Urine injury biomarkers are not associated with kidney transplant failure

Abstract

Background

Kidneys transplanted from deceased-donors with serum creatinine-defined acute kidney injury (AKI) have similar allograft survival as non-AKI kidneys but are discarded at a higher rate. Urine injury biomarkers are sensitive markers of structural kidney damage and may more accurately predict graft outcomes.

Materials and Methods

In the 2010–2013 multicenter Deceased Donor Study of 2430 kidney transplant recipients from 1298 donors, we assessed the association of donor urine injury biomarkers microalbumin, NGAL, KIM-1, IL-18, and L-FABP with graft failure (GF) and death-censored GF (dcGF) using Cox proportional hazard models (median follow-up 4 years). We examined if serum creatinine-defined donor AKI modified this association to assess the relationship between subclinical donor AKI (elevated biomarkers without creatinine-defined AKI) and graft failure. Through chart review of a sub-cohort (1137 recipients), we determined associations between donor injury biomarkers and a 3-year composite outcome of GF, mortality, or eGFR <=20mL/min/1.73m².

Results

Risk of GF, dcGF, and 3-year composite outcome did not vary with donor injury biomarker concentrations after adjusting for donor, transplant, and recipient characteristics (adjusted HR ranged from 0.96 to 1.01 per log-2 increase in biomarker). Subclinical injury in transplanted kidneys without AKI was not associated with GF.

Conclusions

AKI measured using injury biomarkers was not associated with posttransplant graft outcomes (at median 4 years posttransplant). When assessing posttransplant graft viability, clinicians can prioritize other donor and recipient factors over donor kidney injury, measured by either serum creatinine or urine injury biomarkers.

Introduction

From 2000 to 2015, 36700 kidneys (17% of all kidneys which were procured) were discarded in the United States.^1,2 Many of these discarded organs may be viable for transplantation.^1–5 With over 90000 people on the kidney transplant waiting list, expanding the pool of viable donor organs will save lives. Kidney transplants increase survival and quality of life and decrease cost compared to prolonged dialysis;^6–8 unfortunately, approximately 5000 people on the waitlist die each year waiting for a transplant.⁹

While increasing transplant volume is a goal for some transplant centers, transplant center programs may also be reluctant to accept kidneys perceived as lower quality because of concerns about regulatory penalties or the financial costs of managing transplant complications.^10–13 All transplant centers receive evaluations on 1-year post-transplant mortality and graft failure rates from the Scientific Registry for Transplant Recipients. These reports adjust for donor organ characteristics, but many transplant centers perceive the adjustment as insufficient. Organs are often viewed unfavorably and discarded when associated with poor biopsy findings, poor function, prolonged ischemia, or poor donor medical history.¹ Donor kidneys may be injured over time from chronic kidney disease (CKD) or more acutely from disseminated intravascular coagulopathy, hemodynamic instability, or other causes of acute kidney injury (AKI). The Kidney Donor Risk Index (KDRI), which comprises 10 donor characteristics and intends to function as a summary measure of graft quality, is supplied to transplant centers with every organ offer.¹⁴

Transplant centers may be able to utilize more donor organs with AKI to expand the supply of deceased-donor kidneys if posttransplant outcomes are acceptable. Our group previously found that AKI donor kidneys (increasing admission-to-terminal serum creatinine by AKI Network Stages) are discarded at a higher rate than non-AKI donor kidneys (30% versus 18%), but AKI donor kidneys that are selected for transplant have similar long-term graft survival (median follow-up = 4 years) to non-AKI donor kidneys.^15,16 Since serum creatinine is widely available, it remains the primary means by which donor AKI is identified. However, serum creatinine is a relatively insensitive marker of AKI or acute tubular injury (ATI). Urine injury biomarkers such as microalbumin, neutrophil gelatinase-associated lipocalin (NGAL), kidney injury molecule-1 (KIM-1), IL-18, and liver-type fatty acid binding protein (L-FABP) are evident within 1–4 hours of acute tubular injury.^17,18 These biomarkers can help identify subclinical AKI – elevated kidney injury without serum creatinine-defined AKI.^19–22 Among kidneys from hospitalized deceased donors without serum creatinine-defined AKI, our group found rates of mild and severe acute tubular injury measured by urine injury biomarkers to be 18% and 9%, respectively.¹⁹

Our group previously found that donor urine injury biomarkers NGAL and L-FABP were associated with lower 6-month eGFR only among recipients without posttransplant DGF but the relationship with medium to long-term graft survival was unclear.²³ This study investigated the association between donor urine injury biomarkers microalbumin, NGAL, KIM-1, IL-18, and L-FABP and longer-term posttransplant allograft survival (median follow-up 4 years).

