Jäntti et al. Ann. Intensive Care (2021) 11:25 https://doi.org/10.1186/s13613-021-00814-8 RESEARCH Open Access Predictive value of plasma proenkephalin and neutrophil gelatinase‑associated lipocalin in acute kidney injury and mortality in cardiogenic shock Toni Jäntti1* , Tuukka Tarvasmäki1, Veli‑Pekka Harjola2, Kari Pulkki3, Heidi Turkia3, Tuija Sabell1, Heli Tolppanen1, Raija Jurkko1, Mari Hongisto2, Anu Kataja4, Alessandro Sionis5, Jose Silva‑Cardoso6, Marek Banaszewski7, Salvatore DiSomma8, Alexandre Mebazaa9, Mikko Haapio10, Johan Lassus1 and for the CardShock investigators Abstract Background: Acute kidney injury (AKI) is a frequent form of organ injury in cardiogenic shock. However, data on AKI markers such as plasma proenkephalin (P-PENK) and neutrophil gelatinase-associated lipocalin (P-NGAL) in cardio‑ genic shock populations are lacking. The objective of this study was to assess the ability of P-PENK and P-NGAL to predict acute kidney injury and mortality in cardiogenic shock. Results: P-PENK and P-NGAL were measured at different time points between baseline and 48 h in 154 patients from the prospective CardShock study. The outcomes assessed were AKI defined by an increase in creatinine within 48 h and all-cause 90-day mortality. Mean age was 66 years and 26% were women. Baseline levels of P-PENK and P-NGAL (median [interquartile range]) were 99 (71–150) pmol/mL and 138 (84–214) ng/mL. P-PENK > 84.8 pmol/mL and P-NGAL > 104 ng/mL at baseline were identified as optimal cut-offs for AKI prediction and independently associ‑ ated with AKI (adjusted HRs 2.2 [95% CI 1.1–4.4, p = 0.03] and 2.8 [95% CI 1.2–6.5, p = 0.01], respectively). P-PENK and P-NGAL levels at baseline were also associated with 90-day mortality. For patients with oliguria < 0.5 mL/kg/h for > 6 h before study enrollment, 90-day mortality differed significantly between patients with low and high P-PENK/P-NGAL at baseline (5% vs. 68%, p < 0.001). However, the biomarkers provided best discrimination for mortality when meas‑ ured at 24 h. Identified cut-offs of P-PENK24h > 105.7 pmol/L and P-NGAL24h > 151 ng/mL had unadjusted hazard ratios of 5.6 (95% CI 3.1–10.7, p < 0.001) and 5.2 (95% CI 2.8–9.8, p < 0.001) for 90-day mortality. The association remained significant despite adjustments with AKI and two risk scores for mortality in cardiogenic shock. Conclusions: High levels of P-PENK and P-NGAL at baseline were independently associated with AKI in cardiogenic shock patients. Furthermore, oliguria before study inclusion was associated with worse outcomes only if combined with high baseline levels of P-PENK or P-NGAL. High levels of both P-PENK and P-NGAL at 24 h were found to be strong and independent predictors of 90-day mortality. Trial registration: NCT01374867 at www.clini​caltr​ials.gov, registered 16 Jun 2011—retrospectively registered Keywords: Cardiogenic shock, Acute kidney injury, AKI, Mortality, Prognosis, Proenkephalin, PENK, NGAL Background *Correspondence: toni.jantti@fimnet.fi 1 Department of Cardiology, Heart and Lung Center, Helsinki University Cardiogenic shock is a severe state of hypoperfusion, Hospital, University of Helsinki, 00029 HUS Helsinki, Finland caused by low cardiac output, resulting in end-organ Full list of author information is available at the end of the article hypoperfusion and congestion [1]. Despite recent The list of CardShock investigators is provided in the Acknowledgments © The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat​iveco​ mmons​.org/licen​ses/by/4.0/. Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 2 of 15 advances in treatment, cardiogenic shock still carries a shock. Written informed consent was obtained from the high mortality rate of 40–60% [2]. Acute kidney injury patient or next of kin if the patient was unable to give (AKI) is a frequent form of organ injury in cardiogenic consent. To be included in the study, the patients needed shock, affecting ~ 30% of patients surviving the initial to have (1) an acute cardiac cause for the shock, (2) sys- stage [3]. Traditionally, serum creatinine and urine out- tolic blood pressure < 90 mmHg (after adequate fluid put are used to define AKI [4], but these markers are challenge) for 30 min or a need for vasopressor therapy limited by delayed changes following kidney injury and to maintain systolic blood pressure > 90 mmHg, and (3) have low sensitivity and specificity [5]. Prediction of AKI signs of hypoperfusion (one or several of the following: with biomarkers such as neutrophil gelatinase-associated altered mental status, cold periphery, oliguria < 0.5 mL/ lipocalin (NGAL) and proenkephalin (PENK) have been kg/h for the previous 6 h, or blood lactate > 2 mmol/L). studied in critically ill patients [6, 7], but similar studies Exclusion criteria were shock caused by ongoing hemo- among cardiogenic shock patients are lacking. dynamically significant arrhythmia or shock after car- PENK is a small endogenous opioid peptide, which is diac or non-cardiac surgery. Baseline characteristics and cleaved from the common endogenous opioid precursor, previous medical history were recorded. Biochemical and thus can be used as a surrogate marker for the activ- and clinical findings as well as hemodynamic param- ity of the endogenous opioid system [8]. The opioid sys- eters were documented at detection of shock and at pre- tem has been shown to have depressive effects on cardiac specified time points up to 48 h after inclusion. Patients and renal function, and has been implicated in the prog- were treated according to local practice and treatment nosis of myocardial infarction [9]. PENK and the opioid and procedures were registered. Local investigators were system have also been shown to be associated with worse responsible for determining the etiology of cardiogenic outcomes both in acute [10, 11] and chronic heart fail- shock. Acute coronary syndrome etiology was defined ure [11]. PENK has been shown to have a strong nega- as shock caused by myocardial infarction (with or with- tive correlation with eGFR measured by iohexol and is out ST-elevation). Echocardiography was performed per therefore a kidney biomarker of glomerular function [12]. protocol at study inclusion. Coronary angiography was PENK was selected as biomarker for this study as it has carried out in 128 patients, of which 95 (74%) were done also been shown to predict AKI in patients with severe before study inclusion, and 24 (19%) within 12 h after sepsis or shock [13]. study inclusion. NGAL is a member of the lipocalin protein family which is expressed in the kidney tubular structures, as Test methods well as other tissues [14]. NGAL is rapidly released in Plasma samples were collected at baseline in 178/219 response to tubular damage [15, 16], and is one of the patients enrolled in the CardShock study. Additionally, most