Neutrophil Gelatinase Associated Lipocalin Predicts Acute Kidney Injury in Patients Undergoing Liver Transplantation

Transcription

1 LIVER TRANSPLANTATION 16: , 2010 ORIGINAL ARTICLE Neutrophil Gelatinase Associated Lipocalin Predicts Acute Kidney Injury in Patients Undergoing Liver Transplantation Andrew J. Portal, 1 Mark J. W. McPhail, 1,3 Matthew Bruce, 1 Iona Coltart, 1 Andrew Slack, 1 Roy Sherwood, 2 Nigel D. Heaton, 1 Debbie Shawcross, 1 Julia A. Wendon, 1 and Michael A. Heneghan 1 1 Institute of Liver Studies; 2 Department of Medical Biochemistry, Kings College Hospital NHS Foundation Trust, Denmark Hill, London, England; and 3 Department of Hepatology, Division of Medicine, Imperial College London, 10th Floor QEQM Wing, St Mary s Hospital Campus, London, England Postoperative acute kidney injury (AKI) increases morbidity and mortality after liver transplantation (LT). Novel methods of assessing AKI including cystatin C (CyC) and neutrophil gelatinase associated lipocalin (NGAL) have been identified as potential markers of AKI. We compare the ability of standard renal markers (serum creatinine [scr], estimated glomerular filtration rate [egfr] and intensive therapy unit organ failure scores with CyC and NGAL to predict AKI within the first 48 hours after LT. 95 patients (median age 50 [interquartile range ¼ 41-59], 60% male) underwent LT (25% with acute liver failure). AKI was defined according to the Acute Kidney Injury Network criteria. Severe AKI was classified as stage 2. NGAL (urine [u] and plasma [p]) and CyC concentrations taken immediately after transplantation on admission to the Liver Intensive Care Unit were compared with standard markers of renal function. Predictive ability was assessed using the area under the curve generated by receiver operator characteristic analysis (AUROC) and logistic regression. Day 0 scr, ungal, pngal, CyC, and egfr predicted AKI as did SOFA (Sequential Organ Failure Assessment) and APACHE II (Acute Physiology and Chronic Health Evaluation II) scores. APACHE II and pngal were the most powerful predictors of severe AKI (APACHE II AUROC ¼ 0.87 [ ], P < 0.001; pngal AUROC ¼ 0.87 [ ], P < 0.001). Using multivariate logistic regression, APACHE II (odds ratio 1.64/point [95% confidence interval ¼ , P ¼ 0.001] and pngal [odds ratio ¼ 1.01/ng/mL [95% confidence interval ¼ ], P ¼ 0.002) retained independent significance. A renal risk score using APACHE II > 13 and pngal > 258 ng/ml was calculated with a score of 1 having a 100% sensitivity and 76% specificity for severe AKI. In conclusion, a combination of NGAL and APACHE II predicts AKI with high sensitivity and specificity after LT. Liver Transpl 16: , VC 2010 AASLD. Received March 14, 2010; accepted July 15, Renal dysfunction is common after liver transplantation (LT). The incidence of acute renal failure complicating the posttransplant period varies between 48% and 94% 1 and affects both short-term and long-term outcome. Mortality in those requiring renal replacement therapy (RRT) may be as high as 40% at 90 days, rising to 54% at 1 year. 2 Although many risk factors for developing renal dysfunction posttransplant have been investigated, the greatest impact on outcome is in patients who develop de novo renal impairment, especially in those who require RRT. 2 Patients with low glomerular filtration rates (GFR) at 1 month post-lt are also at risk of developing severe renal dysfunction at 5 years post-lt 3 and the need for dialysis in LT recipients has been reported to be as high as 18% at 5 years. 4 Abbreviations: AKI, acute kidney injury; APACHE II, Acute Physiology and Chronic Health Evaluation II; AST, aspartate aminotransferase; AUC, area under the curve; AUROC, area under the receiver operating characteristic curve; CI, confidence interval; CyC, cystatin C; egfr, enhanced glomerular filtration rate; ITU, intensive therapy unit; LT, liver transplantation; MAP, mean arterial pressure; NGAL, neutrophil gelatinase associated lipocalin; p, plasma; RRT, renal replacement therapy; SOFA, Sequential Organ Failure Assessment score; scr, serum creatinine; u, urine. Address reprint requests to Michael A. Heneghan, M.D., MmedSc, King s College Hospital NHS Foundation Trust, Denmark Hill, London SE5 9RS, England. Telephone: þ ; FAX: þ ; michael.heneghan@kch.nhs.uk DOI /lt View this article online at wileyonlinelibrary.com. LIVER TRANSPLANTATION.DOI /lt. Published on behalf of the American Association for the Study of Liver Diseases VC 2010 American Association for the Study of Liver Diseases.

