- Email: [email protected]
Urinary Neutrophil Gelatinase-Associated Lipocalin and Acute Kidney Injury After Cardiac Surgery
Urinary Neutrophil Gelatinase-Associated Lipocalin and Acute Kidney Injury After Cardiac Surgery Gebhard Wagener, MD,1 Gina Gubitosa, BA,1 Shuang Wang, PhD,2 Niels Borregaard, PhD,3 Mihwa Kim, BS,1 and H. Thomas Lee, MD, PhD1 Background: Neutrophil gelatinase-associated lipocalin (NGAL) is proposed as an early marker of kidney injury. We report the association of urinary NGAL with indexes of intraoperative renal hypoperfusion (cardiopulmonary bypass time and aortic cross-clamp time) and acute kidney injury (AKI) after adult cardiac surgery. Study Design: Diagnostic test accuracy. Setting & Participants: Adult cardiac surgical patients (n ⫽ 426) in a single center from 2004 to 2006. Index Test: Urinary NGAL immediately and 3, 18, and 24 hours after cardiac surgery, using an enzyme-linked immunosorbent assay. Reference Test or Outcome: Serum creatinine– based definition for AKI (increase in serum creatinine from preoperative values by ⬎50% or ⬎0.3 mg/dL within 48 hours). Results: Mean urinary NGAL level was 165 ⫾ 663 (SD) ng/mL preoperatively, peaked immediately after cardiac surgery at 1,490 ⴞ 102 ng/mL, and remained significantly higher 3, 18, and 24 hours after surgery. 85 patients (20%) developed AKI. Areas under the receiver operating characteristic curve for urinary NGAL immediately after and 3, 18, and 24 hours later as a predictor for AKI were 0.573 (95% confidence interval [CI], 0.506 to 0.640), 0.603 (95% CI, 0.533 to 0.674), 0.611 (95% CI, 0.544 to 0.679), and 0.584 (95% CI, 0.510 to 0.657), respectively. Urinary NGAL, but not serum creatinine, level correlated significantly with cardiopulmonary bypass and aortic cross-clamp times. Areas under receiver operating characteristic curves for cardiopulmonary bypass time and aortic cross-clamp time to predict AKI were 0.592 (95% CI, 0.518 to 0.666) and 0.593 (95% CI, 0.523 to 0.665), respectively. Limitations: Limited sensitivity of changes in serum creatinine levels for kidney injury. Conclusions: Urinary NGAL has limited diagnostic accuracy to predict AKI defined by change in serum creatinine after cardiac surgery. Am J Kidney Dis 52:425-433. © 2008 by the National Kidney Foundation, Inc. INDEX WORDS: Cardiac surgery; renal failure; biomarkers; neutrophil gelatinase-associated lipocalin (NGAL).
Editorial, p. 395
R
enal failure is a major cause of morbidity and mortality after cardiac surgery.1,2 Acute kidney injury (AKI), defined as a postoperative increase in serum creatinine level by more than 50%, occurs in about 20% of patients3 and is associated with a mortality rate of 8% compared with 0.9% in patients without AKI. AKI requiring hemodialysis in the postoperative period is rare (⬃1% to 5%), but associated with a very high mortality rate of 30% to 60%.4 Although cardiac surgery (with or without cardiopulmonary bypass [CPB]) causes renal injury, this injury often remains undetected until several days after surgery because the currently used marker of renal injury (serum creatinine level) is extremely insensitive. There has been no significant progress with regard to renoprotective strategies in nearly 50 years, mainly because there are no biomarkers that allow early, sensitive, and specific identification of renal injury.5,6 Serum creatinine level
increases too late to identify patients at risk and allow clinical application of early renoprotective interventions that may have been found to be beneficial in preclinical studies. By the time renal dysfunction is detected by using serum creatinine level, renal injury is nearly completed,7 and interventions may not be possible to correct the renal dysfunction. Serum creatinine level also is too insensitive to be useful as an end point in clinical studies evaluating nephrotoxic From the 1Department of Anesthesiology, College of Physicians and Surgeons, and 2Department of Biostatistics, Mailman School of Public Heath, Columbia University, New York, NY; and 3Department of Hematology, Rigshospitalet, Copenhagen, Denmark. Received January 14, 2008. Accepted in revised form May 12, 2008. Originally published online as doi: 10.1053/j.ajkd.2008.05.018 on July 25, 2008. Address correspondence to Gebhard Wagener, MD, Assistant Professor, Department of Anesthesiology, Columbia University, P&S Box 46 (PH-5), 630 West 168th St, New York, NY 10032-3784. E-mail: [email protected] © 2008 by the National Kidney Foundation, Inc. 0272-6386/08/5203-0008$34.00/0 doi:10.1053/j.ajkd.2008.05.018
American Journal of Kidney Diseases, Vol 52, No 3 (September), 2008: pp 425-433
425
426
Wagener et al
and ischemic insults and to evaluate possible protective strategies. Additionally, there is no uniform definition of AKI that allows a comparison of different trials.8 Urinary neutrophil gelatinase-associated lipocalin (NGAL) is a novel biomarker for which levels are increased within hours after a nephrotoxic or ischemic insult.9 The increase in urinary NGAL levels correlates with the duration and severity of renal injury, and urinary NGAL is easily detectable using either quantitative enzyme-linked immunosorbent assay (ELISA) or semiquantitative immunoblot techniques. Mishra et al10 showed that urinary NGAL level 2 hours after pediatric cardiac surgery had 100% sensitivity and 98% specificity to predict AKI. We previously measured urinary NGAL in a small adult cardiac surgical population (n ⫽ 81) by using immunoblotting and found that urinary NGAL level was also predictive for AKI in adult patients,11 but with greater variability compared with the result by Mishra et al.10 This study therefore aims to evaluate urinary NGAL levels after cardiac surgery in a large cardiac surgical patient population using ELISA. Our goal is to evaluate whether: (1) urinary NGAL level reflects the extent of renal injury sustained during cardiac surgery through prolonged CPB time and aortic cross-clamp time (AXT) in comparison to serum creatinine level and (2) increased urinary NGAL level is associated with AKI defined by increases in serum creatinine level.
METHODS Patients All adult patients undergoing cardiac surgery at ColumbiaUniversity Medical Center, New York, NY, were eligible for inclusion in this study. The Institutional Review Board of Columbia University waived the requirement to obtain informed consent because urine collection is considered minimal risk by New York State and US Federal regulations. From July 26, 2004, to January 11, 2006, a total of 426 patients were enrolled. Approximately 2 to 3 patients per day were arbitrarily selected in the morning from the daily operating room schedule by the research assistant (G.G.) without a specific randomization scheme. There was no significant difference in the frequency of surgical procedures comparing the study population with the approximately 2,000 cardiac surgeries performed at Columbia University Medical Center during this time. Of these, 200 patients with preoperative end-stage renal failure requiring hemodialysis therapy were excluded from the study.
Sample Collection and Processing Five milliliters of urine was collected after induction of anesthesia and insertion of a urinary catheter before surgical incision. Subsequently, 5 mL of urine was collected from the urinary catheter immediately after CPB (or after completion of the last graft in case of off-pump coronary artery bypass grafting), then 3, 18, and 24 hours later. Urine was centrifuged at 1,000g for 15 minutes, and the supernatant was frozen at ⫺20°C. Urinary NGAL levels were determined by using ELISA. No protease inhibitors were added to the samples because NGAL is resistant to proteases.12,13
Enzyme-Linked Immunosorbent Assay NGAL level was determined by using a commercially available ELISA kit (Antibodyshop, Gentofte, Denmark) according to the manufacturer’s instructions, as described previously.14 The limit of detection for this assay is 0.5 to 4.0 pg/mL, and intra-assay variation in urine is 2.1% (range, 1.3% to 4.0%).15,16
Data Collection Basic demographic data were collected and Parsonnet scores were calculated for risk stratification in all patients.17 We determined the number of intensive care unit (ICU)- and hospital-free days, defined as the number of days a patient was not in the ICU or the hospital within the first 30 days after surgery. We also determined the number of patient deaths (patients with no ICU- and hospital-free days). Serum creatinine was measured by the central laboratory of Columbia University Medical Center. Creatinine clearance was estimated using the method described by Cockcroft and Gault.18
Definition of AKI AKI was defined as an increase in serum creatinine level by either more than 50% or more than 0.3 mg/dL (compared with preoperative values) during the first 48 hours after surgery (serum creatinine in mg/dL may be converted to mol/L by multiplying by 88.4). This definition of AKI has been suggested by the Acute Kidney Injury Network (AKIN)19 and reflects recent studies that suggest that even small changes in serum creatinine levels increase morbidity and mortality after cardiac surgery.20 However, patients who were oliguric, but did not have a change in serum creatinine level, were not included as having AKI because urine output after surgery is often increased with the use of diuretics. Serum creatinine level was determined at least twice daily during the first 2 postoperative days, then daily for the duration of the hospital stay by the laboratory of Columbia University Medical Center as part of routine patient care.
Statistics Values are presented as mean ⫾ SEM. Comparisons between groups were made using unpaired t-test for values with Gaussian distribution and Mann-Whitney (Wilcoxon rank) test or Spearman test for correlation for continuous variables without normal distribution. Gaussian distribution was determined by using the Kolmogorov-Smirnov test. P values are 2 tailed, and P less than 0.05 is considered significant.