extensively studied biomarkers used for AKI pre- serial blood samples were collected at 12 h, 24 h, 48 h diction [17]. (all ± 3 h) and at discharge from ICU (5–10 days) (see Our aim was to assess the predictive value of P-PENK Additional file 1: Fig. 1 for a diagram). A baseline sample and P-NGAL for AKI within 48 h after admission and and at least one additional sample within 48 h were avail- 90-day all-cause mortality in patients with cardiogenic able for 154 patients, which constitute the study popula- shock. tion in this report. After collection, the blood sample was centrifuged, and separated plasma was immediately fro- Methods tions in plasma were determined using the ­Sphingotest® zen in aliquots and stored at − 80 °C. PENK concentra- Study design The CardShock study (NCT01374867 at www.clini​caltr​ penKid immunoassay (SphingoTec GmbH, Hennigsdorf, ials.gov) was a European prospective, observational, mul- Germany). Plasma NGAL concentrations were deter- ticenter and multinational study on cardiogenic shock. mined with a commercially available particle-enhanced Patients were recruited between October 2010 and turbidimetric immunoassay (BioPorto Diagnostics A/S, December 2012 from emergency departments, cardiac Hellerup, Denmark). Creatinine, C-reactive protein and intensive care units, as well as catheterization labora- (CRP), high-sensitivity troponin T, N-terminal pro-B- tories from nine tertiary hospitals in eight countries. For type natriuretic peptide (NT-proBNP), alanine ami- a detailed design and description of the study population notransferase, alkaline phosphatase, total bilirubin and please see Harjola et al. [18]. cystatin C were analyzed from the plasma samples using commercially available standard kits (Abbott Laborato- Participants ries, Abbott Park, IL, USA for cystatin C, Roche Diag- Consecutive patients older than 18 years were enrolled nostics, Basel, Switzerland for all other tests) at a central in the study within 6 h from identification of cardiogenic accredited laboratory (ISLAB, Kuopio, Finland). Arterial Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 3 of 15 blood gas analysis (including arterial pH and lactate), multivariable Cox regression models to calculate likelihood hemoglobin and leucocytes were analyzed by local labo- ratios, using a significance of > 0.10 for elimination and a ratories. Estimated glomerular filtration rate (eGFR) was significance of < 0.05 for retention. The additive value of a calculated from creatinine values using the Chronic Kid- variable in the multivariable Cox regression models was ney Disease Epidemiology Collaboration equation [19]. assessed using likelihood ratio test for nested models. The AKI was defined and staged according to the Kidney Dis- proportionality of hazards assumption was tested using ease: Improving Global Outcomes (KDIGO) criteria [4]. log-minus-log plots. Hazard ratios (HR) for Cox regres- For AKI staging at baseline, a recently described staging sion analyses are shown with 95% confidence intervals which includes biomarker levels was used [5]. Urine out- (CIs). Discriminative capability of P-PENK and P-NGAL put was recorded at 6, 12, 18, and 24 h and used for urine at different time points was assessed using the area under output-based definitions of AKI. Main outcomes investi- receiver-operating characteristics curve (AUC). The statis- gated in this study were AKI defined by an increase cre- tical significance of changes between different time points atinine of more than 26,5 μmol/L within 48 h of study in P-PENK and P-NGAL levels was assessed using Wil- inclusion ­(AKIcrea48h) and all-cause 90-day mortality. coxon signed ranks test. Differences in P-PENK, P-NGAL, Assessment of ­AKIcrea48h was based on creatinine levels and creatinine trajectories between different groups were from baseline until 48 h, and the highest increase within assessed using linear mixed modeling. Two-sided p val- this time was used for staging. AKI staging by urine out- ues < 0.05 were regarded as statistically significant. Data put was categorized according to the lowest urine out- were analyzed using the SPSS statistical package, version 23 put for a time interval within the first 24 h. Subclinical (IBM Corp, Armonk, NY, USA). STATA (StataCorp. 2019. AKI was defined as positive biomarker (either P-PENK Stata Statistical Software: Release 16. College Station, TX: or P-NGAL > cut-off ) but no ­AKIcrea48h. Vital status of StataCorp LLC) was used to calculate c-statistics for the the patients during follow-up was determined through different Cox regression models. direct contact with the patient or next of kin, or through population and hospital registers. Patients with missing Results plasma samples were left out of the analysis. Two patients Patient characteristics were lost to follow-up and were left out of the survival Baseline characteristics of patients (n = 154) stratified by analyses. The study was approved by local ethics commit- P-PENK and P-NGAL quartiles at baseline are presented tees and conducted in accordance with the Declaration in Table 1. The mean age of all patients in the cohort was of Helsinki. Two published risk scores for cardiogenic 66 years and 26% of the patients were women. The main shock, CardShock risk Score [18] and IABP SHOCK cause of cardiogenic shock was acute coronary syndrome, II score [20] were calculated for every patient to assess which accounted for 81% of the cases. The overall 90-day whether P-PENK or P-NGAL provided additive value in mortality was 38%. AKI was observed in 47/154 (31%) risk prediction. patients within 48 h of baseline. Most patients (n = 29) had AKI already within 12 h from baseline. Median base- Statistical analysis line P-PENK (n = 152) was 105 (IQR 71–167) pmol/mL, Results are presented as number (n) and percentage (%), and median baseline P-NGAL (n = 146) was 138 ng/mL mean and standard deviation (SD), or median and inter- (IQR 84–214). Patients with P-PENK > median and patients quartile range (IQR) as appropriate. Group comparisons with P-NGAL > median levels had lower eGFR and hemo- were performed using Fisher’s exact test for categorical globin at baseline, and had higher levels of creatinine, variables, and Student’s t test or Mann–Whitney U test for NT-proBNP, alkaline phosphatase and lactate. Oliguria continuous variables, as appropriate. Associations between prior to enrollment was more frequent in patients with continuous variables were assessed using Spearman cor- biomarker levels > median and they had higher CardShock relations. Youden index was used to select the cut-offs of and IABP II Shock risk score points. Patients with PENK/ P-PENK and P-NGAL used for AKI and mortality predic- NGAL > median had a higher incidence of AKI within 48 h tion. To determine variables independently associated with and their mortality was