SEARCH

SEARCH BY CITATION

Keywords:

  • Deceased donor;
  • expanded criteria donor;
  • graft survival;
  • kidney;
  • renal transplantation;
  • risk factors

Abstract

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

The quality of the deceased donor organ clearly is one of the most crucial factors in determining graft survival and function in recipients of a kidney transplant. There has been considerable effort made towards evaluating these organs culminating in an amendment to allocation policy with the introduction of the expanded criteria donor (ECD) policy.

Our study, from first solitary adult deceased donor transplant recipients from 1996 to 2002 in the National Scientific Transplant Registry database, presents a donor kidney risk grade based on significant donor characteristics, donor–recipient matches and cold ischemia time, generated directly from their risk for graft loss. We investigated the impact of our donor risk grade in a naïve cohort on short- and long-term graft survival, as well as in subgroups of the population.

The projected half-lives for overall graft survival in recipients by donor risk grade were I (10.7 years), II (10.0 years), III (7.9 years), IV (5.7 years) and V (4.5 years). This study indicates that there is great variability in the quality of deceased donor kidneys and that the assessment of risk might be enhanced by this scoring system as compared to the simple two-tiered system of the current ECD classification.


Introduction

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

As early as 1980, studies have documented that the quality of organs from deceased donors in kidney transplantation represents one of the most crucial factors affecting graft survival (1). Subsequently, there have been numerous research papers evaluating various potential donor risk factors for graft loss in transplantation highlighting the importance of the organ characteristics independent of the transplant recipient (2–10). Studies have addressed donor disease history, anatomical characteristics of transplanted organs, utilizing biopsy results, donor and recipient matching and donor demographic characteristics. Beyond the individual findings, the prodigious body of research on this topic underlines the significance to the field of assessing the quality of the donated organ in kidney transplantation.

The Expanded Criteria Donor (ECD) designation has served major importance to the transplant community by labeling those deceased donor kidneys with a “high” relative risk for graft loss as well as shortening waiting times for patients consented to receive these organs (6,11). Perhaps even more paramount, the ECD policy was considered to lower discard rates of those organs with a perceived higher risk of graft loss (11). Clearly for the purpose of the utilization of more organs, the ECD definition has an importance in the allocation process. On the other hand, the binary nature of the ECD definition may underappreciate the variability of the quality of the organ, and a more granulated scoring system that is easy to implement may be important to physicians and transplant recipients to make crucial decisions at the time of transplant. The ECD label includes those organs which had an associated relative risk of >1.7 (excluding pediatric donations) from the model generated by the Scientific Transplant Registry (SRTR) for the outcome of overall graft loss (6). This model incorporated donor age, donor history of hypertension, donor serum creatinine level and donor cause of death. By examining the model output used to generate the ECD classification (Figure 1), we can initially observe the tremendous impact of the deceased donated organ on graft survival associated with the magnitude of the hazard ratios; however, we can also observe the variability in the ECD organs [adjusted hazard ratios (AHR) ranging from 1.74 to 2.69] as well as in the non-ECD organs (AHR from 1.00 to 1.66), which translate to substantial expected survival differences.

image

Figure 1. SRTR model output utilized in formulation of ECD.

Download figure to PowerPoint

Nyberg et al. have addressed a more granulated scoring system for kidneys using their local center data as well as from a national database perspective (5,12). Their study examining deceased donor kidneys from the UNOS registry from 1994 to 1999 provided a quantitative approach to evaluating these organs. Their study utilized the four donor variables applicable to the ECD designation along with Human Leukocyte Antigen (HLA) matching. The model for their scoring system was based on 6-month renal function, and from this model they developed a risk scale of 0–40 pertaining to the quality of the organ.

