Hindawi Disease Markers Volume 2019, Article ID 8480925, 9 pages https://doi.org/10.1155/2019/8480925 Research Article Does Whole-Blood Neutrophil Gelatinase-Associated Lipocalin Stratify Acute Kidney Injury in Critically Ill Patients? M. Cuartero ,1 A. J. Betbesé,1 K. Núñez,1 J. Baldirà,1 and J. Ordonez-Llanos2 1 Intensive Care Department and Institut d’Investigacions Biomèdiques, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain 2 Biochemistry Department and Institut d’Investigacions Biomèdiques, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain Correspondence should be addressed to M. Cuartero; mireia.cuartero@gmail.com Received 29 October 2018; Revised 11 February 2019; Accepted 3 March 2019; Published 2 May 2019 Guest Editor: Dominique Guerrot Copyright © 2019 M. Cuartero et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Purpose. To analyse the capacity of whole-blood NGAL (wbNGAL) to stratify AKI in critically ill patients with and without sepsis. Methods. Whole-blood NGAL was measured with a point-of-care device at admission and 48 hours later in patients admitted to a general ICU. Patients were classified by the AKIN and KDIGO classifications at admission and 24 and 48 hours. We performed an ROC curve analysis. wbNGAL values at admission were compared in patients with sepsis and septic shock. Results. The study included 100 consecutively admitted patients (40 female) with mean age 59 1 ± 17 8 years. Thirty-three patients presented AKI at admission, and 10 more developed it in the next 48 h. Eighteen patients had AKI stage 3, 14 of them at admission. Nine patients required renal replacement therapy. According to KDIGO at admission, wbNGAL values were 78 μg/L (60-187) in stage 0 (n = 67), 263 μg/L (89-314) in stage 1 (n = 8), 484 μg/L (333-708) in stage 2 (n = 11), and 623 μg/L (231-911) in stage 3 (n = 14), p = 0 0001 for trend. Ten patients did not complete 48 hours of study: 6 of 10 were discharged (initial wbNGAL 130 μg/L (60- 514)) and 4 died (773 μg/L (311-1010)). The AUROC curve of wbNGAL to predict AKI was 0.838 (95% confidence interval 0.76-0.92, p = 0 0001), with optimal cut-off value of 178 μg/L (sensitivity 76.7%, specificity 78.9%, p < 0 0001). At admission, twenty-nine patients had sepsis, of whom 20 were in septic shock. wbNGAL concentrations were 81 μg/L (60-187) in patients without sepsis, 481 (247-687) in those with sepsis, and 623.5 μg/L (361-798) in the subgroup of septic shock (p < 0 0001). Conclusions. Whole-blood NGAL concentration at ICU admission was a good stratifier of AKI in critically ill patients. However, wbNGAL concentrations were higher in septic patients irrespective of AKI occurrence. 1. Introduction Neutrophil gelatinase-associated lipocalin (NGAL) is a glycoprotein first isolated from specific granules of human Acute kidney injury (AKI) is a disease often diagnosed in leukocytes [9]. NGAL was found elevated in bacterial infec- intensive care unit (ICU) patients. AKI incidence can be as tions when compared to viral infections [10]. NGAL exists high as 24% depending on its definition [1]. ICU-related as a monomeric form of 25 kDa, a homodimer linked by a AKI is associated with an in-hospital mortality that can reach disulphide bridge of 45 kDa and a heterodimer with matrix 60% [2, 3]. Therefore, early AKI detection is crucial to pre- metalloproteinase-9 (MMP-9, gelatinase) with an intermo- vent or stop the natural course of renal dysfunction and lecular disulphide bridge of 135 kDa [11]. NGAL is also improve its morbi-mortality. Serum creatinine is the gold formed in other cells apart from leukocytes. In response to standard biomarker in all AKI definitions [4–6]. However, several injuries, NGAL is expressed in kidney, hepatic, or epi- creatinine is a late biomarker for AKI, as values peak after thelial cells [11–13]. In the kidneys, NGAL is freely filtered by 24-48 hours of renal injury. Hence, several biomarkers have the glomerulus and reabsorbed in the proximal tubule [14]. been proposed for early AKI detection [7, 8]. After a tubular injury, NGAL is overexpressed in the diseased 2 Disease Markers endothelium [15–17]. In this setting, tubular reabsorption is with the Mann–Whitney U or Kruskal-Wallis tests. Categor- reduced, and blood and urine NGAL concentrations ical data are reported as percentage and were compared by increase. In AKI, NGAL increase is observed 24-48 hours the chi-square test or Fisher exact test. Reporting of results earlier than plasma creatinine peak [18, 19]. However, the followed the STARD (Standards for Reporting Diagnostic specificity of NGAL to detect AKI could be limited by overex- Accuracy Studies) statement. Whole-blood NGAL predictive pression in other tissues [20]. This could be relevant in criti- values were evaluated by receiver operating characteristic cally ill patients, in whom sepsis could promote NGAL (ROC) curve analysis. We defined an area under the ROC release from tissues other than renal epithelium [21, 22]. (AUROC) curve of 0.60-0.69 