higher. STARD flow diagrams of P-PENK and P-NGAL levels univariable general linear study participants are provided as Additional file 1: Fig. 2. models were constructed using log-normalized P-PENK and P-NGAL levels as dependent variables. Differences in Time‑related changes in P‑PENK and P‑NGAL levels survival between groups were assessed comparing Kaplan– in patients with/without A ­ KIcrea48h and survivors/ Meier survival curves using log-rank test. Cox regres- nonsurvivors sion models were used to assess associations of variables Figure 1 shows the trajectories of P-PENK, P-NGAL and with AKI and mortality in uni- and multivariable models. creatinine between baseline and 48 h in patients with Forward and backward selection of variables was used in and without A­ KIcrea48h and Fig. 2 shows P-PENK and Table 1 Patient characteristics stratified by quartiles of P-PENK and P-NGAL at baseline All P-PENK Q1 P-PENK Q2 P-PENK Q3 P-PENK Q4 p P-NGAL Q1 P-NGAL Q2 P-NGAL Q3 P-NGAL Q4 p N = 154 N = 38 N = 38 N = 38 N = 38 N = 36 N = 37 N = 37 N = 36 Age, years; (SD) 66 (12) 62 (10) 62 (15) 69 (10) 72 (11) < 0.001 62 (13) 67 (12) 69 (11) 68 (12) 0.09 Women 40 (26%) 8 (21%) 9 (24%) 10 (26%) 13 (34%) 0.66 5 (14%) 11 (30%) 11 (30%) 10 (28%) 0.34 BMI, kg/m2; (SD) 27.1 (4.1) 27.5 (4.5) 27.1 (3.6) 26.6 (4.1) 26.9 (4.3) 0.84 27.3 (4.0) 25.6 (3.5) 28.0 (4.6) 27.3 (4.1) 0.09 Jäntti et al. Ann. Intensive Care Medical history Hypertension 96 (62%) 20 (53) 20 (53) 28 (64%) 26 (68%) 0.14 16 (44%) 21 (57%) 30 (81%) 25 (69%) 0.008 Coronary artery 51 (33%) 6 (16%) 11 (29%) 15 (40%) 19 (50%) 0.01 10 (28%) 10 (27%) 16 (43%) 15 (42%) 0.31 disease Previous 39 (25%) 6 (16%) 9 (24%) 13 (24%) 11 (29%) 0.32 9 (25%) 7 (19%) 12 (32%) 11 (31%) 0.55 (2021) 11:25 myocardial infarction or CABG Heart failure 24 (16%) 2 (5%) 5 (13%) 8 (21%) 9 (24%) 0.13 2 (6%) 5 (14%) 6 (16%) 11 (31%) 0.03 Diabetes mel‑ 43 (28%) 7 (18%) 13 (34%) 11 (29%) 12 (32%) 0.48 7 (19%) 5 (14%) 17 (46%) 13 (36%) 0.007 litus Renal insuf‑ 17 (11%) 1 (3%) 0 (0%) 5 (13%) 11 (29%) < 0.001 0 (0%) 1 (3%) 6 (16%) 10 (28%) 0.001 ficiency Smoking 62 (41%) 17 (45%) 22 (58%) 14 (38%) 9 (24%) 0.02 18 (50%) 15 (42%) 18 (49%) 8 (22%) 0.06 Medications in use at admis‑ sion ACEI or ARB 63 (41%) 14 (37%) 11 (29%) 19 (50%) 18 (49%) 0.20 14 (39%) 10 (27%) 18 (49%) 17 (49%) 0.19 Diuretics 43 (28%) 5 (13%) 10 (26%) 12 (32%) 15 (41%) 0.06 6 (17%) 7 (19%) 15 (41%) 15 (43%) 0.02 Clinical presenta‑ tion at baseline Oliguria 79 (52%) 18 (49%) 15 (40%) 19 (50%) 26 (72%) 0.04 14 (40%) 16 (43%) 20 (54%) 28 (80%) 0.003 CardShock risk 4.2 (1.8) 3.1 (1.4) 4.1 (1.7) 4.2 (1.8) 5.5 (1.8) < 0.001 3.1 (1.6) 3.8 (1.6) 4.6 (1.5) 5.6 (1.7) < 0.001 score, points; mean (SD) IABP II SHOCK 2.2 (1.7) 1.5 (1.3) 2.1 (1.7) 2.4 (1.8) 2.8 (1.5) 0.04 1.1 (1.2) 2.1 (1.7) 2.6 (1.8) 3.4 (1.2) < 0.001 risk score, points; mean (SD) ACS etiology of 124 (81%) 31 (82%) 31 (82%) 32 (84%) 28 (74%) 0.77 29 (81%) 30 (81%) 32 (87%) 25 (69%) 0.33 cardiogenic shock Resuscitated 44 (29%) 10 (26%) 8 (22%) 12 (32%) 14 (37%) 0.51 7 (19%) 12 (32%) 8 (22%) 14 (39%) 0.24 Page 4 of 15 Table 1 (continued) All P-PENK Q1 P-PENK Q2 P-PENK Q3 P-PENK Q4 p P-NGAL Q1 P-NGAL Q2 P-NGAL Q3 P-NGAL Q4 p N = 154 N = 38 N = 38 N = 38 N = 38 N = 36 N = 37 N = 37 N = 36 Mean arterial 57 (11) 59 (10) 56 (11) 56 (11) 58 (10) 0.63 57 (10) 59 (12) 57 (10) 58 (11) 0.85 pressure, mmHg; mean (SD) Jäntti et al. Ann. Intensive Care LVEF, % (SD) 33 (13) 35 (15) 30 (11) 36 (13) 30 (13) 0.10 36 (14) 32 (11) 32 (11) 31 (15) 0.29 Laboratory test results at baseline eGFR, mL/ 65 (28) 90 (22) 70 (20) 61 (24) 39 (22) < 0.001 89 (17) 72 (24) 59 (22) 37 (22) < 0.001 (2021) 11:25 min/1.73 m2; mean (SD) Creatinine mg/ 1.12 (0.87–1.54) 0.76 (0.63–1.01) 1.10 (0.90–1.31) 1.18 (0.93–1.47) 1.69 (1.29–2.56) < 0.001 0.85 (0.67–1.01) 1.00 (0.86–1.26) 1.22 (1.04–1.41) 1.93 (1.48–2.61) < 0.001 dL; median (IQR) NT-proBNP, 2247 (550–8253) 1125 (242–3930) 2582 (630–8414) 1701 (251–7390) 6949 (2342– < 0.001 1666 (371–5434) 1611 (425–5558) 2475 (678–7162) 8811 (2485– < 0.001 ng/L; median 27,361) 30,442) (IQR) hs-TnT, ng/L; 2275 (379–5985) 2571 (531–6643) 2439 (441–4933) 1635 (310–6820) 1857 (124–8061) 0.92 1187 (327–5336) 2862 (573–9346) 2601 (603–7689) 1597 (136–5616) 0.39 median (IQR) CRP, mg/L; 13 (4–48) 7 (4–45) 18 (5–73) 7 (3–46) 25 (9–61) 0.13 8 (4–40) 6 (3–37) 19 (5–46) 27 (9–101) 0.02 median (IQR) Leukocytes 13.6 (5.2) 14.2 (6.1) 14.4 (4.7) 13.4 (4.5) 12.7 (5.4) 0.48 11.2 (4.2) 14.9 (5.1) 14.1 (4.6) 13.3 (6.0) 0.02 (10E9); mean (SD) Hemoglobin, 13.0 (2.4) 13.2 (1.9) 13.2 (2.6) 13.6 (2.0) 11.9 (2.5) 0.008 12.9 (2.3) 13.5 (1.9) 13.3 (2.5) 11.8 (2.5) 0.008 (g/dL); mean (SD) Alanine 42 (20–87) 29 (17–63) 43 (21–101) 42 (19–81) 54 (20–164) 0.18 24 (14–63) 46 (25–77) 43 (20–85) 83 (17–260) 0.04 aminotrans‑ ferase, (IU/L); median (IQR) Alkaline 61 (49–81) 55 (42–71) 64 (53–87) 62 (49–78) 65 (20–112) 0.048 53 (40–74) 62 (49–81) 70 (59–91) 64 (53–85) 0.02 phosphatase, (IU/L); median (IQR) Lactate, 2.6 (1.6–5.2) 1.6 (1.0–2.7) 2.5 (2.0–4.6) 2.7 (1.7–5.9) 3.9 (2.6–8.4) < 0.001 1.8 (1.1–2.5) 2.6 (1.5–3.9) 2.6 (2.1–6.0) 5.2 (2.7–8.6) < 0.001 mmol/L; median (IQR) Page 5 of 15 Jäntti et al. Ann. Intensive Care (2021) 11:25 Table 1 (continued) All P-PENK Q1 P-PENK Q2 P-PENK Q3 P-PENK Q4 p P-NGAL Q1 P-NGAL Q2 P-NGAL Q3 P-NGAL Q4 p N = 154 N = 38 N = 38 N = 38 N = 38 N = 36 N = 37 N = 37 N = 36 P-PENK, pmol/ 99 (71–150) 54 (42–63) 81 (75–88) 120 (108–136) 211 (180–271) < 0.001 68 (55–85) 102 (68–133) 103 (80–139) 211 (136–261) < 0.001 mL; median (IQR) P-NGAL, ng/ 138 (84–214) 78 (55–113) 132 (81–169) 146 (99–182) 287 (139–452) < 0.001 61 (51–72) 118 (97–126) 164 (150–176) 323 (259–462) < 0.001 mL; median (IQR) AKI by 48 h 47 (31%) 5 (13%) 12 (31%) 17 (45%) 12 (30%) 0.03 4 (11%) 13 (35%) 13 (35%) 16 (44%) 0.02 90-day mortality; 58 (38%) 7 (18%) 13 (35%) 15 (41%) 22 (58%) 0.005 7 (19%) 13 (35%) 17 (49%) 21 (58%) 0.005 N = (%) Results shown as n (%) for categorical and mean (SD) or median (IQR) for continuous variables ACEI angiotensin-converting enzyme inhibitor, ACS acute coronary syndrome, AKI acute kidney injury, ARB angiotensin receptor blocker, BMI body mass index, CABG coronary artery bypass grafting, CRP C-reactive protein, eGFR estimated glomerular filtration rate, hs-TnT high-sensitivity troponin T, IQR interquartile range, LVEF left ventricular ejection fraction, NGAL neutrophil gelatinase-associated lipocalin, NT-proBNP N-terminal pro-b-type natriuretic peptide, PENK proenkephalin, SD standard deviation Page 6 of 15 Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 7 of 15 P-NGAL trajectories in survivors and nonsurvivors. In Association of baseline levels of P‑PENK and P‑NGAL nonsurvivors, the increase from baseline to 24 h was sta- with 90‑day mortality tistically significant for P-NGAL (p = 0.003), but not for The relationships between AKI, 90-day mortality and P-PENK (p = 0.58). There were statistically significant high P-PENK0h and/or P-NGAL0h are depicted as a differences in both P-PENK and P-NGAL levels with Venn diagram in Fig. 4. 