Urinary NGAL in Adult Cardiac Surgery
427
For the purpose of a pretrial power analysis, we assumed that urinary NGAL level with AKI would be 50% greater than without AKI (mean NGALnoAKI ⫽ 1,000 ng/mL, mean NGALAKI ⫽ 1,500 ng/mL, common SD ⫽ 1,000 ng/mL). We concluded that we would require 63 subjects per group to achieve 80% power with ␣ ⫽ 0.05. Assuming an incidence of 20% for AKI, we therefore would have to include 378 patients in this study. SAS, version 9.1 (SAS Inc, Cary, NC); SPSS, version 11.0.4 (SPSS Inc, Chicago, IL); and Graphpad Prism, version 4.0 (San Diego, CA) software was used for the statistical analysis.
RESULTS
by more than 50% or more than 0.3 mg/dL within 48 hours after surgery. Patients who developed postoperative AKI (either definition) were older; had higher Parsonnet scores, greater body mass indexes, and longer CPB times and AXTs; and received aprotinin more frequently as an antifibrinolytic (Table 1). They also remained longer in the ICU and hospital and had greater in-hospital mortality. Serum Creatinine After Cardiac Surgery
Acute Kidney Injury Eighty-five patients (20.0%) developed AKI, defined as an increase in serum creatinine level
Serum creatinine levels decreased immediately after surgery, even if patients subsequently developed AKI. In patients who developed AKI,
Table 1. Patient Characteristics
Preoperative Women Age (y) Body mass index (kg/m2) Serum creatinine (mg/dL) Parsonnet score Operation CABG with CPB CABG without CPB Single valve Multiple valve CABG ⫹ valve Cardiac transplant Ventricular assist device Other Reoperation Intraoperative CPB time (min) AXT (min) Aprotinin use Postoperative ICU-free days within 30 d Hospital-free days within 30 d Peak serum creatinine (mg/dL) ⌬ Serum creatinine (mg/dL) Continuous hemodialysis Mortality Urinary NGAL Preoperative (ng/dL) Immediately postoperative (ng/dL) 3 Hours (ng/dL) 18 Hours (ng/dL) 24 Hours (ng/dL)
Total (N ⫽ 426)
AKI (n ⫽ 85; 20.0%)
No AKI (n ⫽ 341; 80.0%)
P
144 (34) 63 ⫾ 15 28 ⫾ 16 1.08 ⫾ 0.45 9.2 ⫾ 7.5
25 (29.4) 69.5 ⫾ 13.1 29.5 ⫾ 7.8 1.21 (⫾ 0.47) 11.9 ⫾ 8.5
119 (34.9) 61.7 ⫾ 14.9 27.8 ⫾ 17.3 1.05 (⫾ 0.45) 8.5 ⫾ 7.1
0.4 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001
69 (16.2) 56 (13.1) 136 (31.9) 21 (4.9) 74 (17.4) 10 (2.3) 13 (3.1) 47 (11.0) 58 (13.6)
12 (14.1) 11 (12.9) 25 (29.4) 4 (4.7) 20 (23.5) 2 (2.4) 5 (5.9) 6 (7.1) 13 (15.3)
57 (16.7) 45 (13.2) 111 (32.6) 17 (5.0) 54 (15.8) 8 (2.3) 8 (2.3) 41 (12.0) 45 (13.2)
0.7 0.9 0.6 0.9 0.1 0.9 0.1 0.2 0.6
123 ⫾ 48 86 ⫾ 35 208 (49)
141 ⫾ 60 97 ⫾ 38 52 (61.2)
119 ⫾ 43 83 ⫾ 33 156 (45.8)
0.02 0.02 0.02
26.9 ⫾ 5.3 20.6 ⫾ 7.6 1.33 ⫾ 0.75 0.24 ⫾ 0.55 8 (1.9) 16 (3.8)
24.8 ⫾ 7.6 17.1 ⫾ 9.8 2.03 ⫾ 1.02 0.82 ⫾ 0.83 3 (3.5) 8 (9.4)
27.5 ⫾ 4.4 21.5 ⫾ 6.7 1.15 ⫾ 0.55 0.10 ⫾ 0.32 5 (1.5) 16 (1.8)
165 ⫾ 663 1,490 ⫾ 2,103 1,139 ⫾ 2,038 381 ⫾ 921 494 ⫾ 1,124
156 ⫾ 527 1,786 ⫾ 2,129 1,599 ⫾ 2,356 473 ⫾ 987 437 ⫾ 712
167 ⫾ 693 1,417 ⫾ 2,093 1,031 ⫾ 1,944 356 ⫾ 903 509 ⫾ 1,210
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.001 0.002 0.1 0.03 0.006 0.002 0.03
Note: Values expressed as mean ⫾ SD or number (percent). Serum creatinine in mg/dL may be converted to mol/L by multiplying by 88.4. Abbreviations: CABG, coronary artery bypass grafting; CPB, cardiopulmonary bypass; AXT, aortic cross-clamp time; NGAL, neutrophil gelatinase-associated lipocalin; ICU, intensive care unit.
428
Wagener et al
serum creatinine levels increased on the first postoperative day and peaked on postoperative day 2. Serum creatinine levels were statistically different between patients with and without AKI from postoperative days 0 through 10 (Fig 1). Urinary NGAL Before and After Cardiac Surgery in All Patients Studied Figure 2 shows preoperative and postoperative urinary NGAL levels in all patients studied. Before surgery, mean urinary NGAL level in all patients studied was 165 ⫾ 32 ng/mL. Preoperative urinary NGAL level was not different in patients with preoperative renal dysfunction (estimated creatinine clearance ⬍ 60 mL/min; Cockcroft-Gault formula18) compared with patients with normal preoperative renal function (178 ⫾ 756 versus 136 ⫾ 340 ng/mL; P ⫽ 0.6, respectively). Urinary NGAL level peaked immediately after cardiac surgery by more than 9-fold (1,490 ⫾ 102 ng/mL) and remained significantly greater 3, 18, and 24 hours after surgery compared with preoperative values (Fig 2). Urinary NGAL and AKI Preoperative urinary NGAL levels were similar in patients who later developed AKI com-
Figure 1. Percentage of change in serum creatinine levels after cardiac surgery with and without acute kidney injury (AKI; mean ⫾ SEM). *P ⬍ 0.05 compared with preoperative values. Serum creatinine levels decreased immediately after surgery in all patients and but increased to significantly greater levels in patients with AKI (defined as an increase in serum creatinine ⬎50% or ⬎0.3 mg/dL within 48 hours) and peaked on postoperative (postOP) day 2. Serum creatinine in mg/dL may be converted to mol/L by multiplying by 88.4.
Figure 2. Urinary neutrophil gelatinase-associated lipocalin (NGAL) levels before (PreOP) and after (PostOP) cardiac surgery in all patients (mean ⫾ SEM). *P ⬍ 0.05 compared with preoperative values. Urinary NGAL levels increased immediately after cardiac surgery and remained increased for 24 hours.
pared with patients without postoperative AKI (156 ⫾ 57 versus 167 ⫾ 38 ng/mL; P ⫽ 0.1]). Immediately after surgery, patients with postoperative AKI had significantly greater urinary NGAL levels (1,786 ⫾ 232 versus 1,417 ⫾ 114 ng/mL; P ⫽ 0.04). Urinary NGAL levels remained different throughout the 24-hour postoperative period between groups. The largest and most significant difference between patients with and without AKI was 3 hours after surgery (1,599 ⫾ 276 versus 1,031 ⫾ 110 ng/mL; P ⫽ 0.006). At 24 hours, patients who developed AKI had slightly lower urinary NGAL levels than patients without AKI (437 ⫾ 85 versus 509 ⫾ 75 ng/mL; P ⫽ 0.03; Fig 3).
Figure 3. Urinary neutrophil gelatinase-associated lipocalin (NGAL) levels before (PreOP) and after (PostOP) cardiac surgery (hours) in patients with or without acute kidney injury (AKI; mean ⫾ SEM). *P ⬍ 0.05 comparing patients with and without AKI.
Urinary NGAL in Adult Cardiac Surgery
Receiver Operating Characteristic Curves Receiver operating characteristic (ROC) curves were generated for all times to evaluate the ability of urinary NGAL to predict AKI. Areas under the curves (AUCs) were low at any time. AUC was highest 18 hours after surgery (AUC ⫽ 0.611; 95% CI, 0.544 to 0.679; P ⬍ 0.001). The cutoff value at this time (defined as the point on the ROC curve closest to sensitivity ⫽ 1 ⫺ specificity ⫽ 1) was 65 ng/mL (Fig 4; Table 2). AUCs of the ROC curve for CPB time and AXT to predict AKI were 0.592 (95% CI, 0.518 to 0.666; P ⫽ 0.01) and 0.593 (95% CI, 0.523 to 0.665; P ⫽ 0.01), respectively (Fig 4). Urinary NGAL and CPB Time/AXT Increased duration of aortic cross-clamping or CPB increases the severity of intraoperative renal injury,21,22 but may not necessarily be detected by an increase in postoperative serum creatinine levels. To assess whether we are able to detect this injury by using urinary NGAL level, we calculated the correlation coefficients between CPB time and AXT with urinary NGAL and postoperative creatinine (peak and change ⌬). Urinary NGAL levels immediately after CPB (P ⬍ 0.001 for CPB time and AXT) and at 3 (P ⬍ 0.001 for CPB time and AXT), 18 (P ⫽ 0.02 for CPB time and AXT), and 24 hours (P ⫽ 0.02 for CPB time and P ⫽ 0.03 for AXT) correlated significantly with CPB time/AXT. However, CPB time and AXT correlated better with urinary NGAL levels measured immediately after CPB and at 3 hours than with later levels of urinary NGAL (18 and 24 hours). Peak and ⌬ serum creatinine levels
Sensitivity
A
1 - Specificity
B
Sensitivity
Urinary NGAL and Urinary NGAL/Urine Creatinine Ratio We measured urine creatinine in 59 patients to evaluate the effect of postoperative hemodilution on urinary NGAL levels. Urinary creatinine levels decreased immediately after surgery and 3 hours later, then increased 18 and 24 hours after surgery. Urinary NGAL levels correlated well with urinary NGAL/urine creatinine ratio (Pearson r ⫽ 0.842; 95% confidence interval [CI], 0.795 to 0.879; P ⬍ 0.001), and change (⌬) in urinary NGAL/urine creatinine ratio in the postoperative period was very similar to the course of urinary NGAL.