2 1258 PORTAL ET AL. LIVER TRANSPLANTATION, November 2010 Many studies have attempted to determine risk factors for the development of renal dysfunction after undergoing transplantation. Factors such as preoperative serum creatinine (scr), baseline GFR < 70 ml/minute/ 1.73 m 2, severe early graft dysfunction, low preoperative albumin, retransplantation, postoperative sepsis and duration of dopamine therapy have all been implicated in the development of renal dysfunction post-lt. 5-8 No unifying factor has, however, been yet been identified to accurately predict renal dysfunction after LT. Acute renal failure has now been reclassified as acute kidney injury (AKI) by international consensus. 9 These criteria proposed by the Acute Kidney Injury Network (AKIN) were devised from the previously issued RIFLE (Risk, Injury, Failure, Loss, and End-stage kidney disease) criteria. 10 The changes were made for several reasons: first the RIFLE criteria needed a baseline GFR or scr in order to stratify patients. This was often estimated from expected values of estimated GFR (egfr) calculated according to patient age and sex, but assuming normal renal function. This was inherently flawed, but particularly so in patients with liver disease, given that scr is such a poor marker of renal dysfunction. 11 Also, recent evidence has highlighted that even minor changes in scr are associated with adverse outcomes. 9 Although the AKIN criteria need 2 scr values, they recognize the fact that small changes in scr affect can significantly affect outcome. The search for novel markers of renal function has intensified over recent years and several promising markers have been identified. Cystatin C (CyC) is a ubiquitous protein that is freely filtered by the kidney and then metabolized by the tubules. It has been extensively studied and compares favorably with standard markers of renal function. 12 Moreover, CyC significantly outperformed both serum creatinine and creatinine clearance and detected impairment of GFR earlier than serum creatinine. CyC also has a role in the pretransplant setting and may obviate the need for isotope studies. 13 CyC has also been used to predict AKI immediately post-lt with greater accuracy than scr. 14 Synthesis of neutrophil gelatinase associated lipocalin (NGAL) is massively up-regulated after AKI, as it is released in response to any tubular injury. 15 Levels are elevated in serum, plasma and also urine. As serum or plasma levels of NGAL can be elevated in other conditions such as sepsis, it is thought that urinary NGAL may be a more pure predictor of AKI. Mishra et al. recently reported on the development of AKI in children undergoing cardiopulmonary bypass procedures. 16 Serum and urine NGAL levels were significantly elevated in patients who developed AKI and this rise preceded any change in scr by up to 48 hours. These findings have subsequently been corroborated in several other studies using adult rather than pediatric patients NGAL has also been shown to predict delayed graft function and the need for RRT in renal transplantation. 20 Because AKI is a common occurrence after LT, we hypothesize that NGAL may predict AKI with high accuracy. To test this theory, we collected plasma and urine samples from patients undergoing LT and compared NGAL and CyC levels against standard markers of renal function for the development of AKI and 90- day renal morbidity. We also examined their role in predicting the need for RRT in the cohort. PATIENTS AND METHODS Study Population We recruited 95 adult patients who underwent LT at King s College Hospital NHS Foundation Trust, London, England. Ethical approval was granted from the local Research and Ethics committee prior to commencement. Clinical practice was not changed during the study and all patients were treated according to established protocols. RRT was commenced for one or more of the following reasons: presence of oliguria (<100 ml of urine for a duration of more than 6 hours); hyperkalemia >6.5 mmol/l refractory to medical treatment; fluid overload compromising cardiovascular or respiratory function; severe acidosis refractory to volume replacement. egfr was calculated using the 4-parameter Modification of Diet in Renal Disease (MDRD) equation found on the UK CKD eguide ( Immunosuppression was commenced according to established protocol with steroid induction given perioperatively and tacrolimus commenced using a twice daily dosing regimen to achieve levels of between 5-10 lmol/l by the end of the first week posttransplant. Alternatively, calcineurin inhibitor sparing regimens were used in patients identified as being at high risk of calcineurin inhibitor induced toxicity. Clinical Outcomes We assessed standard markers of renal function on admission to our intensive care unit, immediately after LT (day 0 scr and day 0 egfr), and compared them with novel biomarkers (CyC and NGAL). Day 0 scr was used rather than urine output. This was primarily because patients remained on the intensive care unit for different durations and urine output measurement was not reliably measured on an hourly basis when transferred to the ward environment. 21 AKI within the first 48 hours after LT was defined according to the Acute Kidney Injury Network criteria (Table 1). 9 Patients who developed AKI of any severity (ie, stages 1-3) within 48 hours were classified as having day 2 AKI. Patients with more severe AKI (stages 2 or 3) within the first 48 hours were classified as having severe AKI. Patients requiring RRT were classified as having AKI stage 3. The systemic inflammatory response syndrome (SIRS) was defined according to Thabut et al. 22,23 Abilities of standard and novel markers to predict the need for RRT and mortality were evaluated. Measurement of Biomarkers Plasma NGAL (pngal) was collected within 24 hours of LT and within the first 12 hours of arrival to the