90-day mortality differed sig- respect to time and group, and an interaction between nificantly between AKI stages at admission (7% for the groups with time was observed. stage 0, 19% for stage 1S, 5% for stage 1A and 68% for stage 1B, p < 0.001). Both P-PENK0h > 84.8 pmol/ mL and P-NGAL0h > 104 ng/mL were associated with Association of baseline P‑PENK and P‑NGAL levels with AKI, higher 90-day mortality (Table 3). Figure 5 shows the renal outcomes, and interventions survival curves for patients with and without A ­ KIcrea48h AUCs were calculated at each time point for A ­ KIcrea48h separated by baseline P-PENK (panel A) and baseline and 90-day all-cause mortality to assess the discrimina- P-NGAL (panel B) higher (subclinical AKI) or lower tory capabilities of both biomarkers (Additional file 1: than the optimal cut-off for AKI prediction (84.8 pmol/ Table 1). At baseline, the predictive value of P-PENK0h mL for P-PENK0h and 104 ng/mL for P-NGAL0h). Using and P-NGAL0h for ­AKIcrea48h was only moderate (AUCs the cut-offs predictive of AKI, P-NGAL0h > 104 ng/mL 0.621 and 0.664, respectively) (Additional file 1: Table 1). was able to further stratify patients with and without Compared to baseline measurements, biomarker values ­AKIcrea48h further into high or low mortality risk groups at later time points had better performance in prediction (Fig. 5). of ­AKIcrea48h. However, since most AKI cases occurred within 12 h from baseline, P-PENK and P-NGAL meas- urements in time points later than baseline were not considered for AKI prediction. The optimal biomarker Utility of P‑PENK and P‑NGAL at later time points cut-off value at baseline for predicting ­ AKIcrea48h was for mortality risk stratification in cardiogenic shock found to be 84.8 pmol/mL for P-PENK0h and 104 ng/mL Although high levels of both P-PENK and P-NGAL were for P-NGAL0h. In univariable Cox regression analysis, associated with higher mortality at all time points exam- both P-PENK0h > 84.8 pmol/mL and P-NGAL0h > 104 ng/ ined, the AUC for mortality was highest for both bio- mL were associated with the development of AKI markers measured at 24 h (Additional file 1: Table 2). The (Table 2). Multivariable regression modeling identi- AUCs for 90-day mortality for CardShock risk score and fied high P-PENK0h (adj HR 2.2 [95% CI 1.1–4.4] and IABP SHOCK II risk score at 24 h were 0.778 and 0.707. P-NGAL0h (adj HR 2.8 [95% CI 1.2–6.5]) as being inde- The optimal cut-off for mortality was 105.7 pmol/ pendently associated with A ­ KIcrea48h (Table 2). Table 3 mL for P-PENK24h, and 151 ng/mL for P-NGAL24h. shows renal outcomes and differences in interventions Survival curves using these cut-offs showed that for using these P-PENK0h and P-NGAL0h cut-offs. There patients with P-PENK24h > 105.7 pmol/mL 90-day mor- were 51 (34%) patients, who fulfilled the criteria for AKI tality was 68.2% compared with 17.4% for patients with stage 1S. In patients with oliguria at baseline but with- P-PENK24h < 105.7 pmol/mL. Similarly, 90-day mortality out high PENK or NGAL (stage 1A) only 2/16 (13%) of for patients with P-NGAL24h > 151 ng/mL was 63.5% com- patients developed AKIcrea48h, compared with 13/51 pared with 17.7% for patients with P-NGAL24h < 151 ng/ (26%) of patients in stage 1S and 33/63 (48%) in stage 1B mL (p < 0.001 for all, Additional file 1: Fig. 2). (p = 0.001). Only 1/21 (5%) of patients in stage 0 devel- In univariable Cox regression, both oped AKIcrea48h. High P-PENK0h was associated with P-PENK24h > 105.7 pmol/mL and P-NGAL24h > 151 ng/ very early AKI (< 24 h), whereas high P-NGAL0h was mL were strongly associated with mortality (Table 4). also associated with AKI detected at 24–48 h (p < 0.05 In multivariable analysis, the associations of both for comparisons between subsets). Both high P-PENK0h P-PENK24h > 105.7 pmol/mL and P-NGAL24h > 151 ng/ and high P-NGAL0h were associated with AKI as defined mL with higher mortality were found to be independent by low urine output according to the KDIGO criteria, of both CardShock risk score and IABP II Shock score. whereas only high P-NGAL0h was associated with AKI Further adjustment of the multivariable models with as defined by an increase in Cystatin C. High P-NGAL0h ­AKIcrea48h did not change the independent association was also associated with the use of renal replacement of P-PENK24h with 90-day all-cause mortality, whereas therapy, whereas high P-PENK0h was not. Figure 3 shows P-NGAL24h was no longer significantly associated with Kaplan–Meier curves for the development of ­AKIcrea48h 90-day mortality after adjusting for both CardShock risk for high P-PENK and high P-NGAL groups. Cross tabu- score and ­AKIcrea48h. Cross tabulation of P-PENK24h and lation of high P-PENK and P-NGAL levels with respect P-NGAL24h with respect to 90-day mortality can be seen to ­AKIcrea48h can be seen in Additional file 1: Table 3. in Additional file 1: Table 3. Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 8 of 15 Fig. 1 Biomarker medians at different time points separated by occurrence of AKI within 48 h. a P-PENK. b P-NGAL. c Creatinine. Error bars = 95% confidence interval Determinants of P‑PENK and P‑NGAL in cardiogenic shock and 16.4; p < 0.001 for P-NGAL). In addition, body mass Results from a Spearman correlation analysis with base- index (F-statistic 6.0, p = 0.02) and alkaline phosphatase line variables and levels of P-PENK and P-NGAL are (F-statistic 5.0, p = 0.03 were independent associates shown in Additional file 1: Table 2. Markers of renal of P-NGAL levels. eGFR alone accounted for 36% of function had moderate to strong correlations with both the observed variance of P-PENK and 39% of P-NGAL P-PENK and P-NGAL at baseline. P-PENK and P-NGAL variance. were also intercorrelated. General linear model analysis showed that the biomarkers had strongest associations with eGFR (F-statistic 82.8 for P-PENK and 84.9 for Discussion P-NGAL; p < 0.001 for both) and log-normalized lactate This multicenter cohort study is the first to investigate levels at baseline (F-statistic 5.0; p = 0.03 for P-PENK P-PENK and P-NGAL in cardiogenic shock of various etiologies. We describe the early kinetics of P-PENK Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 9 of 15 Fig. 2 P-PENK and P-NGAL median levels at different time points separated by 90-day mortality. a P-PENK. b P-NGAL. Error bars = 95% confidence interval Table 2 Multivariable Cox regression models with P-PENK and P-NGAL at baseline for prediction of A ­ KIcrea within 48 h Hazard ratio 95% CI p value p value c-statistic c-statistic for addition* with PENK/NGAL without PENK/ NGAL Model 1 P-PENK > 84.8 pmol/mL at baseline 2.2 1.1–4.4 