis patients has been shown to roughly double after kidney transplantation [6, 7]. Given the increased demand for transplants and the decreasing supply of donations, it is imperative that organ allocation be as efficient and effective as possible to ensure optimal health outcomes for the candidate population.

Just over half (56.3%) of the single kidneys donated in 2011 were from deceased donors [8]. The risks associated with deceased donor kidneys have been well documented [8-11]. In an effort to meet the growing need for kidney organ donations, expanded criteria donor (ECD) kidneys are used in spite of an approximately 70% greater risk of graft failure compared to transplants from the lowest risk non-ECD donors [12].

Classifying kidneys as ECD versus non-ECD has elucidated the effects of measured differences in donor characteristics on outcomes and has highlighted the importance for physicians and patients to consider these differences in the transplant process. A continuous kidney donor risk index (KDRI) was also developed to measure the spectrum of risk associated with various factors known to influence graft failure [10]. The KDRI, comprised of 10 donor and 4 transplant characteristics that are significantly and independently associated with increased risk of failure of deceased donor kidney transplants, is a useful tool to assist physicians and transplant candidates when these kidneys become available. This more specific index incorporates a large number of characteristics influencing graft outcome into one metric resulting in an improvement over the less accurate ECD classification system [10].

Although the developed classification systems may greatly assists physicians and candidates in assessing their options when a deceased donor kidney becomes available, designation of donor risk may often be assumed to be uniform for recipients. The ECD versus non-ECD classification and the KDRI can provide estimated relative risks and posttransplant survival associated with the quality of the donor organ, however, there may exist significant interactions of the effect of donor kidney quality by recipient characteristics which influence transplant outcomes. The aim of this study was to evaluate the interaction of two donor risk metrics, the ECD designation and the KDRI, with recipient characteristics on the risk for outcomes of graft loss.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. Reference

Data for this retrospective cohort study came from the national Scientific Registry of Transplant Recipients (SRTR) database that contains patient-level information supplemented with data from Social Security and the Centers for Medicare and Medicaid Services. Between January 01, 1995 and December 31, 2010, there were 144 984 deceased kidney donor transplants. Exclusion criteria for recipients removed sequentially included: cold ischemia times less than 1 h (n = 156) or greater than 60 h (n = 378); donors listed as less than 1 year or greater than 80 years of age (n = 602); recipients with a creatinine level greater than the 99th percentile of the sample equivalent to a value of 4.0 (n = 1340); missing donor height (n = 323), donor weight (n = 3) and donor creatinine (n = 618); donors with a BMI less than 13 or greater than 50 (n = 3403); recipients aged less than 18 years (n = 5979); recipients with previous transplants (kidney n = 16 437, other organ n = 3088); and multiorgan transplants (n = 3265). The final study population consisted of 109 392 deceased donor kidney transplant recipients aged 18 and older.

The primary outcome of interest in this study was overall graft loss defined as either return to dialysis or death. In addition, models with patient death and with death-censored graft failure as separate outcomes were generated. Follow-up began from the time of kidney transplant until graft failure, loss to follow-up, or the end of the cohort observation period (March 01, 2011). Overall, 39 432 graft failure events occurred including 28 307 deaths (71.8%). At the end of follow-up 69 960 (64.0%) recipients had a functioning graft.

Two metrics, the ECD definition and the KDRI, were used to assess donor risk. The KDRI calculated for this study comprised the 10 factors associated with donors including: age, height, weight, race, history of hypertension and diabetes, serum creatinine, cause of death, donation after cardiac death (DCD) and hepatitis C to be consistent with the Organ Procurement and Transplantation Network (OPTN) KDRI calculator [13]. The KDRI score is a measure of the estimated risk of graft failure relative to a reference donor with KDRI = 1.00 with characteristics as specified by Rao et al. [10].

The interaction of each risk metric with recipient characteristics was evaluated in multivariable survival models to determine the impact on overall graft failure. Initially, Kaplan-Meier curves were generated for the categorical predictors to observe the shape of the survival function for each group and provide insight into whether or not the groups were proportional. Tests of equality across strata including the log-rank test for categorical variables and univariate Cox proportional hazard regression for continuous variables were used to determine which predictors should be considered for the final model (p-value < 0.2). In addition to the donor risk metrics, recipient characteristics considered for inclusion in the final model included: age; race; gender; primary diagnosis; history of diabetes; serum panel reactive antibody (PRA) percent; primary insurance; income status; obesity; level of education; years on dialysis; center effect (percentage of ECD transplants at each center) and DCD for the ECD model (Table 1).

