Deceased donor neutrophil gelatinase-associated lipocalin and delayed graft function after kidney transplantation: a prospective study

The present study was performed at Helsinki University Hospital, which provides organ transplant service for Finland, which has a population of 5.2 million. For this study, we prospectively enrolled 99 consecutive, deceased, heartbeating donors and their 176 adult kidney recipients between August 2007 and December 2008. The study protocol was approved by the Helsinki University Hospital Ethics Committee and the hospital's Department of Surgery. Written informed consent was obtained from the recipients before enrollment.

Altogether 198 kidneys were obtained from the 99 donors. One kidney was not transplanted because of a vascular lesion. Twenty-one kidneys were not included in the study: six were used for pediatric recipients, two were used for recipients who underwent combined kidney and liver transplantation and one was used for a combined kidney and lung transplantation. Nine kidneys were shipped to the other Nordic countries according to the Scandiatransplant exchange rules. Three patients did not consent to participate in the study. The recipients of the remaining 176 kidneys were included in this study.

Donor clinical history data were obtained from the hospital records. The following variables were gathered: age, gender, history of hypertension, need for cardiopulmonary resuscitation, need for intracranial surgery, use of vasopressor support, use of antidiuretic hormone (ADH), plasma creatinine level, length of hospital stay before brain death diagnosis, cause of death and multiorgan or kidney-only donation. Estimated glomerular filtration rate (eGFR) was calculated using the modification of diet in renal disease equation for glomerular filtration rate (MDRD equation) [28] in 96 adult donors. In three donors who were under 18 years of age (ages 9, 16 and 17 years), eGFR was calculated using the Schwartz formula [29]. ECDs were defined according to the criteria described by Port et al.[7], which include all donors older than 60 years of age, or donors older than 50 years of age with at least two of the following: plasma creatinine concentration above 132 μmol/L (1.5 mg/dL), cerebrovascular accident as the cause of death or a history of hypertension.

Intravenous steroids were given to all donors before undergoing the organ retrieval operation, and they were given mannitol before in situ perfusion was initiated. The University of Wisconsin solution was used for in situ perfusion and cold storage preservation of the kidneys. A biopsy for histological evaluation was taken from the donor kidney before the initiation of in situ perfusion. The biopsies were examined later and scored using the Banff 97 criteria [30] and the Chronic Allograft Damage Index (CADI) [31] to quantify renal allograft histology. In both scoring systems, different components in the biopsy are semiquantitatively evaluated and then summarized. The CADI score may have a value between 0 and 18, and it is obtained from individual component scores (0 to 3) for glomerular sclerosis, vascular intimal proliferation, interstitial inflammation, mesangial matrix increase, tubular atrophy and interstitial fibrosis.

Recipient clinical data were obtained from the patients' hospital records and the Finnish Kidney Transplant Registry database. Plasma creatinine concentration was recorded daily after transplantation during the recipient's stay in the transplant unit, then at 3 months and 1 year after transplantation. eGFR was calculated using the MDRD equation [28] at 3 months and 1 year after transplantation. Our standard immunosuppressive regimen was used as previously described [27].

The primary recipient outcome variable was onset of graft function after transplantation. DGF was defined as described by Halloran et al.[32]: oliguria less than 1 L/24 hours for more than 2 days, or plasma creatinine concentration greater than 500 μmol/L throughout the first week after transplantation, or more than one dialysis session needed during the first week after transplantation. In the analyses examining DGF duration, we divided the transplantations into three groups: EGF (n = 106), short DGF lasting less than 14 days (n = 43) and prolonged DGF lasting 14 days or longer (n = 27).

Serum samples for NGAL analyses were taken for logistical reasons in the donor hospital simultaneously with blood samples for HLA determination. The serum sample was drawn before the diagnosis of brain death in 36 cases (mean 9.1 hours, range 0.5 to 24.5) and after that in 63 cases (mean 1.4 hours, range 0.03 to 5.6). The serum samples were taken before steroid administration in 77 donors and after that in 22 donors. Urine samples were taken by the transplant team at the beginning of donor surgery. Thus all donors had already received the steroids before their urine samples were taken. All samples were immediately centrifuged at 2,500 rpm at 4°C for 10 minutes, and after that the serum and urine supernatant were divided into tubes and frozen at -70°C. No additives were used.

The S-NGAL assays were performed using a commercial enzyme-linked immunosorbent assay (ELISA) kit (BioPorto Diagnostics A/S, Gentofte, Denmark) as recommended by the manufacturer. The measurements were performed in duplicate and blinded to sample sources and clinical outcomes. Serum samples were available for NGAL analyses from 95 donors. In four cases, S-NGAL levels could not be analyzed because of inadequate (n = 2) or incorrectly processed (n = 2) sampling.

The U-NGAL assays were performed using a standardized clinical platform (ARCHITECT analyzer; Abbott Diagnostics, Abbott Park, IL, USA) as previously described [33]. Urine samples from 95 donors were available for NGAL analyses. Donor U-NGAL levels could not be determined in four cases because of inadequate (n = 1) or incorrectly processed (n = 3) sampling.

We divided the donors using the mean NGAL concentrations as cutoffs into a high NGAL group (S-NGAL ≥214 ng/mL, n = 38; U-NGAL ≥18 ng/mL, n = 26) and a low NGAL group (S-NGAL < 214 ng/mL, n = 57; U-NGAL < 18 ng/mL, n = 69).