0.03 0.02 0.689 0.677 Prior use of diuretics 2.4 1.2–4.9 0.01   Arterial pH 0.7 0.6–0.9 0.01 Prior use of ACEI or ARB 0.4 0.2–0.8 0.01 Model 2 P-NGAL > 104 ng/mL at baseline 2.8 1.2–6.5 0.01 0.007 0.684 0.634 Arterial pH 0.9 0.7–1.1 0.14 ACEI angiotensin-converting enzyme inhibitor, ARB angiotensin receptor blocker, NGAL neutrophil gelatinase-associated lipocalin, PENK proenkephalin * 2 log-likelihood test for additive value of P-PENK/P-NGAL in the model and P-NGAL and their associations with markers of risk scores designed to assess the prognosis of patients in kidney function, AKI and outcomes. The main findings cardiogenic shock. can be summarized as follows. Firstly, both P-PENK and P-NGAL at baseline correlated with markers of kidney P‑PENK and P‑NGAL as predictors of AKI function. Secondly, high baseline levels of both studied Elevated levels of P-PENK and P-NGAL were both markers were associated with the development of AKI associated with the development of A ­ KIcrea48h and had within 48 h. Thirdly, subclinical AKI at baseline was asso- comparable discriminatory properties in AUC analysis. ciated with increased mortality both in patients with Interestingly, we found that ­AKIcrea48h occurred almost oliguria before study enrollment as well as in patients solely if oliguria (< 0.5 mL/kg/h for > 6 h) before study without ­AKIcrea48h. Finally, although high levels of inclusion was combined with subclinical AKI at baseline. P-PENK and P-NGAL were associated with worse out- P-NGAL has been shown to predict AKI after cardiac comes at all studied time points, P-PENK and P-NGAL surgery [21], in critically ill children [22] and adults [23] levels at 24 h had the best discriminatory capabilities on with suggested cut-offs ranging from 100 to 270 nmol/mL survival. They were also associated with mortality inde- [24]. In a recent meta-analysis comparing plasma/serum pendently of the development of AKI and two separate NGAL, urine NGAL and serum cystatin C, plasma/ Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 10 of 15 Table 3 Differences in renal outcomes, interventions, and mortality by P-PENK and P-NGAL cut-offs at baseline All PENK PENK p value P-NGAL P-NGAL p value at 0 h < 84.8 pmol/ at 0 h > 84.8 pmol/ at 0 h < 104 ng/ at 0 h > 104 ng/ mL mL mL mL (N = 154) (N = 61) (N = 91) (N = 49) (N = 97) AKI staging at ­baselinea Stage 0 21 (14%) 21 (14%) 0 (0%) < 0.001 17 (12%) 0 (0%) < 0.001 Stage 1S 51 (34%) 12 (8%) 38 (26%) 13 (9%) 36 (25%) Stage 1A 16 (11%) 16 (11%) 0 (0%) 15 (10%) 0 (0%) Stage 1B 63 (42%) 11 (7%) 51 (34%) 3 (2%) 60 (42%) Detection of AKI by increase in creati‑ nine after baseline AKI by 12 h 29 (19%) 5 (8%) 23 (25%) 0.05 5 (10%) 24 (25%) 0.006 AKI 12–24 h 6 (4%) 2 (3%) 4 (4%) 1 (2%) 5 (5%) AKI 24–36 h 10 (7%) 4 (7%) 6 (7%) 0 (0%) 9 (9%) AKI 36–48 h 2 (1%) 0 (0%) 2 (2%) 1 (2%) 1(1%) AKI 48 h-discharge from ICU 6 (4%) 2 (3%) 4 (4%) 0.73 3 (6%) 3 (3%) 0.38 (5–10 days)b AKI CysC 49 (33%) 17 (29%) 32 (36%) 0.32 9 (18%) 40 (42%) 0.005 AKI severity by urine output after baseline No 75 (50%) 37 (61%) 38 (42%) 0.04 31 (63%) 42 (44%) 0.02 Stage 1 33 (22%) 13 (21%) 20 (22%) 11 (22%) 19 (20%) Stage 2 or 3 43 (29%) 11 (18%) 32 (36%) 7 (14%) 35 (37%) AKI severity within 48 h of baseline by creatinine 0.008 0.005 (RRT excluded from staging) No AKI 106 (70%) 50 (82%) 56 (61%) 42 (86%) 58 (60%) Stage 1 27 (18%) 9 (15%) 18 (20%) 5 (10%) 22 (23%) Stage 2 or 3 (RRT excluded) 19 (12%) 2 (3%) 17 (19%) 2 (4%) 17 (17%) History of renal insufficiency 17 (11%) 1 (2%) 16 (18%) 0.003 0 (0%) 17 (18%) 0.002 Interventions Renal replacement therapy 22 (14%) 7 (12%) 15 (17%) 0.49 3 (6%) 19 (20%) 0.048 Ultrafiltration 11 (8%) 2(4%) 9 (11%) 0.20 2 (5%) 9 (11%) 0.33 Hemodialysis 7 (5%) 0 (0%) 7 (9%) 0.04 0 (0%) 7 (8%) 0.10 Use of vasopressors 126 (83%) 50 (82%) 76 (84%) 0.83 41 (84%) 81 (84%) 0.99 Use of adrenaline 21 (14%) 7 (12%) 14 (15%) 0.63 2 (4%) 18 (19%) 0.02 Use of noradrenaline 114 (75%) 44 (72%) 70 (77%) 0.57 39 (80%) 73 (75%) 0.68 Use of dobutamine 25 (16%) 11 (18%) 14 (15%) 0.82 5 (10%) 18 (19%) 0.23 History of renal insufficiency 17 (11%) 1 (2%) 16 (18%) 0.003 0 (0%) 17 (18%) 0.002 Coronary angiography 128 (83%) 55 (90%) 71 (78%) 0.08 45 (92%) 75 (77%) 0.04 Amount of contrast used; mL (SD) 189 (104) 192 (106) 186 (105) 0.72 183 (113) 197 (102) 0.29 Amount of contrast used per eGFR; 3.4 (2.4) 2.5 (1.6) 4.1 (3.0) 0.005 2.3 (1.8) 4.2 (3.6) < 0.001 mL/eGFR (SD) IABP use 86 (57%) 39 (64%) 47 (52%) 0.18 27 (55%) 56 (58%) 0.86 Intubated 93 (62%) 36 (60%) 57 (63%) 0.73 28 (57%) 61 (64%) 0.47 LVAD or ECMO 6 (4%) 2 (3%) 4 (4%) 0.99 1 (2%) 4 (4%) 0.66 90-day mortality 58 (38%) 15 (25%) 42 (47%) 0.006 9 (18%) 49 (52%) < 0.001 Results shown as n (%) for categorical and mean (SD) or median (IQR) for continuous variables AKI acute kidney injury, CysC cystatin C, ECMO extracorporeal membrane oxygenation, eGFR estimated glomerular filtration rate, IABP intra-aortic balloon pump, ICU intensive care unit, IQR interquartile range, LVAD left ventricular assist device, NGAL neutrophil gelatinase-associated lipocalin, PENK proenkephalin, RRT​renal replacement therapy, SD standard deviation a AKI staging on admission based on urine output and biomarkers, as data on creatinine prior to baseline were unavailable b KDIGO definition of increase in serum creatinine to 1.5 times baseline or more Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 11 of 15 Fig. 3 Kaplan–Meier curves for occurrence of acute kidney injury by P-PENK and P-NGAL levels at baseline. a Lines separated by P-PENK > 84.8 pmol/mL at baseline. b Lines separated by P-NGAL > 104 ng/mL at baseline serum NGAL was found to be the earliest marker pre- was associated with renal replacement therapy and AKI dicting contrast-induced nephropathy [25]. Of the two assessed by changes in cystatin C plasma concentrations, markers investigated in the present study, only P-NGAL suggesting P-NGAL may be a more robust indicator of AKI and other renal outcomes. For most patients who developed AKI, it occurred very early, which makes it most likely related to the acute state of cardiogenic shock and not secondary to other events in-hospital. Thus, based on our findings, the ability to predict the occur- rence and severity of AKI using biomarkers measured at presentation opens a narrow window of opportunity for targeted interventions. Biomarker trajectories in AKI and nonsurvivors Overall, levels of P-PENK and P-NGAL decreased from 0 to 48 h. In nonsurvivors and patients who developed AKI within 48 h the levels were higher at all time points for both P-PENK and P-NGAL compared to survivors and patients without AKI. While the absolute levels dif- fered, the trajectories of P-PENK over time were similar in survivors and nonsurvivors. In contrast, for P-NGAL, the levels were increasing up to 48 h in nonsurvivors, but were stable or declined