Materials and Methods

The Deceased Donor Study is a multicenter, observational analysis of deceased donors and corresponding kidney transplant recipients from five organ procurement organizations (OPOs) between May 2010 and December 2013. Donors that did not lead to transplantation, donors with age less than 16 years, and donors without urine samples were excluded. We also generated a sub-cohort by performing direct chart reviews of recipients from donors at 13 transplant centers, which allowed us to ascertain detailed posttransplant outcomes (e.g., poor graft function). The five participating OPOs were Gift of Life Donor Program, Philadelphia, PA; New Jersey Sharing Network, New Providence, NJ; Gift of Life Michigan, Ann Arbor, MI; New York Organ Donor Network, New York, NY; and New England Organ Bank, Waltham, MA. This study was approved by the scientific review boards from participating OPOs and the institutional review boards from the 13 transplant centers. A thorough description of data quality and accuracy efforts in this study is provided in Supplemental Methods S1.

This study used data from the Organ Procurement and Transplantation Network (OPTN), which was supplemented through our study’s database of clinical and biomarker data. The OPTN data system includes data on all donor, wait-listed candidates, and transplant recipients in the U.S., submitted by the members of the OPTN, and has been described elsewhere.^24,25 The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services provides oversight to the activities of the OPTN contractor. The analyses are based on OPTN data as of July 31, 2017 and may be subject to change due to future data submission or correction by transplant centers.

Donor urine injury biomarker data

Prior to organ procurements, 10mL of fresh donor urine was collected using an indwelling urinary catheter tube, transferred on ice, and then frozen (median time to freezing was 6.3 hours, interquartile range 4.9–8.0 hours). The urine was stored at −80 degrees Celsius until monthly shipments to the Yale University Biorepository. Samples were then thawed, centrifuged, and separated into 1-mL aliquots and then stored at −80 degrees Celsius. We measured the following urine injury biomarkers from stored samples: microalbumin (mg/dL), NGAL (ng/dL), KIM-1 (pg/mL), IL-18 (pg/mL), and L-FABP (pg/mL). Urine creatinine was measured using the RxDaytona immunoturbidimetry assay (Randox Laboratories, Boston, MA). Biomarker measurements are described more completely in the supplemental materials (Supplemental Methods S2) and in our group’s previous work.²³

Primary outcomes

In the full study cohort of recipients, we assessed posttransplant all-cause graft failure (GF) and death-censored graft failure (dcGF). All-cause GF was defined as all-cause mortality, return to dialysis, or re-transplantation with median posttransplant follow-up of 4 (IQR = 3,5) years. In the chart review sub-cohort of recipients, we measured a 3-year composite outcome of graft failure, death, or estimated glomerular filtration rate (eGFR) <=20 mL/min/1.73m², which was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation.²⁶ The threshold of 20 mL/min/1.73m² was used because the OPTN uses this value in determining eligibility for re-transplantation (OPTN policy 8.4).²⁷ Outcomes were ascertained using data from OPTN until July 31, 2017 and chart reviews until November 4, 2016.

Statistical analysis

For donor and recipient descriptive statistics, means with standard deviations and frequencies with percentages were reported. We used Cox proportional hazard models (clustered by donor²⁸) to analyze the association between log-transformed (base 2) urine injury biomarker concentrations and primary outcomes. We report hazard ratios of both unadjusted and adjusted models accounting for KDRI, donor urine creatinine, cold ischemia time, number of human leukocyte antigen mismatches, and the following recipient variables: age, black race, sex, previous kidney transplant, diabetes as the cause of end-stage kidney disease, panel reactive antibody, body mass index (BMI), and pre-emptive transplant status. Historically, the expanded-criteria donor (ECD) designation is defined as donors >= 60 years old, or age 50–59 years with at least two of the following: cerebrovascular accident as the cause of death, preexisting hypertension, or terminal serum creatinine >1.5 mg/dL. The KDRI^29,30 was computed for each donor. We used Kolmogorov-type supremum tests³¹ to evaluate proportional hazards assumptions.

