Increasing the Use of Kidneys From Unconventional and High-Risk Deceased Donors

Introduction

The imbalance between patients with end-stage renal disease (ESRD) who are awaiting renal transplant and the number of deceased donors has widened over the past decade. Several studies have demonstrated that extended criteria donors (ECDs) 1, 2 and public health service increased-risk kidneys 3 offer a survival advantage to patients with ESRD; however, many centers have been reluctant to use organs with perceived high risk. The lone exception has been the successful utilization of kidneys from donation after cardiac death (DCD). Many kidneys are discarded because of increased donor age, acute kidney injury (AKI), diabetes, hypertension, biopsy findings and prolonged ischemia times. Many of these kidneys would likely provide acceptable function and graft survival and thus provide important survival benefits to many patients with ESRD awaiting transplant. This review studied the literature on donor kidneys that are currently used at relatively low rates.

Donors With AKI

According to the 2012 Scientific Registry of Transplant Recipients (SRTR) annual report, 32.5% of kidneys from deceased donors with terminal creatinine >1.5 mg/dL were discarded compared with 15.5% from donors with terminal creatinine ≤1.5 mg/dL; however, this report did not distinguish between donors with AKI and those with stable impaired renal function. This distinction is important because donors with chronically impaired renal function may not be suitable donors, whereas those with AKI may have reversible injury that would not limit long-term outcomes. Several studies attempted to address this issue. In an analysis of the SRTR database, Kayler et al found that terminal donor creatinine >2.0 mg/dL was associated with a 7.04-fold (6.5- to 7.6-fold, adjusted for baseline differences) higher risk of discard. The transplant rate was 71% from donors with terminal creatinine >2.0 mg/dL compared with 92% for kidneys from donors with terminal creatinine ≤1.5 mg/dL 4. Many of these kidneys likely had AKI; this was not directly addressed but points out the magnitude of the problem. Although this analysis may have included some patients with chronic kidney disease, it is reassuring to note that donor terminal creatinine was not a risk factor for graft loss of standard criteria donor (SCD) kidneys. In the ECD cohort, graft survival was slightly lower for kidneys from donors with higher terminal creatinine 4. This study strongly suggested that for non-ECD kidneys with AKI (or elevated terminal creatinine), excellent graft survival and function can be obtained.

Several studies looked directly at the outcome of transplanting kidneys from deceased donors with AKI (Table 1) 4-10. In a prospective study, Hall et al reported outcomes of transplanting kidneys from deceased donors with AKI from multiple organ procurement organizations (OPOs) 6. They reported transplanting 3264 kidneys form 1632 donors. The donors were classified using Acute Kidney Injury Network (AKIN) criteria (stage 0–3, no AKI to severe AKI). Overall, 27.1% of the donors had AKI (AKIN stage 1–3). The discard rate was 30% for kidneys from donors with AKI and 18% for donors without AKI. The relative risk of delayed graft function (DGF) was 1.50 (interquartile range 1.31–1.70) for the AKI kidney group; however, the estimated GFR (eGFR) at 6 mo was similar across all AKI categories and kidneys without AKI.

In a single-center study, Heilman et al retrospectively analyzed the outcomes of 162 kidneys from deceased donors (139 SCDs and 23 ECDs) with AKI, which was defined as terminal creatinine ≥2.0 mg/dL or requirement of acute renal replacement therapy prior to procurement 5. DGF occurred in 66% of the AKI kidney group versus 27% of the control group (p < 0.001); however, eGFR and chronic changes on protocol biopsy at 1 year were similar in AKI and control cohorts. Graft survival at 3 years was 91.4% for the AKI kidney group, similar to the control group. In this analysis, it was estimated that 31% of SCD and 22% of ECD AKI kidneys that were discarded could be acceptable for transplantation. Despite the higher rate of DGF, through a process of outpatient management of DGF, this study showed similar hospital length of stay. This surrogate marker for hospital cost might suggest that transplanting kidneys from donors with AKI may not substantially increase hospital cost; however, addressing this question would require more comprehensive cost analysis.

