Key words : Deceased donation, Renal transplantation, Resource utilization, Single center cohort study

Introduction

A significant shortfall remains between the current availability of kidneys and those in need of a renal transplant, with over 6500 patients on the United Kingdom (UK) waiting list.¹ Recent strategies attempting to address this have focused on expanding the use of extended criteria donors, all while maintaining efforts to increase awareness and availability of the living donor sharing scheme. There have been positive results, with a 21% decrease in wait list numbers over 10 years; however this achievement was enabled by a 115% increase in donations after circulatory death (DCD) and the use of “higher risk kidneys,” which have increased by 119% over the same time period.^2-4 Countering these efforts, the prevalence of diabetes as a cause of kidney failure continues to rise, and there are an increasing number of elderly people with comorbid conditions requiring dialysis and/or kidney transplant as a treatment for end-stage renal disease (ESRD).^5-7 Another key priority is maximizing the functioning lifespan of a kidney graft, achievable through better matching between patient and graft life expectancy. The previous UK national kidney allocation scheme, in place since 2006, successfully reduced wait times for those most difficult to match and improved equity for Black, Asian, and minority ethnic patients but conversely has seen an increase in median waiting time, as well as an increase in the number of discarded grafts.⁸

In March 2018, the National Health Service Blood and Transplant (NHSBT) published a proposed kidney matching scheme, implemented in September 2019, with the following principal aims: to better match patient and graft expected survival, to improve wait times for difficult to match patients, and to decrease offer decline rates. Furthermore, donations after brain stem death (DBDs) and DCDs were to be combined into a single allocation scheme. There remains a tiered system where the most difficult to match or those with a wait list duration over 7 years are considered “Tier A,” with remaining patients in “Tier B” subjected to the revised scoring system. The scheme utilizes new donor and recipient risk indices, which rank donors from one to four (D1-D4) and similarly recipients (R1-R4), with the aim of matching recipient and donor (ie, to increase the proportion of R4 recipients receiving a D4 graft, for example). Furthermore, targets for an age difference between the donor and recipient were proposed, with fewer than 8% receiving a kidney with >25-year age difference and 20% receiving a kidney with a 15- to 25-year age difference.⁹

Although it was accepted that 5-year patient and graft survival may be impacted by the proposed scheme, the aim of better matching patient and graft survival was deemed necessary with the theoretical reduced rate of retransplant as a corollary. However, unanswered questions remain as to how this new scheme will impact already overstretched clinical transplant units. Furthermore, the revised allocation scheme was formulated using modeling data from an historic cohort of transplants performed during a period of center-based recipient selection (ie, choosing suitable recipients for offered organs) and therefore may not appropriately account for this clinician discretion and uncaptured bias. Matching the most vulnerable patient group with the poorest quality donor grafts may pose particular challenges with prolonged admissions, higher rates of delayed graft function (DGF), and higher need for reintervention in the early posttransplant period. Here, we evaluated a retrospective cohort, restratified as per the proposed risk index, with the aim of understanding how the revised allocation scheme affects these concerns.

Materials and Methods

A retrospective analysis was performed on prospectively acquired data on all deceased donor kidney transplants performed between January 2015 and December 2016 at the Queen Elizabeth University Hospital, Glasgow (Scotland, UK), a unit that provides transplant services to a population of approximately 2.5 million. Living donor kidney transplants were excluded from the analysis. Both recipients and donors were reclassified using the new quartiles (ie, D1-D4 and R1-R4).¹⁰ The donor risk index considered donor age, height, length of hospital stay, cytomegalovirus (CMV) status, estimated glomerular filtration rate (eGFR), female sex, and past medical history of hypertension (see Table 1). Recipients were stratified as per age, diabetes, and dialysis status and duration. All factors were differentially weighted per a computed algorithm. Because of the retrospective nature of this study, the institutional review board did not require research ethics approval, and reporting of results was completed in accordance with the “Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)” guidelines.

The following patient outcomes were collected: length of stay, readmission, frequency of radiological investigation (eg, computed tomography and transplant ultrasonography scan), and patient and graft survival. The number of inpatient days was calculated as the index admission length, in addition to any readmission days within the 2-year follow-up period. To accommodate censored data (ie, adjusted for censoring events of death or graft loss), we calculated hospitalization as days per 1000 patient days, thus providing data on inpatient bed use that reflects only time with functioning transplant.