as poor, 0.70-0.79 as fair, NGAL has been described as a predictor of AKI and the need 0.80-0.89 as good, and 0.90-1.00 as excellent in terms of pre- of renal replacement therapies (RRT) in patients admitted to dictive value. A p < 0 05 was considered significant. general intensive care units (ICU) [21–24]. However, as a systematic review by Hjortrup et al. [24] pointed out, the 3. Results NGAL role is not fully understood: literature shows that NGAL might predict AKI with a wide range AUROC, from 3.1. Clinical Characteristics. We recruited 100 consecutive 0.54 to 0.98. The aims of this study were (1) to evaluate the patients fulfilling the admission criteria (Figure 1). Ten of capacity of whole-blood NGAL at ICU admission to predict them did not complete the 48 h follow-up due to discharge AKI development and (2) to analyse the effect of sepsis on (6 cases) or death (4 cases). Mean age was 59 1 ± 17 8 years. its predictive capacity. 60% of cases were male. The causes of ICU admission were medical 54% (respiratory 26%, cardiovascular 7%), postsur- 2. Methods gical care 39% (gastrointestinal 21%, neurosurgery 12%), and miscellaneous 7%. At ICU admission, twenty-nine were 2.1. Patients. The study protocol was approved by the Ethic septic, and twenty of them had septic shock. ICU length of Committee Board at Hospital de la Santa Creu i Sant Pau stay was 10 3 ± 9 6 days, and ICU mortality was 22%. (Barcelona, Spain). We obtained informed consent from par- ticipants or their guardians. The study prospectively included 3.2. Whole-Blood NGAL and AKI. Forty-three patients pre- 100 patients consecutively admitted during 8 months (June sented AKI, 33 at admission (8 stage 1, 11 stage 2, and 14 2010-February 2011) in a general ICU. Inclusion criteria stage 3) and 10 more within 48 h of the ICU stay; 4 of the were age over 18 years old and the expected ICU stay of at latest group of 10 developed renal failure (stage 3). Nine least 48 hours. Patients were excluded if they had already patients required renal replacement therapies. wbNGAL been admitted in the hospital for more than 24 hours, had values were 78 μg/L (IQR 60-187 μg/L) in non-AKI patients any degree of preexisting chronic kidney disease, or were and 263 μg/L (IQR 89-314 μg/L), 484 μg/L (IQR 333- not expected to survive for at least 24 hours due to a nonre- 708 μg/L), and 623 μg/L (IQR 231-911 μg/L) in those with versible clinical condition. Physicians attending patients were stage 1, stage 2, and stage 3, respectively (p = 0 0001 for blinded to NGAL results throughout the study. trend) (Figure 2). Four of 33 patients with AKI at admis- sion were diagnosed of AKI solely because of urine output 2.2. Clinical and Laboratory Data Collection. Clinical data criterion (one patient had stage 1, one patient had stage 2, included admission diagnosis, demographics, ICU severity and two patients had stage 3); 10 of 33 patients were diag- scores, haemodynamic parameters, and urine output. nosed of AKI based on changes in sCr, and the remaining Patients were initially classified following the AKIN defini- 19 patients were diagnosed based in both sCr and drop of tion [5] and KDIGO [6] definition for AKI. Both definitions urine output. The incidence and severity of AKI were the were based on serum creatinine and urine output obtained same when AKIN classification was applied within the first since admission to 6 hours later, at 24 and 48 hours of admis- 48 h. sion. Baseline serum creatinine was taken from patients’ pre- In the group of 6 patients discharged before 24 h, admis- admission records whenever possible and used to estimate sion wbNGAL and plasma creatinine were of 130 μg/L (IQR eGFR before ICU admission using the Cockcroft–Gault for- 60-514) and 78 μmol/L (IQR 54-123), respectively; whereas mula. According to standard criteria [25], patients were also in the 4 patients who died, wbNGAL was of 773 μg/L (IQR classified as having sepsis and septic shock. 311-1010) and plasma creatinine 165 μmol/L (IQR 59-577). NGAL was measured at admission and 24 and 48 hours We did not find statistical differences in wbNGAL or serum in EDTA-anticoagulated whole-blood using the Triage® creatinine between subgroups of patients who were dis- NGAL Test (Alere Diagnostics, formerly Inverness Medical charged or passed away. Innovations). All samples were analysed in the same batch Whole-blood NGAL values were predictive of AKI both to avoid between-batch variability. at admission and within 48 h of the ICU stay (Supplemental Table 1). The area under the ROC curve (AUROC) of 2.3. Statistical Analysis. SPSS® version 18 (SPSS Inc., wbNGAL for AKI prediction within 48 hours of ICU Chicago, IL) was used. Variables with normal distribution admission was 0.838 (95% confidence interval (CI) 0.760- are reported as mean ± standard deviation and were com- 0.917, p = 0 0001). The wbNGAL optimal cut-off for AKI pared with Student’s t-test or one-way analysis of variance. within 48 h of ICU admission was 178 μg/L, with sensitivity