3 LIVER TRANSPLANTATION, Vol. 16, No. 11, 2010 PORTAL ET AL TABLE 1. Classification/Staging System for Acute Kidney Injury Using scr Criteria 9 Stage Serum Creatinine Criteria 1 Increase in serum creatinine of 0.3 mg/dl (26.4 lmol/l) or increase to 150%-200% (1.5-fold to 2-fold) from baseline 2 Increase in serum creatinine to >200%-300% (>2-fold to 3-fold) from baseline 3 Increase in serum creatinine to >300% (>3-fold) from baseline (or serum creatinine of 4.0 mg/dl [354 lmol/l] with an acute increase of at least 0.5 mg/dl [44 lmol/l]) AKI is defined over a 48-hour time period. intensive therapy unit (ITU) in all patients. This was defined as day 0, and samples were again obtained at 48 hours (day 2). Urinary NGAL (ungal) and CyC were collected in approximately 50% of patients at the same time points as for pngal. Samples were stored at 80 C until time of testing. Plasma was stored for measurement of NGAL whereas serum was used for CyC. Plasma NGAL was measured using a commercially available ELISA kit (Bioporto Diagnostics, Gentofte, Denmark). Briefly, pre-prepared microwells containing horseradish peroxidase conjugated monoclonal antibodies against human NGAL were filled with calibrators and samples (diluted according to the manufacturer s specification). A chromogenic peroxidase substrate was then added to the wells. Horseradish peroxidase bound to the detection antibodies reacted with the substrate to form a colored product. This reaction was then stopped chemically, and the color intensity was then read at 450 nm in a plate reader. Results from the calibration wells were read to allow a calibration curve to be constructed, which in turn allowed the concentrations of NGAL in the specimens to be calculated. All samples were measured in duplicate and in a blinded fashion, by the same operator. CyC was measured by an immunoturbidimetric assay produced by DakoCytomation (distributed by Bio-Stat, Ltd., Stockport). Samples were mixed with immunoparticles and the resultant complexes measured by turbidimetry at 550 nm wavelength at 37 C. The signal generated was then correlated with the concentration of CyC in the sample. By interpolation on a standard curve the concentration of CyC in the sample was calculated. APACHE II and SOFA scores were calculated using the Riyadh Intensive Care (RIP) programme Data Collection and Statistical Analysis Data was collected prospectively on admission to the intensive care unit, using the Riyadh Intensive Care Programme (RIP). Data regarding renal function was collected up to 3 months after transplantation. Measures of central tendency were compared using the Mann-Whitney U test as all clinical and biochemical data were shown to be nonparametric and thus are expressed as median and interquartile range. A 2-sided P value of <0.05 was considered significant. Correlations between novel biomarkers and other variables were evaluated using Spearman s correlation coefficient. Mann-Whitney U and Kruskal Wallis tests were used to compare medians between two and multiple groups respectively. Ability of a biomarker to predict an outcome was assessed using the area under the curve generated by receiver operator characteristic analysis. The area under the receiver operating curve (AUROC) was used to assess the ability of the continuous variable to distinguish categorical state using the formula of Hanley and McNeil. Optimal cutoffs were also determined from the ROC curves. An AUROC value of indicated excellent, good, fair, poor and no useful performance for discrimination of the outcome under assessment. 19 Data were analyzed using SPSS (SPSS version 16.0, Chicago, IL) and MedCalc version 10 (MedCalc Software, Mariakerke, Belgium). RESULTS Patient Characteristics Data from 95 adult liver transplant recipients were analyzed. 15 patients had been on RRT for more than 24 hours prior to LT. They were subsequently excluded from the AKI analysis but were included in the other analyses. The median age of patients was 50 years with 22 patients having undergone LT for acute liver failure. Three patients were urgently retransplanted for either primary graft dysfunction or hepatic artery thrombosis, whereas the remainder underwent transplantation due to a variety of chronic liver disease (alcohol-related liver disease n ¼ 21; viral liver disease [hepatitis B or hepatitis C virus infection] n ¼ 19; cholestatic liver disease n ¼ 12; autoimmune hepatitis n ¼ 5; other [Alagille syndrome, cryptogenic cirrhosis, alpha 1 antitrypsin deficiency, Wilson disease], n ¼ 16). All 95 patients had pngal, day 0 scr, egfr, and APACHE II recorded. A total of 54 and 46 had CyC and ungal measured, respectively. There were no significant differences between the median levels of renal biomarkers according to etiology or in relation to urgency of transplantation. The 15 patients who required preoperative RRT were then excluded and the data reanalyzed. In patients who subsequently went on to develop AKI, there were no significant differences between median values of day 0 renal biomarkers between those transplanted for acute or chronic liver failure. However, the median values of day