The use of day-case facilities was defined as a prearranged outpatient attendance to a facility adjacent to our transplant ward that did not require an overnight stay in the hospital, such as for routine ureteric stent removal; these were not included in inpatient day analyses. Data were obtained from the prospectively maintained Strathclyde Electronic Renal Patient Record (Vitalpulse).

Cost analysis was performed using the individual healthcare item costs available from the National Schedule of Reference Costs (2016-2017).¹¹ Values used are the average unit costs to the NHS of providing the secondary health care items to patients.

Delayed graft function was defined as need for dialysis within 7 days posttransplant, excluding a single session of dialysis within the first 24 hours for hyperkalemia. The total follow-up period was from date of first transplant in the series (January 2015) up to and including December 2019, with admission days only reported for the first 2 years posttransplant. The current and active wait lists for the West of Scotland were also accessed via local transplant coordinators and restratified as per the aforementioned recipient risk indices.

Statistical analyses
Statistical analysis was performed with R version 3.6 for Linux with survival, KMsurv, car, and dplyr and ggplot2 packages.^12,13 Differences between groups were assessed with the Pearson chi-square test. Comparisons between donor and recipient types for nonparametric data were assessed with the Kruskal-Wallis with Dunn post hoc test. Graft function was analyzed using the Tukey honestly significant difference test. The log-rank test was used for patient and graft survival analyses. Cost differences were assessed using Wilcoxon rank sum test with continuity correction.

Results

Of the 195 deceased donor transplants performed during the study period, 82 (42.1%) were DCD and 113 (57.9%) were DBD. Male-to-female ratio was 1.1:1. Mean donor age and mean recipient age were 51 and 51.7 years, respectively. Median wait time from dialysis to transplant was 681 days. Mean recipient body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) was 27.1. With regard to age differences, 2% had an age difference between donor and recipient of >25 years and 12.8% had an age difference of 15 to 25 years, well within the NHSBT proposed figures of 7.8% and 20.4%, respectively. Regarding inpatient bed days, median index admission was 8 days for all transplants, and median total inpatient days was 15 days (mean of 26.8 days; range, 5-158 days) over the 2-year follow-up period (see Table 2).

Recipient and donor risk combination in this cohort
When stratified according to the revised allocation scheme donor risk index, 39 of 195 donors (20.0%) were categorized as D1, 54 (27.6%) were categorized as D2, 36 (18.5%) were categorized as D3, and 66 (33.8%) were categorized as D4. The DBD-to-DCD ratio for each donor category was as follows: 2.25:1 for D1, 1.35:1 for D2, 1.57:1 for D3, and 1:1 for D4. Therefore, the proportion of DCDs was higher in the D4 cohort (50% vs 38.0% in D1-D3); however, this difference was not significant (chi-square test = 3.9, P = .11).

In contrast, the recipient risk index was unevenly distributed, with 5 of 195 (2.6%) categorized as R1, 14 (7.2%) categorized as R2, 32 (16.4%) categorized as R3, and 144 (73.8%) categorized as R4 (Figure 1). In this cohort, R4 recipients received a relatively equal allocation among the donor risk categories, with 24 of 144 (16.6%) receiving a D1 graft, 39 (27.1%) receiving a D2 graft, 26 (18.1%) receiving a D3 graft, and 55 (38.1%) receiving a D4 graft. The R4-D4 combination, therefore, made up 28.2% of all transplants (n = 55).

Graft and patient survival
Recipient survival was 100% for the R1, R2, and R3 groups at 4 years. One-year survival for the R4 cohort was 97.2% (4 deaths), and 4-year survival was 93.8% (9 deaths). When we analyzed the R4 cohort alone, we found no difference in patient survival dependent on donor category: 4-year survival was 95.8%, 94.9%, 96.2%, and 90.9% for D1, D2, D3, and D4, respectively (Figure 1).