Table 1. Recipient characteristics (n = 109 392)
Variablen (%)Median (IQR)
  1. a

    Percentages may total <100 due to missing values.

  2. b

    Missing values n = 499 for insurance added to “no” category.

  3. c

    Missing values n = 31 added to “no” category.

Age
18–3414112 (12.9) 
35–4933010 (30.2)52 (42–61)
50–6445443 (41.5) 
65+16827 (15.4) 
Gender
Male66413 (60.7) 
Female42979 (39.3) 
Race
White52955 (48.4) 
African American34625 (31.7) 
Multiracial174 (0.2) 
American Indian/Alaska native1236 (1.1) 
Asian5839 (5.3) 
Hawaiian/other Pacific Islander591 (0.5) 
Hispanic/Latino13970 (12.8) 
Education
None/grade school7017 (6.4) 
High school/GED42102 (38.5) 
Some college/bachelor degree31959 (29.2) 
Graduate degree4663 (4.3) 
Serum panel reactive antibody (%)
079978 (73.1) 
1–3017347 (15.9)0.0 (0–0)
31–805806 (5.3) 
≥812670 (2.4) 
Primary source of paymentb
Medicare69206 (63.3) 
Private32359 (29.6) 
Other (primarily Medicaid)7827 (7.2) 
Polycystic kidney diseasec10531 (9.6) 
History of diabetesc28643 (26.2) 
Working for income12188 (11.1) 
Obese (BMI >30)28597 (26.1) 
Death after cardiac donation7439 (6.8) 
ECD kidney transplant19917 (18.2) 

All predictor variables with a p-value < 0.2 in univariate analyses were initially included in the multivariable Cox regression models in addition to interactions between each recipient characteristic and the donor risk metric (ECD or KDRI). DCD status was also included in the ECD model. Interaction terms with a p-value > 0.1 were removed from the models. Subanalyses were conducted to determine whether the selection of recipients for transplant differed before and after the ECD allocation algorithm in 2002. In addition, paired-kidney models were run to mitigate potential selection biases in which higher risk donor kidneys were allocated to less healthy patients not captured by available data. The paired analysis matched recipients based on discordant characteristics for age (60 and over vs. less than 60), race (AA vs. non-AA) or diabetes status (yes vs. no) with the same donor (and KDRI score).

The proportionality assumption of the models was tested by observing the Kaplan-Meier curves for each time fixed covariate. Survival and adjusted hazard functions were graphed to compare recipients receiving ECD versus SCD kidneys and high versus low KDRI donors. Analyses were performed using SAS software, version 9.2 and statistical significance was determined at an α = 0.05 level. The study was approved by the Cleveland Clinic Institutional Review Board.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. Reference

The overall study population included 109 392 kidney transplant recipients with complete donor risk information that had a median age of 52 years at the time of transplant, 60.7% of who were male. Just over one fourth (26.2%) of recipients had a history of diabetes and 34 625 (31.7%) were African American (Table 1). Recipients receiving an ECD kidney comprised 18.2% of the cohort. Approximately one third (33.5%) of the recipients had a college and/or graduate degree (Table 1).

The median age of the donors was 40, 59.3% were male and 12.3% were African American (AA). Almost one-fourth (24.2%) of donors had a history of hypertension and 5.4% had a history of diabetes. The median terminal creatinine value was 1.00 mg/dL and 40.9% of the donors died of stroke. A small percentage of the recipients were transplanted with a DCD kidney (6.8%) and 2.5% of the donor population was hepatitis C positive. The median KDRI score comprised of the 10 donor-related variables was 1.04 (IQR = 0.87–1.25) with a minimum value of 0.58 and a maximum value of 2.71.

The hazard for overall graft loss for recipients receiving an ECD transplant was 1.81(95%CI 1.73-1.90) times greater compared to recipients with a standard criteria donor (SCD) transplant, adjusting for recipient age, race, gender, PKD status, diabetes status, serum PRA percent, income, primary insurance, obesity status and interactions between ECD status and recipient characteristics (Table 2). The median survival time for recipients with an ECD transplant was 6.3 years compared to 10.2 years for recipients with a non-ECD transplant. For SCD transplants, recipients aged 60 and over had 1.50(95%CI 1.46-1.54) hazard for graft loss compared to recipients under 60 years of age, holding all other variables in the model constant. For recipients aged 60 and over who received an ECD transplant, the hazard for graft loss was 1.34(95%CI 1.28-1.39) times greater than for recipients less than 60 receiving an ECD transplant (Table 2a).