In a secondary analysis, we assessed whether the association between donor urinary injury biomarkers and all-cause graft failure differed between kidneys with and without creatinine-defined AKI. To do so, in unadjusted Cox proportional hazard models we included an interaction term between log-transformed donor biomarkers and AKI, defined as Stage 1 or higher using the AKI Network criteria.³² A significant interaction term indicated effect modification by serum creatinine-defined AKI and unadjusted hazard ratios were estimated separately by donor AKI status. As our group previously used injury biomarkers to demonstrate the presence of mild and severe acute tubular injury among kidneys without creatinine-defined AKI,¹⁹ we assessed the relationship between subclinical AKI (elevated injury without creatinine-defined AKI) and graft outcomes through this analysis. Stage 1 donor AKI was defined as a terminal serum creatinine that increased from the admission value by greater than or equal to 0.3 mg/dL or greater than 1.5 times the admission serum creatinine.

All statistical tests and confidence intervals were two-sided with a significance level of 0.05. We used SAS 9.4 software for Windows (SAS Institute, Cary, NC).

Results

Cohort and sub-cohort characteristics

Of the 1679 donors who contributed urine samples to the study, 1298 donors were eligible for the study and main analyses (Figure 1). The full DDS cohort consisted of 2430 kidney transplant recipients from the 1298 donors. Our study sub-cohort included 1137 recipients (from 862 donors) who underwent transplantation at one of our study’s 13 transplant centers. The donor cohort had the following characteristics: mean age of 41 ± 15 years, 60% male, 16% black, mean KDRI of 1.29, and mean admission to procurement time of 5 days. The recipient cohort had the following characteristics: mean age of 53 ± 15 years, 61% male, 39% black, and mean of 46 ± 38 months on dialysis prior to transplantation. Complete donor and recipient characteristics are described in Table 1. Sub-cohort characteristics are presented in Supplemental Table S1. Supplemental Table S2 compares donor characteristics of kidneys that were transplanted versus discarded. Transplanted kidneys had lower mean KDRI (1.27 ± 0.40 versus 1.85 ± 0.52) and were more likely to not have AKI (76% versus 63%) compared to discarded kidneys.

Figure 1.

Flowchart for study enrollment. OPO, organ procurement organization.

Table 1:

Donor, recipient, and transplant characteristics

Deceased Donor Characteristics		(N=1298)	Recipient and transplant characteristics		(N=2430)
Age, years		41 ± 15	Age, years		53 ± 15
Male		784 (60%)	Male		1492 (61%)
Black race		205 (16%)	Black race		956 (39%)
Hispanic ethnicity		171 (13%)	Hispanic ethnicity		279 (11%)
BMI, kg/m2		28 ± 7	BMI, kg/m2		28.0 ± 5.8
Hypertension		399 (31%)	Cause ESRD	Diabetes	726 (30%)
Diabetes		130 (10%)		Hypertension	643 (26%)
Cause of death	Head trauma	396 (31%)		Glomerulonephritis	391 (16%)
	Anoxia	425 (34%)		Graft failure	174 (7%)
	Stroke	427 (34%)		Other or unknown	496 (20%)
	Other	18 (1%)	Preemptive transplant		274 (11%)
Hepatitis C seropositive		48 (4%)	Dialysis duration, months		46 ± 38
ECD		246 (19%)	Previous kidney transplant		315 (13%)
DCD		206 (16%)	Pre-transplant blood transfusion		438 (18%)
KDRI		1.29 ± 0.41	PRA	Continuous, %	21 ± 35
KDPI, %		48 ± 27		0%	1545 (64%)
Admission to procurement, days		5 ± 7		1–20%	178 (7%)
Admission SCr, mg/dL		1.10 ± 0.61		21–80%	326 (13%)
Terminal SCr, mg/dL		1.17 ± 0.85		>80%	381 (16%)
Urine creatinine, mg/dL		49.09 ± 48.41	Cold ischemia time, hours		15.3 ± 7.1
Number of individual kidneys transplanted	1	166 (13%)	HLA Mismatch	Continuous	4.2 ± 1.5
	2	1132 (87%)		0	153 (6%)
Donor AKI stage	No AKI	976 (75%)		1	21 (1%)
	Stage 1	211 (16%)		2	82 (3%)
	Stage 2	62 (5%)		3	292 (12%)
	Stage 3	49 (4%)		4	642 (26%)
				5	820 (34%)
				6	414 (17%)