429
1 - Specificity Figure 4. Receiver operating characteristic curves for urinary neutrophil gelatinase-associated lipocalin (NGAL) and cardiopulmonary bypass (CPB) time and aortic crossclamp time (AXT) as a predictor of acute kidney injury (AKI). (A) Urinary NGAL immediately after surgery (area under the curve [AUC] ⫽ 0.573; 95% confidence interval [CI], 0.506 to 0.640; P ⫽ 0.04) and 3 (AUC ⫽ 0.603; 95% CI, 0.533 to 0.674; P ⫽ 0.006), 18 (AUC ⫽ 0.611; 95% CI, 0.544 to 0.679; P ⫽ 0.003), and 24 hours after surgery (AUC ⫽ 0.584; 95% CI, 0.510 to 0.657; P ⫽ 0.03). (B) CPB time (AUC ⫽ 0.592; 95% CI, 0.518 to 0.666; P ⫽ 0.01) and AXT (AUC ⫽ 0.593; 95% CI, 0.522 to 0.665; P ⫽ 0.01) as a predictor for AKI.
430
Wagener et al Table 2. AUCs for Receiver Operating Characteristics of NGAL
NGAL postoperative NGAL 3 h postoperative NGAL 18 h postoperative NGAL 24 h postoperative
AUC
P
Cutoff (ng/mL)
Sensitivity (%)
Specificity (%)
Distance
0.573 (0.506-0.640) 0.603 (0.533-0.674) 0.611 (0.544-0.679) 0.584 (0.510-0.657)
0.04 0.006 0.003 0.03
⬎23.5 ⬎18.1 ⬎15.6 ⬎21.2
31.0 (26.1-36.2) 37.7 (32.3-43.4) 39.2 (33.6-45.1) 39.4 (33.5-45.6)
81.0 (70.9-88.7) 78.1 (66.9-86.9) 78.2 (67.4-86.8) 78.6 (67.1-87.4)
71.6 66.0 64.5 64.3
Note: Values in parentheses indicate 95% confidence intervals. Acute renal dysfunction is defined as an increase in serum creatinine level by more than 50% or 0.3 mg/dL (26.52 mol/L) within 48 hours (Acute Kidney Injury Network criteria). Abbreviations: AUC, area under the curve; NGAL, neutrophil gelatinase-associated lipocalin.
did not correlate with CPB time, and ⌬ serum creatinine level correlated significantly, but less robustly (P ⫽ 0.02), with AXT, whereas peak serum creatinine level did not (P ⫽ 0.06; Table 3). Additionally, we divided patients into groups with and without prolonged CPB time (⬎120 minutes; n ⫽ 176) and prolonged AXT (⬎90 minutes; n ⫽ 138). Prolonged CPB time (⬎120 minutes) was associated with significantly greater urinary NGAL levels immediately after surgery (mean difference between groups, 866 ng/mL (95% CI, 427 to 1,305; P ⬍ 0.001); at 3 hours, 1,002 ng/mL (95% CI, 545 to 1,459; P ⬍ 0.001); at 18 hours, 224 ng/mL (95% CI, 9 to 440; P ⫽ 0.04); and at 24 hours, 285 ng/mL (95% CI, 13 to 557; P ⫽ 0.04), but there was no difference with regard to ⌬ or peak serum creatinine level (P ⫽ 0.5 and P ⫽ 0.4, respectively). Prolonged AXT (⬎90 minutes) also was associated with increased urinary NGAL level immediately after surgery (mean difference, 1,012 ng/mL; 95% CI, 531 to 1,492; P ⬍ 0.001) and 3 (mean difference, 1,053 ng/mL; 95% CI, 533 to 1,572; P ⬍ 0.001) and 24 hours later (mean difference, 152 ng/mL; 95% CI, 58 to 657; P ⫽ 0.02), but not with peak or ⌬ serum creatinine and urinary NGAL level 18 hours after surgery (P ⫽ 0.1 and P ⫽ 0.3, respectively).
DISCUSSION We were able to show with this study that urinary NGAL levels correlate better than serum creatinine levels with indexes of intraoperative renal hypoperfusion (CPB time and AXT). Furthermore, patients who had long CPB times and AXTs showed increased NGAL levels, but had no significant change in serum creatinine levels. Urinary NGAL level immediately after surgery was significantly greater in patients with AKI, defined
by using increases in serum creatinine levels compared with patients who did not develop AKI. Urinary NGAL is a 25-kDa protein originally described in neutrophils.23,24 In 2003, Mishra et al25 used microassay techniques to show that NGAL is also expressed in proximal tubular cells of the kidney and increases substantially after ischemic injury in mice. Subsequent, but small, clinical studies showed that urinary NGAL is also detectable in high concentrations in humans after ischemic11,26 and other nephrotoxic27,28 injuries. Renal injury is frequent after cardiac surgery because of renal hypoperfusion, reperfusion injury, and inflammatory responses. AKI with or without the need for hemodialysis is associated with significant increases in morbidity and mortality after cardiac surgery.1 Unfortunately, the currently used marker of renal function (or dysfunction), serum creatinine level, is unacceptably insensitive. Subsequently, by the time renal dysfunction is indicated by an increase in serum creatinine levels, satisfactory clinical intervention is not possible. What is worse, even small increases in serum creatinine levels (⬃0.3 mg/ dL) have been associated with drastic increases in mortality and morbidity after cardiac surgery.20 Therefore, it is clear that serum creatinine level is an unacceptably insensitive and late marker of renal dysfunction. In addition, serum creatinine level is affected by the dilutional effect of increased total-body water in the early postoperative period after cardiac surgery. These factors greatly limit our ability to intervene during the early stage of renal injury to potentially prevent its progression to established acute renal failure, with its even greater morbidity and mortality. The purpose of this study is to test the hypothesis that urinary NGAL level is better suited to
0.173 0.054 to 0.288 0.03
431
Abbreviations: CPB, cardiopulmonary bypass; AXT, aortic cross-clamp time; NGAL, neutrophil gelatinase-associated lipocalin.