in survivors. These findings most Fig. 4 Venn diagram of P-PENK, P-NGAL at baseline, acute kidney likely represent differences in the production and kinet- injury and 90-day mortality. PENK high = P-PENK < 84.8 pmol/mL ics of the two biomarkers. In a mouse renal ischemia– at baseline. P-NGAL high = P-NGAL > 104 ng/mL at baseline. For reperfusion injury model, P-NGAL expression has been illustrative purposes. Areas not proportional, unable to show all shown to be upregulated in the proximal kidney tubu- overlaps lar cells early in response to renal ischemia with a peak Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 12 of 15 Fig. 5 Kaplan–Meier survival curves for patients with and without AKI separated by P-PENK and P-NGAL levels. a P-PENK higher or lower than 84.8 pmol/mL at baseline. b P-NGAL higher or lower than 104 ng/mL at baseline. §p = 0.07, *p < 0.05, **p < 0.001. AKI acute kidney injury Table 4 Hazard ratios for multivariable models of 90-day mortality for P-PENK24h > 105.7 pmol/mL and P-NGAL24h > 151 ng/mL Hazard ratio 95% CI p value* c-statistic c-statistic without PENK/ NGAL P-PENK24h > 105.7 pmol/mL at 24 h Univariable 5.8 3.1–10.7 < 0.001 0.705 Adjusted model 1 (CardShock risk score) 4.5 2.3–8.7 < 0.001 0.778 0.726 Adjusted model 2 (IABP II SHOCK score) 4.2 2.0–9.0 < 0.001 0.750 0.702 Adjusted model 3 (AKIcrea48h) 3.5 1.7–6.9 < 0.001 0.750 0.671 Adjusted model 4 (CardShock risk score + AKIcrea48h) 2.7 1.3–5.6 0.01 0.800 0.781 Adjusted model 5 (IABP II SHOCK score + AKIcrea48h) 2.7 1.2–5.9 0.01 0.773 0.749 P-NGAL24h > 151 ng/mL at 24 h Univariable 5.2 2.8–9.8 < 0.001 0.632 Adjusted model 1 (CardShock risk score) 3.4 1.7–6.8 0.001 0.733 0.726 Adjusted model 2 (IABP II SHOCK score) 4.2 2.0–8.9 < 0.001 0.724 0.702 Adjusted model 3 (AKIcrea48h) 3.2 1.6–6.5 0.001 0.716 0.671 Adjusted model 4 (CardShock risk score + AKIcrea48h) 2.0 0.9–4.2 0.07 0.781 0.781 Adjusted model 5 (IABP II SHOCK score + AKIcrea48h) 2.5 1.1–5.9 0.03 0.755 0.749 AKIcrea48h acute kidney injury by creatine increase within 48 h, CI confidence interval, NGAL neutrophil gelatinase-associated lipocalin, PENK proenkephalin * 2 log-likelihood test for additive value of P-PENK/P-NGAL in the model at 12 h post-ischemia [26], which could explain the ini- higher in those with more severe clinical presentation, tial increase observed for P-NGAL in nonsurvivors and but be more stable over time. P-PENK could also cause patients who developed AKI and also had higher levels a deterioration of renal function through cardiodepres- of lactate. P-PENK, on the other hand, is a marker of sion [27], with high initial levels causing a more marked the endogenous opioid system activity, which could be decline in kidney function. These pathophysiological Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 13 of 15 mechanisms may be seen to be in concordance with the very limited and more research is warranted. The results results of our study that patients with high levels of either of our study should be validated in another cohort of car- P-PENK or P-NGAL were at significantly higher risk for diogenic shock patients to confirm our findings. mortality and for developing AKI within 48 h. Study limitations Association of P‑PENK and P‑NGAL with mortality We did not have plasma samples available for all the Card- Both P-PENK and P-NGAL showed good discrimina- Shock study participants, and also were unable to measure tory capabilities for 90-day mortality at all time points P-NGAL for some of the patients in this study. However, this up to 48 h, with the highest AUCs for both occurring is one of the largest cohorts of biomarker studies in patients at 24 h. High levels of both P-PENK24h and P-NGAL24h with cardiogenic shock and despite the general challenges were associated independently with 90-day mortality with serial sampling in acute cardiac care, we consider the even after adjusting for the CardShock risk score, IABP available samples to be representative of the cohort. One of II SHOCK risk score and AKI. In fact, the AUCs for the study limitations is the lack of data on creatinine prior to both P-PENK24h and P-NGAL24h were similar to those study inclusion. On the other hand, the European Renal Best of the risk scores alone. The independent association of Practice position statement recommends using baseline cre- P-PENK24h and P-NGAL24h with 90-day mortality sug- atinine values instead of historical values or back-calculated gests that they are not just markers for AKI, but may value based on an assumed GFR of 75 mL/min/1.73 m2 also contribute to other unidentified causes of increased [31]. Although adjustments were made for several variables mortality in the CardShock study population that are not in multivariable analyses, there may have been other con- included in the relevant risk scores. A recent study in founding factors we were unable to account for leading to an critically ill patients suggests that high levels of P-PENK overestimation of the independent association of P-PENK in patients without AKI may be a marker of subclinical and P-NGAL with AKI and mortality. AKI, which was associated with a risk of death close to patients with AKI [28]. Similarly in our study, patients with subclinical AKI (elevated biomarkers at baseline but Conclusions without ­AKIcrea48h) had higher 90-day mortality. Inter- Our study shows that in cardiogenic shock, P-PENK estingly, P-NGAL was also able to stratify patients with and P-NGAL levels differed between patients who did ­AKIcrea48h into high or low mortality risk groups. It could and did not develop AKI within 48 h as well as between be hypothesized that this might be due to less severe AKI, 90-day survivors and nonsurvivors. High levels of both where despite an increase in creatinine the renal injury is P-PENK and P-NGAL at presentation were associated smaller. Using the recently described AKI staging using with AKI within 48 h. Subclinical AKI at baseline was biomarkers on admission [5], we were also able to show associated with increased mortality both in patients with that low urine output (< 0.5 mL/kg/h for > 6 h) before oliguria before study enrollment as well as in patients study inclusion was associated with worse outcomes without ­AKIcrea48h. Adjusting for AKI and two risk scores only if combined with high baseline levels of P-PENK or validated for cardiogenic shock patient populations, high P-NGAL. levels of P-PENK and P-NGAL at 24 h were still indepen- Elevated P-NGAL may not be specific for acute kidney dently associated with higher 90-day all-cause mortal- injury, as NGAL is also expressed at low levels in several ity. P-PENK and P-NGAL seem useful biomarkers in the human tissues, including lung, stomach, and colon, as early prediction of outcomes in cardiogenic shock popu- well as neutrophils [14]. Serum NGAL levels have been lations, and