We proposed to build upon this concept by further exploring quantitative assessments of deceased donor quality incorporating a distinct methodology most importantly by looking at the ultimately most important outcome of graft survival, to provide clinicians and patients with a algorithm that can combine several characteristics of differential impact into simple categories to understand the intrinsic risk associated with a particular organ. We intended to demonstrate an objective measure relating to the quality of deceased donor organs in the most comprehensive, accurate and adaptable manner. We further investigated several ramifications of the more granulated donor risk stratification in the context of other clinical outcomes. From data provided by the SRTR on transplants performed between 1996 and 2002, we constructed an index based on all significant risk factors specifically for the outcome of graft loss.

Methods

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

We evaluated all first solitary deceased donor transplant recipients transplanted between 1996 and 2002 from the SRTR database with last follow-up data available from May 1st 2003. Retransplants, multiple organ and pediatric transplant recipients were excluded from our study. For the purpose of generating a donor risk index, we constructed a multivariate model for the outcome of overall graft survival adjusted for recipient and donor characteristics, recipient primary diagnosis, recipient PRA level, recipient–donor HLA (HLA-A, -B and -DR) matching and waiting time on dialysis. Missing variables for the variable cold ischemia time were categorized as an additional level and utilized in the model construction, missing values of pretransplant dialysis time were considered preemptive transplants, for the variable of recipient weight in the Cockgroft-Gualt clearance estimation, we utilized the average weight observed in the immediate follow-up periods if it was missing, or in several cases, when the straddling weights were also not available, we utilized the transplant weight indication alone. From the output of this model, we extracted the parameter estimates of significant “donor” risk factors. We tested for several interaction terms within the structure of the model. We then generated a risk index from the summation of the parameter estimates that were applicable for individuals. We examined the distribution of this index score, and using cluster analysis, generated intervals that best defined the natural groupings of risk scores. We used disjoint clustering analysis, maximizing the least squares distances between cluster centers, and observed the approximate expected coefficient of determination with each additional cluster. The final determination of the number of groups was a result of examining the relative incremental gain with the inclusion of additional groups and determining where additional groups fail to contribute significant gain.

In order to validate our findings, and avoid overparameterization, we created a model and test sample from our defined population. We randomly assigned individuals to the respective groups at an approximate 2:1 ratio with the model sample constituting 67.2% of the entire cohort. We replicated our analysis procedures in the model group alone and, based on this output, observed the results in the test sample. To further illustrate the independent impact of donor quality, we evaluated outcomes in a chosen homogenous sample. We selected this specific test sample from the most frequently populated recipient age, primary diagnosis and ethnicity combination, in order to more completely extricate hidden selection biases on organ allocation that may be masked in a retrospective analysis. Additionally, we compared outcomes of our donor risk grades with transplants from living donors and from six antigen-matched donations (excluding those from the risk grade groups) to assess the relationship with these transplant conditions.

In order to compare our model with other donor risk formulations, we tested alternative models on the same cohort directly for the outcome of overall graft loss adjusted for the same recipient and transplant covariates. We used the difference in the Log Likelihood statistic (−2*Log Likelihood) of the three models, which is approximately Chi-square distributed, generated from using the ECD categorization, the Nyberg model and our model as comparators (13). We also reported the Akaike Information Criteria (AIC) which incorporates the residual sum of squares of the regression model with a penalty incorporated for additional parameter inclusion (14).

Delayed graft function was defined as a need for dialysis in the first week or failure to produce urine (>40 mL) within 24 h posttransplant. We calculated projected life years for recipients of the applicable donors by extrapolating the univariate survival plots grafted on to the end using data after 2 years (in order to obtain the long-term attrition rate) and calculating the time by which 50% survival was achieved from this projection. The projected survival rates were both graphically and analytically (r2 values >0.99 for the time period with complete data) a strong fit for the projected slopes by donor group strata.

We utilized Kaplan–Meier survival plots for univariate analysis and Cox proportional hazard models adjusted for possible confounding factors (excluding those utilized to generate our risk index) to examine the association of the donor risk grade and graft and patient survival. Chi-square tests were utilized for testing the independence of categorical variables. We used Cochrane–Armitage trend tests to evaluate any significant linear association of rates for binary variables. A type-I error probability of 0.05 was the threshold of significance for our analyses. All analyses were performed on SAS v.9.0 (Cary, NC).