0.123 0.008 to 0.234 0.01 0.233 0.124 to 0.337 ⬍0.0001 0.125 0.019 to 0.229 0.02 0.101 ⫺0.006 to 0.205 0.06
0.240 0.137 to 0.339 ⬍0.0001
0.143 0.028 to 0.255 0.02 0.138 0.022 to 0.240 0.01 0.224 0.118 to 0.326 ⬍0.0001 0.227 0.127 to 0.323 ⬍0.0001 0.08415 ⫺0.0193 to 0.069 0.1 0.0314 ⫺0.072 to 0.134 0.5
Correlation with CPB time Pearson r 95% Confidence interval P (2 tailed) Correlation with AXT Pearson r 95% Confidence interval P (2 tailed)
18 h After Surgery 3 h After Surgery Peak Creatinine
Change in Creatinine
Postoperative
Urinary NGAL Serum Creatinine
Table 3. Correlation Between CBP Time or AXT and Postoperative Creatinine and Urinary NGAL at Different Times
24 h After Surgery
Urinary NGAL in Adult Cardiac Surgery
detect intraoperative renal injury after cardiac surgery than serum creatinine level. In addition, we want to confirm our previous findings that used less precise semiquantitative immunoblotting to measure NGAL in a small number of patients (n ⫽ 81) with a more accurate and quantitative ELISA in a large adult patient population undergoing a large variety of cardiac surgeries. We previously studied urinary NGAL (measured by means of immunoblotting technique) after cardiac surgery in 81 patients and found that urinary NGAL level as early as immediately after surgery was significantly greater in patients with postoperative AKI compared with those without AKI.11 We then found that use of aprotinin during cardiac surgery was associated with significantly greater postoperative NGAL levels (and a greater incidence of AKI) independent of possible confounders than the use of -amino caproic acid.14 With the present study, we confirmed our earlier findings that AKI is associated with increased urinary NGAL levels by using a more quantitative ELISA. However, we also observed much greater variability in urinary NGAL levels in adult patients than described by Mishra et al10 in pediatric patients. Serum creatinine levels decreased immediately after surgery regardless of subsequent AKI because of perioperative hemodilution and, in patients who developed AKI, peaked 2 days after surgery. Therefore, intraoperative renal ischemia reflected by an increase in urinary NGAL levels may not necessarily cause serum creatinine levels to increase several days later. It is well established that CPB time and AXT correlate with the development of AKI.21 We found a highly significant correlation between postoperative urinary NGAL levels and AXT and CPB time (Table 3). Conversely, peak serum creatinine level did not correlate with AXT and CPB time, and ⌬ serum creatinine correlated only weakly with AXT (P ⫽ 0.02). Additionally, long CPB times and long AXTs were associated with high urinary NGAL levels, but not increased serum creatinine levels. We therefore propose that urinary NGAL level is better suited to detect intraoperative renal injury than serum creatinine level: urinary NGAL level reflects the acute injury sustained through long CBP times and AXTs, whereas changes in serum creatinine lev-
432
els that occur days later reflect not only intraoperative, but also postoperative, renal ischemia. There is no uniform definition of AKI.29 We used the definition suggested by AKIN19 for stage 1 AKI. This definition is more inclusive than most other definitions, reflecting recent evidence that even small changes in serum creatinine levels are associated with increased morbidity and mortality in critically ill patients.30,31 For example, Lassnigg et al20 found that even a small change in serum creatinine level was predictive of increased 30-day mortality after cardiac surgery. However, the AKIN definition also has limitations in the immediate postoperative period because it restricts the period of observation to 48 hours and therefore may not be able to identify kidney injury and increases in serum creatinine level that become detectable only after the early period of postoperative hemodilution.32 The immediate postoperative period after cardiac surgery also does not necessarily constitute an optimal state of hydration required to apply the AKIN criteria.19 Additionally, patients frequently receive diuretics in this period. We therefore did not include the oliguria component of the AKIN definition of AKI (decrease in urine output to ⬍0.5 mL/h for 6 hours). Urinary NGAL level increased immediately after surgery and remained increased during the 3-hour postoperative period in patients with AKI compared with patients without AKI. At 18 and 24 hours after surgery, urinary NGAL levels were similar in both groups. The ability of urinary NGAL level to predict AKI was far less compelling (indicated by the small AUCs of the ROC curves) than previously described: the highest AUC of the ROC curve was 0.611 compared with the AUC of 0.998 observed by Mishra et al10 in the pediatric cardiac patient population. Unlike the pediatric cardiac surgery population,10 many patients in our study had such significant preoperative comorbidities as hypertension or diabetes and underwent highcomplexity surgeries, evidenced by a mean Parsonnet score of 9.2 ⫾ 7.5 and overall mortality rate of 3.8%. Almost all our patients had an increase in urinary NGAL levels after cardiac surgery, and only the degree of change differed in patients with and without AKI compared with the pediatric population described by Mishra et al.10 Serum creatinine levels peaked days after the surgery in patients who developed AKI and thus
Wagener et al
reflected not only intraoperative injury, but also hemodynamic instability, high vasopressor requirements, or other nephrotoxic or ischemic insults in the postoperative period.33 Serum creatinine levels decreased immediately after surgery regardless of subsequent AKI because of perioperative hemodilution. Therefore, a minor intraoperative renal injury reflected by an increase in urinary NGAL levels may not necessarily cause serum creatinine levels to increase 48 hours later in this very heterogeneous patient population. Exposure of blood to CPB circuits induces activation of neutrophils, which could lead to a release of significant amounts of intracellular NGAL that is then filtrated in urine. The urinary NGAL we measured therefore may not (only) represent NGAL released from renal tubuli. However, plasma NGAL levels are approximately 3 to 10 times lower than urinary levels after cardiac surgery.10,14 It therefore is very unlikely that urinary NGAL consists of any significant amount of plasma NGAL released from activated neutrophils. This study is limited by its end point, serum creatinine level, which may not be suited to detect renal injury as accurately and as sensitively as urinary NGAL level. This limitation most likely led to lower AUCs from the ROCs generated. Future studies will have to evaluate whether urinary NGAL level correlates with levels of other markers of renal injury, such as kidney injury molecule 1, serum cystatin C, or N-acetyl--D-glucosaminidase.34,35 In conclusion, we believe that urinary NGAL level is an earlier and a more sensitive marker of renal injury compared with serum creatinine level.26 Urinary NGAL level may allow earlier and more successful renal interventions and thus serve as a useful marker of renal injury in clinical studies investigating renoprotective strategies and allow early identification and treatment of patients at risk.
ACKNOWLEDGEMENTS Support: This work was funded by the intramural grant support from the Department of Anesthesiology, Columbia University College of Physicians and Surgeons. We would like to thank Antibodyshop, Entofte, Denmark, for supplying the NGAL ELISA kits. Financial Disclosure: None.
Urinary NGAL in Adult Cardiac Surgery
REFERENCES 1. Conlon PJ, Stafford-Smith M, White WD, et al: Acute renal failure following cardiac surgery. Nephrol Dial Transplant 14:1158-1162, 1999 2. Corwin HL, Sprague SM, DeLaria GA, et al: Acute renal failure associated with cardiac operations. A casecontrol study. J Thorac Cardiovasc Surg 98:1107-1112, 1989 3. Kuitunen A, Vento A, Suojaranta-Ylinen R, et al: Acute renal failure after cardiac surgery: Evaluation of the RIFLE classification. Ann Thorac Surg 81:542-546, 2006 4. Chertow GM, Lazarus JM, Christiansen CL, et al: Preoperative renal risk stratification. Circulation 95:878884, 1997 5. Conger JD: Interventions in clinical acute renal failure: What are the data? Am J Kidney Dis 26:565-576, 1995 6. Star RA: Treatment of acute renal failure. Kidney Int 54:1817-1831, 1998 7. Bagshaw SM, Bellomo R: Early diagnosis of acute kidney injury. Curr Opin Crit Care 13:638-644, 2007 8. O’Leary MJ, Bihari DJ: Preventing renal failure in the critically ill. There are no magic bullets—Just high quality intensive care. BMJ 322:1437-1439, 2001 9. Nguyen MT, Devarajan P: Biomarkers for the early detection of acute kidney injury. Pediatr Nephrol 2007 Mar 30 [Epub ahead of print] 10. Mishra J, Dent C, Tarabishi R, et al: Neutrophil gelatinase-associated lipocalin (NGAL) as a biomarker for acute renal injury after cardiac surgery. Lancet 365:12311238, 2005 11. Wagener G, Jan M, Kim M, et al: Association between increases in urinary neutrophil gelatinase-associated lipocalin and acute renal dysfunction after adult cardiac surgery. Anesthesiology 105:485-491, 2006 12. Kjeldsen L, Johnsen AH, Sengelov H, et al: Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase. J Biol Chem 268:1042510432, 1993 13. Peitsch MC, Boguski MS: The first lipocalin with enzymatic activity. Trends Biochem Sci 16:363, 1991 14. Wagener G, Gubitosa G, Wang S, et al: Increased incidence of acute kidney injury with aprotinin use during cardiac surgery detected with urinary NGAL. Am J Nephrol 28:576-582, 2008 15. Antibodyshop: NGAL ELISA Kit (KIT 036) product inlay. Available at: http://www.antibodyshop.com/content/ download/1130/34986/version/22/file/Inlay⫹NGAL⫹ELISA⫹ Kit⫹%28KIT⫹036%29.pdf. Accessed April 1, 2008 16. Kjeldsen L, Koch C, Arnljots K, et al: Characterization of two ELISAs for NGAL, a newly described lipocalin in human neutrophils. J Immunol Methods 198:155-164, 1996 17. Parsonnet V, Dean D, Bernstein AD: A method of uniform stratification of risk for evaluating the results of surgery in acquired adult heart disease. Circulation 79:I3I12, 1989 18. Cockcroft DW, Gault MH: Prediction of creatinine clearance from serum creatinine. Nephron 16:31-41, 1976