may have role in the prediction of AKI. shown to increase also with infection, inflammation and ischemia [29]. The association of NGAL with inflamma- Supplementary Information tion and ischemia might be a drawback in acute kidney The online version contains supplementary material available at https​://doi. injury prediction, but might help explain the high dis- org/10.1186/s1361​3-021-00814​-8. criminatory capability in 90-day mortality in cardiogenic shock, where systemic inflammatory response syndrome Additional file 1 : Table S1. Areas under the receiver-operating char‑ and end-organ ischemia play an important role. P-PENK acteristic curve for acute kidney injury occurring within 48 hours and 90-day all-cause mortality for P-PENK and P-NGAL at different time points. and P-NGAL could be useful in assessing mortality risk Table S2. Spearman correlations between P-PENK and P-NGAL and other of cardiogenic shock patients at later time points for variables at baseline. Table S3. Cross tabulation of index test results. Fig‑ which few risk markers have so far been assessed [30]. ure S1. Diagram of study sampling times. Figure S2. STARD flow diagram of study participants. A: AKI B: 90-day mortality. Figure S3. Kaplan–Meier However, although this study offers new insight to pre- survival curves stratified by a) P-PENK >105.7 pmol/mL at 24 hours and b) diction of AKI and mortality in cardiogenic shock pop- P-NGAL >151 ng/mL at 24 hours. ulations, overall data on biomarkers in this area are still Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 14 of 15 Abbreviations University of Helsinki, Helsinki, Finland. 3 HUSLAB Diagnostic Services, Helsinki AKI: Acute kidney injury; AUC​: Area under the receiver-operating character‑ University Hospital and University of Helsinki, Helsinki, Finland. 4 Internal Medi‑ istics curve; CRP: C-reactive protein; eGFR: Estimated glomerular filtration cine, Department of Internal Medicine and Rehabilitation, Helsinki University fraction; HR: Hazard ratio; IABP SHOCK II: Intra-aortic balloon counterpulsation Hospital, University of Helsinki, Helsinki, Finland. 5 Intensive Cardiac Care Unit, in acute myocardial infarction complicated by cardiogenic shock trial; IQR: Cardiology Department, Hospital de La Santa Creu I Sant Pau, Biomedical Interquartile range; KDIGO: Kidney disease: improving global outcomes; NGAL: Research Institute IIB‐SantPau, Universidad Autónoma de Barcelona, Barce‑ Neutrophil gelatinase-associated lipocalin; NT-proBNP: N-terminal Pro-b-type lona, Spain. 6 CINTESIS, Department of Cardiology, São João Hospital Center, natriuretic peptide; PENK: Proenkephalin; SD: Standard deviation. and Porto Medical School, University of Porto, Porto, Portugal. 7 Intensive Car‑ diac Therapy Clinic, National Institute of Cardiology, Warsaw, Poland. 8 Depart‑ Acknowledgements ment of Medical Sciences and Translational Medicine, Sant’Andrea Hospital, List of CardShock investigators: Veli-Pekka Harjola (chair), Marek Banaszewski, University of Rome Sapienza, Rome, Italy. 9 INSERM U942, Department of Anes‑ Lars Køber, Johan Lassus, Alexandre Mebazaa, Marco Metra, John Parissis, Jose thesia and Critical Care, Hôpital Lariboisière, APHP, University Paris Diderot, Silva-Cardoso, Alessandro Sionis, Salvatore Di Somma, Jindrich Spinar, Katerina Paris, France. 10 Nephrology, Department of Nephrology, Abdominal Center, Koniari, Astrinos Voumvourakis, Apostolos Karavidas, Jordi Sans-Rosello, Helsinki University Hospital, University of Helsinki, Helsinki, Finland. Montserrat Vila, Albert Duran-Cambra, Michela Bulgari, Valentina Lazzarini, Jiri Parenica, Roman Stipal, Ondrej Ludka, Marie Palsuva, Eva Ganovska, Petr Received: 31 August 2020 Accepted: 20 January 2021 Kubena, Matias G. Lindholm, Christian Hassager, Tom Bäcklund, Raija Jurkko, Kristiina Järvinen, Tuomo Nieminen, Kari Pulkki, Leena Soininen, Reijo Sund, Ilkka Tierala, Jukka Tolonen, Marjut Varpula, Tuomas Korva, Anne Pitkälä, Ros‑ sella Marino, Alexandra Sousa, Carla Sousa, Mariana Paiva, Inês Rangel, Rui Almeida, Teresa Pinho, Maria Júlia Maciel, Janina Stepinska, Anna Skrobisz, Piotr References Góral. The study was performed in collaboration with the GREAT network. 1. van Diepen S, Katz JN, Albert NM, Henry TD, Jacobs AK, Kapur NK, et al. Contemporary management of cardiogenic shock: a scientific statement Authors’ contributions from the American Heart Association. Circulation. 2017;136:e232–68. All authors have contributed to the final version of the manuscript as follows. 2. Puymirat E, Fagon JY, Aegerter P, Diehl JL, Monnier A, Hauw Berlemont C, Study concept and design: TJ, JL, VPH. Acquisition, analysis, or interpretation et al. Cardiogenic shock in intensive care units: evolution of prevalence, of data: all authors. Drafting of the manuscript: TJ, JL. Statistical analyses: TJ. All patient profile, management and outcomes, 1997–2012. Eur J Heart Fail. authors critically revised the manuscript and gave final approval for the ver‑ 2017;19:192–200. sion to be published. All authors read and approved the final manuscript. 3. Tarvasmäki T, Haapio M, Mebazaa A, Sionis A, Silva-Cardoso J, Tolp‑ panen H, et al. Acute kidney injury in cardiogenic shock: definitions, Funding incidence, haemodynamic alterations, and mortality. Eur J Heart Fail. This study was supported by grants from the Finnish Foundation for Cardio‑ 2017;10:193–210. vascular Research and Aarne Koskelo Foundation. Laboratory kits for analysis 4. Kellum JA, Lameire N, Aspelin P, Barsoum RS, Burdmann EA, Goldstein SL, of NT-proBNP and hsTnT were kindly provided by Roche Diagnostics, Basel, et al. Kidney disease: Improving global outcomes (KDIGO) acute kidney Switzerland. Role of the Funder/Sponsor: the funding sources had no role in injury work group. KDIGO clinical practice guideline for acute kidney the design and conduct of the study; collection, management, analysis, and injury. Kidney Int Suppl. 2012;2:1–138. interpretation of the data; preparation, review, or approval of the manuscript; 5. Ostermann M, Zarbock A, Goldstein S, Kashani K, Macedo E, Murugan or decision to submit the manuscript for publication. R, et al. Recommendations on acute kidney injury biomarkers from the acute disease quality initiative consensus conference. JAMA Netw Open. Availability of data and materials 2020;3:e2019209–17. The datasets used and analyzed during the current study are available from 6. Basu RK, Gist K, Wheeler DS. Improving acute kidney injury diagnostics the corresponding author on reasonable request. using predictive analytics. Curr Opin Crit Care. 2015;21:473–8. 