In summary, the published evidence suggests that a significant opportunity exists to increase the use of kidneys from donors with AKI. Although AKI increases the rate of DGF, other clinical outcomes with AKI kidneys are the same as those for kidneys from donors without AKI. In addition, the risk of primary nonfunction does not appear to be higher with kidneys transplanted from donors with AKI 5, 10.

Donors With Diabetes

According to the 2013 annual SRTR report, 40% of donors had diabetes mellitus, and the proportion of donors with diabetes has been increasing 11. In the analysis by Rao et al on which the Kidney Donor Risk Index (KDRI) is based, the hazard ratio (HR) for graft failure in recipients of transplants from donors with diabetes was 1.14 (95% confidence ratio [CI] 1.04–1.24, p = 0.004) 12. Given this HR, one would expect only a small absolute independent impact on graft survival if one assumes a 15% graft loss rate at 3 years (85% vs. 83%). Despite this relatively low absolute risk, the discard rate for kidneys from donors with diabetes was 44.8% compared with 15.5% for donors without diabetes.

Mohan et al used the United Network of Organ Sharing (UNOS) database to assess outcomes of 1982 kidneys from donors with diabetes 13. The outcome of transplants from SCDs with diabetes was inferior to that from SCDs without diabetes (log rank = 70.9, p < 0.001), but graft survival for kidneys from SCDs with diabetes was better than that for kidneys from ECDs without diabetes (log rank = 22, p < 0.001). In another UNOS database analysis, Ahmad et al showed that transplanting kidneys from donors with diabetes was associated with inferior graft survival 14; however, after doing a propensity analysis, they demonstrated that the absolute difference was small (HR 1.11, 95% CI 1.02–1.22), and there was no difference in patient survival (HR 1.06, 95% CI 0.94–1.18).

Data from small case reports suggest that donor-derived diabetes-related glomerular changes may be reversible when these kidneys are transplanted into recipients without diabetes 15. In addition, the outcome of transplanting a kidney from a donor with diabetes may be influenced by the recipient's diabetes status. In a UNOS database analysis of 9074 recipients of kidneys from donors with diabetes, the kidneys from these donors had an increased risk of graft failure (HR 1.21, 95% CI 1.16–1.26) and death (HR 1.19, 95% CI 1.13–1.24) 16. This analysis, however, demonstrated a significant interaction between donor and recipient diabetic status, suggesting adverse recipient selection bias. Consequently, the authors performed an analysis of recipients of kidneys from the same donor when one kidney went to a recipient with diabetes and the other kidney went to a recipient without diabetes. In this discordant pair analysis, diabetic recipients of kidneys from diabetic donors had a higher risk of all-cause graft loss (HR 1.27, 95% CI 1.05–1.53) compared with recipients without diabetes.

These studies, although limited by either small size or lack of granularity, still suggest that transplanting kidneys from donors with diabetes (particularly those with a short history of documented diabetes, good baseline renal function and lack of severe systemic disease) is associated with a modestly increased risk of allograft failure, particularly in recipients without diabetes, but the absolute difference is small, and the results are far better than those for ECDs and thus confer a relevant survival advantage over dialysis 1.

Donors With Hypertension

The 2013 annual SRTR report showed that 38% of donors had hypertension, and the proportion of donors with hypertension appears to be increasing 11. The kidney discard rate was 35.2% from donors with hypertension compared with 10.6% for donors without hypertension. Donor hypertension is one of the donor factors in the KDRI model. In the KDRI model, donor hypertension is associated with an HR of 1.13 (95% CI 1.08–1.19, p < 0.0001) for graft failure 12. Once again, these data are aggregate, and further study is needed to assess those kidneys from donors with hypertension that could be used without a clinically important impact on graft survival or graft function. As with diabetes, the HR indicates that the independent association of hypertension and graft loss in absolute terms would still likely confer a survival advantage over dialysis.