Graft survival was lower in the D4 cohort at 1 year (87.9%) compared with that shown for the D1 cohort (97.4%) and for the D2 and D3 cohorts (both 94.4%); however, this difference was not statistically significant (log-rank test, P = .2). Overall graft survival at 4 years was 88% (all donor types). When we assessed only the R4 cohort, we found no difference in graft survival between donor risk categories, which was 83.6% at 4 years with D4 graft and 87.5%, 87.2%, and 88.5% for D1, D2, and D3 grafts, respectively (Figure 1).

Delayed graft function
Overall rate of DGF was 26.8% for all transplants. The rate of DGF was highest in the D4 group at 39.1%, compared with 21.1% for D1 to D3 grafts (P = .014). Within the R4 cohort, overall DGF rate was 33.0%; the R4-D4 combination had a DGF rate of 44.4% compared with 26.4% for R4 recipients with D1 to D3 grafts (P = .003). Mean duration of DGF was 4.0 days in the R4-D4 group (maximum of 28 days) compared with 1.9 days for D1, 2.4 days for D2, and 1.5 days for D3 in R4 recipients (no statistical difference between individual groups). The incidence of DGF demonstrated a relationship with kidney function at 1 year. Among R4 recipients, median eGFR at 1 year was 28.5 mL/min/1.73 m² in recipients with DGF compared with 52.9 mL/min/1.73 m² in patients without DGF.

The effect of donor type (ie, DCD vs DBD) was assessed. The DBD grafts categorized as D1 to D3 had a DGF rate of 17.5% (14 of 80) compared with 25.8% (8 of 31) for DBD grafts categorized as D4, although this difference was not statistically significant (P = .47). The DCD grafts categorized as D1 to D3 had a DGF rate of 27% (14 of 48) compared with 51.5% (17 of 33) for DCD grafts categorized as D4, although this difference was also not statistically significant (P = .072).

Short-term and medium-term function
Function at 7 days was lower with D4 grafts, with a median eGFR of 11.85 mL/min/1.73 m² (mean of 21.1 mL/min/1.73 m²), compared with 36.9, 28.2, and 30.9 mL/min/1.73 m2 with D1, D2, and D3 grafts, respectively. Observed differences only reached statistical significance for D4 compared with D1 (P < .001). The linear pattern of inferior function with D1 to D4 grafts remained at 14 days, with mean eGFR ranging from 51.3 (D1) to 31.0 (D4) mL/min/1.73 m². Excretory renal function at 1 year was worse for recipients of D4 grafts (median eGFR of 36.5 mL/min/1.73 m²) compared with 72.7, 53.7, and 55.0 mL/min/1.73 m² for recipients of D1, D2, and D3 grafts, respectively (P < .05 between all groups) (Figure 3).

Within the R4 cohort, function at all time points was lower with a D4 graft compared with all other donor risk categories. Median function with an R4-D4 combination at 7 days was 10.6 (mean, 17.5) mL/min/1.73 m², improving to 22.3 at 14 days, 35.0 at 1 year, and 36.7 mL/min/1.73 m² at 2 years. A D4 graft achieved a higher median eGFR if given to an R1, R2, or R3 recipient (48.4 mL/min/1.73 m2) versus an R4 recipient (35.0 mL/min/1.73 m²) (Figure 4).

Index admission length
Mean index admission length for all transplants was 11.3 days (median of 8 days; range, 5-158 days). For R4 recipients with all donor types, mean index admission length was longer (12.4 vs 8.1 days for R1-R3; P = .002). Within the R4 group, index admission was longest with a D4 graft (mean/median of 13.5/11 days). For recipients categorized as R4 with D1, D2, and D3 grafts, median index admission length was 8 days (mean of 11.9, 12.6 and 10.2 days, respectively), with P = .038 for D4 versus D3, P = .047 for D4 versus D1, and no significance (P = .07) for D4 versus D2. Furthermore, for R1 to R3 recipients who received a D4 graft, mean index admission length was 7.9 days (median of 7.0 days) versus 13.5 days for R4 recipients (P = .001).

When we assessed DBD transplants alone in all recipient types, median index admission was longer with a D4 graft versus a D1 to D3 graft (median/mean of 9/9.6 vs 7/11.2 days; P = .014). In the R4 DBD cohort, this difference was also present, with median index admission of 7 days for D1 to D3 compared with 10 days for D4 graft (mean 10.8 vs 12.4 days; P = .014).