7. Kim H, Hur M, Lee S, Marino R, Magrini L, Cardelli P, et al. Proenkephalin, Ethics approval and consent to participate neutrophil gelatinase-associated lipocalin, and estimated glomerular The study was approved by all local ethics committees and conducted in filtration rates in patients with sepsis. Ann Lab Med. 2017;37:388–410. accordance with the Declaration of Helsinki. Written informed consent was 8. Beunders R, Struck J, Wu AHB, Zarbock A, Di Somma S, Mehta RL, et al. obtained from the patient or next of kin if the patients were unable to give the Proenkephalin (PENK) as a novel biomarker for kidney function. J Appl consent on admission. Lab Med. 2017;2:400–12. 9. Ng LL, Sandhu JK, Narayan H, Quinn PA, Squire IB, Davies JE, et al. Consent for publication Proenkephalin and prognosis after acute myocardial infarction. J Am Coll Not applicable. Cardiol. 2014;63:280–9. 10. Ng LL, Squire IB, Jones DJL, Cao TH, Chan DCS, Sandhu JK, et al. Proen‑ Competing interests kephalin, renal dysfunction, and prognosis in patients with acute heart VPH: advisory board fees from Roche Diagnostics, research grant from Abbott, failure: a great Network study. J Am Coll Cardiol. 2017;69:56–69. speaker fees from Orion. KP: advisory board fees from Roche Diagnostics. 11. Matsue Y, ter Maaten JM, Struck J, Metra M, O’Connor CM, Ponikowski P, JSC has consulted and received speaker fees, or advisory boards’ participa‑ et al. Clinical correlates and prognostic value of proenkephalin in acute tion fees, or investigational grants for Abbott, Astra-Zeneca Pharmaceuticals, and chronic heart failure. J Cardiac Fail. 2017;23:231–9. Bial, Boehringer Ingelheim, Menarini, Merck Serono, Merck Sharp & Dohme, 12. Beunders R, VanGroenendael R, Leijte G, Kox M, Pickkers P. Proenkephalin Novartis, Orion, Pfizer, Sanofi, Servier and Vifor. AM: lecture fees from Novartis, compared to conventional methods to assess kidney function in critically Orion, and Abbott, research grants from Roche, consultant fees from Servier ill sepsis patients. Shock. 2020;54:308–14. and Sanofi and fees as a member of the advisory board and/or steering com‑ 13. Caironi P, Latini R, Struck J, Hartmann O, Bergmann A, Bellato V, et al. mittee from Cardiorentis, Adrenomed, Sphingotec, Sanofi, Roche, Abbott, and Circulating proenkephalin, acute kidney injury, and its improvement in Bristol-Myers Squibb. JL: Speakers bureau and consultancy fees: Astra-Zeneca, patients with severe sepsis or shock. Clin Chem. 2018;64:1361–9. Bayer, Boehringer-Ingelheim, Novartis, Orion, Pfizer, Roche Diagnostics, and 14. Cowland JB, Borregaard N. Molecular characterization and pattern of tis‑ ViforPharma. All other authors report that they have no relationships with sue expression of the gene for neutrophil gelatinase-associated lipocalin industry to disclose. from humans. Genomics. 1997;45:17–23. 15. Parikh CR, Devarajan P. New biomarkers of acute kidney injury. Crit Care Author details Med. 2008;36:S159–65. 1 Department of Cardiology, Heart and Lung Center, Helsinki University Hospi‑ 16. Di Grande A, Giuffrida C, Carpinteri G, Narbone G, Pirrone G, Di Mauro A, tal, University of Helsinki, 00029 HUS Helsinki, Finland. 2 Emergency Medicine, et al. Neutrophil gelatinase-associated lipocalin: a novel biomarker for the Department of Emergency Medicine and Services, Helsinki University Hospital, Jäntti et al. Ann. Intensive Care (2021) 11:25 Page 15 of 15 early diagnosis of acute kidney injury in the emergency department. Eur 26. Mishra J, Ma Q, Prada A, Mitsnefes M, Zahedi K, Yang J, et al. Identifica‑ Rev Med Pharmacol Sci. 2009;13:197–200. tion of neutrophil gelatinase-associated lipocalin as a novel early urinary 17. Klein SJ, Brandtner AK, Lehner GF, Ulmer H, Bagshaw SM, Wiedermann biomarker for ischemic renal injury. J Am Soc Nephrol. 2003;14:2534–43. CJ, et al. Biomarkers for prediction of renal replacement therapy in acute 27. Emmens JE, ter Maaten JM, Damman K, van Veldhuisen DJ, de Boer RA, kidney injury: a systematic review and meta-analysis. Intens Care Med. Struck J, et al. Proenkephalin, an opioid system surrogate, as a novel com‑ 2018;44:323–36. prehensive renal marker in heart failure. Circ Heart Fail. 2019;12:e005544. 18. Harjola V-P, Lassus J, Sionis A, Køber L, Tarvasmäki T, Spinar J, et al. Clinical 28. Dépret F, Hollinger A, Cariou A, Deye N, Vieillard-Baron A, Fournier picture and risk prediction of short-term mortality in cardiogenic shock. M-C, et al. Incidence and outcome of sub-clinical acute kidney injury Eur J Heart Fail. 2015;17:501–9. using penKid in critically ill patients. Am J Respir Crit Care Med. 19. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, et al. 2020:rccm.201910–1950OC. A new equation to estimate glomerular filtration rate. Ann Intern Med. 29. Sivalingam Z, Larsen SB, Grove EL, Hvas A-M, Kristensen SD, Magnusson 2009;150:604–12. NE. Neutrophil gelatinase-associated lipocalin as a risk marker in cardio‑ 20. Pöss J, Köster J, Fuernau G, Eitel I, de Waha S, Ouarrak T, et al. Risk stratifi‑ vascular disease. Clin Chem Lab Med (CCLM). 2017;56:211–4. cation for patients in cardiogenic shock after acute myocardial infarction. 30. Tolppanen H, Rivas-Lasarte M, Lassus J, Sans-Roselló J, Hartmann O, J Am Coll Cardiol. 2017;69:1913–20. Lindholm M, et al. Adrenomedullin: a marker of impaired hemodynam‑ 21. Mishra J, Dent C, Tarabishi R, Mitsnefes MM, Ma Q, Kelly C, et al. Neutrophil ics, organ dysfunction, and poor prognosis in cardiogenic shock. Ann gelatinase-associated lipocalin (NGAL) as a biomarker for acute renal Intensive Care. Springer Paris. 2016; 1–10. injury after cardiac surgery. Lancet. 2005;365:1231–8. 31. The ad-hoc working group of ERBP, Fliser D, Laville M, Covic A, Fouque 22. Wheeler DS, Devarajan P, Ma Q, Harmon K, Monaco M, Cvijanovich N, D, Vanholder R, et al. A European Renal Best Practice (ERBP) position et al. Serum neutrophil gelatinase-associated lipocalin (NGAL) as a marker statement on the Kidney Disease Improving Global Outcomes (KDIGO) of acute kidney injury in critically ill children with septic shock. Crit Care Clinical Practice Guidelines on Acute Kidney Injury: Part 1: definitions, Med. 2008;36:1297–303. conservative management and contrast-induced nephropathy. Nephrol 23. Hang C-C, Yang J, Wang S, Li C-S, Tang Z-R. Evaluation of serum neutrophil Dial Transplant. 2012;27:4263–72. gelatinase-associated lipocalin in predicting acute kidney injury in criti‑ cally ill patients. J Int Med Res. 2017;45:1231–44. 24. Haase M, Bellomo R, Devarajan P, Schlattmann P, Haase-Fielitz A, NGAL Publisher’s Note Meta-analysis Investigator Group. Accuracy of neutrophil gelatinase-asso‑ Springer Nature remains neutral with regard to jurisdictional claims in pub‑ ciated lipocalin (NGAL) in diagnosis and prognosis in acute kidney injury: lished maps and institutional affiliations. a systematic review and meta-analysis. Am J Kidney Dis. 2009;54:1012–24. 25. He Y, Deng Y, Zhuang K, Li S, Xi J, Chen J. Predictive value of cystatin C and neutrophil gelatinase-associated lipocalin in contrast-induced nephropathy: a meta-analysis. PLoS ONE. 2020;15:e0230934.