Variables with non-Gaussian distribution are reported as of 76.7% and specificity of 78.9% (Figure 3). The AUROC median and interquartile range (IQR) and were compared of admission serum creatinine for AKI within the first 48 h Disease Markers 3 Potentially eligible participants 205 patients were excluded n = 305 because of previous hospital admission for more than 24 h, chronic kidney disease, nonreversible cause of ICU admission, Eligible participants and lack of consent. n = 100 wbNGAL test at admission n = 100 Test-negative Test-positive wbNGAL < 178 휇g/L wbNGAL ≥ 178 휇g/L n = 67 n = 33 Inconclusive n = 0 Inconclusive n = 0 Reference standard: Reference standard: AKIN at admission AKIN at admission No AKI n = 0 No AKI n = 67 Stage 1 n = 8 AKI = 0 Stage 2 n = 11 Stage 3 n = 14 Figure 1: Diagram of whole-blood NGAL to predict AKI, study enrolment and inclusion/exclusion criteria. ROC curve 1500 623 (231-911) 1.0 p = 0.0001 for trend wbNGAL at ICU admission (휇g/L) 1000 0.8 484 (333-708) 78 (60-187) ⁎⁎ Sensitivity 0.6 ⁎ 500 263 (89-314) 0.4 0 0 Stage 1 Stage 2 Stage 3 0.2 Cut-off 178 휇g/L 0.029 0.002 0.024 Sensivity 76.7%; specificity 78.9% <0.0001 AUROC: 0.838 (0.760-0.917; p < 0.0001) <0.0001 0.93 0.0 0.0 0.2 0.4 0.6 0.8 1.0 Figure 2: Boxplot comparing whole-blood NGAL (wbNGAL) 1 − specificity concentrations (μg/L) and KDIGO score at admission. Boxplots indicate the median and 25th and 75th percentiles. Whiskers Figure 3: ROC curve for whole-blood NGAL value at ICU indicate the 5th and 95th percentiles. Statistical significance (p) admission to predict AKI. Diagnostic and overall accuracies given comparing wbNGAL (μg/L) with KDIGO categories is given at the with sensitivity, specificity, and 95% confidence interval compared bottom of the figure. with gold standard KDIGO classification. 4 Disease Markers Table 1: Clinical characteristics of the study patients depending on the cut-off obtained for AKI prediction. <178 μg/L ≥178 μg/L p value n 55 45 — Age (years) 54 9 ± 18 5 64 ± 15 5 0.009 Male 37 (67) 23 (51) 0.075 Characteristics at ICU admission SAPS II 37 8 ± 14 3 43 15 ± 15 5 0.095 APACHE II 15 ± 6 9 18 ± 8 2 0.059 SOFA 5 2±2 7 7 7±3 6 <0.0001 AKI admission (%) 6 (10.9) 27 (15.6) <0.0001 KDIGO stage 3 at admission (%) 2 (3.6) 12 (26.7) 0.001 RRT admission (%) 2 (3.6) 3 (6.7) 0.400 Creatinine clearance (mL/min) 109 ± 58 57 ± 52 <0.0001 Cockcroft–Gault 104 ± 47 58 ± 38 <0.0001 Characteristics during ICU stay Mechanical ventilation (%) 46 (82) 36 (80) 0.410 Vasopressor requirement (%) 6 (11) 22 (49) <0.0001 ICU length of stay (days) 11 1 ± 9 3 9 4 ± 10 0.360 ICU mortality (%) 10 (18.2) 12 (26.7) 0.210 Sepsis 3 (5.5) 26 (57.8) 0.007 Septic shock 2 (3.6) 18 (40) <0.0001 AKI development (%) 2 (3.6) 6 (13.3) 0.070 KDIGO stage 3 development (%) 2 (3.6) 16 (35.6) <0.0001 RRT after admission (%) 0 (0) 4 (8.9) 0.038 RRT total (%) 2 (3.6) 7 (15.6) 0.040 Values expressed as either % per column or mean ± standard deviation. p: value of statistical significance. SAPS II: Simplified Acute Physiology Score II; APACHE II: Acute Physiology and Chronic Health Evaluation II; SOFA: Sequential Organ Failure Assessment score; AKI: acute kidney injury; KDIGO: Improving Global Outcomes AKI group classification; RRT: Renal replacement therapies; ICU: intensive care unit. of ICU admission was 0.904 (95% CI 0.841-0.967, p = 0 0001). There were no statistical differences between wbNGAL and 1400 481 (247-681) plasma creatinine AUROC. wbNGAL at admission (휇g/L) Forty-five patients had wbNGAL > 178 μg/L. They were 1200 more likely to be older, have higher SOFA at admission, 1000 81 (60-187) have higher incidence of AKI development, stage 3, sepsis, ⁎ or septic shock, and have requirement of vasopressor 800 ⁎ drugs and renal replacement therapies during the ICU stay ⁎ (Table 1). Six of eighteen patients without AKI at admission 600 but wbNGAL > 178 μg/L developed AKI within the next 48 h. Accordingly, wbNGAL identified at admission extra 14.6% of 400 AKI patients not diagnosed by the serum creatinine criterion. 200 3.3. Whole-Blood NGAL and Sepsis. wbNGAL concentrations 0 were 81 μg/L (IQR 60-187) and 481 μg/L (IQR 247-681) in 71 No sepsis Sepsis patients without sepsis and 29 patients with sepsis, respec- tively (p < 0 0001) (Figure 4). wbNGAL was 623.5 μg/L Figure 4: Whole-blood NGAL concentrations (μg/L) according to (IQR 361-798) in those 20 of 29 septic patients who also the presence of sepsis. Boxplots indicate the median and 25th and 75th percentiles of wbNGAL in ng/L. Whiskers indicate the 5th had shock. The incidence of AKI in sepsis and septic shock and 95th percentiles. Statistical significance (p). Patients without was 28.6% and 65%, respectively. sepsis n = 71; patients with sepsis n = 29. Serial measurement of wbNGAL did not improve AKI prediction in septic patients compared to nonseptic patients (data not shown). In nonseptic patients, AKI con- in AKI patients, p < 0 0001. However, when comparing centrations appeared dependent on AKI status: 62 μg/L