DCD Donors

Overall, 20% of all kidneys transplanted nationally are from DCD donors, but variation in the use of DCD kidneys across transplant programs suggests opportunities for improving population-based outcomes in ESRD 17. Transplant programs have reported increased resource utilization and costs with these grafts related to DGF 18, but national trends over the past 15 years demonstrated growing comfort with the use of these kidneys. This use has been driven by increasing evidence that DGF alone does not predict inferior long-term outcomes 19. In terms of survival benefit of DCD kidney transplant for the ESRD population, it is important to understand what additional factors affect long-term graft survival. Donor age, terminal AKI and donor warm ischemia time are important considerations in selecting these kidneys.

Locke et al demonstrated that transplanting DCD kidneys from donors who were aged <50 years or who had short cold ischemia time (CIT) had similar death-censored graft survival at 5 years 20. For the average recipient, kidneys from older DCD donors (aged > 50) had higher incidence of DGF and worse long-term survival compared with kidneys from younger DCD donors and donation after brain death counterparts. Preexisting AKI in a DCD donor is also an important common consideration 5, 21, 22. These studies indicated that well-selected DCD-AKI kidneys can be used with excellent clinical outcomes, even at 10 years after transplantation. Donor warm ischemia time (DWIT) has conventionally been delineated as the time from withdrawal to cold perfusion, but it is difficult to interpret based on the timing of when true malperfusion occurs. The local OPO sets a ceiling of time from withdrawal, typically 60–90 min, but it can be as long as 120 min. DWIT <60 min is generally accepted in most programs, and >60 min is judged on a case-by-case basis 23. It is important to know that these thresholds originally emerged from preclinical animal experiments and must be interpreted as such.

DCD kidney transplantation has become common, but significant geographic variation remains in the use of these grafts. Further study is required to determine barriers to appropriate use; several single-center and registry-based studies have identified both benefits and risks.

Long CIT

The association between CIT and DGF is well established, but the effect of this injury on long-term outcomes is less well known. Salahudeen et al evaluated the long-term implications of cold ischemia using 6-year follow-up data from the UNOS and Organ Procurement and Transplantation Network registry for kidneys transplanted in 1995 24. Prolonged CIT was associated with a monotonic increase in the relative risk of graft loss with each 10-h increment, but this was statistically significant only when comparing CIT >30 h and <10 h (relative risk 1.32, p = 0.01). Long-term graft survival, however, was acceptable in the prolonged CIT group at 70% at 6 years in a cohort from the 1990s. In a well-designed modern study evaluating kidneys with differing CITs from single ECDs, Kayler et al demonstrated similar long-term graft survival between the groups with short and long CIT, even when kidneys from the same donor had a difference in CIT of >15 h after adjusting for other factors 25. In terms of survival benefit, the reasons for prolonged CIT must be considered when accepting kidneys with CIT >30 h; CIT may be a proxy for other factors that affected decisions to decline locally and that may or may not be related to organ quality.

Pediatric En Bloc Transplant

Transplanting kidneys en bloc from small pediatric donors (weighing <15–20 kg) has been shown to result in excellent outcomes, minimizing short-term graft loss from vascular thrombosis and possibly improving long-term graft survival by preventing hyperfiltration injury. Despite these excellent results, kidneys were not recovered from ≈32% of pediatric donors weighing ≤20 kg, and 15% of the recovered kidneys were discarded 26. In addition, transplanting kidneys en bloc into one recipient as opposed to performing single transplants into two recipients results in fewer transplants, exacerbating the disparity between the waiting list and the number of organs available. In a recent publication analyzing UNOS data, the proportion of en bloc versus single-kidney transplants was lower as the donor weight increased: 84% versus 16%, respectively, in donors weighing <8 kg compared with 19% versus 81%, respectively, in donors weighing 20 kg 26. This study demonstrated that en bloc kidney transplantation is associated with superior 1-year graft survival at the expense of one rather than two transplants; however, single-kidney transplants from donors weighing >10–12 kg have been shown to yield excellent outcomes when performed at experienced centers (>90% 1-year graft survival rate). Using a decision analysis model, Laurence et al demonstrated a greater gain in overall life expectancy using single-kidney transplants as opposed to en bloc kidney transplants except in recipients from donors weighing <10 kg 27.