433 19. Mehta RL, Kellum JA, Shah SV, et al: Acute Kidney Injury Network: Report of an initiative to improve outcomes in acute kidney injury. Crit Care 11:R31, 2007 20. Lassnigg A, Schmidlin D, Mouhieddine M, et al: Minimal changes of serum creatinine predict prognosis in patients after cardiothoracic surgery: A prospective cohort study. J Am Soc Nephrol 15:1597-1605, 2004 21. Fischer UM, Weissenberger WK, Warters RD, et al: Impact of cardiopulmonary bypass management on postcardiac surgery renal function. Perfusion 17:401-406, 2002 22. Solomon R, Werner C, Mann D, et al: Effects of saline, mannitol, and furosemide to prevent acute decreases in renal function induced by radiocontrast agents. N Engl J Med 331:1416-1420, 1994 23. Bundgaard JR, Sengelov H, Borregaard N, et al: Molecular cloning and expression of a cDNA encoding NGAL: A lipocalin expressed in human neutrophils. Biochem Biophys Res Commun 202:1468-1475, 1994 24. Kjeldsen L, Johnsen AH, Sengelov H, et al: Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase. J Biol Chem 268:1042510432, 1993 25. Mishra J, Ma Q, Prada A, et al: Identification of neutrophil gelatinase-associated lipocalin as a novel early urinary biomarker for ischemic renal injury. J Am Soc Nephrol 14:2534-2543, 2003 26. Mori K, Nakao K: Neutrophil gelatinase-associated lipocalin as the real-time indicator of active kidney damage. Kidney Int 71:967-970, 2007 27. Mishra J, Mori K, Ma Q, et al: Neutrophil gelatinaseassociated lipocalin: A novel early urinary biomarker for cisplatin nephrotoxicity. Am J Nephrol 24:307-315, 2004 28. Bachorzewska-Gajewska H, Malyszko J, Sitniewska E, et al: NGAL (neutrophil gelatinase-associated lipocalin) and cystatin C: Are they good predictors of contrast nephropathy after percutaneous coronary interventions in patients with stable angina and normal serum creatinine? Int J Cardiol 127:290-291, 2008 29. Ronco C, Levin A, Warnock DG, et al: Improving outcomes from acute kidney injury (AKI): Report on an initiative. Int J Artif Organs 30:373-376, 2007 30. Levy MM, Macias WL, Vincent JL, et al: Early changes in organ function predict eventual survival in severe sepsis. Crit Care Med 33:2194-2201, 2005 31. Praught ML, Shlipak MG: Are small changes in serum creatinine an important risk factor? Curr Opin Nephrol Hypertens 14:265-270, 2005 32. Moran SM, Myers BD: Course of acute renal failure studied by a model of creatinine kinetics. Kidney Int 27:928937, 1985 33. Andersson LG, Ekroth R, Bratteby LE, et al: Acute renal failure after coronary surgery—A study of incidence and risk factors in 2009 consecutive patients. Thorac Cardiovasc Surg 41:237-241, 1993 34. Coca SG, Yalavarthy R, Concato J, et al: Biomarkers for the diagnosis and risk stratification of acute kidney injury: A systematic review. Kidney Int 73:1008-1016, 2008 35. Han WK, Bailly V, Abichandani R, et al: Kidney injury molecule-1 (KIM-1): A novel biomarker for human renal proximal tubule injury. Kidney Int 62:237-244, 2002
Editorial, p. 395
R
enal failure is a major cause of morbidity and mortality after cardiac surgery.1,2 Acute kidney injury (AKI), defined as a postoperative increase in serum creatinine level by more than 50%, occurs in about 20% of patients3 and is associated with a mortality rate of 8% compared with 0.9% in patients without AKI. AKI requiring hemodialysis in the postoperative period is rare (⬃1% to 5%), but associated with a very high mortality rate of 30% to 60%.4 Although cardiac surgery (with or without cardiopulmonary bypass [CPB]) causes renal injury, this injury often remains undetected until several days after surgery because the currently used marker of renal injury (serum creatinine level) is extremely insensitive. There has been no significant progress with regard to renoprotective strategies in nearly 50 years, mainly because there are no biomarkers that allow early, sensitive, and specific identification of renal injury.5,6 Serum creatinine level
increases too late to identify patients at risk and allow clinical application of early renoprotective interventions that may have been found to be beneficial in preclinical studies. By the time renal dysfunction is detected by using serum creatinine level, renal injury is nearly completed,7 and interventions may not be possible to correct the renal dysfunction. Serum creatinine level also is too insensitive to be useful as an end point in clinical studies evaluating nephrotoxic From the 1Department of Anesthesiology, College of Physicians and Surgeons, and 2Department of Biostatistics, Mailman School of Public Heath, Columbia University, New York, NY; and 3Department of Hematology, Rigshospitalet, Copenhagen, Denmark. Received January 14, 2008. Accepted in revised form May 12, 2008. Originally published online as doi: 10.1053/j.ajkd.2008.05.018 on July 25, 2008. Address correspondence to Gebhard Wagener, MD, Assistant Professor, Department of Anesthesiology, Columbia University, P&S Box 46 (PH-5), 630 West 168th St, New York, NY 10032-3784. E-mail: [email protected] © 2008 by the National Kidney Foundation, Inc. 0272-6386/08/5203-0008$34.00/0 doi:10.1053/j.ajkd.2008.05.018
American Journal of Kidney Diseases, Vol 52, No 3 (September), 2008: pp 425-433
425
426
Wagener et al
and ischemic insults and to evaluate possible protective strategies. Additionally, there is no uniform definition of AKI that allows a comparison of different trials.8 Urinary neutrophil gelatinase-associated lipocalin (NGAL) is a novel biomarker for which levels are increased within hours after a nephrotoxic or ischemic insult.9 The increase in urinary NGAL levels correlates with the duration and severity of renal injury, and urinary NGAL is easily detectable using either quantitative enzyme-linked immunosorbent assay (ELISA) or semiquantitative immunoblot techniques. Mishra et al10 showed that urinary NGAL level 2 hours after pediatric cardiac surgery had 100% sensitivity and 98% specificity to predict AKI. We previously measured urinary NGAL in a small adult cardiac surgical population (n ⫽ 81) by using immunoblotting and found that urinary NGAL level was also predictive for AKI in adult patients,11 but with greater variability compared with the result by Mishra et al.10 This study therefore aims to evaluate urinary NGAL levels after cardiac surgery in a large cardiac surgical patient population using ELISA. Our goal is to evaluate whether: (1) urinary NGAL level reflects the extent of renal injury sustained during cardiac surgery through prolonged CPB time and aortic cross-clamp time (AXT) in comparison to serum creatinine level and (2) increased urinary NGAL level is associated with AKI defined by increases in serum creatinine level.
METHODS Patients All adult patients undergoing cardiac surgery at ColumbiaUniversity Medical Center, New York, NY, were eligible for inclusion in this study. The Institutional Review Board of Columbia University waived the requirement to obtain informed consent because urine collection is considered minimal risk by New York State and US Federal regulations. From July 26, 2004, to January 11, 2006, a total of 426 patients were enrolled. Approximately 2 to 3 patients per day were arbitrarily selected in the morning from the daily operating room schedule by the research assistant (G.G.) without a specific randomization scheme. There was no significant difference in the frequency of surgical procedures comparing the study population with the approximately 2,000 cardiac surgeries performed at Columbia University Medical Center during this time. Of these, 200 patients with preoperative end-stage renal failure requiring hemodialysis therapy were excluded from the study.
Sample Collection and Processing Five milliliters of urine was collected after induction of anesthesia and insertion of a urinary catheter before surgical incision. Subsequently, 5 mL of urine was collected from the urinary catheter immediately after CPB (or after completion of the last graft in case of off-pump coronary artery bypass grafting), then 3, 18, and 24 hours later. Urine was centrifuged at 1,000g for 15 minutes, and the supernatant was frozen at ⫺20°C. Urinary NGAL levels were determined by using ELISA. No protease inhibitors were added to the samples because NGAL is resistant to proteases.12,13
Enzyme-Linked Immunosorbent Assay NGAL level was determined by using a commercially available ELISA kit (Antibodyshop, Gentofte, Denmark) according to the manufacturer’s instructions, as described previously.14 The limit of detection for this assay is 0.5 to 4.0 pg/mL, and intra-assay variation in urine is 2.1% (range, 1.3% to 4.0%).15,16
Data Collection Basic demographic data were collected and Parsonnet scores were calculated for risk stratification in all patients.17 We determined the number of intensive care unit (ICU)- and hospital-free days, defined as the number of days a patient was not in the ICU or the hospital within the first 30 days after surgery. We also determined the number of patient deaths (patients with no ICU- and hospital-free days). Serum creatinine was measured by the central laboratory of Columbia University Medical Center. Creatinine clearance was estimated using the method described by Cockcroft and Gault.18
Definition of AKI AKI was defined as an increase in serum creatinine level by either more than 50% or more than 0.3 mg/dL (compared with preoperative values) during the first 48 hours after surgery (serum creatinine in mg/dL may be converted to mol/L by multiplying by 88.4). This definition of AKI has been suggested by the Acute Kidney Injury Network (AKIN)19 and reflects recent studies that suggest that even small changes in serum creatinine levels increase morbidity and mortality after cardiac surgery.20 However, patients who were oliguric, but did not have a change in serum creatinine level, were not included as having AKI because urine output after surgery is often increased with the use of diuretics. Serum creatinine level was determined at least twice daily during the first 2 postoperative days, then daily for the duration of the hospital stay by the laboratory of Columbia University Medical Center as part of routine patient care.
Statistics Values are presented as mean ⫾ SEM. Comparisons between groups were made using unpaired t-test for values with Gaussian distribution and Mann-Whitney (Wilcoxon rank) test or Spearman test for correlation for continuous variables without normal distribution. Gaussian distribution was determined by using the Kolmogorov-Smirnov test. P values are 2 tailed, and P less than 0.05 is considered significant.
Urinary NGAL in Adult Cardiac Surgery
427
For the purpose of a pretrial power analysis, we assumed that urinary NGAL level with AKI would be 50% greater than without AKI (mean NGALnoAKI ⫽ 1,000 ng/mL, mean NGALAKI ⫽ 1,500 ng/mL, common SD ⫽ 1,000 ng/mL). We concluded that we would require 63 subjects per group to achieve 80% power with ␣ ⫽ 0.05. Assuming an incidence of 20% for AKI, we therefore would have to include 378 patients in this study. SAS, version 9.1 (SAS Inc, Cary, NC); SPSS, version 11.0.4 (SPSS Inc, Chicago, IL); and Graphpad Prism, version 4.0 (San Diego, CA) software was used for the statistical analysis.