Results

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

Calculation of the donor risk score

The risk index was created by the summation of all significant donor risk factors that were applicable to the individual from a multivariate model also adjusted for recipient characteristics. The parameter estimates used for creation of the risk score (greater magnitude indicates greater risk) are summarized in Table 1.

Table 1.  Adjusted parameter estimates for calculation of risk score
Donor characteristic Level of variable Parameter estimatePercentage of donors with characteristic
  1. 1Missing values excluded from frequency distribution.

Donor/Recipient CMV matchD+/R−0.13623.1
 Other combinations076.9
Donor raceAfrican American0.16511.1
 Non-African American088.9
Donor age0–60.3783.7
 7–110.3033.1
 12–29031.2
 29–390.13814.8
 40–490.26820.0
 50–590.42217.7
 60–690.6768.0
 70+0.7701.5
Cause of deathCerebrovascular0.08941.0
 Other059.0
HLA-A mismatches0–1057.7
 20.06942.3
HLA-B mismatches0–1060.1
 20.11139.9
HLA-DR mismatches0032.1
 10.08542.3
 20.16325.6
Cold ischemia time1 (h)0–9014.3
 10–180.10853.5
 19–290.15227.6
 30+0.2614.5
Donor history of hypertensionNo080.2
 Yes0.13819.8
Donor history of diabetesNo096.3
 Yes0.1563.7

Donor risk grades were determined by cluster analysis and an evaluation of the incremental gain in information by utilizing additional groups. The expected coefficient of determination for the number of clusters were three (88.9%), four (93.8%), five (96.0%) and six (97.2%). We selected the five group model, which appeared to have strong discriminatory power while still maintaining clinically relevant sample sizes. The group ranges for the overall risk score are summarized in Table 2.

Table 2.  Donor grade statistics
Donor gradeRelative frequency n (%)Risk score rangeAHRs1 0–1 yearsAHRs1 1–5 years
  1. 195% CI for AHRs in parentheses.

I5084 (11.1%) [0–0.234]ReferenceReference
II14881 (32.5%)(0.234–0.524]1.34 (1.18, 1.53)1.05 (0.94–1.18)
III15782 (34.4%)(0.524–0.853]1.85 (1.63, 2.10)1.36 (1.27–1.51)
IV7782 (17.0%)(0.853–1.17]2.55 (2.24, 2.90)2.00 (1.79–2.24)
V2321 (5.1%)>1.173.72 (3.22, 4.31)2.22 (1.92–2.57)

In general, we observed very similar results in the test sample as in the results encompassing all observations. The relative frequencies of the donor grades in the model and test samples were I (11.1%, 11.1%), II (32.6%, 32.3%), III (34.6%, 34.1%), IV (16.7%, 17.5%) and V (5.1%, 5.0%). The parameter estimates from the model group alone were also similar to the final estimates, 5-year overall graft survival rates in the overall model by donor grade (shown in Figure 2) were also similar to results in our test sample as follows: I (75.1%), II (73.6%), III (65.6%), IV (54.2%) and V (43.3%).

image

Figure 2. Kaplan–Meier Plot of overall graft survival by donor grade.

Download figure to PowerPoint

The recipient characteristics of the entire cohort by donor grade are summarized in Table 3. There was a statistically significant trend towards African Americans and recipients on dialysis greater than 36 months receiving a lower-quality kidney. There was also a slight increase in grade IV and V kidneys over the period of study, between the 1996 and 2002, the rate of grade IV donations increased from 16.0% to 17.7% and the rate of grade V donations increased from 4.9% to 5.8% over the same period. To simulate certain organ selection decisions, we added 12 and 18 h separately to the cold ischemia time, as a result 20.3% and 35% of the organs, respectively, were reclassified as a more severe risk grade with the additional ischemia time.