Table 2. Recipient characteristics including interaction with donor risk (ECD) and hazard ratios for graft failure
Predictor  HRa95% C.I.Chi-Squaredfp-value
  1. a

    Hazard ratios and 95% C.I.s generated for ECD kidney and serum PRA% assume no interaction and ECD = 0 for recipient characteristic

ECD kidney1.81(1.73–1.90)675.211<.0001
Age (≥ 60 years)(see Table 2a)929.511<.0001
Race (AA)(see Table 2a)759.01<.0001
Diabetes(see Table 2a)658.931<.0001
Gender (Female)(see Table 2a)67.601<.0001
PKD0.75(0.72–0.78)200.131<.0001
DCD1.01(0.95–1.07)0.0710.7893
Serum PRA% (Ref = 0%)     
1–301.04(1.00–1.07)5.0610.0245
31–801.14(1.09–1.21)24.241<.0001
≥ 811.21(1.12–1.31)23.021<.0001
Insurance (Ref = Medicaid&other)     
Medicare1.10(1.06–1.14)22.081<.0001
Private0.87(0.84–0.91)39.511<.0001
Working for Income0.82(0.77–0.87)46.871<.0001
Obese (BMI > 30)1.13(1.10–1.16)94.521<.0001
Education (Ref = None/Grade School)     
High School/GED1.09(1.04–1.14)14.1310.0002
Some College/Bachelor Degree0.96(0.92–1.01)2.6410.1044
Graduate Degree0.95(0.89–1.02)2.3110.1282
ECD*Ageyes≥ 60yrs(see Table 2a)21.311<.0001
ECD*RaceyesAA(see Table 2a)39.681<.0001
ECD*Diabetesyesyes(see Table 2a)12.7210.0004
ECD*Genderyesfemale(see Table 2a)6.9010.0086
ECD*DCDyesDCD1.23(1.08–1.40)9.4310.0021
ECD*PRAyes1–30%1.02(0.95–1.08)0.2210.6366
ECD*PRAyes31–801.08(0.95–1.22)1.3710.2417
ECD*PRAyes≥ 811.26'(1.03–1.55)4.8410.0278
Table 2a. Contrast estimation for donor risk (ECD) by age, race, diabetes status and gender
ContrastHR estimate95%CICI Chi-squarep-Value
Age    
60 vs. < 60 years, SCD kidney1.50(1.46-1.54)  
60 vs. < 60 years, ECD kidney1.34(1.28-1.39)  
ECD effect, < 60 years1.64(1.59-1.69)953.59<0.0001
ECD effect, 60 years1.47(1.41-1.52)392.83<0.0001
Race    
AA vs. non-AA, SCD kidney1.40(1.37-1.44)  
AA vs. non-AA, ECD kidney1.19(1.14-1.25)  
ECD effect, non-AA1.65(1.60-1.70)1093.58<0.0001
ECD effect, AA1.42(1.36-1.48)279.80<0.0001
Disease status    
Diabetic vs. nondiabetic, SCD kidney1.39(1.36-1.43)  
Diabetic vs. nondiabetic, ECD kidney1.27(1.21-1.33)  
ECD effect, nondiabetic1.62(1.57-1.67)1029.60<0.0001
ECD effect, diabetic1.47(1.41-1.53)320.92<0.0001
Gender    
Female vs. male, SCD kidney0.90(0.88-0.93)  
Female vs. male, ECD kidney0.85(0.81-0.88)  

There was a 40.1% increase in the adjusted hazard of graft failure for AAs compared to all other races receiving SCD kidneys as opposed to an increase of 19.2% for AAs compared to non-AAs receiving ECD kidneys (Table 2a). Significant interactions of donor risk were also present for recipient gender and a history of diabetes (Table 2). Compared to male recipients, female recipients who received ECD transplants had a greater reduction in their rates of graft failure compared to those who did not. The rate of graft failure decreased by 15.5% for female recipients compared to male recipients with an ECD transplant as opposed to a reduction of 9.6% in graft failure rate for female recipients compared to male recipients with a SCD transplant.

Recipients with a history of diabetes who received a SCD transplant had a 39.2% hazard for graft failure compared to recipients without diabetes while the hazard was 27.0% higher for diabetic recipients who received an ECD transplant compared to nondiabetic recipients with an ECD transplant. There were no significant interactions of donor risk by ECD status for serum PRA percent, education level, obesity status, income status, type of insurance or history of PKD (Table 2). Although the interaction between ECD status and receiving a DCD kidney was significant, adjusting for interactions overall resulted in a nonsignificant association between this predictor and the hazard of graft failure. The likelihood ratio (LR) statistic comparing models with and without interaction was 92.0 with 9 degrees of freedom and p < 0.001, indicating that the model with the interaction terms is a better overall fit.