RESULTS
by more than 50% or more than 0.3 mg/dL within 48 hours after surgery. Patients who developed postoperative AKI (either definition) were older; had higher Parsonnet scores, greater body mass indexes, and longer CPB times and AXTs; and received aprotinin more frequently as an antifibrinolytic (Table 1). They also remained longer in the ICU and hospital and had greater in-hospital mortality. Serum Creatinine After Cardiac Surgery
Acute Kidney Injury Eighty-five patients (20.0%) developed AKI, defined as an increase in serum creatinine level
Serum creatinine levels decreased immediately after surgery, even if patients subsequently developed AKI. In patients who developed AKI,
Table 1. Patient Characteristics
Preoperative Women Age (y) Body mass index (kg/m2) Serum creatinine (mg/dL) Parsonnet score Operation CABG with CPB CABG without CPB Single valve Multiple valve CABG ⫹ valve Cardiac transplant Ventricular assist device Other Reoperation Intraoperative CPB time (min) AXT (min) Aprotinin use Postoperative ICU-free days within 30 d Hospital-free days within 30 d Peak serum creatinine (mg/dL) ⌬ Serum creatinine (mg/dL) Continuous hemodialysis Mortality Urinary NGAL Preoperative (ng/dL) Immediately postoperative (ng/dL) 3 Hours (ng/dL) 18 Hours (ng/dL) 24 Hours (ng/dL)
Total (N ⫽ 426)
AKI (n ⫽ 85; 20.0%)
No AKI (n ⫽ 341; 80.0%)
P
144 (34) 63 ⫾ 15 28 ⫾ 16 1.08 ⫾ 0.45 9.2 ⫾ 7.5
25 (29.4) 69.5 ⫾ 13.1 29.5 ⫾ 7.8 1.21 (⫾ 0.47) 11.9 ⫾ 8.5
119 (34.9) 61.7 ⫾ 14.9 27.8 ⫾ 17.3 1.05 (⫾ 0.45) 8.5 ⫾ 7.1
0.4 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001
69 (16.2) 56 (13.1) 136 (31.9) 21 (4.9) 74 (17.4) 10 (2.3) 13 (3.1) 47 (11.0) 58 (13.6)
12 (14.1) 11 (12.9) 25 (29.4) 4 (4.7) 20 (23.5) 2 (2.4) 5 (5.9) 6 (7.1) 13 (15.3)
57 (16.7) 45 (13.2) 111 (32.6) 17 (5.0) 54 (15.8) 8 (2.3) 8 (2.3) 41 (12.0) 45 (13.2)
0.7 0.9 0.6 0.9 0.1 0.9 0.1 0.2 0.6
123 ⫾ 48 86 ⫾ 35 208 (49)
141 ⫾ 60 97 ⫾ 38 52 (61.2)
119 ⫾ 43 83 ⫾ 33 156 (45.8)
0.02 0.02 0.02
26.9 ⫾ 5.3 20.6 ⫾ 7.6 1.33 ⫾ 0.75 0.24 ⫾ 0.55 8 (1.9) 16 (3.8)
24.8 ⫾ 7.6 17.1 ⫾ 9.8 2.03 ⫾ 1.02 0.82 ⫾ 0.83 3 (3.5) 8 (9.4)
27.5 ⫾ 4.4 21.5 ⫾ 6.7 1.15 ⫾ 0.55 0.10 ⫾ 0.32 5 (1.5) 16 (1.8)
165 ⫾ 663 1,490 ⫾ 2,103 1,139 ⫾ 2,038 381 ⫾ 921 494 ⫾ 1,124
156 ⫾ 527 1,786 ⫾ 2,129 1,599 ⫾ 2,356 473 ⫾ 987 437 ⫾ 712
167 ⫾ 693 1,417 ⫾ 2,093 1,031 ⫾ 1,944 356 ⫾ 903 509 ⫾ 1,210
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.001 0.002 0.1 0.03 0.006 0.002 0.03
Note: Values expressed as mean ⫾ SD or number (percent). Serum creatinine in mg/dL may be converted to mol/L by multiplying by 88.4. Abbreviations: CABG, coronary artery bypass grafting; CPB, cardiopulmonary bypass; AXT, aortic cross-clamp time; NGAL, neutrophil gelatinase-associated lipocalin; ICU, intensive care unit.
428
Wagener et al
serum creatinine levels increased on the first postoperative day and peaked on postoperative day 2. Serum creatinine levels were statistically different between patients with and without AKI from postoperative days 0 through 10 (Fig 1). Urinary NGAL Before and After Cardiac Surgery in All Patients Studied Figure 2 shows preoperative and postoperative urinary NGAL levels in all patients studied. Before surgery, mean urinary NGAL level in all patients studied was 165 ⫾ 32 ng/mL. Preoperative urinary NGAL level was not different in patients with preoperative renal dysfunction (estimated creatinine clearance ⬍ 60 mL/min; Cockcroft-Gault formula18) compared with patients with normal preoperative renal function (178 ⫾ 756 versus 136 ⫾ 340 ng/mL; P ⫽ 0.6, respectively). Urinary NGAL level peaked immediately after cardiac surgery by more than 9-fold (1,490 ⫾ 102 ng/mL) and remained significantly greater 3, 18, and 24 hours after surgery compared with preoperative values (Fig 2). Urinary NGAL and AKI Preoperative urinary NGAL levels were similar in patients who later developed AKI com-
Figure 1. Percentage of change in serum creatinine levels after cardiac surgery with and without acute kidney injury (AKI; mean ⫾ SEM). *P ⬍ 0.05 compared with preoperative values. Serum creatinine levels decreased immediately after surgery in all patients and but increased to significantly greater levels in patients with AKI (defined as an increase in serum creatinine ⬎50% or ⬎0.3 mg/dL within 48 hours) and peaked on postoperative (postOP) day 2. Serum creatinine in mg/dL may be converted to mol/L by multiplying by 88.4.
Figure 2. Urinary neutrophil gelatinase-associated lipocalin (NGAL) levels before (PreOP) and after (PostOP) cardiac surgery in all patients (mean ⫾ SEM). *P ⬍ 0.05 compared with preoperative values. Urinary NGAL levels increased immediately after cardiac surgery and remained increased for 24 hours.
pared with patients without postoperative AKI (156 ⫾ 57 versus 167 ⫾ 38 ng/mL; P ⫽ 0.1]). Immediately after surgery, patients with postoperative AKI had significantly greater urinary NGAL levels (1,786 ⫾ 232 versus 1,417 ⫾ 114 ng/mL; P ⫽ 0.04). Urinary NGAL levels remained different throughout the 24-hour postoperative period between groups. The largest and most significant difference between patients with and without AKI was 3 hours after surgery (1,599 ⫾ 276 versus 1,031 ⫾ 110 ng/mL; P ⫽ 0.006). At 24 hours, patients who developed AKI had slightly lower urinary NGAL levels than patients without AKI (437 ⫾ 85 versus 509 ⫾ 75 ng/mL; P ⫽ 0.03; Fig 3).
Figure 3. Urinary neutrophil gelatinase-associated lipocalin (NGAL) levels before (PreOP) and after (PostOP) cardiac surgery (hours) in patients with or without acute kidney injury (AKI; mean ⫾ SEM). *P ⬍ 0.05 comparing patients with and without AKI.
Urinary NGAL in Adult Cardiac Surgery
Receiver Operating Characteristic Curves Receiver operating characteristic (ROC) curves were generated for all times to evaluate the ability of urinary NGAL to predict AKI. Areas under the curves (AUCs) were low at any time. AUC was highest 18 hours after surgery (AUC ⫽ 0.611; 95% CI, 0.544 to 0.679; P ⬍ 0.001). The cutoff value at this time (defined as the point on the ROC curve closest to sensitivity ⫽ 1 ⫺ specificity ⫽ 1) was 65 ng/mL (Fig 4; Table 2). AUCs of the ROC curve for CPB time and AXT to predict AKI were 0.592 (95% CI, 0.518 to 0.666; P ⫽ 0.01) and 0.593 (95% CI, 0.523 to 0.665; P ⫽ 0.01), respectively (Fig 4). Urinary NGAL and CPB Time/AXT Increased duration of aortic cross-clamping or CPB increases the severity of intraoperative renal injury,21,22 but may not necessarily be detected by an increase in postoperative serum creatinine levels. To assess whether we are able to detect this injury by using urinary NGAL level, we calculated the correlation coefficients between CPB time and AXT with urinary NGAL and postoperative creatinine (peak and change ⌬). Urinary NGAL levels immediately after CPB (P ⬍ 0.001 for CPB time and AXT) and at 3 (P ⬍ 0.001 for CPB time and AXT), 18 (P ⫽ 0.02 for CPB time and AXT), and 24 hours (P ⫽ 0.02 for CPB time and P ⫽ 0.03 for AXT) correlated significantly with CPB time/AXT. However, CPB time and AXT correlated better with urinary NGAL levels measured immediately after CPB and at 3 hours than with later levels of urinary NGAL (18 and 24 hours). Peak and ⌬ serum creatinine levels
Sensitivity
A
1 - Specificity
B
Sensitivity
Urinary NGAL and Urinary NGAL/Urine Creatinine Ratio We measured urine creatinine in 59 patients to evaluate the effect of postoperative hemodilution on urinary NGAL levels. Urinary creatinine levels decreased immediately after surgery and 3 hours later, then increased 18 and 24 hours after surgery. Urinary NGAL levels correlated well with urinary NGAL/urine creatinine ratio (Pearson r ⫽ 0.842; 95% confidence interval [CI], 0.795 to 0.879; P ⬍ 0.001), and change (⌬) in urinary NGAL/urine creatinine ratio in the postoperative period was very similar to the course of urinary NGAL.