Total inpatient days and day-case use
With the consideration that poorer quality kidneys (D4) and higher risk recipients (R4) may be independently or mutually associated with rates of hospital admission, both factors were analyzed. At 90 days, recipient types R1 to R3 (including all donor categories, censored for graft loss/death) accrued a median of 8 inpatient days compared with a median of 10.3 days for recipient type R4 (mean of 11.1 vs 14.1 days; P = .013, Mann-Whitney test). The D1 to D3 donor categories (including all recipient categories) had a median of 9.2 versus 11.2 days for D4 (mean of 13.5 vs 16.9 days; P = .005). By 2 years, only the donor category retained statistical significance as a discriminator with regard to inpatient days. Recipients categorized as R4 (of all donor types) demonstrated a median 14.7 days compared with 12.0 days for R1 to R3 recipients (mean 22.2 vs 19.9 days; P = .185, Mann-Whitney test), with recipients with D4 grafts accruing a median of 16.5 days versus 12.6 days for D1 to D3 grafts (mean of 24.1 vs 20.3 days; P = .022). In the 90-day posttransplant period, within the R4 group, the R4-D4 combination demonstrated additional inpatient days compared with R4 recipients receiving D1 to D3 grafts (median of 13.6 vs 8.9 days; P = .006, Mann-Whitney test).

During the early posttransplant period (first 90 days), the rate of hospital admission days per 1000 patient days for R1 to R3 recipients was 122.9, compared with 185.7 for R4 recipients (P < .001, chi-square test) (Figure 5). Per donor categories, the rate was 154.1 for D1 to D3 and 196.7 for D4 (P = .014). By 2 years of follow-up, rates were comparable between categories, at 27.4/1000 patient days for R1 to R3 and 38.7/1000 patient days for R4 (P = .198). At both 90 days and at 2 years, the R4-D4 combination had approximately twice the rate of hospital inpatient days than recipients categorized as R1 to R3, showing 223.8/1000 patient days for the R4-D4 combination versus 122.9/1000 patient days for R1 to R3 recipients at 90 days (P < .001, chi-square test) and 47.1 versus 27.4/1000 patient days at 2 years (P = .027).

Regarding day-case attendances, the mean for all transplants was 5.6 attendances (range, 0-28). This value was higher for R3 recipients (6.2 attendances vs 1.8 for R1, 5.1 for R2, and 5.7 for R4); observed differences were not statistically significant (P = .27). Use of day-case facilities averaged 5.2 times for the R4-D4 combination versus 5.8 times for the R4-D3 and 6.7 times for the R4-D2 combination (not significant; P = .8).

With regard to use of ultrasonography, D4 grafts of all recipient types had the highest use at 5.6 scans during the follow-up period, with the highest use in for the R4-D4 combination at 5.9 scans. Use was lowest with D1 grafts (3.8 scans); however, differences observed were not statistically significant. Mean use of computed tomography was 0.9 scans per patient. There were no apparent differences observed between donor or recipient risk types.

Cost
The estimated cost of a deceased donor kidney transplant is £12167 and £12888 for DCD and DBD, respectively.11 An excess bed day beyond an expected admission length (8 days for DBD, 9 days for DCD) is costed at £510 for DCD and £549 for DBD transplants. The total cost of an R4 transplant was £28?754 compared with £21?619 for an R1 to R3 transplant (P = .001). The R4-D4 transplant cost the most, at £31?213 or €35?271 (at time of writing) compared with the R4-D1 (£29?104; €32?887), the R4-D2 (£27?677; €31?273), and R4-D3 combinations (£24?845; €28?074). However, because of significant variance within groups, observed differences were not significant (P = .1). As previously mentioned, index admission length for an R4-D4 transplant was longer compared with other donor types (mean of 13.5 days vs 11.7 days for R4 with D1-D3 grafts; P = .001). This additional inpatient stay of 1.8 days (mean) equates to a £918 to £988 (DCD and DBD) increase in index admission cost per transplant, not including cost of investigations or interventions during admission. When we assessed D4 grafts alone, the mean excess bed days (beyond average index admission) was 0.8 when transplanted to R1 to R3 recipients and 5.3 for R4 recipients, with mean associated cost of £435.0 compared with £2772.1.