wbNGAL in septic patients with (632 μg/L (IQR 344-1060)) (IQR 60-99) in non-AKI and 297 μg/L (IQR 123-502) or without (414 μg/L (IQR 214-552)) AKI, there was no Disease Markers 5 AKI+ AKI− n = 33 n = 67 p < 0.0001 Sepsis– n = 71 297 (1.23-502)& 62 (60-99)&& (n = 18) (n = 53) p = 0.017 p < 0.0001 632 (344-1060)&& 414 (214-552)& (n = 15) (n = 14) Sepsis+ n = 29 p = 0.46 Figure 5: Distribution of admission whole-blood NGAL (μg/L) values at ICU admission depending on AKI and sepsis. AKI+ vs. AKI- and sepsis+ vs. sepsis- represents the presence or absence of either AKI or sepsis upon admission, respectively. Values show median and percentiles 25-75. p represents the statistical intragroup differences. wbNGAL values given in μg/L. &Statistical difference between subgroups of nonseptic AKI vs. septic non-AKI patients (p = 0 676). &&Statistical difference between subgroups of nonseptic non-AKI vs. septic and AKI patients (p < 0 0001). AKI: acute kidney injury. statistical difference (p = 0 46). Septic patients with and with- transplantation [18], or chronic kidney disease [29]. Fewer out AKI presented higher wbNGAL than those with the same studies were done in general critically ill population, both renal status without sepsis (p < 0 0001 for both compari- paediatric [30, 31] and adult patients [19, 21, 32, 33]. There sons). wbNGAL values in septic non-AKI patients were is still debate whether NGAL is a good predictor of AKI in undistinguishable of those of nonseptic AKI patients (p = general ICU patients [24]. Our AUROC of wbNGAL is sig- 0 676). Nine of the 11 non-AKI patients with wbNGAL higher nificantly higher than that found by Parikh et al. [18] in the than the cut-off >178 μg/L presented sepsis (Figure 5). In postoperative care in cardiac surgery and by Haase et al. our study, there were 20 patients with septic shock: 3 had [22] in a paediatric critically ill population. In our study, no AKI (median wbNGAL of 481 μg/L, IQR 142-481 μg/L), wbNGAL stratified AKI severity, showing increasing concen- 2 had stage 1 (251, IQR 234-269 μg/L), 4 had stage 2 (562, trations with an increased KDIGO stage from median values IQR 421-681 μg/L), and 11 had stage 3 (685, IQR 526- of 71 μg/L in stage 1 and 186 μg/L in stage 2 to 381 μg/L in 1130 μg/L). stage 3 (Figure 2). This increasing pattern in line with the severity of AKI has also been described by other investigators 4. Discussion in paediatric and adult patients [24, 25]. Although wbNGAL concentrations in non-AKI patients are similar to those The main finding of our study is that admission wbNGAL is a found by Singer et al. [25], AKI severity categories were lower good stratifier of AKI within 48 hours of ICU admission in a in our study. In view of our results, we would suggest clini- heterogeneous group of critically ill patients (AUROC 0.838 cians to consider wbNGAL as a complementary tool in (95% CI 0.760-0.917, p < 0 0001)). The cut-off value higher patients at AKI risk or receiving nephrotoxic drugs. Despite than 178 μg/L defines a group of patients with higher severity NGAL’s promising results in general ICU population, the and higher probability of developing AKI. However, NGAL AKIKI [34], ELAIN [35], and the most recent IDEAL-ICU concentrations were affected by sepsis status irrespective of [36] trials showed that wbNGAL combined with KDIGO the AKI presence (p = 0 46). Thus, NGAL measures are not staging did not improve the timing of renal replacement ther- useful to evaluate kidney function in patients with sepsis. apies or patient’s prognosis. Some studies suggested that NGAL in blood or urine In our cohort, the AUROC curve of wbNGAL for AKI could be a useful biomarker of AKI [26]. Most of those initial prediction showed an optimal cut-off value of 178 μg/L. studies were done in patients after cardiac surgery [7, 27], in Our cut-off is close to the ones described in other studies, context of contrast-induced nephropathy [28], kidney around 150 μg/L [22, 27, 37]. In critically ill patients, the 6 Disease Markers exact time when acute tubular damage occurs is often it has not yet occurred. In these cases, urinary NGAL, which unknown. The differences in AUROC described in literature is not submitted to the influence of inflammatory mediators, could also be explained by their study designs. Unlike our could have better performance in AKI prediction. Besides, study, those set up in cardiac surgery or radiocontrast admin- differentiation of monomeric and dimeric plasma NGAL iso- istration had an exact time of a potential onset of renal injury. forms could be crucial to recognise the NGAL concentration We found significant differences in admission wbNGAL secondary to renal damage or inflammation [44]. Our data between the groups of patients with AKI and those without. also suggest that NGAL may be a biomarker of illness sever- We also identified a subgroup of patients who presented high ity. In an article by Shapiro et al. [8], NGAL was a good pre- levels of wbNGAL with no increase of creatinine. This