Values are mean (SD) or n (%). ECD, expanded-criteria donor; DCD, donation after cardiovascular determination of death; KDRI, kidney donor risk index; KDPI, kidney donor profile index; SCr serum creatinine; eGFR, estimated glomerular filtration rate; ESRD, end-stage renal disease; PRA; panel reactive antibody; HLA, human leukocyte antigen.

Urine injury biomarkers and graft failure among full and sub-cohort

The total combined follow-up of the cohort was 9479 patient-years with a median follow-up of 4 (IQR: 3.0–5.0) years. Of the 2430 recipients in the full cohort, 623 (26%) experienced all-cause graft failure during the entire follow-up period. There were 313 death-censored graft failures and 402 deaths, with 92 deaths occurring after graft failure. The overall event rate for all-cause and death-censored graft failure among the full cohort was 66.2 (95% CI=61.1, 71.7) and 33.3 (95% CI=29.7, 37.3) per 1000 person-years, respectively. Of the 1137 recipients in the study sub-cohort, 214 (19%) recipients experienced the 3-year composite outcome of graft failure, death, or eGFR <= 20 mL/min/1.73m², with an event rate of 76.1 (95% CI=66.3, 87.4) per 1000 person-years (Table 2).

Table 2.

Association between urinary injury biomarkers and all-cause and death-censored graft failure (full cohort) and composite eGFR ≤ 20, graft failure, or death (subcohort)

	Full cohort				Sub-cohort
	All-Cause Graft Failure (n=2430)		Death Censored GF (n=2430)		Composite eGFR≤20, GF, or death (n=1137)
	Event rate per 1000 py 66.2 (61.1, 71.7)		Event rate per 1000 py 33.3 (29.7, 37.3)		Event rate per 1000 py 76 (66, 87)
Log2 Transformed Urine Biomarker	Hazard Ratios (95% Confidence Interval)				Hazard Ratios (95% Confidence Interval)
	Unadjusted	Adjusted	Unadjusted	Adjusted	Unadjusted	Adjusted
Microalbumin	1.02 (0.97, 1.06)	1.00 (0.96, 1.06)	0.99 (0.93, 1.06)	0.98 (0.91, 1.05)	0.98 (0.91, 1.07)	0.96 (0.88, 1.05)
NGAL	1.02 (0.99, 1.05)	1.00 (0.97, 1.03)	1.01 (0.97, 1.05)	0.99 (0.95, 1.04)	0.99 (0.95, 1.04)	0.99 (0.94, 1.04)
KIM-1	0.99 (0.94, 1.03)	0.97 (0.92, 1.02)	1.00 (0.94, 1.06)	0.99 (0.93, 1.07)	1.00 (0.93, 1.08)	1.01 (0.92, 1.11)
IL-18	0.99 (0.95, 1.03)	0.98 (0.93, 1.02)	0.97 (0.92, 1.03)	0.96 (0.9, 1.03)	0.96 (0.89, 1.03)	0.96 (0.89, 1.05)
L-FABP	1.00 (0.96, 1.03)	0.98 (0.95, 1.01)	0.99 (0.95, 1.04)	0.98 (0.93, 1.03)	0.97 (0.92, 1.03)	0.97 (0.91, 1.02)

Adjusted for KDRI, Urine Creatinine, cold ischemia time, Recipient age, black race, sex, previous kidney transplant, diabetes as the cause of end-stage kidney disease, number of human leukocyte antigen mismatches, panel reactive antibody, BMI, pre-emptive transplant

The donor urine injury biomarkers microalbumin, NGAL, KIM-18, IL-18, and L-FABP were neither associated with graft failure in the full cohort nor the sub-cohort. Unadjusted and adjusted hazard ratios are presented in Table 2. In these models, KDRI was significantly associated with increased graft failure (not shown). In a supplementary analysis combining all biomarkers into one model, there was no significant association between biomarkers and graft failure outcomes (not shown).