Figure 1 depicts the hazard of transplant characteristics by donor kidney quality for overall graft loss. Recipients who were 60 and older, AA or diabetic had significantly lower adjusted hazard rates associated with ECD kidney transplants compared to their counterparts; the effect of ECD transplants is shown to significantly differs with respect to recipient characteristics. The difference in survival rates 10 years after transplant was 18% higher for non-AAs who received a SCD kidney compared to those who received an ECD kidney while AAs who received a SCD kidney had a 13% higher survival rate compared to AA recipients with an ECD transplant (p-value < 0.0001). The difference in survival rates at 10 years posttransplant among recipients less than 60 by ECD status was 5% greater than the difference in survival rates among recipients 60 years or older by ECD status (p-value < 0.0001).

image

Figure 1. Hazard rates of transplant characteristics by donor quality for overall graft loss.

Download figure to PowerPoint

As KDRI increased by one unit the rate of graft failure increased by 248% (Table 3). Recipients aged 60 years and over had an increased hazard rate of 1.80 (95%CI 1.67-1.95) for graft failure compared to recipients less than 60 assuming a KDRI score of 0. AAs had a 62.9% increased adjusted rate of graft failure compared to recipients of other races assuming a KDRI score of 0 and recipients with a history of diabetes and a KDRI = 0 had a hazard rate of 1.59 (95%CI 1.46-1.72) compared to recipients without diabetes (Table 3a). For recipients 60 and older with a KDRI score of 1.00, the hazard rate was reduced to 1.49 compared to recipients less than 60 while the hazard for a recipient 60 years or more with a KDRI score of 1.50 was reduced to 1.35 compared to those less than 60. AA recipients with a score of KDRI = 1.00 had a 38.2% increased rate of graft failure compared to all other races and the rate was further reduced to 27.3% compared to non-AAs with a KDRI = 1.50. Recipients with diabetes and a KDRI = 1 had a reduced hazard rate of 1.39 and those with a KDRI = 1.50 had a hazard rate of 1.31 compared to nondiabetic recipients (Table 3a).

Table 3. Recipient characteristics including interaction with donor risk (KDRI) and hazard ratios for graft failure
Predictor  HRa95% C.IChi-Squaredfp-value
  1. a

    Hazard ratios and 95% C.I.s generated for predictors including KDRI, gender, and serum PRA% assume no interaction (and KDRI = 0 for recipient characteristics)

KDRI2.48(2.33–2.63)847.771<.0001
Age (≥ 60 years)(see Table 3a)214.241<.0001
Race (AA)(see Table 3a)148.851<.0001
Diabetes(see Table 3a)124.061<.0001
Gender (Female)1.01(0.93–1.09)0.0310.8576
PKD(see Table 3a)39.351<.0001
Serum PRA% (Ref = 0)     
1–300.96(0.87–1.06)0.7810.3783
31–801.04(0.86–1.26)0.1510.6949
≥ 810.91(0.68–1.23)0.3610.5504
Insurance (Ref = Medicaid & other)     
Medicare1.11(1.06–1.15)24.591<.0001
Private0.88(0.84–0.91)37.081<.0001
Working for Income0.82(0.77–0.87)47.151<.0001
Obese (BMI > 30)1.14(1.12–1.17)115.311<.0001
Education (Ref = None/Grade School)     
High School/GED1.09(1.05–1.14)15.971<.0001
Some College/Bachelor Degree0.97(0.93–1.02)1.4310.2317
Graduate Degree0.95(0.89–1.02)1.8810.1707
KDRI*Age1.00≥ 60yrs(see Table 3a)34.191<.0001
KDRI*Race1.00AA(see Table 3a)23.301<.0001
KDRI*Diabetes1.00yes(see Table 3a)13.8310.0002
KDRI*Gender1.00female0.89(0.83–0.95)12.0010.0005
KDRI*PKD1.00yes(see Table 3a)6.7010.0096
KDRI*PRA1.001–30%1.08(0.99–1.17)3.0510.0806
KDRI*PRA1.0031—801.12(0.95–1.32)1.7910.1804
KDRI*PRA1.00≥ 811.35(1.04–1.75)4.9410.0262
Table 3a. Hazard Ratios for KDRI score by recipient characteristics
DescriptionPoint Estimate95% Wald Confidence Limits
Age   
≥ 60yrs vs. < 60 yrs, KDRI = 01.801.671.95
≥ 60yrs vs. < 60 yrs, KDRI = 0.501.641.561.72
≥ 60yrs vs. < 60 yrs, KDRI = 1.001.491.451.52
≥ 60yrs vs. < 60 yrs, KDRI = 1.501.351.311.39
Race   
AA vs. non-AA, KDRI = 01.631.511.76
AA vs. non-AA, KDRI = 0.501.501.431.57
AA vs. non-AA, KDRI = 1.001.381.351.41
AA vs. non-AA, KDRI = 1.501.271.231.32
Diabetes Status   
Diabetes vs. no Diabetes, KDRI = 01.591.461.72
Diabetes vs. no Diabetes, KDRI = 0.501.491.421.56
Diabetes vs. no Diabetes, KDRI = 1.001.391.361.43
Diabetes vs. no Diabetes, KDRI = 1.501.311.261.35
PKD Status   
PKD vs. no PKD, KDRI = 0.0.630.550.73
PKD vs. no PKD, KDRI = 0.500.680.620.74
PKD vs. no PKD, KDRI = 1.000.740.710.77
PKD vs. no PKD, KDRI = 1.500.800.760.85