Kidneys With High Kidney Donor Profile Index

Rao et al proposed the KDRI for deceased donor kidneys to quantify graft failure risk and to support the complex decision of organ acceptance at the time of organ offer 12. The following donor characteristics are used to calculate the KDRI: age, height, weight, ethnicity, history of hypertension, history of diabetes, cause of death, serum creatinine, hepatitis C status and DCD status. The Kidney Donor Profile Index (KDPI) maps the KDRI to the cumulative percentage scale compared with a reference population of donors (all donors in the United States from whom a kidney was recovered during the prior calendar year) to rank order the quality of kidneys. In the new kidney allocation system introduced in December 2014, KDPI >85% replaced the ECD status, and these kidneys are allocated on a regional basis to increase utilization. Approximately 14% of all kidneys recovered from deceased donors have KDPI >85%, and the discard rate for these kidneys is 56% compared with 17% for the kidneys with lower KDPI. This high discard rate has not changed significantly in the first year following the new kidney allocation system. It should be noted that the C statistic for the KDRI (and thus the KDPI) is very modest (0.63) and, from a statistical perspective, does not provides sufficient discrimination as a reliable metric for an individual decision to accept or turn down an organ.

Analyzing UNOS data, Klair et al and Tanriover et al demonstrated that kidneys with higher KDPIs (KDRI >2.2; KDPI ≈95% and >90%, respectively) have graft and patient survival benefits when transplanted as dual kidneys in a single recipient compared with single kidneys in two recipients 28, 29. Gandolfini et al also showed that pretransplant donor biopsy–based allocation of marginal grafts led to a limited discard rate of 15% for kidneys with KDPI 80–90% and 37% for kidneys with a KDPI 91–100% 29. The role of preimplantation biopsy in the process of deciding to use kidneys as single or dual transplants versus discard has not been adequately studied.

Thousands of patients are waiting for kidney transplants, and transplanting kidneys with high KDPI (i) does not necessarily put the program at risk for flagging by SRTR, (ii) is a poor predictor of cost and resource utilization, and (iii) provides a survival benefit over dialysis or waiting for a lower KDPI kidney in certain groups of patients. In addition, KDPI in itself cannot be used to determine whether to accept or discard a kidney. The evidence would suggest that greater use of these kidneys should be encouraged if we hope to maximize the scarce resource of donor kidneys.

Other Donor Anatomical Challenges

Donor kidney anatomy may present challenges to surgeons and potentially increase the risk of organ discard. Anatomical challenges vary and may involve duplicated vasculature, renal masses, embryological abnormalities or other collecting system irregularities. With further expansion of living donor kidney transplantation, the success of direct anastomosis of multiple donor renal arteries to recipient aorta or iliacs has enabled surgeons to handle vascular abnormalities with and without the use of conduits, arterial patch reconstructions or direct anastomosis to traditional and nontraditional sites, such as the inferior epigastric artery or internal iliac artery 30, 31. Several reports have identified potential increases to the donor pool by transplanting kidneys with small renal cell carcinomas (RCCs) 32. Category T1a RCCs can be excised during backbench preparation of the kidney as a partial nephrectomy with 5-mm margins. Repair of the kidney involves reapproximation of the collecting system with absorbable sutures, reapproximation of the parenchyma with pledget reinforcement and application of adhesive sealants 32. These kidneys have been documented to have excellent short- and long-term results with low risk of surgical complications, but the oncological implications of high tumor grade, category T1b, aberrant histology and positive margins also must be considered 32. Numerous case reports have been published on transplanting horseshoe kidneys from deceased donors 33. These transplants have been described in a variety of ways: with and without division of the renal isthmus, as en bloc transplantation with vascular anastomoses to the donor aorta and inferior vena cava or with direct anastomoses of hilar vessels. It is important to realize that these kidneys typically have multiple vascular and collecting systems, and careful backbench preparation and surgical planning for the recipient is critical. Kidney transplantation with allograft ureter duplication is the subject of many case series in the literature, with good outcomes, involving separate ureteroneocystotomy anastomoses or joining the two ureters together prior to bladder anastomosis using the Lich–Gregor technique with temporary ureteral stenting 34. Deceased donor kidneys with nephrolithiasis are commonly discarded, but techniques applied from the living donor kidney transplant domain have been applied. Technical success has been achieved with stone retrieval using ex vivo ureteroscopy or postoperative treatment 35-37.