Supplemental Table S3 shows urine injury biomarker concentrations by all-cause graft failure, death-censored graft failure, and the 3-year composite of graft failure, death, or eGFR <= 20 ml/min/1.73m².

Urine injury biomarkers and graft failure by serum creatinine-defined donor AKI

Of the 2430 kidney transplant recipients in our cohort, 1845 (76%) received kidneys from donors without serum creatinine-defined donor AKI. Among the recipient cohort without donor AKI, 475 (26%) recipients had all-cause graft failure. The event rate for all-cause graft failure was 65.4 (95% CI=55.7, 76.8) per 1000 person-years in the AKI group and 65.8 (60.2, 72.0) per 1000 person-years in the non-AKI group. In a secondary analysis, associations between urine biomarker and all-cause graft failure did not differ by serum creatinine-defined donor AKI except for LFABP (p-value for interaction <0.05; Figure 2). In donors with AKI, elevated LFABP was modestly protective of all-cause graft failure (unadjusted HR = 0.94, 95% CI = 0.88, 1.00). Overall, lack of significant effect modification of AKI on the association between biomarker concentration and graft failure suggests no relationship between subclinical AKI and graft failure. Figure 2 presents the full results and hazard ratios of this secondary analysis.

Figure 2.

Urine biomarker association with all-cause graft failure does not differ by donor AKI status (full cohort). Unadjusted hazard ratio (95% CI) per log-2-transformed urine biomarkers presented in figure for all recipients (full cohort, n = 2430) in blue, followed by HR by donor AKI strata in black. *P for interaction term between donor AKI and urine biomarkers was only significant for L-FABP (P < 0.05). AKI defined as stage 1 or higher using the AKI Network criteria. Event rate per 1000 person-y: for recipients from donors without AKI (n = 1845) was 65.8 (60.2, 72.0) and recipients from donors with AKI (n = 585) was 65.4 (55.7, 76.8). AKI, acute kidney injury; CI, confidence interval; HR, hazard ratio; KIM-1, kidney injury molecule-1; L-FABP, liver-type fatty acid binding protein; NGAL, neutrophil gelatinase-associated lipocalin.

Discussion

This multicenter observational analysis of long-term graft survival among kidney transplant recipients from deceased donors revealed that elevated donor urine injury biomarkers microalbumin, NGAL, KIM-1, IL-18, and L-FABP prior to organ procurement were not associated with long-term graft survival. Risk of all-cause graft and death-censored graft failure did not change with elevated urine injury biomarkers. Additionally, in our sub-cohort analysis, donor injury biomarkers were not associated with the 3-year composite outcome. Lastly, subclinical donor AKI defined by elevated urine injury biomarkers and normal serum creatinine was not associated with graft failure.

Given that urine injury biomarkers are more sensitive measures of AKI than serum creatinine,^17–21 these findings are generally consistent with and enhance existing literature showing that, among kidneys that are transplanted, clinical donor AKI by serum creatinine is not associated with worse recipient outcomes.^16,33–50 A recent meta-analysis of fourteen studies with 15345 deceased donors concluded that there is no significant difference in recipient allograft survival from AKI versus non-AKI donor groups at 12 months through 120 months post-transplantation.⁵¹ A review by Heilman et al. concluded that kidneys with biopsy-confirmed cortical necrosis less than 10% and no more than mild chronic changes had comparable long-term outcomes to kidneys without injury.⁵² Recent studies and the meta-analysis by Zheng et al. also suggest higher rates of DGF despite similar long-term recipient graft survival among recipients with AKI versus non-AKI donors.⁵¹ Our previous work showed that biomarkers could identify injury in the absence of creatinine-defined AKI (subclinical AKI).¹⁹ Our current findings show that the lack of association between injury biomarkers and graft failure did not differ by creatinine-defined AKI status (none versus Stage 1 or greater). Thus, subclinical AKI among kidneys without creatinine-defined AKI was not associated with long-term graft survival.