Significant interactions by donor risk using the KDRI were consistent with the ECD model for age, race and a history of diabetes (Table 3). The interaction between KDRI and a primary diagnosis of PKD was also statistically significant (Tables 3 and 3a). Recipients with PKD and a KDRI = 1.00 had an adjusted hazard rate of graft failure reduced by 26.1% compared to recipients without PKD. When the KDRI is increased to 1.50, the hazard rate of graft failure in recipients with PKD was reduced by 20.0% compared to recipients without PKD (Table 3). Although the interactions with gender and serum PRA percent were significant, adjusting for them resulted in nonsignificant associations between the main predictors and the hazard of graft failure (Table 3). There were no significant interactions of donor risk by KDRI for recipient obesity status, income status, type of insurance and education level. The LR statistic comparing models with and without interaction was 103.0 with 9 degrees of freedom and p < 0.001, indicating the full model as a better overall fit.

The adjusted hazard for overall graft loss, patient death and death-censored graft loss for recipient (a) race, (b) diabetes status and (c) age by donor quality interaction measured by KDRI score is depicted in Figure 2. In each graph the reference category is contrasted with recipients who were AA, diabetic or 60 years or older, revealing the interaction associated with each recipient characteristic and KDRI where increasing KDRI results in a decrease in the difference of overall hazard of graft loss between the two recipient groups in each case. The adjusted hazard of patient death for the same recipient characteristics reveals that the interaction associated with each characteristic and KDRI persists. Overall, the hazard of death is greater for diabetics and older recipients while the difference in the hazard rate of death by race group is attenuated and nonsignificant at higher KDRI scores. With death-censored graft loss as the final outcome, the adjusted hazard for the same recipients revealed that the interaction with each characteristic and KDRI persists. For AA individuals, the hazard rate of death-censored graft failure decreased with increasing KDRI score with the overall hazard rates being greater than in the original model (Figure 2A–C).

image

Figure 2. A–C. Adjusted hazard of recipient (A) race, (B) diabetes status and (C) age by donor quality interaction measured by KDRI score for overall graft loss, patient death and death-censored graft loss.

Download figure to PowerPoint

Figure 3 shows the adjusted survival 5 years posttransplant of recipients by (a) race, (b) diabetes status, (c) age and (d) all characteristics with KDRI score interaction. The difference in survival rates between AA and non-AA recipients decreased with increasing KDRI score with 95% confidence limits overlapping at a KDRI score of 2.25 (Figure 3A). Similar trends are displayed in Figures 3B-D with Figure 3D comparing recipients who are AA, diabetic and greater than or equal to 60 years of age to those who are non-AA, nondiabetic and less than 60 years of age.

image

Figure 3. A–D. Adjusted survival 5 years posttransplant of recipient (A) race, (B) diabetes status, (C) age and (D) all characteristics by donor quality interaction with lower and upper 95% confidence limits for survival measured by KDRI score.