sub- dictor of septic shock in a panel of multiple biomarkers and group could be considered false positives or, in our opinion, correlated with survival. Wang et al. also described that high a subgroup of patients who may represent subclinical AKI wbNGAL independently predicted mortality and multiple [22]. Patients with stage 1 showed relatively low wbNGAL organ dysfunction syndrome in sepsis and septic shock [45]. concentrations within the group of patients with AKI. This On the other hand, a study set up in the emergency finding could be attributable to treated reversible causes of department by Wang et al. [46] suggested that combined AKI, like hypovolemia or hypotension. Nickolas et al. [38] NGAL and TIMP-1 (tissue inhibitor of matrix metallopro- described low levels of urine NGAL in patients with prerenal teinases-1, a cell cycle arrest biomarker for AKI) was use- azotaemia. Those patients inadequately resuscitated after a ful for the diagnosis and risk stratification of patients with prerenal azotaemia had increasing urinary NGAL concentra- AKI, including those who also presented with sepsis. tions. Similar to our study, this is a clinically relevant finding. Although cell cycle arrest (CCA) biomarkers are still under Timely treatment of patients with subclinical AKI or stage 1 evaluation in different clinical settings, their sole predictive could avoid AKI progression to failure and tubular damage. power and risk stratification appears to be higher than that NGAL is a protein upregulated after a tubular injury exhibited by NGAL. Since CCA biomarkers were presented [39]. However, it can also be produced by other organs like in Sapphire study [47], there has been an increasing num- the liver or lung [7] or in different inflammatory situations ber of studies in paediatric [48] and adult critical care pop- [40]. In our study, the serial measurement of wbNGAL did ulations, with clinical implementation in rapid response not improve AKI prediction in septic patients compared to teams [49] and KDIGO care bundles [50, 51]. Besides, nonseptic patients. No statistical differences were found unlike NGAL, CCA biomarkers are not determined by between ROC curves obtained at admission and 48 h later; sepsis [52, 53]. both ROC curves produced very similar cut-offs (Supple- Our study has some limitations. First of all, the sample mental Table 1). These results concur with those from size was small, and the study was performed in a single cen- Bagshaw et al. [41], which showed that peak plasma NGAL tre. Secondly, we recruited a cohort that might not represent did not perform better to predict AKI in septic than in an average ICU population. We aimed to study the AKI inci- nonseptic patients. This is clinically relevant because single dence in a population with ideally normal baseline renal sampling at admission decreases costs, and it is easier to function. AKI is a common complication of many nonrenal implement in daily routine. hospitalisations [54–56]. We purposefully excluded patients wbNGAL values were higher in patients with sepsis with a background of chronic disease and/or admitted in and much higher in those with septic shock. These data the hospital for more than 24 hours prior to ICU admission. strongly suggest that wbNGAL not only is a good predictor However, this is also the main strength of our study, because of AKI but also can be considered a good severity score in our unique cohort of patients were not under the effect of patients with inflammatory status. Although Mishra et al. intrahospital risk factors for AKI that could have acted as [27] showed that plasma NGAL was independent of inflam- confounding variables. Finally, the point-of-care test used matory markers like C-reactive protein, Zappitelli et al. [31] to analyse wbNGAL was not able to differentiate between reported that plasma and urine NGAL concentrations at NGAL isoforms, which could have helped to explain the role ICU admission were higher in patients in which AKI was of inflammation in wbNGAL concentrations. due to sepsis than in those in which it was due to nonseptic causes. Like in our study, these authors found that serial 5. Conclusions NGAL measurement did not add predictive power to NGAL concentrations at admission. Other articles also suggested Our study showed that whole-blood NGAL concentrations that the inflammatory status [24] or septic shock could at ICU admission stratified AKI in adult critically ill patients. influence wbNGAL concentrations [22], and this could be Nonetheless, wbNGAL concentrations increased by sepsis found regardless of the presence of AKI [42]. Kim et al. status irrespective of AKI occurrence. Thus, NGAL mea- [43] described that wbNGAL was significantly higher in sep- sures should be avoided to evaluate kidney function in tic patients with AKI regardless of their levels of procalcito- patients with sepsis. nin. In our subgroup of 20 patients with septic shock, 3 of them had no AKI, 2 stage 1, 4 stage 2, and 11 stage 3, with Data