Current waiting list
There are approximately 450 patients on the West of Scotland renal transplant waiting list, 239 of which are currently active. Average age is 53.5 years, BMI is 27.7, and male-to-female patient ratio is 1.4:1. Per the new allocation scheme, those considered Tier A (ie, calculated reaction frequency [cRF] 100%, wait time over 7 years, or matchability score of 10) are not stratified using the recipient risk indices (26 patients). Thus, with the recipient risk criteria as previously described, categorization of the 213 recipients resulted in 140 (65.7%) potential recipients deemed R4, 37 (17.4%) deemed R3, 27 (12.7%) deemed R2, and 9 (4.2%) deemed R1. With regard to cRF, 31 of Tier B patients (14.5%) had a cRF over 85% and were therefore considered highly sensitized.

Discussion

The new UK kidney transplant allocation scheme came into effect in September 2019, remaining nearly unchanged, excluding minor amendments, since 2006.¹⁴ Underpinning the donor/recipient risk stratification was a simulation based on over 7000 patients in a cohort spanning 10 years from 2006 to 2016. The revised scheme aimed to better match patient and graft age, principally to reduce the discrepancy between patient and graft life expectancy, but also to minimize the UK offer decline rates, which ranged from 24% to 69% in 2016/2017. Historically, advancing age has been a barrier to renal transplant, with a quoted 30% reduction in transplant access with each additional decade over 55 years.¹⁵ Although the survival and quality of life benefits accompanying transplant in elderly recipients cannot match younger populations,^16-18 there is a necessary drive to make access to transplant more equitable, particularly given the advancing age of the recipient population. Because of donor organ shortages, there will always be difficult and conflicting prioritization with regard to those who could benefit the most while creating fair and equitable access to all recipients.

The European Seniors Programme initiative, which rolled out in 2001, was the first kidney allocation scheme to give a higher priority to matching donor and recipient ages. Donor grafts over 65 years were allocated to recipients over 65 years, regardless of HLA matching, with consideration of proximity to minimize cold ischemia time. Five-year patient and graft survival rates were 60% and 47%, respectively, which were lower than rates for older recipients receiving younger grafts in a matched cohort.¹⁹ Because of the omission of HLA matching in the model, the rates of acute rejection and immunosup­pression load were increased.²⁰ Furthermore, subsequent studies reported that rates of surgical complications (47%) and return to theatre (28%) were significantly higher in the European Seniors Programme group,²¹ thus highlighting that there may be an additional impact beyond crude patient and graft survival rates.

To facilitate the estimation of graft survival, the kidney donor risk profile (KDRI) is a tool based on donor characteristics to aid the decision-making process in renal transplantation.^3,22 In the United States, the Kidney Allocation Scheme, which utilizes the KDRI and was released in 2014, aimed both to improve equity of access to all recipients and to “maximize the potential survival of every transplanted kidney.”23 Furthermore, the information provided by this risk stratification tool aimed to reduce graft discard rates.²⁴ With a consideration to move away from the binary standard criteria versus extended criteria donor, the allocation scheme implemented the concept of the kidney donor profile index, based on the KDRI and not dissimilar to risk stratification used in the aforementioned UK scheme. There was recognition that this scheme, although it improved utility with improved transplant rates for the highly sensitized and also maximized life years gained, would negatively impact access to transplantation for those over 50 years of age.10 Furthermore, Butler and colleagues recently demonstrated an increase in 2-year mortality with post- versus pre-Kidney Allocation Scheme imple­mentation (6.31% vs 5.91%, respectively). Although overall graft loss appeared unaltered, the study identified that 2 subgroups (age groups 46-55 and 56-65 years) may disproportionately carry an increased risk of mortality and graft loss.²⁵

Two-thirds of patients on the current active waiting list in our center would be deemed the highest risk category (R4). It is unclear whether this patient cohort is representative of the national waiting list as a whole or whether the composition of the ever-changing ESRD patient group alters from that with which this model was produced. Unsurprisingly, our study demonstrated that R4 recipients (receiving all donor types) have a longer index admission, higher rate of DGF, and higher total inpatient days (including readmissions). Of particularly interest is that all of these metrics worsen with the combination of a D4 graft with an R4 recipient, a so-called R4-D4 transplant. However, patient and graft survival rates at 4 years in the R4-D4 group were not affected, showing 90.9% and 83.6%, respectively.