Table 3.  Recipient characteristics by donor grade
VariableDonor grade
IIIIIIIVVp-Value1/direction
  1. 1From Cochrane–Armitage trend test for significant linear trend for binary variables. Direction displayed for significant trends.

Recipient age [mean (s.d.)]48.5 (12.6)47.6 (12.8)48.2 (12.6)50.8 (12.4)55.3 (11.9)n/a
Recipient female%40.639.640.038.540.30.2083
Recipient African American%18.828.630.933.237.2<0.0001 increasing
Dialysis time >36 mos.%27.633.135.439.039.6<0.0001 increasing
Recipient primary diagnosis of diabetes26.823.722.924.826.90.7264
Recipient PRA level >3013.514.213.213.312.50.0233 decreasing

Graft and patient survival by donor risk grade

The projected half-lives by donor grade, calculated utilizing data beyond 2-year posttransplant, were I (10.7 years), II (10.0 years), III (7.9 years), IV (5.7 years) and V (4.5 years). The donor grades were independently significant for the outcome of overall graft survival in the presence of recipient and other confounding factors (Figure 3). In a similar fashion, for the outcome of death censored graft loss, with grade I as reference, the relative risks were as follows: II (AHR = 1.32, 95% CI 1.17–1.48), III (1.87, CI 1.67–2.09), IV (2.93, CI 2.61–3.29) and V (4.19, CI 3.67–4.78). To further confirm this point, we produced the survival plots for donor grade for recipients older and younger than 50, and noted a similar survival pattern with the expectable decreased shift in the older sample.

image

Figure 3. Multivariate estimates for graft loss by donor grade (model additionally adjusted for recipient age, primary diagnosis, peak PRA level, recipient gender, recipient race and waiting time on dialysis). *Grade I donor serves as the reference group in the model.

Download figure to PowerPoint

For the specific homogenous sample, we selected which was chosen as the most common age (44–54), diagnosis (diabetes) and ethnicity (Caucasian). The sample size for this group was 2230 and the results in this group demonstrated a similar effect as in the population. Of the 2230 recipients, the distribution by donor grade was as follows: I (14.4%), II (34.0%), III (32.8%), IV (15.1%) and V (3.8%). Delayed graft function rates were I (14.9%), II (25.8%), III (30.2%), IV (37.5%) and V (41.7%). Overall graft survival rates by donor grade at 5 years were I (82.7%), II (81.1%), III (73.1%), IV (63.4%) and V (46.7%).

Patient survival was also independently associated with donor grade. Increasing donor grade was associated with higher AHRs for overall death (with grade I as the reference group): II (AHR = 1.08, 95% CI 0.98–1.20), III (1.37, CI 1.25–1.52), IV (1.83, CI 1.66–2.03) and V (2.14, CI 1.89–2.41).

Donor risk score compared to alternative classifications

The frequencies of ECD and non-ECD kidneys by our donor grade category are summarized in Table 4. As expected, there was a clear positive relationship between the presence of ECD labeled kidneys in the higher donor grades. However, there was as significant percentage of ECDs (14.0%) in the grade III category, and almost 10% of the non-ECD kidneys in the grade IV and V categories. We also examined the distribution by dividing donors by age greater than 50 only. In this case, we also found significant “crossover,” the percentages of older kidneys by grade were I (0%), II (1.3%), III (32.9%), IV (46.5%) and V(19.3%). For the younger donors, the distribution was I (14.8%), II (43.0%), III (34.9%), IV (7.0%) and V (0.3%). To illustrate this point further, we presented theoretical donor cases and their associated risk scores in Table 5. Of note was that despite not being classified as an ECD by the current definition, the younger donor would have produced a higher-risk score in our model.