429
1 - Specificity Figure 4. Receiver operating characteristic curves for urinary neutrophil gelatinase-associated lipocalin (NGAL) and cardiopulmonary bypass (CPB) time and aortic crossclamp time (AXT) as a predictor of acute kidney injury (AKI). (A) Urinary NGAL immediately after surgery (area under the curve [AUC] ⫽ 0.573; 95% confidence interval [CI], 0.506 to 0.640; P ⫽ 0.04) and 3 (AUC ⫽ 0.603; 95% CI, 0.533 to 0.674; P ⫽ 0.006), 18 (AUC ⫽ 0.611; 95% CI, 0.544 to 0.679; P ⫽ 0.003), and 24 hours after surgery (AUC ⫽ 0.584; 95% CI, 0.510 to 0.657; P ⫽ 0.03). (B) CPB time (AUC ⫽ 0.592; 95% CI, 0.518 to 0.666; P ⫽ 0.01) and AXT (AUC ⫽ 0.593; 95% CI, 0.522 to 0.665; P ⫽ 0.01) as a predictor for AKI.
430
Wagener et al Table 2. AUCs for Receiver Operating Characteristics of NGAL
NGAL postoperative NGAL 3 h postoperative NGAL 18 h postoperative NGAL 24 h postoperative
AUC
P
Cutoff (ng/mL)
Sensitivity (%)
Specificity (%)
Distance
0.573 (0.506-0.640) 0.603 (0.533-0.674) 0.611 (0.544-0.679) 0.584 (0.510-0.657)
0.04 0.006 0.003 0.03
⬎23.5 ⬎18.1 ⬎15.6 ⬎21.2
31.0 (26.1-36.2) 37.7 (32.3-43.4) 39.2 (33.6-45.1) 39.4 (33.5-45.6)
81.0 (70.9-88.7) 78.1 (66.9-86.9) 78.2 (67.4-86.8) 78.6 (67.1-87.4)
71.6 66.0 64.5 64.3
Note: Values in parentheses indicate 95% confidence intervals. Acute renal dysfunction is defined as an increase in serum creatinine level by more than 50% or 0.3 mg/dL (26.52 mol/L) within 48 hours (Acute Kidney Injury Network criteria). Abbreviations: AUC, area under the curve; NGAL, neutrophil gelatinase-associated lipocalin.
did not correlate with CPB time, and ⌬ serum creatinine level correlated significantly, but less robustly (P ⫽ 0.02), with AXT, whereas peak serum creatinine level did not (P ⫽ 0.06; Table 3). Additionally, we divided patients into groups with and without prolonged CPB time (⬎120 minutes; n ⫽ 176) and prolonged AXT (⬎90 minutes; n ⫽ 138). Prolonged CPB time (⬎120 minutes) was associated with significantly greater urinary NGAL levels immediately after surgery (mean difference between groups, 866 ng/mL (95% CI, 427 to 1,305; P ⬍ 0.001); at 3 hours, 1,002 ng/mL (95% CI, 545 to 1,459; P ⬍ 0.001); at 18 hours, 224 ng/mL (95% CI, 9 to 440; P ⫽ 0.04); and at 24 hours, 285 ng/mL (95% CI, 13 to 557; P ⫽ 0.04), but there was no difference with regard to ⌬ or peak serum creatinine level (P ⫽ 0.5 and P ⫽ 0.4, respectively). Prolonged AXT (⬎90 minutes) also was associated with increased urinary NGAL level immediately after surgery (mean difference, 1,012 ng/mL; 95% CI, 531 to 1,492; P ⬍ 0.001) and 3 (mean difference, 1,053 ng/mL; 95% CI, 533 to 1,572; P ⬍ 0.001) and 24 hours later (mean difference, 152 ng/mL; 95% CI, 58 to 657; P ⫽ 0.02), but not with peak or ⌬ serum creatinine and urinary NGAL level 18 hours after surgery (P ⫽ 0.1 and P ⫽ 0.3, respectively).
DISCUSSION We were able to show with this study that urinary NGAL levels correlate better than serum creatinine levels with indexes of intraoperative renal hypoperfusion (CPB time and AXT). Furthermore, patients who had long CPB times and AXTs showed increased NGAL levels, but had no significant change in serum creatinine levels. Urinary NGAL level immediately after surgery was significantly greater in patients with AKI, defined
by using increases in serum creatinine levels compared with patients who did not develop AKI. Urinary NGAL is a 25-kDa protein originally described in neutrophils.23,24 In 2003, Mishra et al25 used microassay techniques to show that NGAL is also expressed in proximal tubular cells of the kidney and increases substantially after ischemic injury in mice. Subsequent, but small, clinical studies showed that urinary NGAL is also detectable in high concentrations in humans after ischemic11,26 and other nephrotoxic27,28 injuries. Renal injury is frequent after cardiac surgery because of renal hypoperfusion, reperfusion injury, and inflammatory responses. AKI with or without the need for hemodialysis is associated with significant increases in morbidity and mortality after cardiac surgery.1 Unfortunately, the currently used marker of renal function (or dysfunction), serum creatinine level, is unacceptably insensitive. Subsequently, by the time renal dysfunction is indicated by an increase in serum creatinine levels, satisfactory clinical intervention is not possible. What is worse, even small increases in serum creatinine levels (⬃0.3 mg/ dL) have been associated with drastic increases in mortality and morbidity after cardiac surgery.20 Therefore, it is clear that serum creatinine level is an unacceptably insensitive and late marker of renal dysfunction. In addition, serum creatinine level is affected by the dilutional effect of increased total-body water in the early postoperative period after cardiac surgery. These factors greatly limit our ability to intervene during the early stage of renal injury to potentially prevent its progression to established acute renal failure, with its even greater morbidity and mortality. The purpose of this study is to test the hypothesis that urinary NGAL level is better suited to
0.173 0.054 to 0.288 0.03
431
Abbreviations: CPB, cardiopulmonary bypass; AXT, aortic cross-clamp time; NGAL, neutrophil gelatinase-associated lipocalin.
0.123 0.008 to 0.234 0.01 0.233 0.124 to 0.337 ⬍0.0001 0.125 0.019 to 0.229 0.02 0.101 ⫺0.006 to 0.205 0.06
0.240 0.137 to 0.339 ⬍0.0001
0.143 0.028 to 0.255 0.02 0.138 0.022 to 0.240 0.01 0.224 0.118 to 0.326 ⬍0.0001 0.227 0.127 to 0.323 ⬍0.0001 0.08415 ⫺0.0193 to 0.069 0.1 0.0314 ⫺0.072 to 0.134 0.5
Correlation with CPB time Pearson r 95% Confidence interval P (2 tailed) Correlation with AXT Pearson r 95% Confidence interval P (2 tailed)
18 h After Surgery 3 h After Surgery Peak Creatinine
Change in Creatinine
Postoperative
Urinary NGAL Serum Creatinine
Table 3. Correlation Between CBP Time or AXT and Postoperative Creatinine and Urinary NGAL at Different Times
24 h After Surgery
Urinary NGAL in Adult Cardiac Surgery
detect intraoperative renal injury after cardiac surgery than serum creatinine level. In addition, we want to confirm our previous findings that used less precise semiquantitative immunoblotting to measure NGAL in a small number of patients (n ⫽ 81) with a more accurate and quantitative ELISA in a large adult patient population undergoing a large variety of cardiac surgeries. We previously studied urinary NGAL (measured by means of immunoblotting technique) after cardiac surgery in 81 patients and found that urinary NGAL level as early as immediately after surgery was significantly greater in patients with postoperative AKI compared with those without AKI.11 We then found that use of aprotinin during cardiac surgery was associated with significantly greater postoperative NGAL levels (and a greater incidence of AKI) independent of possible confounders than the use of -amino caproic acid.14 With the present study, we confirmed our earlier findings that AKI is associated with increased urinary NGAL levels by using a more quantitative ELISA. However, we also observed much greater variability in urinary NGAL levels in adult patients than described by Mishra et al10 in pediatric patients. Serum creatinine levels decreased immediately after surgery regardless of subsequent AKI because of perioperative hemodilution and, in patients who developed AKI, peaked 2 days after surgery. Therefore, intraoperative renal ischemia reflected by an increase in urinary NGAL levels may not necessarily cause serum creatinine levels to increase several days later. It is well established that CPB time and AXT correlate with the development of AKI.21 We found a highly significant correlation between postoperative urinary NGAL levels and AXT and CPB time (Table 3). Conversely, peak serum creatinine level did not correlate with AXT and CPB time, and ⌬ serum creatinine correlated only weakly with AXT (P ⫽ 0.02). Additionally, long CPB times and long AXTs were associated with high urinary NGAL levels, but not increased serum creatinine levels. We therefore propose that urinary NGAL level is better suited to detect intraoperative renal injury than serum creatinine level: urinary NGAL level reflects the acute injury sustained through long CBP times and AXTs, whereas changes in serum creatinine lev-
432
els that occur days later reflect not only intraoperative, but also postoperative, renal ischemia. There is no uniform definition of AKI.29 We used the definition suggested by AKIN19 for stage 1 AKI. This definition is more inclusive than most other definitions, reflecting recent evidence that even small changes in serum creatinine levels are associated with increased morbidity and mortality in critically ill patients.30,31 For example, Lassnigg et al20 found that even a small change in serum creatinine level was predictive of increased 30-day mortality after cardiac surgery. However, the AKIN definition also has limitations in the immediate postoperative period because it restricts the period of observation to 48 hours and therefore may not be able to identify kidney injury and increases in serum creatinine level that become detectable only after the early period of postoperative hemodilution.32 The immediate postoperative period after cardiac surgery also does not necessarily constitute an optimal state of hydration required to apply the AKIN criteria.19 Additionally, patients frequently receive diuretics in this period. We therefore did not include the oliguria component of the AKIN definition of AKI (decrease in urine output to ⬍0.5 mL/h for 6 hours). Urinary NGAL level increased immediately after surgery and remained increased during the 3-hour postoperative period in patients with AKI compared with patients without AKI. At 18 and 24 hours after surgery, urinary NGAL levels were similar in both groups. The ability of urinary NGAL level to predict AKI was far less compelling (indicated by the small AUCs of the ROC curves) than previously described: the highest AUC of the ROC curve was 0.611 compared with the AUC of 0.998 observed by Mishra et al10 in the pediatric cardiac patient population. Unlike the pediatric cardiac surgery population,10 many patients in our study had such significant preoperative comorbidities as hypertension or diabetes and underwent highcomplexity surgeries, evidenced by a mean Parsonnet score of 9.2 ⫾ 7.5 and overall mortality rate of 3.8%. Almost all our patients had an increase in urinary NGAL levels after cardiac surgery, and only the degree of change differed in patients with and without AKI compared with the pediatric population described by Mishra et al.10 Serum creatinine levels peaked days after the surgery in patients who developed AKI and thus
Wagener et al
reflected not only intraoperative injury, but also hemodynamic instability, high vasopressor requirements, or other nephrotoxic or ischemic insults in the postoperative period.33 Serum creatinine levels decreased immediately after surgery regardless of subsequent AKI because of perioperative hemodilution. Therefore, a minor intraoperative renal injury reflected by an increase in urinary NGAL levels may not necessarily cause serum creatinine levels to increase 48 hours later in this very heterogeneous patient population. Exposure of blood to CPB circuits induces activation of neutrophils, which could lead to a release of significant amounts of intracellular NGAL that is then filtrated in urine. The urinary NGAL we measured therefore may not (only) represent NGAL released from renal tubuli. However, plasma NGAL levels are approximately 3 to 10 times lower than urinary levels after cardiac surgery.10,14 It therefore is very unlikely that urinary NGAL consists of any significant amount of plasma NGAL released from activated neutrophils. This study is limited by its end point, serum creatinine level, which may not be suited to detect renal injury as accurately and as sensitively as urinary NGAL level. This limitation most likely led to lower AUCs from the ROCs generated. Future studies will have to evaluate whether urinary NGAL level correlates with levels of other markers of renal injury, such as kidney injury molecule 1, serum cystatin C, or N-acetyl--D-glucosaminidase.34,35 In conclusion, we believe that urinary NGAL level is an earlier and a more sensitive marker of renal injury compared with serum creatinine level.26 Urinary NGAL level may allow earlier and more successful renal interventions and thus serve as a useful marker of renal injury in clinical studies investigating renoprotective strategies and allow early identification and treatment of patients at risk.