Download figure to PowerPoint

The final ECD and KDRI models were further adjusted to determine whether a center effect or the amount of time spent on dialysis had an impact on the hazard ratios associated with particular groups of recipients receiving higher donor kidneys. Adjusting for these factors did not alter the significant interactions or the trends seen in hazard rates among the donor risk and recipient characteristics though the hazard rates for AA recipients were slightly attenuated. In order to determine whether selection of recipients for donor kidneys differed before and after the implementation of the ECD policy in 2002, a subanalysis was conducted in which the final ECD and KDRI models were run separately for the two time periods (Table 4). In the paired kidney analysis for the KDRI models, hazard rates were also similar as in the original models (Table 5).

Table 4. Hazard ratios for KDRI score by recipient characteristics stratified by year of transplant according to ECD allocation algorithm in 2002
 Point estimate95% Wald confidence limits
DescriptionLE 2002GT 2002LE 2002GT 2002
Age
60 vs. < 60 years, KDRI = 01.763.24(1.60-1.94)(2.71-3.87)
60 vs. < 60 years, KDRI = 0.501.652.62(1.56-1.75)(2.35-2.93)
60 vs. < 60 years, KDRI = 1.001.552.12(1.50-1.59)(2.01-2.24)
60 vs. < 60 years, KDRI = 1.501.451.72(1.39-1.51)(1.61-1.84)
Race
AA vs. non-AA, KDRI = 01.571.28(1.43-1.72)(1.06-1.54)
AA vs. non-AA, KDRI = 0.501.461.19(1.38-1.54)(1.06-1.33)
AA vs. non-AA, KDRI = 1.001.361.10(1.32-1.40)(1.04-1.16)
AA vs. non-AA, KDRI = 1.501.271.01(1.22-1.32)(0.95-1.09)
Diabetes status
Diabetes vs. no diabetes, KDRI = 01.662.30(1.50-1.83)(1.93-2.75)
Diabetes vs. no diabetes, KDRI = 0.501.562.02(1.47-1.65)(1.81-2.26)
Diabetes vs. no diabetes, KDRI = 1.001.471.77(1.43-1.51)(1.68-1.87)
Diabetes vs. no diabetes, KDRI = 1.501.381.56(1.32-1.44)(1.46-1.66)
Table 5. Hazard Ratios for KDRI score by recipient characteristics for paired kidney analyses
DescriptionPoint Estimate95% Wald Confidence Limits
Discordant age recipients   
≥ 60yrs vs. < 60 yrs, KDRI = 01.831.572.13
≥ 60yrs vs. < 60 yrs, KDRI = 0.501.671.521.83
≥ 60yrs vs. < 60 yrs, KDRI = 1.001.521.461.59
≥ 60yrs vs. < 60 yrs, KDRI = 1.501.391.311.48
Discordant race recipients   
AA vs. non-AA, KDRI = 01.481.271.73
AA vs. non-AA, KDRI = 0.501.371.251.50
AA vs. non-AA, KDRI = 1.001.261.211.32
AA vs. non-AA, KDRI = 1.501.171.101.24
Discordant diabetes status recipients   
Diabetes vs. no Diabetes, KDRI = 01.601.381.85
Diabetes vs. no Diabetes, KDRI = 0.501.511.381.65
Diabetes vs. no Diabetes, KDRI = 1.001.431.371.49
Diabetes vs. no Diabetes, KDRI = 1.501.351.271.43

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. Reference

The principal findings of our study indicate that donor quality has statistically and clinically significant interactions with recipient characteristics. In particular, the impact of donor risk varies dramatically with respect to recipient race, primary diagnoses and age. Higher donor risk kidneys (e.g. high KDRI donor kidneys) have an attenuated effect on older, diabetic and AA recipients relative to their counterparts whereas the effect of lower donor risk kidneys have a pronounced effect on graft survival between groups. These findings have critically important ramifications for both allocation policy and organ acceptance decisions. Further understanding of the mechanisms of these associations and mechanisms to incorporate these results into the distribution of deceased donor organs are needed.

Research has shown that ECD kidneys may be particularly important options for older recipients, those of minority ethnicities, and recipients with a disease that is associated with shorter survival while on the waiting list, such as diabetes [14-16]. Results from a previous study indicated that the risk for overall graft loss was significantly reduced among older AA and Caucasian recipients when compared to younger recipients in which AA patients had a significantly elevated risk compared to their Caucasian counterparts. It was concluded that among AA transplant recipients, donor risk should not be considered as a uniform risk factor [17]. As of October 31, 2005, 42.6% of the candidates on the transplant waiting list had an ECD designation including older and diabetic candidates being significantly more likely to be listed for ECD in addition to many candidates who will not necessarily benefit from receiving an ECD kidney [14, 15]. Findings in the current study further indicate that AA, diabetic and older recipients are less adversely affected by higher risk donor kidneys compared to their counterparts. Based on present findings, it may be critically important to not simply report a single donor risk score (i.e. KDRI) when an organ becomes available, but to tailor this score to the candidate characteristics (e.g. diabetic status, age and race). This type of personalized information may improve the efficiency of allocation, clinical outcomes and individual decision-making.