There is difficultly in generalizing the findings presented here, owing to the potential differences in recipient cohorts between transplant units, which is a limitation to this study. Furthermore, given the skewed distribution of the R4 cohort, the study size was comparatively small within the R1 and R2 cohorts. The impact of donor type (DCD vs DBD) is important to note. The higher rate of DGF with DCD compared with DBD was not statistically significant; furthermore, the rate of DGF in the D4 group (in both the DCD and DBD subgroups) was increased compared with that shown in the D1 to D3 groups, demonstrating the additional impact of these factors within the risk indices. In addition, index admission length and readmission rates may be influenced by factors beyond the risk indices documented here (eg, frailty, and socio-economic status, which were not accounted for in these data).

With particular regard to our transplant unit, the simulation publicized by NHSBT stated that approximately 70% of R4 recipients would be allocated D4 grafts and therefore potentially 45% of all transplants, as predicted, would be an R4-D4 combination. The sequelae of this could create significant additional strain on health care, particularly as the burden of resource use is front-loaded in the first few months posttransplant. The additional resource strain may be offset by a larger proportion of R1 to R3 patients receiving better quality grafts; however, this is yet to be proven. Furthermore, although we acknowledge that this new scheme may facilitate more transplants and ultimately be cost-effective by removing R4 recipients from dialysis, finances may need to be redirected to the services that will shoulder the strain of this resource shift.

Emerging techniques such as normothermic regional perfusion and ex vivo normothermic perfusion, used to precondition, assess, and mitigate the effects of ischemia-reperfusion injury, may prove crucial in their roles to improve early function, reduce rates of DGF, and dampen the impact of allocating the poorest quality grafts to the most at-risk recipients. Because of the shortage of donor organs, the delicate ethical balance between equity and utility will forever challenge renal transplantation. During this postim-plementation monitoring period, prospective data are a priority to ensure the pendulum has not swung too far toward utility.

References:

1. NHS Blood and Transplant. Statistics About Organ Donation. NHS Organ Donation Register. Accessed December 21, 2020. https://www.organdonation.nhs.uk/helping-you-to-decide/about-organ-donation/statistics-about-organ-donation/
   CrossRef - PubMed
   
2. Gridelli B, Remuzzi G. Strategies for making more organs available for transplantation. N Engl J Med. 2000;343(6):404-410. doi:10.1056/NEJM200008103430606
   CrossRef - PubMed
   
3. 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(2):231-236. doi:10.1097/TP.0b013e3181ac620b
   CrossRef - PubMed
   
4. NHS Blood and Transplant. Transplant Activity Report. NHS Organ Donation Register. Accessed December 21, 2020. https://www.organdonation.nhs.uk/helping-you-to-decide/about-organ-donation/statistics-about-organ-donation/transplant-activity-report/
   CrossRef - PubMed
   
5. Andre M, Huang E, Everly M, Bunnapradist S. The UNOS Renal Transplant Registry: Review of the Last Decade. Clin Transpl. 2014:1-12.
   CrossRef - PubMed
   
6. Levitt M. Could the organ shortage ever be met? Life Sci Soc Policy. 2015;11:6. doi:10.1186/s40504-015-0023-1
   CrossRef - PubMed
   
7. Wu DA, Robb ML, Forsythe JLR, et al. Recipient comorbidity and survival outcomes after kidney transplantation: A UK-wide prospective cohort study. Transplantation. 2020;104(6):1246-1255. doi:10.1097/TP.0000000000002931
   CrossRef - PubMed
   
8. Johnson RJ, Bradbury LL, Martin K, Neuberger J, Registry UKT. Organ donation and transplantation in the UK-the last decade: a report from the UK national transplant registry. Transplantation. 2014;97 Suppl 1:S1-S27. doi:10.1097/01.TP.0000438215.16737.68
   CrossRef - PubMed
   