One analysis by Boffa et al. of 11219 kidneys in the UK differed from other studies with evidence for 20% greater risk of all-cause graft failure within 1 year for kidneys from donors with AKI as well as a slight but significantly worse 5-year graft survival for AKI kidneys compared with non-AKI kidneys (78% vs. 76%, p=0.009).⁵³ The authors of the UK study advised caution for using donor kidneys with the most severe AKI (Stage 3) because of an apparent 3-fold risk for primary non-function compared with non-AKI donor kidneys.⁵³ Among our recipient cohort, only 4% received kidneys with Stage 3 donor AKI versus 9% in the UK cohort, which may contribute to differences in findings between our study and Boffa et al. Our cohort may not have had sufficient power to detect differences in long-term outcomes among those with highest levels of donor kidney injury.

We hypothesize several reasons for our findings of similar long-term graft outcomes despite higher donor kidney injury biomarkers. First, donor AKI in the setting of DGF may induce protective upregulation of anti-ischemic proteins and cell repair processes through a mechanism similar to ischemic preconditioning.^54–56 Second, the most important factors determining graft function may take place after transplantation. Third, recipient mortality driven by factors like infection, cardiovascular disease, and frailty may generate noise in isolating the role of donor injury on long-term graft failure, which may affect our analysis of all-cause graft failure. However, our analysis of death-censored graft failure showed no association with injury biomarkers. With longer follow-up, this study supplements an earlier analysis of our group investigating donor urine injury biomarkers and early posttransplant graft survival.²³ Our earlier study suggested that while donor urine injury biomarkers were associated with donor AKI, they were associated with lower eGFR at 6 months only among recipients without DGF. Our current analysis with several years of follow-up after transplantation suggests that graft survival did not significantly change with increased urine injury biomarker concentrations, despite the increased rates of DGF shown in our previous work.

Some important limitations should be considered when evaluating our study’s findings. While our observational analysis adjusted for a variety of donor and recipient characteristics, including KDRI, residual confounding may exist. Consequently, donor kidneys with injury that are transplanted may differ from those with injury that are discarded and be healthier. Also, we did not measure urine injury biomarkers over time posttransplantation, which may have provided more insight into the mechanism of similar graft survival across donor kidney injury levels. Our definition of AKI is limited as it relies on admission creatinine, which may not be reflective of the donors’ true baseline (e.g., pre-renal injury prior to hospitalization). Also, the use of a single measurement of terminal serum creatinine to define AKI, without accounting for the different creatinine trends throughout the donor’s hospital stay prior to demise, may not accurately capture the true course the kidney has undertaken. Our cohort may not have had sufficient power to assess long-term recipient outcomes from donor kidneys with highest levels of injury (i.e., 4% of our cohort received Stage 3 AKI kidneys). Lastly, given median national graft survival times of approximately 10 years,⁵⁷ our median follow-up of four years may have been inadequate to detect potential effects of donor AKI on CKD at later stages in the post-transplant course.

Future investigations should evaluate mechanisms by which donor kidneys with AKI have similar post-transplant graft survival compared with non-AKI donor kidneys so that clinicians and transplant centers can better assess and predict long-term graft survival. These studies can measure urine injury biomarkers over time, both before and after transplant. We hypothesize that additional injury occurs during implantation and posttransplantation, and thus may make donor kidney injury a less important factor in long-term graft survival. Other donor and recipient factors may better predict posttransplant injury and be more important in predicting long-term allograft success.^58–61 Furthermore, donor AKI may induce protective processes that support long-term recovery and survival (e.g., ischemic preconditioning). Future studies can also investigate the role of other emerging biomarkers of repair like YKL-40 and uromodulin in predicting long-term graft outcomes.

We found no association between injury biomarkers and posttransplant graft survival (at median four years posttransplant). Our results suggest that, when evaluating long-term graft survival, transplant centers and clinicians can prioritize other donor and recipient factors that correlate with long-term outcomes over donor kidney injury, measured by either serum creatinine or urine injury biomarkers.

Abbreviations

AKI

acute kidney injury

ATI

acute tubular injury

BMI

body mass index

CKD

chronic kidney disease

DcGF

death censored graft failure

ECD

expanded-criteria donor

GFR

glomerular filtration rate

GF

graft failure

HRSA

health resources and services administration

KDRI

kidney donor risk index

KIM-1

kidney injury molecule-1

L-FABP

liver-type fatty acid binding protein

NGAL

neutrophil gelatinase-associated lipocalin

OPO

organ procurement organizations

OPTN

organ procurement and transplantation network

UNOS

united network for organ sharing