Availability median wbNGAL of 481 μg/L, 251 μg/L, 562 μg/L, and 685 μg/L, respectively. These data favour the hypothesis that The data used to support the findings of this study are inflammatory status could increase the level of wbNGAL and restricted by the Hospital de la Santa Creu i Sant Pau’s Ethic could be misleadingly interpreted as a tubular damage when Committee Board in order to protect patient privacy. Data Disease Markers 7 are available from authors (Dr. Antoni Jordi Betbese Roig, [7] A. Haase-Fielitz, R. Bellomo, P. Devarajan et al., “Novel and ajbetbese@santpau.cat) for researchers who meet the criteria conventional serum biomarkers predicting acute kidney injury for access to confidential data. in adult cardiac surgery-a prospective cohort study,” Critical Care Medicine, vol. 37, no. 2, pp. 553–560, 2009. [8] N. I. Shapiro, S. Trzeciak, J. E. Hollander et al., “A prospective, Conflicts of Interest multicenter derivation of a biomarker panel to assess risk of organ dysfunction, shock, and death in emergency department KN received congress fees from Abbot to present the prelim- patients with suspected sepsis,” Critical Care Medicine, vol. 37, inary results of this study at ESICM 2011. no. 1, pp. 96–104, 2009. [9] L. Kjeldsen, A. H. Johnsen, H. Sengelov, and N. Borregaard, Authors’ Contributions “Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase,” The Journal of Data collection was done by KN, JB, and MC. The study Biological Chemistry, vol. 268, no. 14, pp. 10425–10432, 1993. design, data analysis, and interpretation were independently [10] S. Y. Xu, K. Pauksen, and P. Venge, “Serum measurements of performed by AJB, JO-Ll, and MC. human neutrophil lipocalin (HNL) discriminate between acute bacterial and viral infections,” Scandinavian Journal of Acknowledgments Clinical and Laboratory Investigation, vol. 55, no. 2, pp. 125– 131, 1995. The study sponsor (Abbot) was responsible for providing the [11] J. B. Cowland, O. E. Sorensen, M. Sehested, and N. Borregaard, portable device (Triage® NGAL Test (Alere Diagnostics, for- “Neutrophil gelatinase-associated lipocalin is up-regulated in merly Inverness Medical Innovations)). The authors would human epithelial cells by IL-1β, but not by TNF-α,” The Jour- like to thank the staff in the ICU and biochemistry depart- nal of Immunology, vol. 171, no. 12, pp. 6630–6639, 2003. ments who generously participated in sample processing [12] J. B. Cowland and N. Borregaard, “Molecular characterization and analysis and whose work is essential to the completion and pattern of tissue expression of the gene for neutrophil of this study. gelatinase-associated lipocalin from humans,” Genomics, vol. 45, no. 1, pp. 17–23, 1997. [13] J. Mishra, K. Mori, Q. Ma et al., “Amelioration of ischemic Supplementary Materials acute renal injury by neutrophil gelatinase-associated lipoca- lin,” Journal of the American Society of Nephrology, vol. 15, Supplemental Table 1: diagnostic and overall accuracy of no. 12, pp. 3073–3082, 2004. wbNGAL for AKI prediction, RRT at admission, and sepsis [14] K. M. Schmidt-Ott, K. Mori, J. Y. Li et al., “Dual action of at admission. (Supplementary Materials) neutrophil gelatinase-associated lipocalin,” Journal of the American Society of Nephrology, vol. 18, no. 2, pp. 407– References 413, 2007. [15] K. Leelawat, S. Narong, J. Wannaprasert, and S. Leelawat, [1] S. Uchino, J. A. Kellum, R. Bellomo et al., “Acute renal failure “Serum NGAL to clinically distinguish cholangiocarcinoma in critically ill patients: a multinational, multicenter study,” from benign biliary tract diseases,” International Journal of Journal of the American Medical Association, vol. 294, no. 7, Hepatology, vol. 2011, Article ID 873548, 6 pages, 2011. pp. 813–818, 2005. [16] A. A. Zabron, V. M. Horneffer-van der Sluis, C. A. Wadsworth [2] VA/NIH Acute Renal Failure Trial Network, P. M. Palevsky, et al., “Elevated levels of neutrophil gelatinase-associated lipo- J. H. Zhang et al., “Intensity of renal support in critically ill calin in bile from patients with malignant pancreatobiliary dis- patients with acute kidney injury,” The New England Journal ease,” The American Journal of Gastroenterology, vol. 106, of Medicine, vol. 359, no. 1, pp. 7–20, 2008. no. 9, pp. 1711–1717, 2011. [3] W. Silvester, R. Bellomo, and L. Cole, “Epidemiology, manage- ment, and outcome of severe acute renal failure of critical ill- [17] T. M. Eagan, J. K. Damås, T. Ueland et al., “Neutrophil ness in Australia,” Critical Care Medicine, vol. 29, no. 10, gelatinase-associated lipocalin: a biomarker in COPD,” Chest, pp. 1910–1915, 2001. vol. 138, no. 4, pp. 888–895, 2010. [4] R. Bellomo, C. Ronco, J. A. Kellum, R. L. Mehta, P. Palevsky, [18] C. R. Parikh, A. Jani, J. Mishra et al., “Urine NGAL and IL-18 and Acute Dialysis Quality Initiative workgroup, “Acute renal are predictive biomarkers for delayed graft function following failure – definition, outcome measures, animal models, fluid kidney transplantation,” American Journal of