ECD transplants have become more appealing due to expanding waiting times to receive a deceased donor organ [15]. The average waiting time for patients on dialysis has increased by about 2 years since 2003 [15, 18]. In the first 18 months after the ECD kidney policy was implemented, ECD kidney transplants increased by 15.0% [19]. The aging population has also increased the proportion of donors characterized as ECD. If clinicians are able to inform patients regarding their appropriateness for a higher risk kidney transplant, the number of ECD kidneys discarded and cold ischemia times for those transplanted may be reduced while simultaneously optimizing personal expected quality adjusted life years for these recipients.

Since 1995, the current kidney allocation system and the aging of candidates on the waitlist have resulted in a decline by 18 months in the lifespan of patients posttransplant [20]. The current system is primarily focused on allocating kidneys based on waiting time for a transplant without accounting for possible differences in survival of recipients by donated organs. The exception to this is the ECD distinction that is not always accurate in assessing the relative survival potential of deceased-donor kidneys [12]. “Survival matching” between donated kidneys and wait-listed kidney-transplant recipients has been suggested to avoid further loss of posttransplantation survival [20]. To facilitate this process, how donor risk varies by recipient characteristics should be taken into account.

The need to improve the current allocation system has been a topic of debate since prior to the introduction of the ECD concept in 2002 [14-16, 19-23]. One suggested approach that may allow for all transplant recipients to benefit would be for patients with shorter predicted lifetimes to list for and receive an ECD kidney earlier whereas patients with longer predicted lifetimes would spend more time on the wait list to receive SCD kidneys or lower risk kidneys [22]. The consistency in results in this study provides a means to further enhance the allocation system with a continuous risk scale. More recently, in the fall of 2012, the OPTN/UNOS kidney transplant committee proposed revisions to the current deceased kidney allocation policy. The proposal seeks to revise the national kidney allocation system primarily to boost posttransplant survival benefit and to increase the utilization of donated kidneys. Comments from community members will be reviewed after the December 14, 2012 deadline.

When the final models in this study were further adjusted to consider a center effect and the time recipients spent on dialysis, significant interactions among donor and recipient factors were not altered. The paired kidney subanalysis revealed that some degree of selection bias may impact the estimated effect of donor kidneys as the hazard ratios were slightly attenuated. However, the qualitative results remained consistent regarding the diminished relative impact of donor kidney risk at higher levels among AAs, diabetics and older recipients. In addition, analyzing the results by pre- and post-ECD designation revealed that the attenuation effects do differ slightly by era [24], which may imply that recipients of marginal donors were selected differently in different time periods. Finally, it is possible that other recipient characteristics may be associated with significant interactions with regards to donor risk and overall graft loss but the statistical power in the models to detect these interactions may be insufficient.

One main limitation to this study is the nature of the design where model adjustments may not account for all possible confounding factors or risks, thus leading to bias. For older and diabetic recipients, the reason for overall graft loss is often death not related to graft failure. These competing risks may help to explain why receiving a donor kidney with a higher KDRI score may not have as much of an impact on elderly recipients. However, the interactions between donor quality and recipient age and diabetic status remained significant when the outcome of interest modeled was patient death. These results should be integrated into patients’ prognoses on dialysis to determine how a prospective transplant would affect overall survival outcomes on an individual recipient basis. Finally, there are limitations to the predictive accuracy of this data and its applicability to individual patients due to generalizations made when grouping recipients. In summary, our study results indicate that the effects of donor risk are not uniform for transplant recipients with respect to overall graft loss. Donor kidney quality has a more pronounced effect in lower risk recipients who are younger, not AA and without a history of diabetes.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. Reference

The data reported here have been supplied by the Minneapolis Medical Research Foundation as the contractor for the Scientific Registry of Transplant Recipients (SRTR). The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy of or interpretation by the SRTR or the US Government.