Table 4.  Distribution of ECD and non-ECD donor kidneys by donor grade
Donor gradeNon-ECD (%)ECD (%)Five-year overall graft survival (%)
I13.2076.7
II38.70.173.6
III38.414.066.3
IV9.456.154.8
V0.329.947.6
Table 5.  Examples of applying donor rating in theoretical patients
Tx caseDonor factors1
CMV D+/R−A.A.AgeCITHLA-AHLA-BHLA-DRHx of hyp.Hx of diab.COD CVScoreGrade
  1. 1Risk factors (from left to right): Donor CMV+/recipient−, donor African American, donor age, estimated cold ischemia time, HLA-A, -B and -DR mismatches, donor history of hypertension and diabetes, cerebrovascular cause of death.

#1NoNo6211210001 
Score000.67600.06900000.0890.834III
#2NoYes4532012101 
Score00.1650.2680.261000.1630.13800.0891.084IV

The degree to which our model described the variability in graft loss relative to the ECD classification or utilizing the Nyberg model can be quantified by examining the Log Likelihood of the models on the same cohort. The −2*Log Likelihood for the models (with lower values indicating better model fit), adjusted for recipients and other transplant characteristics, for the ECD, Nyberg and our models, respectively, were 195,816.5, 195,593.5 and 195,443.1. The differences in these values are approximately Chi-square distributed with degrees of freedom equal to the difference in rank of the models as displayed below:

  • •  
    ECD Model(195,816.5) − Nyberg Model(195,593.5) = 223 ∼χ23 (p-value < 0.001)
  • •  
    Nyberg Model (195593.5) − Schold Model(195443.1) = 150 ∼χ21 (p-value < 0.001)

These results indicate that there is a significant gain of information achieved when comparing the Nyberg model to the ECD model. In addition, there is a significant gain from the model which we present relative to the Nyberg model. To further this point, an alternative model fit statistic, the AIC criteria (models with smaller information criteria are considered to fit the data better), confirms this observation as follows: ECD(195,868), Nyberg(195,649) and our model (195,501).

Deceased donor graft survival by donor risk score compared to living donated kidneys and six antigen matches

Living donor recipients over the same period had a similar survival trajectory as the grade I deceased donor recipients, the 5-year survival rate for the living transplant recipients was 79.6% (as compared to 76.7% in the grade I recipients). This observation was confirmed in multivariate analyses, the AHR for overall graft loss in living transplants, with the grade I deceased donor transplant as the reference group, was 0.95 (95% CI 0.87–1.03) and for death censored graft loss 1.10 (95% CI 0.98–1.23). Six antigen-matched kidneys that were also classified as ECDs had a very similar 5-year survival rate to the grade III kidneys (excluding the six Ag matches). In a similar fashion, in the non-ECD cohort, the six antigen-matched kidneys had an almost identical survival rate at 5 years as the highest quality donor grade I kidneys.

Donor risk score and the association with DGF

The association of the donor gradation was also strongly associated with incidence of delayed graft function. We depicted the rate of delayed graft function in the presence and absence of perfusion pumping (Table 6). The relative decrease in delayed graft function was fairly consistent with pumping; however, the absolute decrease was much more impressive in the high-risk donor grades.

Table 6.  Rates of delayed graft function by use of perfusion pump in donor grades
Donor gradePumped (%)DGF without pump (%)DGF with pump (%)
I10.816.712.6
II12.423.115.4
III11.930.320.5
IV14.639.226.0
V18.246.331.7

Donor risk score and posttransplant renal function and acute rejection

Baseline acute rejection (defined as occurring in the immediate posttransplant period) was also more frequently reported in the lower grade organs, rates by grade were as follows: I (4.3%), II (6.8%), III (8.1%), IV (8.7%) and V (11.5%). For those patients with a minimum of 1-year graft survival, we also found a gradually greater increase in the slope of renal function, as measured by the Cockgroft-Gault estimate, by donor grade, the mean increase in estimated clearance level for grade I donors was 4.6%, while for grade V donors it was 1.1% which was statistically significant (t-test p-value < 0.001). Each donor grade had a gradually decreased slope associated with the donor grade. The average 1-year creatinine levels at 1 year by donor grade were I (1.3), II (1.5), III (1.7), IV (2.0) and V (2.2). This trend was also observed excluding patients with 6-month creatinine levels above 2.0, and using follow-up periods from 1 to 2 years posttransplant.