9. NHS Blood and Transplant. Kidney Offering Scheme. Accessed December 21, 2020. https://www.odt.nhs.uk/odt-structures-and-standards/odt-hub-programme/kidney-offering-scheme/
   CrossRef - PubMed
   
10. US Department of Health and Human Services. Organ Transplantation and Procurement Network. Kidney Allocation System. Accessed December 18, 2020. https://optn.transplant.hrsa.gov/learn/professional-education/kidney-allocation-system/
    CrossRef - PubMed
    
11. NHS Improvement Website. Archived Reference Costs. Accessed August 15, 2020. https://improvement.nhs.uk/resources/reference-costs/
    CrossRef - PubMed
    
12. Therneau TM, Lumley T, Atkinson E, Crowson C. A Package for Survival Analysis in R. R package version 3.2-7. https://CRAN.R-project.org/package=survival2020.
    CrossRef - PubMed
    
13. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag; 2016. https://ggplot2.tidyverse.org
    CrossRef - PubMed
    
14. Wu DA, Watson CJ, Bradley JA, Johnson RJ, Forsythe JL, Oniscu GC. Global trends and challenges in deceased donor kidney allocation. Kidney Int. 2017;91(6):1287-1299. doi:10.1016/j.kint.2016.09.054
    CrossRef - PubMed
    
15. Huang E, Segev DL, Rabb H. Kidney transplantation in the elderly. Semin Nephrol. 2009;29(6):621-635. doi:10.1016/j.semnephrol.2009.07.011
    CrossRef - PubMed
    
16. Humar A, Denny R, Matas AJ, Najarian JS. Graft and quality of life outcomes in older recipients of a kidney transplant. Exp Clin Transplant. 2003;1(2):69-72.
    CrossRef - PubMed
    
17. Shah T, Bunnapradist S, Hutchinson I, et al. The evolving notion of “senior” kidney transplant recipients. Clin Transplant. 2008;22(6):794-802. doi:10.1111/j.1399-0012.2008.00881.x
    CrossRef - PubMed
    
18. 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(23):1725-1730. doi:10.1056/NEJM199912023412303
    CrossRef - PubMed
    
19. Frei U, Noeldeke J, Machold-Fabrizii V, et al. Prospective age-matching in elderly kidney transplant recipients--a 5-year analysis of the Eurotransplant Senior Program. Am J Transplant. 2008;8(1):50-57. doi:10.1111/j.1600-6143.2007.02014.x
    CrossRef - PubMed
    
20. Dreyer GJ, Hemke AC, Reinders ME, de Fijter JW. Transplanting the elderly: Balancing aging with histocompatibility. Transplant Rev (Orlando). 2015;29(4):205-211. doi:10.1016/j.trre.2015.08.003
    CrossRef - PubMed
    
21. Bentas W, Jones J, Karaoguz A, et al. Renal transplantation in the elderly: surgical complications and outcome with special emphasis on the Eurotransplant Senior Programme. Nephrol Dial Transplant. 2008;23(6):2043-2051. doi:10.1093/ndt/gfm912
    CrossRef - PubMed
    
22. Schamberger B, Lohmann D, Sollinger D, Stein R, Lutz J. Association of kidney donor risk index with the outcome after kidney transplantation in the Eurotransplant Senior Program. Ann Transplant. 2018;23:775-781. doi:10.12659/AOT.909622
    CrossRef - PubMed
    
23. Friedewald JJ, Samana CJ, Kasiske BL, et al. The kidney allocation system. Surg Clin North Am. 2013;93(6):1395-1406. doi:10.1016/j.suc.2013.08.007
    CrossRef - PubMed
    
24. Aubert O, Reese PP, Audry B, et al. Disparities in acceptance of deceased donor kidneys between the United States and France and estimated effects of increased US acceptance. JAMA Intern Med. 2019;179(10):1365-1374. doi:10.1001/jamainternmed.2019.2322
    CrossRef - PubMed
    
25. Butler CR, Perkins JD, Johnson CK, et al. Burden of excess mortality after implementation of the new kidney allocation system may be borne disproportionately by middle-aged recipients. PLoS One. 2019;14(1):e0210589. doi:10.1371/journal.pone.0210589
    CrossRef - PubMed