4 1260 PORTAL ET AL. LIVER TRANSPLANTATION, November 2010 TABLE 2. Characteristics of All Patients and Those With/Without Subsequent Acute Kidney Injury Within the First 48 Hours of LT All patients Excluding Patients Undergoing Renal Replacement Therapy Before Transplant (n ¼ 15) With acute Without acute kidney injury kidney injury Characteristic N ¼ 95 N ¼ 30 N ¼ 50 P Value Age 50 (38-57) 54(41-61) 52 (40-56) Sex (M/F) 56/39 23/7 29/ Acute/chronic liver disease 23/72 4/26 8/ Mortality at 30 days 4/95 3/30 0/ ITU stay (days) 5 (3-11) 6 (4-12) 3 (3-4) <0.001 MELD (preoperative) 16 (12-28) 16 (13-28) 14 (10-21) 0.06 APACHE II 12 (9-15) 14 (11-16) 10 (8-12) <0.001 SOFA 8 (5-12) 8 (6-12) 7 (5-9) AST (IU/L) 623 ( ) 616 ( ) 612 ( ) Albumin (g/dl) 22 (18-24) 19 (16-25) 22 (19-24) Bilirubin (lmol/l) 59 (32-89) 74 (47-94) 42 (25-78) Day 1 scr (lmol/l) 116 (89-158) 124 ( ) 96 (84-127) Pre-LT scr (lmol/l) 104 (81-129) 112 (81-139) 91 (73-109) Urine output (ml/24 hours) 970 ( ) 805 ( ) 1295 ( ) <0.001 egfr (ml/minute/1.73 m 2 ) 58 (39-71) 52 (33-63) 67 (52-78) Results are expressed as medians with interquartile ranges. The scr was taken immediately after LT, at time of sampling for novel markers. Pre-LT scr was taken immediately prior to LT. 0 scr and pngal were slightly but significantly higher in those who subsequently developed AKI by day 2 after LT, with regards to their urgency of transplant (day 0 scr in those with no AKI transplanted for ALF or urgent retransplantation 87 [80-89] lmol/l versus 101 [83-133] lmol/l for those with chronic liver disease, P ¼ 0.04; pngal in those with no AKI transplanted for ALF or urgent retransplantation 90.5 [58-133] ng/ml versus 130 [ ] ng/ml, P ¼ 0.02), possibly reflecting more chronic renal dysfunction in those with previous underlying liver disease. Comparison of Novel and Standard Renal Biomarkers in Those With/Without Acute Kidney Injury and Severe Acute Kidney Injury 30 patients developed AKI by the second postoperative day. Patients who developed AKI had a significantly longer stay in intensive care, had higher APACHE II and SOFA scores (Table 2). There were no significant differences in fluid balance status or type of fluid received during the transplant procedure, or in the duration of the procedure (Table 3). Urine output and egfr were significantly lower, and levels of pre-lt scr, day 0 scr, and serum bilirubin were significantly higher in those who developed AKI (Table 2). Novel markers of renal function were also significantly higher in those that developed AKI (Fig. 1). Although there were also significant differences between pre-lt scr in those with and without severe day 2 AKI, interestingly, median values were not elevated beyond the normal ranges for our institution (upper limit of normal 120 lmol/l). For patients with and without severe AKI, respectively, pre LT scr was 116 (90-229) lmol/l and 92 (75-113)) lmol/l). In those with stage 1 AKI, day 0 scr levels were not elevated outside of the normal range and the median value of day 0 scr in more severe AKI was only 158 lmol/l (Fig. 1). All novel markers of kidney injury were elevated outside their reported normal ranges at any stage of AKI. pngal levels were significantly elevated in those patients who required RRT by day 2 post LT (day 0 pngal in patients on RRT [day 2] versus no RRT: 307 [ ] ng/ml versus 140 [ ] ng/ml, P < 0.001). pngal also predicted the need for RRT in the early postoperative period (AUROC 0.84, 95% CI ). Three of 19 patients with severe AKI subsequently died within 30 days. Development of severe AKI was significantly associated with mortality at 30 days (P ¼ 0.021, chi-square test). Prediction of Acute Kidney Injury and Severe Acute Kidney Injury Within 48 Hours of LT pngal predicted the development of AKI within the first 48 hours after LT with high accuracy (Fig. 2), compared with day 0 scr. A cutoff value of pngal >212 ng/ml predicted AKI with a sensitivity of 67% and specificity of 84%, P < For day 0 scr, the cutoff value yielding the best sensitivity and specificity was still within the normal range :cutoff >104 lmol/ L, AUC 0.72 (95% CI ¼ ), P < (upper limit of laboratory normal range ¼ 120 lmol/l). There were no significant differences between pngal and day 0 scr to predict AKI (P ¼ 0.296, Delong method for ROC curve comparison). Only CyC and ungal