Donor risk score and long-term graft survival

Long-term graft survival was associated with donor grade for the cohort of patients with a minimum of 1 year overall graft survival, the unadjusted 5 years survival rates under this condition by donor grade were I (81.7%), II (80.2%), III (74.6%), IV (65.2%) and V (62.3%). We retested this association in patients with creatinine level less than or equal to 2.0 at 1 year and found a similar relationship. We also reconstructed the original model for overall graft survival for only those patients with a minimum of 1-year graft survival. The donor variables that remained significant in this model were history of hypertension, history of diabetes, CMV D−/R+, HLA-DR matching, donor age and extended pretransplant dialysis time. The AHRs by donor grade, separated by follow-up in the first-year posttransplant only and from 1 to 5 years, are summarized in Table 2. The adjusted survival by donor grade for those recipients with 1-year graft survival is shown in Figure 4.

image

Figure 4. Adjusted long-term overall graft survival by donor grade (adjusted Cox model results for patients with a minimum of 1-year of graft survival).

Download figure to PowerPoint

Discussion

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

The quality of the donated organ in kidney transplantation is one of the most crucial factors for graft survival. For the patient suffering from ESRD and seeking a kidney transplant, the decision of how to acquire this organ is critical. For those who choose to seek a deceased donor kidney, the main decision that they are presented with is whether or not to consent to receive an ECD, and potentially avoid extensive waiting periods on dialysis. Although this is a crucial function of the ECD categorization, beyond this decision, patients and patient advocates should be well aware that there is wide variability in graft survival associated with the ECD classification alone. In a similar fashion, organs that do not have the ECD label appear to have a broad survival expectation as well. In addition, as we have demonstrated, there are a substantial number of non-ECD kidneys that are associated with a lower survival rate than some ECD kidneys.

The posttransplant morbidity and mortality as measured by delayed graft function, graft and patient loss in the short- and long-term are strongly affected by the quality of the deceased donor kidney as measured by our evaluation index. For this reason, we think it is providing transplant patient care providers and also patients with a crucial tool in evaluating any deceased donor kidney as opposed to formulating their judgment on the quality of the potential organ on the ECD designation alone.

There are several novel aspects of our model to describe deceased donor kidney quality. By utilizing weights (derived from the applicable parameter estimates from our initial model) associated with specific variables, we have a relatively precise estimate of the impact of each factor. Incorporating all significant risk factors may on one hand offer a slightly more complex model, but on the other hand offers an objective approach for this evaluation and in the scheme of finding the most accurate system avoids making subjective determinations. By developing the risk strata using analytical techniques, we also have helped describe the natural distribution of donor risk and determined distinguishable and relevant risk groups. Even though mathematically somewhat complex, the use of this system is fairly simple when used in a spreadsheet or similar format, and all the information is generally available at the time of transplant.

The consistency of the relative frequencies of the donor ratings in the model and test sample adds assurance that these rates are robust and we may expect similar rates in future cases. Of note in the multivariate model for overall graft survival is that the confidence intervals are each separable, supporting the conclusion that these are independent classifications. We attempted to correct for the fact that high-risk donations are more likely to be given to high-risk recipients, by both correcting for recipient characteristics in the model and by stratifying our results by recipient risk groups. Even in these particularly homogenous samples, chosen in an attempt to mitigate some of the potential subtle allocation selection biases, the effect remained strong and our risk score remained the predominant factor associated with graft loss.

As expected, the donor grades demonstrated a strong association with increased rates of delayed graft function. Donor age and cold ischemia time are major contributors to our donor index and are well-known risk factors of delayed graft function. DGF rates were markedly reduced with the use of perfusion pumping. Pumping was associated with a somewhat greater absolute utility because of the higher incidence of DGF in high-risk organs. On the other hand, there was a similar relative DGF rate reduction in all donor risk categories indicating a substantial effect even in low-risk organs. From a cost–benefit perspective, centers may incur a greater benefit by avoidance of initial dysfunction by pumping higher-risk organs as outlined in Table 6.