SPSS version 18.0 software (SPSS, Inc., Chicago, IL, USA) was used for statistical analyses. All analyzed variables were tested for distribution. Student's t-test and analysis of variance were used to calculate samples with normal distribution, and the Mann-Whitney U and Kruskal-Wallis tests were used for analyses of samples with skewed distribution. χ² and Fisher's exact tests were employed for analyses of contingency tables. To assess DGF predictors, multilogistic regression analyses (forward and conditional) were used. Factors which were significantly different between the DGF and EGF groups in the univariate analyses, as well as for the other clinically relevant factors in this respect, were included in the multivariate analyses. The factors in the multivariate analyses consisted of categorical variables and the covariates of continuous variables. The parametric correlations were assessed using the Pearson correlation coefficient, and the nonparametric correlations were assessed using the Spearman correlation coefficient. Receiver operating characteristic curve (ROC) analysis was performed to assess the potential of NGAL to predict DGF. Positive and negative predictive values were calculated using Bayes' formula. A P value < 0.05 was considered significant.

The mean donor S-NGAL concentration was 212 ng/mL (range 27 to 720 ng/mL, SD 145.1). Donor S-NGAL concentrations correlated directly with donor plasma creatinine levels (R² = 0.35, P = 0.001) and inversely with donor eGFRs (R² = 0.24, P = 0.021).

The mean donor U-NGAL concentration was 18 ng/mL (range 0 to 177, SD 26.1). Donor U-NGAL concentrations correlated directly with donor plasma creatinine levels (R = 0.37, P < 0.0001) and inversely with donor eGFRs (R = 0.24, P = 0.01). Donor U-NGAL concentrations correlated directly with donor S-NGAL concentrations (R = 0.40, P < 0.0001).

Donors treated with ADH had significantly lower mean S-NGAL (188 ng/mL, SD 125.3) and U-NGAL (13 ng/mL, SD 14.3) levels compared to those not treated with ADH (S-NGAL: 249 ng/mL, SD 161.2, P = 0.002; U-NGAL: 26 ng/mL, SD 36.6, P = 0.045). Donor S-NGAL and U-NGAL levels did not correlate with donor age (R = 0.15 and P = NS for S-NGAL and donor age; R = 0.12 and P = NS for U-NGAL and donor age) and were not affected by gender, history of hypertension, use of vasopressors, length of hospital stay, need for cardiopulmonary resuscitation or intracranial surgery before brain death, ECD or standard criteria donor status, and multiorgan or kidney-only donation. In addition, there were no significant differences between donor S-NGAL levels in samples taken before or after brain death or before or after steroid administration (see Additional file 1).

Using the high vs. low NGAL division, we found that mean donor plasma creatinine level was significantly higher and that mean eGFR was lower in the high NGAL groups compared to the low NGAL groups (Table 3).

A representative biopsy for histological evaluation was available from 97 of 99 donors. Of the 97 biopsies, 58 (58.6%) showed normal histology. The mean CADI score of the biopsies was 0.72, ranging from 0 to 5 (Figure 1). Overall, the changes in the kidney biopsies were rare, apart from arterial changes (Table 4). Positive findings in single Banff classification components were not associated with the levels of donor plasma creatinine, eGFR, S-NGAL or U-NGAL (data not shown). However, the donors with high U-NGAL had significantly higher CADI scores than the donors with low U-NGAL (Figure 1 and Table 4).

Mean donor U-NGAL was significantly higher in cases with prolonged DGF (35 ng/mL, SD 49.4) compared to those with short DGF (15 ng/mL, SD 13.7) or EGF (15 ng/mL, SD 19.8) (P = 0.002). Mean donor S-NGAL did not differ significantly between the prolonged DGF (220 ng/mL, SD 141.5), short DGF (234 ng/mL, SD134.6) and EGF (206 ng/mL, SD 150.4) (P = NS) groups. There were no significant differences in mean donor S-NGAL and U-NGAL levels in the DGF (including both short and prolonged DGF) (S-NGAL 229 ng/mL, SD 136.4; U-NGAL 23 ng/mL, SD 33.3) and EGF (S-NGAL 206 ng/mL, SD 150.4, P = NS; U-NGAL 16 ng/mL, SD 19.8, P = 0.058) groups. High donor U-NGAL level was associated with more prolonged DGF and worse 1-year graft survival compared to low donor U-NGAL level (Table 3).

Multivariate analysis was performed to assess the factors predicting DGF and prolonged DGF. We included in the multivariate analysis the factors differing significantly between the DGF and EGF groups (donor age, ECDs vs. standard criteria donors, cold ischemia time and recipient pretransplantation mode of and time on dialysis) in addition to donor S-NGAL, U-NGAL, eGFR and plasma creatinine levels.

None of the included factors appeared to be a significant risk factor for DGF per se. Donor U-NGAL, ECD status and eGFR emerged as independent risk factors for prolonged DGF (Table 5). ROC analysis for donor U-NGAL in predicting DGF (Figure 2) resulted in an area under the curve (AUC) of 0.595 (95% confidence interval (95% CI) 0.506 to 0.749). ROC analysis performed to predict prolonged DGF (Figure 3) resulted in an AUC of 0.616 (95% CI 0.493 to 0.739). Table 6 shows the sensitivities, specificities and positive and negative predictive values at the lowest quartile (4 ng/mL), the median (9 ng/mL), the mean (18 ng/mL) and the highest quartile (20 ng/mL).