5 LIVER TRANSPLANTATION, Vol. 16, No. 11, 2010 PORTAL ET AL TABLE 3. Duration and Fluid Balance Data During LT Operation in Those With and Without Subsequent Acute Kidney Injury Acute Kidney No Acute Kidney Data Injury (n ¼ 30) Injury (n ¼ 48) P Value Anesthetic time (hours) 8 ( ) 7.1 ( ) Cell saver processed (ml) 3830 ( ) 2148 ( ) Cell saver returned (ml) 1118 ( ) 630 ( ) Packed red cells (ml) 1925( ) 1750 ( ) Fresh frozen plasma (ml) 3521 ( ) 2617 ( ) Platelets (ml) 374( ) 275 (0-444) Colloid (ml) 2500 ( ) 2500 ( ) Total blood loss (ml) 5616 ( ) 3935 ( ) Urine output (ml) 300 ( ) 426 ( ) Fluid balance (ml) 4380 ( ) 3533 ( ) Fluid balance is defined as the total fluid given or transfused minus fluid or blood losses. There were no significant differences between the groups. Data from 2 patients were incomplete and were therefore omitted. approached the performance of pngal when predicting AKI, although all markers were statistically significant (cutoff value for CyC > 2.04 mg/l, AUC 0.78 ( , P ¼ 0.006; cutoff value for ungal > 150 ng/ml, AUC 0.76 ( ), P ¼ Pre-LT scr was not a reliable predictor of AKI, cutoff value >108 lmol/l, AUC 0.67 ( ), P ¼ (Table 4). For patients who developed severe AKI, a pngal value of >258 ng/ml was predictive with an AUROC of 0.87 ( ) (sensitivity 74% and specificity 85%, P < (Fig. 2). The day 0 scr cutoff value predicting severe AKI was again within the normal reference range (>104 lmol/l, AUC 0.81 [ ], P < 0.001). APACHE II score was also a powerful predictor of severe AKI. All markers performed better at predicting severe (ie, stages 2 and 3) AKI than at predicting the presence of AKI per se (Table 4). Using univariate regression analysis, we demonstrated that admission values of novel renal markers (CyC, pngal, and ungal), standard markers (day 0 scr and egfr), intensive care organ scores (SOFA and APACHE II) and serum bilirubin were associated with development of AKI (Table 2, Fig. 1). In backwards stepwise multiple logistic regression analysis, only admission pngal and APACHE II were retained as independent predictors of severe AKI (Table 5). We calculated a Renal Risk Score, scoring 1 point for APACHE II > 13 or serum NGAL > 258 ng/ml, to generate a result of 0 (low risk of severe AKI), 1 (intermediate risk of severe AKI), or 2 (high risk of severe AKI). No patients with a score of 0 (n ¼ 46) developed severe AKI, whereas 8 of 20 patients (40%) with a score of 1 developed severe AKI whereas 11 of 14 (79%) patients with a score of 2 developed severe AKI. A score of >0 was associated with 100% sensitivity and 75% specificity for the development of severe AKI. DISCUSSION In this prospective cohort study of patients undergoing LT, a single measurement of pngal taken within 24 hours of LT predicted AKI and severe AKI with a high degree of accuracy. pngal is also elevated in those that require RRT. Furthermore, pngal was superior to day 0 SCr at determining which patients are at risk of developing AKI especially when combined with APACHE II. NGAL has been studied in a variety of different patient groups. Mishra et al. reported that a serum NGAL cutoff value of >25 ng/ml predicted the development of AKI with the best sensitivity and specificity in a cohort of children undergoing cardiopulmonary bypass (CPB). 16 Other studies have reported somewhat different cutoff levels, suggesting certain heterogeneity between the studies. Dent et al. reported that a pngal >150 ng/ml in children 2 hours post bypass predicted AKI with an AUROC of 0.96 (sensitivity 84%, specificity 96%). 27 Similarly, in an adult population of 100 patients undergoing CPB, pngal >150 ng/ml, taken on arrival to the intensive care unit, was again highly predictive of AKI (AUROC 0.8, sensitivity 79%, specificity 78%). 19 Although our cutoff levels are somewhat different to previously published data, a recent meta-analysis of the utility of NGAL suggested that some of the heterogeneity of thresholds may relate to the assay used in addition to the heterogeneity of patient groups evaluated. 28 In this study, we used an ELISA-based antibody method whereas other studies have used either larger scale laboratory methods or point of care testing devices. However, our cutoffs of 212 ng/ml and 258 ng/ml fall well within the ranges previously published ( ng/ml) as do our AUROC values of 0.79 and 0.87 for any AKI or severe AKI respectively (AUROC meta-analysis range from To date, there are few data available describing utility of NGAL in patients with liver disease. Niemann et al. have recently published a report of NGAL and its use in LT. 29 They found that NGAL levels measured both 2 and 24 hours postreperfusion were elevated in those who subsequently had a >50% increase in their scr, fulfilling the criteria for AKI as