Transplantation, therapy and information technology needs: the Second Inter- vol. 6, no. 7, pp. 1639–1645, 2006. national Consensus Conference of the Acute Dialysis Quality [19] D. N. Cruz, M. de Cal, F. Garzotto et al., “Plasma neutrophil Initiative (ADQI) Group,” Critical Care, vol. 8, no. 4, article gelatinase-associated lipocalin is an early biomarker for acute R204, 2004. kidney injury in an adult ICU population,” Intensive Care [5] R. L. Mehta, J. A. Kellum, S. V. Shah et al., “Acute Kidney Medicine, vol. 36, no. 3, pp. 444–451, 2010. Injury Network: report of an initiative to improve outcomes [20] Q. Liu and M. Nilsen-Hamilton, “Identification of a new acute in acute kidney injury,” Critical Care, vol. 11, no. 2, article phase protein,” Journal of Biological Chemistry, vol. 270, R31, 2007. no. 38, pp. 22565–22570, 1995. [6] Kidney Disease: Improving Global Outcomes (KDIGO) Acute [21] E. D. Siew, L. B. Ware, T. Gebretsadik et al., “Urine neutrophil Kidney Injury Work Group, “KDIGO clinical practice guide- gelatinase-associated lipocalin moderately predicts acute line for acute kidney injury,” Kidney International Supplement, kidney injury in critically ill adults,” Journal of the American vol. 2, pp. 1–138, 2012. Society of Nephrology, vol. 20, no. 8, pp. 1823–1832, 2009. 8 Disease Markers [22] M. Haase, P. Devarajan, A. Haase-Fielitz et al., “The outcome [36] S. Barbar, R. Clere-Jehl, A. Bourredjem et al., “Timing of renal- of neutrophil gelatinase-associated lipocalin-positive subclin- replacement therapy in patients with acute kidney injury and ical acute kidney injury. A multicenter pooled analysis of sepsis,” The New England Journal of Medicine, vol. 379, prospective studies,” Journal of the American College of Cardi- no. 15, pp. 1431–1442, 2018. ology, vol. 57, no. 17, pp. 1752–1761, 2011. [37] C. L. Dent, Q. Ma, S. Dastrala et al., “Plasma neutrophil [23] P. Kümpers, C. Hafer, A. Lukasz et al., “Serum neutrophil gelatinase-associated lipocalin predicts acute kidney injury, gelatinase-associated lipocalin at inception of renal replace- morbidity and mortality after pediatric cardiac surgery: a pro- ment therapy predicts survival in critically ill patients with spective uncontrolled cohort study,” Critical Care, vol. 11, acute kidney injury,” Critical Care, vol. 14, no. 1, article R9, no. 6, article R127, 2007. 2010. [38] T. L. Nickolas, M. J. O'Rourke, J. Yang et al., “Sensitivity and [24] P. B. Hjortrup, N. Haase, M. Wetterslev, and A. Perner, “Clin- specificity of a single emergency department measurement of ical review: predictive value of neutrophil gelatinase-associated urinary neutrophil gelatinase-associated lipocalin for diagnos- lipocalin for acute kidney injury in intensive care patients,” ing acute kidney injury,” Annals of Internal Medicine, vol. 148, Critical Care, vol. 17, no. 2, p. 211, 2013. no. 11, pp. 810–819, 2008. [25] M. Singer, C. S. Deutschman, C. W. Seymour et al., “The third [39] K. Damman, D. J. van Veldhuisen, G. Navis, A. A. Voors, and international consensus definitions for sepsis and septic shock H. L. Hillege, “Urinary neutrophil gelatinase associated lipoca- (sepsis-3),” Journal of the American Medical Association, lin (NGAL), a marker of tubular damage, is increased in vol. 315, no. 8, pp. 801–810, 2016. patients with chronic heart failure,” European Journal of Heart [26] M. Haase, R. Bellomo, P. Devarajan, P. Schlattmann, Failure, vol. 10, no. 10, pp. 997–1000, 2008. A. Haase-Fielitz, and NGAL Meta-analysis Investigator [40] B. S. Nielsen, N. Borregaard, J. R. Bundgaard, S. Timshel, Group, “Accuracy of neutrophil gelatinase-associated lipocalin M. Sehested, and L. Kjeldsen, “Induction of NGAL synthesis (NGAL) in diagnosis and prognosis in acute kidney injury: a in epithelial cells of human colorectal neoplasia and inflam- systematic review and meta-analysis,” American Journal of matory bowel diseases,” Gut, vol. 38, no. 3, pp. 414–420, Kidney Diseases, vol. 54, no. 6, pp. 1012–1024, 2009. 1996. [27] J. Mishra, C. Dent, R. Tarabishi et al., “Neutrophil gelatinase- [41] S. M. Bagshaw, M. Bennett, M. Haase et al., “Plasma and urine associated lipocalin (NGAL) as a biomarker for acute renal neutrophil gelatinase-associated lipocalin in septic versus non- injury after cardiac surgery,” The Lancet, vol. 365, no. 9466, septic acute kidney injury in critical illness,” Intensive Care pp. 1231–1238, 2005. Medicine, vol. 36, no. 3, pp. 452–461, 2010. [28] R. Hirsch, C. Dent, H. Pfriem et al., “NGAL is an early predic- [42] J. Vanmassenhove, G. Glorieux, N. Lameire et al., “Influence of tive biomarker of contrast-induced nephropathy in children,” severity of illness on neutrophil gelatinase-associated lipocalin Pediatric Nephrology, vol. 22, no. 12, pp. 2089–2095, 2007. performance as a marker of acute kidney injury: a prospective cohort study of patients with sepsis,” BMC Nephrology, vol. 16, [29] D. Bolignano, G. Coppolino, A. Romeo, A. Lacquaniti, and no. 1, p. 18, 2015. M. Buemi, “Neutrophil gelatinase-associated