A Potential Role for Preimplantation Biopsy

The role for preimplantation histopathology is an area of great debate and is beyond the scope of this review. There are unsettled issues regarding which technique to use (needle vs. wedge biopsy) 38, who should evaluate the pathology (renal vs. general pathologist) 39, and which histological parameters (glomerulosclerosis, interstitial fibrosis, arteriosclerosis, and hyaline arteriolosclerosis) are most important for graft function and survival.

Preimplantation biopsies may show unexpected pathology such as glomerular fibrin thrombi (GFT). Despite several reports of successful use of donor kidneys with GFT 40-42, many centers decline these organs. Batra et al demonstrated good short- and long-term outcomes using GFT donors and prompt resolution of most GFT by 1 mo with excellent graft survival and function 43.

Summary

As noted by Garonzik-Wang et al, there is marked variability among UNOS regions in use of high-risk kidneys 44. Higher utilization was associated with wait list size, organ shortage and waiting time. This variability among regions suggests a significant opportunity to improve the utilization of these organs. Kidneys from donors with AKI seem to have outcomes equivalent to those from donors without AKI, provided one can rule out significant cortical necrosis. Kidneys from donors with preexisting diabetes or hypertension may have marginally lower aggregate survival but still provide a significant benefit over remaining on the wait list.

The KDPI derives only an aggregate association with survival, with a very modest C statistic; therefore, the data indicated that KDPI should not be the sole reason to discard a kidney except perhaps in patients with extremely low estimated posttransplant survival scores. Even if the C statistic were high, this would not justify discarding a kidney that might increase survival for an individual patient. Regulatory worries are a deterrent; however, given the low C statistic of KDPI and the overall low correlation coefficient of the risk model, regulatory concerns may be less due to reversion to the mean. Low observed and expected outcomes are just as likely with low and high expectations; in other words, high expected outcomes may be counterproductive for avoiding regulatory sanctions because perfection leaves no room for the random variability of our current risk adjustment models. In a recent analysis of the SRTR database, Snyder et al reported the impact on individual program quality reports of increasing rates of transplant of high-KDPI kidneys. Despite a clear relationship of increasing KDPI and patient and graft survival, there was no relationship of an individual program's utilization of high-KDPI kidneys and the program's performance report after risk adjustment 45. An analysis by Hirth et al might serve as a warning that the process of labeling can have a detrimental effect on transplant rate. This study showed that labeling as an ECD kidney resulted in a measureable drop in the transplant rate of ECD kidneys 46.

The decision to accept a kidney for an individual patient is complex and multifaceted. It is difficult to quantify the interaction of all factors including donor characteristics, recipient characteristics and center resources. Given the growing number of patients on the wait list, broadening our approach to kidney acceptance could have a important impact on the ESRD population awaiting transplant. The data presented in this review indicated that many lives could be prolonged by considering use of kidneys that are often discarded.

Disclosure

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