6 1262 PORTAL ET AL. LIVER TRANSPLANTATION, November 2010 Figure 1. (A) Median (interquartile range) values of day 0 plasma NGAL, urinary NGAL, serum CyC and serum creatinine stratified by AKI stage. The Kruskal-Wallis test for each marker was; plasma NGAL (P < 0.001), urinary NGAL (P ¼ 0.006), CyC (P ¼ 0.003), and creatinine (P ¼ 0.003). The 2-group Mann-Whitney test significances are denoted on the figure. (B) Median values for each of the markers, classified as those without AKI, stage 1 AKI, or stage 2 or greater AKI are listed. defined by the RIFLE criteria. 30 Although it is already known that development of AKI post LT affects graft survival and long-term outcome, NGAL was only of use in diagnosing mild to moderate AKI. 31 Our data goes beyond this and supports the concept that not only is NGAL of use in diagnosing AKI, but that it has in addition good predictive utility for determining which patients may develop such severe kidney injury as might require RRT. In other settings, NGAL has also been studied in critically ill children. In one study, the ability of serum NGAL to predict AKI was poor when compared with previous studies (sensitivity 86%, specificity 39%). 32 However, renal dysfunction was classified using either blood urea, scr level >177 lmol/l or the need for RRT. It is unsurprising therefore that NGAL performs less well, because it is primarily most useful as an early marker of AKI, Moreover, it is likely that NGAL was

7 LIVER TRANSPLANTATION, Vol. 16, No. 11, 2010 PORTAL ET AL Figure 2. Receiver operating curves for performance of day 0 plasma NGAL (solid line) or day 0 creatinine (scr) (dashed line) to predict (A) any AKI or (B) severe AKI. Day 0 samples were taken immediately after transplantation for scr, or within 12 hours for pngal. In (A), the AUROC for NGAL is 0.79 (95% CI ¼ ) and for creatinine (scr) 0.72 (95% CI ¼ ); P ¼ 0.296, Delong method for ROC curve comparison). For (B), the AUROC for NGAL is 0.87 (95% CI ¼ ) and for creatinine (scr) 0.81 (95% CI ¼ ); P ¼ 0.359, Delong method for ROC curve comparison). TABLE 4. Ability of Renal Markers and Intensive Care Outcome Scores to Predict AKI (Defined as Those With Any Stage of AKI) and Severe AKI (Defined as Any Patient With 2 AKI) Within the First 48 Hours Following LT Admission Variable Cutoff AUC (95% CI) P Value Cutoff AUC (95% CI) P Value Pre-LT scr (lmol/l) > ( ) > ( ) Day 0 scr (lmol/l) > ( ) < ( ) <0.001 Day 0 egfr (ml/minute/1.73 m 2 ) < ( ) < ( ) <0.001 Day 0 Plasma NGAL (ng/ml) > ( ) <0.001 > ( ) <0.001 Day 0 Urinary NGAL (ng/ml) > ( ) > ( ) Day 0 Cystatin C (mg/l) > ( ) > ( ) Day 0 APACHE II > ( ) <0.001 > ( ) <0.001 Day 0 SOFA > ( ) > ( ) Cutoffs are quoted and refer to the value of the marker yielding the highest sensitivity and specificity. AKI Severe AKI significantly elevated for some time prior to these children fulfilling the criteria for renal injury. Another pediatric study, comprising 140 children admitted into a single intensive care unit, found that ungal was a fair predictor of AKI (AUROC 0.78), with ungal levels increasing 6-fold up to 48 hours before a rise in scr. 33 In adult critically ill patients. Cruz et al. found that not only was AKI common in an adult ITU (44% developing AKI during the course of their admission), but that pngal was an independent predictor of AKI and correlated with severity of AKI as well as severity of illness. 34 Our findings are very similar to these: pngal was higher in those with severe AKI, and was associated with the need for RRT. Although Cruz used the RIFLE criteria for determining AKI, the methodology was otherwise comparable. Our decision to use the AKIN revised criteria was based primarily on the basis of how poor a marker of renal dysfunction scr is in patients with cirrhosis and how small rises may reflect significant loss of functional nephrons. 9 The AKIN criteria have also been reported to improve the sensitivity of diagnosing AKI as compared with the

8 1264 PORTAL ET AL. LIVER TRANSPLANTATION, November 2010 TABLE 5. Logistic Regression Analysis of Independent Predictors of Severe AKI Predictor Odds Ratio 95% CIP Value Univariate Age Sex (M/F) APACHE II SOFA MELD pngal ungal Serum creatinine egfr Cystatin C (mg/ml) Bilirubin INR HE Glasgow coma scale Lactate Dacluzimab use Vasopressor use Multivariate APACHE II 1.64/point 1.22 to pngal 1.01/ng/mL1.00 to For the final model, the c-statistic is (95% CI ¼ , 91% of cases correctly classified). RIFLE criteria, although there is little difference in their ability to predict outcome. 30 Recent studies have shown that SIRS is associated with renal dysfunction in both acute and chronic liver failure. 22,35 Bagshaw et al. recently found that pngal and ungal were elevated in critically ill patients with septic AKI as compared with non septic AKI, findings which had previously been reported in a pediatric population. 32,36 The question that arises is whether the elevated NGAL reflects a worse AKI associated with either SIRS or sepsis, or alternatively, whether NGAL is disproportionately elevated due to the presence of inflammation? Bagshaw et al. hypothesized that sepsis induces a greater renal injury than in non septic causes of AKI but acknowledged that NGAL expression may be up-regulated in inflammatory conditions, as previously published In our patients, both higher pngal values and presence of SIRS were associated with more severe AKI, but our data do not allow us to determine whether it is the presence of inflammation or the severity of the AKI that drives the rise in pngal. Our data do, however, support the notion that pngal relates to disease severity, as it was significantly elevated in those with severe AKI, those requiring RRT and indirectly associated with a worse outcome (increased mortality in those with