ACKNOWLEDGEMENTS Support: This work was funded by the intramural grant support from the Department of Anesthesiology, Columbia University College of Physicians and Surgeons. We would like to thank Antibodyshop, Entofte, Denmark, for supplying the NGAL ELISA kits. Financial Disclosure: None.
Urinary NGAL in Adult Cardiac Surgery
REFERENCES 1. Conlon PJ, Stafford-Smith M, White WD, et al: Acute renal failure following cardiac surgery. Nephrol Dial Transplant 14:1158-1162, 1999 2. Corwin HL, Sprague SM, DeLaria GA, et al: Acute renal failure associated with cardiac operations. A casecontrol study. J Thorac Cardiovasc Surg 98:1107-1112, 1989 3. Kuitunen A, Vento A, Suojaranta-Ylinen R, et al: Acute renal failure after cardiac surgery: Evaluation of the RIFLE classification. Ann Thorac Surg 81:542-546, 2006 4. Chertow GM, Lazarus JM, Christiansen CL, et al: Preoperative renal risk stratification. Circulation 95:878884, 1997 5. Conger JD: Interventions in clinical acute renal failure: What are the data? Am J Kidney Dis 26:565-576, 1995 6. Star RA: Treatment of acute renal failure. Kidney Int 54:1817-1831, 1998 7. Bagshaw SM, Bellomo R: Early diagnosis of acute kidney injury. Curr Opin Crit Care 13:638-644, 2007 8. O’Leary MJ, Bihari DJ: Preventing renal failure in the critically ill. There are no magic bullets—Just high quality intensive care. BMJ 322:1437-1439, 2001 9. Nguyen MT, Devarajan P: Biomarkers for the early detection of acute kidney injury. Pediatr Nephrol 2007 Mar 30 [Epub ahead of print] 10. Mishra J, Dent C, Tarabishi R, et al: Neutrophil gelatinase-associated lipocalin (NGAL) as a biomarker for acute renal injury after cardiac surgery. Lancet 365:12311238, 2005 11. Wagener G, Jan M, Kim M, et al: Association between increases in urinary neutrophil gelatinase-associated lipocalin and acute renal dysfunction after adult cardiac surgery. Anesthesiology 105:485-491, 2006 12. Kjeldsen L, Johnsen AH, Sengelov H, et al: Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase. J Biol Chem 268:1042510432, 1993 13. Peitsch MC, Boguski MS: The first lipocalin with enzymatic activity. Trends Biochem Sci 16:363, 1991 14. Wagener G, Gubitosa G, Wang S, et al: Increased incidence of acute kidney injury with aprotinin use during cardiac surgery detected with urinary NGAL. Am J Nephrol 28:576-582, 2008 15. Antibodyshop: NGAL ELISA Kit (KIT 036) product inlay. Available at: http://www.antibodyshop.com/content/ download/1130/34986/version/22/file/Inlay⫹NGAL⫹ELISA⫹ Kit⫹%28KIT⫹036%29.pdf. Accessed April 1, 2008 16. Kjeldsen L, Koch C, Arnljots K, et al: Characterization of two ELISAs for NGAL, a newly described lipocalin in human neutrophils. J Immunol Methods 198:155-164, 1996 17. Parsonnet V, Dean D, Bernstein AD: A method of uniform stratification of risk for evaluating the results of surgery in acquired adult heart disease. Circulation 79:I3I12, 1989 18. Cockcroft DW, Gault MH: Prediction of creatinine clearance from serum creatinine. Nephron 16:31-41, 1976
433 19. Mehta RL, Kellum JA, Shah SV, et al: Acute Kidney Injury Network: Report of an initiative to improve outcomes in acute kidney injury. Crit Care 11:R31, 2007 20. Lassnigg A, Schmidlin D, Mouhieddine M, et al: Minimal changes of serum creatinine predict prognosis in patients after cardiothoracic surgery: A prospective cohort study. J Am Soc Nephrol 15:1597-1605, 2004 21. Fischer UM, Weissenberger WK, Warters RD, et al: Impact of cardiopulmonary bypass management on postcardiac surgery renal function. Perfusion 17:401-406, 2002 22. Solomon R, Werner C, Mann D, et al: Effects of saline, mannitol, and furosemide to prevent acute decreases in renal function induced by radiocontrast agents. N Engl J Med 331:1416-1420, 1994 23. Bundgaard JR, Sengelov H, Borregaard N, et al: Molecular cloning and expression of a cDNA encoding NGAL: A lipocalin expressed in human neutrophils. Biochem Biophys Res Commun 202:1468-1475, 1994 24. Kjeldsen L, Johnsen AH, Sengelov H, et al: Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase. J Biol Chem 268:1042510432, 1993 25. Mishra J, Ma Q, Prada A, et al: Identification of neutrophil gelatinase-associated lipocalin as a novel early urinary biomarker for ischemic renal injury. J Am Soc Nephrol 14:2534-2543, 2003 26. Mori K, Nakao K: Neutrophil gelatinase-associated lipocalin as the real-time indicator of active kidney damage. Kidney Int 71:967-970, 2007 27. Mishra J, Mori K, Ma Q, et al: Neutrophil gelatinaseassociated lipocalin: A novel early urinary biomarker for cisplatin nephrotoxicity. Am J Nephrol 24:307-315, 2004 28. Bachorzewska-Gajewska H, Malyszko J, Sitniewska E, et al: NGAL (neutrophil gelatinase-associated lipocalin) and cystatin C: Are they good predictors of contrast nephropathy after percutaneous coronary interventions in patients with stable angina and normal serum creatinine? Int J Cardiol 127:290-291, 2008 29. Ronco C, Levin A, Warnock DG, et al: Improving outcomes from acute kidney injury (AKI): Report on an initiative. Int J Artif Organs 30:373-376, 2007 30. Levy MM, Macias WL, Vincent JL, et al: Early changes in organ function predict eventual survival in severe sepsis. Crit Care Med 33:2194-2201, 2005 31. Praught ML, Shlipak MG: Are small changes in serum creatinine an important risk factor? Curr Opin Nephrol Hypertens 14:265-270, 2005 32. Moran SM, Myers BD: Course of acute renal failure studied by a model of creatinine kinetics. Kidney Int 27:928937, 1985 33. Andersson LG, Ekroth R, Bratteby LE, et al: Acute renal failure after coronary surgery—A study of incidence and risk factors in 2009 consecutive patients. Thorac Cardiovasc Surg 41:237-241, 1993 34. Coca SG, Yalavarthy R, Concato J, et al: Biomarkers for the diagnosis and risk stratification of acute kidney injury: A systematic review. Kidney Int 73:1008-1016, 2008 35. Han WK, Bailly V, Abichandani R, et al: Kidney injury molecule-1 (KIM-1): A novel biomarker for human renal proximal tubule injury. Kidney Int 62:237-244, 2002