Disclosure

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. Reference

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

Reference

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. Reference
  • 1
    United Network for Organ Sharing. Current U.S. Waiting List. Available at: http://optn.transplant.hrsa.gov/data/ Accessed May 05, 2012.
  • 2
    OPRN/SRTR Annual Report. Table 1.5 Time to Transplant. New Waiting List Registrations, 2000–2009. Available at: http://www.srtr.org/annual_reports/2010/105_dh.htm. Accessed May 05, 2012.
  • 3
    United Network for Organ Sharing. Donors Recovered in the U.S. by Donor Type. Available at: http://optn.transplant.hrsa.gov/latestData/rptData.asp. Accessed May 05, 2012.
  • 4
    Lysaght MJ. Maintenance dialysis population dynamics: Current trends and long-term implications. J Am Soc Nephrol 2002; 13: S37S40.
  • 5
    Meier-Kriesche H, Port FK, Ojo AO, et al. Effect of waiting time on renal transplant outcome. Kidney Int 2000; 58: 13111317.
  • 6
    Ojo AO, Hansen JA, Meier-Kriesche H, et al. Survival in recipients of marginal cadaveric donor kidneys compared with other recipients and wait-listed transplant candidates. J Am Soc Nephrol 2001; 12: 589597.
  • 7
    Wolfe RA, Ashby VB, Milford EL, et al. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med 1999; 341: 17251730.
  • 8
    Akkina SK, Asrani SK, Peng Y, Stock P, Kim WR, Israni AK. Development of organ-specific donor risk indices. Liver Transpl 2012; 18: 395404.
  • 9
    Audard V, Matignon M, Dahan K, Lang P, Grimbert P. Renal transplantation from extended criteria cadaveric donors: Problems and perspective overview. Transplant Int 2008; 21: 117.
  • 10
    Rao PS, Schaubel DE, Guidinger MK, et al. A comprehensive risk quantification score for deceased donor kidneys: The kidney donor risk index. Transplantation 2009; 88: 231236.
  • 11
    Schold JD, Kaplan B, Baliga RS, Meier-Kriesche H. The broad spectrum of quality in deceased donor kidneys. Am J Transplant 2005; 5: 757765.
  • 12
    Port FK, Bragg-Gresham JL, Metzger RA, et al. Donor characteristics associated with reduced graft survival: An approach to expanding the pool of kidney donors. Transplantation 2002; 74: 12811286.
  • 13
    OPTN KDPI/KDRI Calculator. Available at: http://optn.transplant.hrsa.gov/resources/allocationcalculators.asp?index=81 Accessed June 20, 2012.
  • 14
    Merion RM, Ashby VB, Wolfe RA, et al. Deceased-donor characteristics and the survival benefit of kidney transplantation. JAMA 2005; 294: 27262733.
  • 15
    Schnitzler MA, Whiting JF, Brennan DC, et al. The expanded criteria donor dilemma in cadaveric renal transplantation. Transplantation 2003; 75: 19401945.
  • 16
    Schold JD, Meier-Kriesche H. Which renal transplant candidates should accept marginal kidneys in exchange for a shorter waiting time on dialysis? Clin J Am Soc Nephrol 2006; 1: 532538.
  • 17
    Schold JD, Srinivas TR, Braun WE, Shoskes DA, Nurko S, Poggio ED. The relative risk of overall graft loss and acute rejection among African American renal transplant recipients is attenuated with advancing age. Clin Transplant 2011; 25: 72130.
  • 18
    Understanding the Organ Transplant Waiting List. Gift of Life Donor Program. Available at: http://www.donors1.org/patients/waitinglist/. Accessed August 17, 2012.
  • 19
    Sung RS, Guidinger MK, Lake CD, et al. Impact of the expanded criteria donor allocation system on the use of expanded criteria donor kidneys. Transplantation 2005; 29: 12571261.
  • 20
    Leichtman AB, McCullough KP, Wolfe RA. Improving the allocation system for deceased-donor kidneys. N Engl J Med 2011; 364: 12871289.
  • 21
    Wolfe RA, McCullough KP, Schaubel DE, et al. Calculating life years from transplant (LYFT): Methods for kidney and kidney-pancreas candidates. Am J Transplant 2008; 8: 9971011.
  • 22
    Schold JD, Hall YN. Enhancing the expanded criteria donor policy as an intervention to improve kidney allocation: Is it actually a “net-zero” model? Am J Transplant 2010; 10: 25822585.
  • 23
    Sung RS, Guidinger MK, Leichtman AB, et al. Impact of expanded criteria donor allocation system on candidates for and recipients of expanded criteria donor kidneys. Transplantation 2007; 84: 11381144.
  • 24
    Hirth RA, Pan Q, Schaubel DE, Merion RM. Efficient utilization of the expanded criteria donor (ECD) deceased donor kidney pool: An analysis of the effect of labeling. Am J Transplant 2010; 10: 304309.