AKI). ungal also appears to be a powerful predictor of AKI. Our results show that it is a better predictor than scr and pre-lt scr, especially in those with severe AKI, but appears not to be as good as pngal. However, we only managed to collect urine on approximately half of our patients, in part at least due to the presence of anuria. Despite these limitations, ungal performed well. Whether ungal would have outperformed pngal remains unanswered, and indeed in the literature, no consensus exists regarding this. In theory, ungal is perhaps a more specific marker of AKI, as it is produced in the tubules, and should not be subject to other conditions that may elevate NGAL levels for nonrenal reasons. Despite this, meta-analysis suggests that NGAL regardless of its derivation performs equally well. 28 CyC in this study did not perform as well as NGAL in predicting AKI. CyC is more commonly thought of as a marker of glomerular filtration, rather than a marker of AKI. However, there are data to suggest that it performs as well as scr in identifying AKI and this is reflected in our findings. 40 This study has several limitations. First, it is a single-center study and the numbers are relatively small. Because of this, we were unable to show that the development of AKI in the immediate post operative period led to ongoing renal dysfunction. Exclusion of those with preexisting renal dysfunction further reduced our numbers. However, these patients had already fulfilled criteria for AKI prior to NGAL measurements being made and so had to be excluded from the AKI analysis. Predictive performance of pngal actually increases if these 15 patients are included. Another reason for excluding these patients is that it is not certain if NGAL is activated by extracorporeal blood circuits as used in hemofiltration, although we feel this is unlikely as NGAL levels did not increase or significantly decrease during the first 24 hours of RRT in 2 patients in whom serial measurements were made (unpublished data). Also, we only used a single measurement of NGAL, and serial NGAL levels may in fact be more useful in the early prediction of AKI. A baseline NGAL measurement allowing the calculation of the change in NGAL values pre and post LT may have added additional sensitivity and specificity. The timing of the NGAL sampling was also variable, with up to 12 hours elapsing in some cases from the time of LT to the time of sample. This heterogeneity in the timing of NGAL measurement is both a weakness and a potential strength of the study. Previous studies have shown that NGAL levels reach their peak at approximately 2 hours post kidney injury. 16 Despite this, our data demonstrates that pngal is a very powerful predictor of AKI, even up to 12 hours after the transplant procedure. It is also associated with the need for RRT. Larger multicenter studies are needed to confirm its utility. REFERENCES 1. Yalavarthy R, Edelstein CL, Teitelbaum I. Acute renal failure and chronic kidney disease following liver transplantation. Hemodial Int 2007;11(Suppl 3):S7-S Gonwa TA, Mai ML, Melton LB, Hays SR, Goldstein RM, Levy MF, et al. Renal replacement therapy and orthotopic liver transplantation: the role of continuous veno-venous hemodialysis. Transplantation 2001;71:

9 LIVER TRANSPLANTATION, Vol. 16, No. 11, 2010 PORTAL ET AL Herlenius G, Fistouris J, Olausson M, Felldin M, Backman L, Friman S. Early renal function post-liver transplantation is predictive of progressive chronic kidney disease. Scand J Gastroenterol 2008;43: Ojo AO, Held PJ, Port FK, Wolfe RA, Leichtman AB, Young EW, et al. Chronic renal failure after transplantation of a nonrenal organ. N Engl J Med 2003;349: Cabezuelo JB, Ramirez P, Rios A, Acosta F, Torres D, Sansano T, et al. Risk factors of acute renal failure after liver transplantation. Kidney Int 2006;69: Bilbao I, Charco R, Balsells J, Lazaro JL, Hidalgo E, Llopart L, et al. Risk factors for acute renal failure requiring dialysis after liver transplantation. Clin Transplant 1998;12: Pawarode A, Fine DM, Thuluvath PJ. Independent risk factors and natural history of renal dysfunction in liver transplant recipients. Liver Transpl 2003;9: Afonso RC, Hidalgo R, Zurstrassen MP, Fonseca LE, Pandullo FL, Rezende MB, et al. Impact of renal failure on liver transplantation survival. Transplant Proc 2008;40: Mehta RL, Kellum JA, Shah SV, Molitoris BA, Ronco C, Warnock DG, et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007;11:R Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P. Acute renal failure - definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8:R204-R Portal AJ, Austin M, Heneghan MA. Novel approaches to assessing renal function in cirrhotic liver disease. Hepatol Res 2007;37: Laterza OF, Price CP, Scott MG. Cystatin C: an improved estimator of glomerular filtration rate? Clin Chem 2002; 48: Samyn M, Cheeseman P, Bevis L, Taylor R, Samaroo B, Buxton-Thomas M, et al. Cystatin C, an easy and reliable marker for assessment of renal dysfunction in children with liver disease and after liver transplantation. Liver Transpl 2005;11: Ling Q, Xu X, Li JJ, Chen J, Shen JW, Zheng SS. Alternative definition of acute kidney injury