The risk factors that we utilized for our donor risk index have all been well documented in the literature previously. Without argument, the most predominant risk factor available in the national databases for graft loss is donor age. Some in the transplant community have made the argument to use donor age alone as a marker of quality, and while this approach has the advantage of simplicity. Our data suggest that this approach alone may not always reflect the most accurate estimate of quality of the organ and that much additional information can be obtained by incorporating all significant information available. We found cold ischemia time to be a significant risk factor for graft loss, and as such a significant component of the decision-making process for a particular organ. According to our results, over 20% of organs increase their risk grade with an additional 12 h of ischemia time which mimics a common scenario for centers of whether to use an organ locally or accept an organ from a less proximate source. We chose not to incorporate donor creatinine into our model for two reasons. One being that we felt this measure was somewhat variable depending on the time of measurement and the concomitant pathology, and secondly that it was missing with a relatively high rate in the database.

There was not any obvious association between the average recipient age and donor risk grade, whether this will change after the implementation of the ECD allocation policy will have to be investigated. There also does not appear to be any systematic differentiation by recipient gender, which is reassuring. One of the disturbing results was the heightened allocation of lower-quality donations and African American recipients. Although it has been noted previously that HLA-matching benefits Caucasians with a greater chance to get transplanted, the almost two-fold rate of African Americans receiving the grade V organs over grade I merits further investigation. This might be in part due to higher PRA and longer dialysis times in this population. In fact, greater time waiting on dialysis also appears associated with receiving a lower-quality organ.

One of the possible ramifications of the extremely low prognosis in the grade V donor group is the more tempered overall survival benefit that may be achieved relative to remaining on dialysis for these candidates. Ojo et al. and Merion et al. demonstrated significant survival benefits to receiving marginal organ relative to the candidate remaining on the wait list overall (15,16). With the addition to policy changes reducing waiting time that is associated with receiving these organs, which has been shown to be a significant risk factor following transplantation, this benefit may in fact have increased (4). However, whether this benefit can be applied to the grade V donations representing the worst of the ECDs has yet to be demonstrated, and given the drastic immediate graft loss rate may at the least incur a much extended time to “survival equilibrium.”

On the other end of the spectrum, our findings suggest that the graft survival in the grade I donor quality group was similar to the average recipient of living donor. The observation that a certain proportion of deceased donor transplants that may incur a greater benefit than some living transplants was also identified by Mandal et al. (17). While living donated organs innately have their own variability in quality, there may be cases in which recipients may be better served to accept a grade I, or high-quality deceased donor organ if offered and available without any delay, such as the case of some six antigen-matched deceased donor transplants.

The trend for continuing accelerated renal function deterioration with higher donor risk grade may suggest an association of donor quality and chronic renal disease progression. This notion is strengthened by the observation that this was true also in a subpopulation of patients with good baseline renal function, for even in these patients there was an accelerated slope of renal function loss. In fact, our results somewhat underestimated this effect as we only examined those grafts that survived over the applicable follow-up periods not accounting for the additional grafts that were lost by higher donor grade. This is somewhat in contrast with data recently published that suggested that donor age did not affect the subsequent renal function slope (18). From our observations, not only did recipients of lesser-quality organs have a lower intercept of renal function at the measured time points, but additionally that their trajectory of renal function was affected by the quality of the organ they received.

In summary, we have demonstrated a novel approach to evaluating deceased donor kidneys and described the important impact of donor quality on transplant outcomes. Patients and transplant care givers can both benefit from an objective way of assessing the quality of an available transplant kidney.

Acknowledgments

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

The data reported here have been supplied by the University Renal Research and Education Association (URREA) 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 U.S. Government. IRB approval or exemption determination is the responsibility of the authors as well.

References

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