lipocalin levels [43] H. Kim, M. Hur, D. N. Cruz, H. W. Moon, and Y. M. Yun, in chronic haemodialysis patients,” Nephrology, vol. 15, no. 1, “Plasma neutrophil gelatinase-associated lipocalin as a bio- pp. 23–26, 2010. marker for acute kidney injury in critically ill patients with [30] D. S. Wheeler, P. Devarajan, Q. Ma et al., “Serum neutrophil suspected sepsis,” Clinical Biochemistry, vol. 46, no. 15, gelatinase-associated lipocalin (NGAL) as a marker of acute pp. 1414–1418, 2013. kidney injury in critically ill children with septic shock,” Crit- [44] N. J. Glassford, A. G. Schneider, S. Xu et al., “The nature and ical Care Medicine, vol. 36, no. 4, pp. 1297–1303, 2008. discriminatory value of urinary neutrophil gelatinase- [31] M. Zappitelli, K. K. Washburn, A. A. Arikan et al., “Urine neu- associated lipocalin in critically ill patients at risk of acute kid- trophil gelatinase-associated lipocalin is an early marker of ney injury,” Intensive Care Medicine, vol. 39, no. 10, pp. 1714– acute kidney injury in critically ill children: a prospective 1724, 2013. cohort study,” Critical Care, vol. 11, no. 4, article R84, 2007. [45] B. Wang, G. Chen, J. Zhang, J. Xue, Y. Cao, and Y. Wu, [32] K. Makris, N. Markou, E. Evodia et al., “Urinary neutrophil “Increased neutrophil gelatinase associated lipocalin is associ- gelatinase-associated lipocalin (NGAL) as an early marker of ated with mortality and multiple organ dysfunction syndrome acute kidney injury in critically ill multiple trauma patients,” in severe sepsis and septic shock,” Shock, vol. 44, no. 3, Clinical Chemistry and Laboratory Medicine, vol. 47, no. 1, pp. 234–238, 2015. pp. 79–82, 2009. [46] M. Wang, Q. Zhang, X. Zhao, G. Dong, and C. Li, “Diagnostic [33] J.-M. Constantin, E. Futier, S. Perbet et al., “Plasma neutrophil and prognostic value of neutrophil gelatinase-associated lipo- gelatinase-associated lipocalin is an early marker of acute kid- calin, matrix metalloproteinase-9, and tissue inhibitor of ney injury in adult critically ill patients: a prospective study,” matrix metalloproteinases-1 for sepsis in the emergency Journal of Critical Care, vol. 25, no. 1, pp. 176.e1–176.e6, 2010. department: an observational study,” Critical Care, vol. 18, [34] S. Gaudry, D. Hajage, F. Schortgen et al., “Initiation strategies no. 6, p. 634, 2014. for renal-replacement therapy in the intensive care unit,” The [47] K. Kashani, A. al-Khafaji, T. Ardiles et al., “Discovery and New England Journal of Medicine, vol. 375, no. 2, pp. 122– validation of cell cycle arrest biomarkers in human acute kid- 133, 2016. ney injury,” Critical Care, vol. 17, no. 1, article R25, 2013. [35] A. Zarbock, J. A. Kellum, C. Schmidt et al., “Effect of early vs [48] J. H. Westhoff, B. Tönshoff, S. Waldherr et al., “Urinary tissue delayed initiation of renal replacement therapy on mortality inhibitor of metalloproteinase-2 (TIMP-2) • insulin-like in critically ill patients with acute kidney injury the ELAIN growth factor-binding protein 7 (IGFBP7) predicts adverse randomized clinical trial,” Journal of the American Medical outcome in pediatric acute kidney injury,” PLoS One, vol. 10, Association, vol. 315, no. 20, pp. 2190–2199, 2016. no. 11, article e0143628, 2015. Disease Markers 9 [49] L. M. Rizo-Topete, M. H. Rosner, and C. Ronco, “Acute kidney injury risk assessment and the nephrology rapid response team,” Blood Purification, vol. 43, no. 1-3, pp. 82–88, 2017. [50] M. Meersch, C. Schmidt, A. Hoffmeier et al., “Prevention of cardiac surgery-associated AKI by implementing the KDIGO guidelines in high risk patients identified by biomarkers: the PrevAKI randomized controlled trial,” Intensive Care Medi- cine, vol. 43, no. 11, pp. 1551–1561, 2017. [51] I. Göcze, D. Jauch, M. Götz et al., “Biomarker-guided interven- tion to prevent acute kidney injury after major surgery: the prospective randomized BigpAK Study,” Annals of Surgery, vol. 267, no. 6, pp. 1013–1020, 2018. [52] P. M. Honore, H. B. Nguyen, M. Gong et al., “Urinary tissue inhibitor of metalloproteinase-2 and insulin-like growth factor-binding protein 7 for risk stratification of acute kidney injury in patients with sepsis,” Critical Care Medicine, vol. 44, no. 10, pp. 1851–1860, 2016. [53] M. Cuartero, J. Ballús, J. Sabater et al., “Cell-cycle arrest bio- markers in urine to predict acute kidney injury in septic and non-septic critically ill patients,” Annals of Intensive Care, vol. 7, no. 1, p. 92, 2017. [54] S. H. Hou, D. A. Bushinsky, J. B. Wish, J. J. Cohen, and J. T. Harrington, “Hospital-acquired renal insufficiency: a prospec- tive study,” The American Journal of Medicine, vol. 74, no. 2, pp. 243–248, 1983. [55] K. Nash, A. Hafeez, and S. Hou, “Hospital-acquired renal insufficiency,” American Journal of Kidney Diseases, vol. 39, no. 5, pp. 930–936, 2002. [56] H. E. Wang, P. Muntner, G. M. Chertow, and D. G. Warnock, “Acute kidney injury and mortality in hospitalized patients,” American Journal of Nephrology, vol. 35, no. 4, pp. 349–355, 2012.