following liver transplantation: based on serum creatinine and cystatin C levels. Transplant Proc 2007;39: Mishra J, Ma Q, Prada A, Mitsnefes M, Zahedi K, Yang J, et al. Identification of neutrophil gelatinase-associated lipocalin as a novel early urinary biomarker for ischemic renal injury. J Am Soc Nephrol 2003;14: Mishra J, Dent C, Tarabishi R, Mitsnefes MM, Ma Q, Kelly C, et al. Neutrophil gelatinase-associated lipocalin (NGAL) as a biomarker for acute renal injury after cardiac surgery. Lancet 2005;365: Koyner JL, Bennett MR, Worcester EM, Ma Q, Raman J, Jeevanandam V, et al. Urinary cystatin C as an early biomarker of acute kidney injury following adult cardiothoracic surgery. Kidney Int 2008;74: Wagener G, Gubitosa G, Wang S, Borregaard N, Kim M, Lee HT. Urinary neutrophil gelatinase-associated lipocalin and acute kidney injury after cardiac surgery. Am J Kidney Dis 2008;52: Haase-Fielitz A, Bellomo R, Devarajan P, Story D, Matalanis G, Dragun D, et al. Novel and conventional serum biomarkers predicting acute kidney injury in adult cardiac surgery-a prospective cohort study. Crit Care Med 2009;37: Parikh CR, Jani A, Mishra J, Ma Q, Kelly C, Barasch J, et al. Urine NGAL and IL-18 are predictive biomarkers for delayed graft function following kidney transplantation. Am J Transplant 2006;6: Garcia-Tsao G, Parikh CR, Viola A. Acute kidney injury in cirrhosis. Hepatology 2008;48: Thabut D, Massard J, Gangloff A, Carbonell N, Francoz C, Nguyen-Khac E, et al. Model for end-stage liver disease score and systemic inflammatory response are major prognostic factors in patients with cirrhosis and acute functional renal failure. Hepatology 2007;46: American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992;20: Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985;13: Gildea TR, Cook WC, Nelson DR, Aggarwal A, Carey W, Younossi ZM, et al. Predictors of long-term mortality in patients with cirrhosis of the liver admitted to a medical ICU. Chest 2004;126: Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996;22: Dent CL, Ma Q, Dastrala S, Bennett M, Mitsnefes MM, Barasch J, et al. Plasma neutrophil gelatinase-associated lipocalin predicts acute kidney injury, morbidity and mortality after pediatric cardiac surgery: a prospective uncontrolled cohort study. Crit Care 2007;11:R Haase M, Bellomo R, Devarajan P, Schlattmann P, Haase-Fielitz A. Accuracy of neutrophil gelatinase-associated lipocalin (NGAL) in diagnosis and prognosis in acute kidney injury: a systematic review and meta-analysis. Am J Kidney Dis 2009;54: Niemann CU, Walia A, Waldman J, Davio M, Roberts JP, Hirose R, et al. Acute kidney injury during liver transplantation as determined by neutrophil gelatinase-associated lipocalin. Liver Transpl 2009;15: Lopes JA, Fernandes P, Jorge S, Goncalves S, Alvarez A, Costa e Silva Z, et al. Acute kidney injury in intensive care unit patients: a comparison between the RIFLE and the Acute Kidney Injury Network classifications. Crit Care 2008;12:R Barri YM, Sanchez EQ, Jennings LW, Melton LB, Hays S, Levy MF, et al. Acute kidney injury following liver transplantation: definition and outcome. Liver Transpl 2009; 15: Wheeler DS, Devarajan P, Ma Q, Harmon K, Monaco M, Cvijanovich N, et al. Serum neutrophil gelatinase-associated lipocalin (NGAL) as a marker of acute kidney injury in critically ill children with septic shock. Crit Care Med 2008;36: Zappitelli M, Washburn KK, Arikan AA, Loftis L, Ma Q, Devarajan P, et al. Urine neutrophil gelatinase-associated lipocalin is an early marker of acute kidney injury in critically ill children: a prospective cohort study. Crit Care 2007;11:R Cruz DN, de Cal M, Garzotto F, Perazella MA, Lentini P, Corradi V, et al. Plasma neutrophil gelatinase-associated lipocalin is an early biomarker for acute kidney injury in an adult ICU population. Intensive Care Med 2010;36: Leithead JA, Ferguson JW, Bates CM, Davidson JS, Lee A, Bathgate AJ, et al. The systemic inflammatory response syndrome is predictive of renal dysfunction in patients with non-paracetamol-induced acute liver failure. Gut 2009;58:

10 1266 PORTAL ET AL. LIVER TRANSPLANTATION, November Bagshaw SM, Bennett M, Haase M, Haase-Fielitz A, Egi M, Morimatsu H, et al. Plasma and urine neutrophil gelatinase-associated lipocalin in septic versus non-septic acute kidney injury in critical illness. Intensive Care Med 2010;36: Fjaertoft G, Foucard T, Xu S, Venge P. Human neutrophil lipocalin (HNL) as a diagnostic tool in children with acute infections: a study of the kinetics. Acta Paediatr 2005; 94: Flo TH, Smith KD, Sato S, Rodriguez DJ, Holmes MA, Strong RK, et al. Lipocalin 2 mediates an innate immune response to bacterial infection by sequestrating iron. Nature 2004;432: Shapiro NI, Trzeciak S, Hollander JE, Birkhahn R, Otero R, Osborn TM, et al. A prospective, multicenter derivation of a biomarker panel to assess risk of organ dysfunction, shock, and death in emergency department patients with suspected sepsis. Crit Care Med 2009;37: Perianayagam MC, Seabra VF, Tighiouart H, Liangos O, Jaber BL. Serum cystatin C for prediction of dialysis requirement or death in acute kidney injury: a comparative study. Am J Kidney Dis 2009;54: