Level of systemic inflammation and endothelial injury is associated with cardiovascular dysfunction and vasopressor support in post-cardiac arrest patients

Accepted Manuscript Title: Level of systemic inflammation and endothelial injury is associated with cardiovascular dysfunction and vasopressor support in post cardiac arrest patients Authors: John Bro-Jeppesen, P¨ar I. Johansson, Jesper Kjaergaard, Michael Wanscher, Sisse R. Ostrowski, Mette Bjerre, Christian Hassager PII: DOI: Reference:

S0300-9572(17)30631-7 http://dx.doi.org/10.1016/j.resuscitation.2017.09.019 RESUS 7322

To appear in:

Resuscitation

Received date: Revised date: Accepted date:

4-8-2017 17-9-2017 22-9-2017

Please cite this article as: Bro-Jeppesen John, Johansson P¨ar I, Kjaergaard Jesper, Wanscher Michael, Ostrowski Sisse R, Bjerre Mette, Hassager Christian.Level of systemic inflammation and endothelial injury is associated with cardiovascular dysfunction and vasopressor support in post cardiac arrest patients.Resuscitation http://dx.doi.org/10.1016/j.resuscitation.2017.09.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Level of systemic inflammation and endothelial injury is associated with cardiovascular dysfunction and vasopressor support in post cardiac arrest patients John Bro-Jeppesen MD PhD1 Pär I. Johansson MD DMSc MPA2, 3, Jesper Kjaergaard MD PhD DMSc1, Michael Wanscher MD PhD4, Sisse R. Ostrowski MD PhD DMSc2, Mette Bjerre PhD5 Christian Hassager MD DMSc1 1

Department of Cardiology, The Heart Centre, Rigshospitalet, Copenhagen University Hospital, Denmark.

2

Section for Transfusion Medicine, Capital Region Blood Bank, Rigshospitalet, Copenhagen University Hospital, Denmark. 3

Department of Surgery, Division of Acute Care Surgery, Centre for Translational Injury Research (CeTIR), University of Texas Medical School at Houston, TX, US. 4

Department of Cardiothoracic Anesthesiology, The Heart Centre, Rigshospitalet, Copenhagen University Hospital, Denmark. 5

The Medical Research Laboratory, Department of Clinical Medicine, Aarhus University, Denmark.

Address for correspondence: John Bro-Jeppesen Department of Cardiology, The Heart Centre, Copenhagen University Hospital Rigshospitalet, Blegdamsvej 9 DK-2100 Copenhagen, Denmark Telephone: +45 35453545 Fax number: +45 35452648 E-mail: [email protected]

Words: Abstract: 315 Total words excluding abstract, references and tables: 3344

Abstract: Aim: Post-cardiac arrest syndrome (PCAS) is characterized by a sepsis-like inflammatory response and hemodynamic instability. We investigated the associations between systemic inflammation, endothelial damage and hemodynamic parameters including vasopressor support in patients with out-of-hospital cardiac arrest (OHCA).

Methods: In this post-hoc study, we analysed data from 163 comatose patients included at a single center in the Target Temperature Management (TTM) trial, randomly assigned to TTM at 33°C or 36°C for 24 hours. Inflammatory biomarkers (interleukin (IL)-6, IL-10, procalcitonin and Tumor Necrosis Factor-α (TNF-α)) and endothelial biomarkers (thrombomodulin, sE-selectin, syndecan-1 and VEcadherin) were measured at randomization and 24, 48 and 72 hours after OHCA. Corresponding hemodynamic status, heart rate (HR), mean arterial pressure (MAP) and Cumulative Vasopressor Index (CVI) was reported.

Results: At randomization, level of IL-6 correlated negatively with MAP (r= -0.19,p=0.03) and positively with HR (r= 0.29,p=0.0002). Serial IL-6 levels correlated consistently with CVI at 24h: (r= 0.19,p=0.02) 48h: (r= 0.31,p=0.0001) and 72h: (r= 0.39,p<0.0001). Thrombomodulin (r= 0.23,p=0.004) and syndecan-1 (r= 0.27,p=0.001) correlated with CVI at 48h. All inflammatory markers excerpt IL-10 and all endothelial markers correlated with CVI at 72 h. Multivariable regression models adjusting for potential confounders confirmed that IL-6 (β=0.2 (95%CI:0.06-0.3),p=0.004) and TTM-group (TTM36: β=-0.5 (95%CI:-0.9- -0.1),p=0.01) were associated with CVI at 48h. At 72h after OHCA, IL-6 (β=0.3 (95%CI:0.03-0.6),p<0.0001), TNF-α (β=0.4 (95%CI:-0.5- -0.2),p<0.0001) and TTM-group (TTM36: β=-0.4 (95%CI:-0.8- -0.1),p=0.008) were associated with CVI. An overall two-fold increase in levels of IL-6 (β= 0.2 (95%CI:0.1-0.3),p<0.0001)) and IL-10 (β= -0.2 (95%CI:-0.3- -0.06),p=0.005)) within 72 h after OHCA were significantly associated with CVI. TTMgroup modified the interaction between CVI and IL-6 (pinteraction=0.008), but not with IL-10 (pinteraction=0.23).

Conclusions: In comatose survivors after OHCA, increasing systemic inflammation and endothelial injury was associated with increased need of vasopressor support. Systemic inflammation, in particular IL-6, was consistently associated with vasopressor support, however endothelial injury may also play a role in PCAS associated cardiovascular dysfunction after OHCA. Clinical Trial Registration: URL: clinicaltrials.gov/ct2/show/NCT01020916. Unique identifier: NCT01020916

Key words: cardiac arrest, hypothermia, hemodynamics, inflammation, endothelial injury, cytokines, post-cardiac arrest syndrome

Introduction An important clinical presentation after out-of-hospital cardiac arrest (OHCA) is development of the post-cardiac arrest syndrome (PCAS) characterized by brain injury, a sepsis-like inflammatory response and hemodynamic instability with impaired vasoregulation.[1] Severity of PCAS is associated with exposure to whole-body ischemia-reperfusion during cardiac arrest and after return of spontaneous circulation (ROSC), but recent studies have also identified systemic inflammation and endothelial dysfunction as important factors.[2-4] Possible mechanisms for this interaction between systemic inflammation and hemodynamics have not been fully characterized but direct effects of inflammatory mediators on the contractility of the myocardium and vascular smooth muscle have been suggested.[5, 6] However, the vascular endothelium, as the largest organ system exposed for ischemia-reperfusion injury, has recently been suggested to play an important role by causing increased vascular permeability and endothelial mediated inflammation leading to vasodilation.[7-9] The endothelium is an highly active “organ” and controls several important physiological functions as vasomotor tone, coagulations hemostasis and exchange of nutrients and immune cells.[10] Pro-inflammatory cytokines as interleukin-6 (IL-6) and Tumor Necrosis Factor (TNF)-α are thought to exhibit their vasoactive effects either by interacting with the activated endothelium directly or by the nitric oxide pathway.[6, 11-13] The presence of endotoxin, release of various cytokines and endothelial injury in PCAS share many similarities with sepsis and clinically both conditions are characterized by major alterations in cardiovascular function.[14, 15] In the early phase of PCAS, hemodynamic instability and shock is frequently present due to cardiac dysfunction and vasodilation, which often requires vasopressor support.[16-18] Furthermore, cardiovascular dysfunction may continue and exacerbate within the first 24-48 h after OHCA despite volume restoration and continuous vasopressor support.[19]

However, our current knowledge is limited in terms of PCAS related cardiovascular dysfunction and the possible associations between systemic inflammation, endothelial dysfunction and hemodynamics. This study aimed to explore the hemodynamic effects of PCAS associated inflammatory response and endothelial activation and damage within the first 72 hours post arrest evaluated by serial blood samples and clinical measures of hemodynamic status and vasopressor support.

Methods This study was an exploratory post-hoc analysis from a single centre (Rigshospitalet, Copenhagen) including patients participating in the prospective investigator-initiated, multi-centre, randomized, parallel-group, and assessor-blinded Target Temperature Management (TTM) trial (ClinicalTrials.gov number NCT01020916).[20] Patients were randomized in a 1:1 fashion to targeted temperature management at 33°C (TTM33) or 36°C (TTM36) for 24 hours after cardiac arrest. The TTM protocol, including the standard use of pulmonary artery catheter in all patients in this sub-study for research purposes, was approved by the Ethics Committee of the Capital Region Copenhagen (H-1-2010-059) and by the Danish Data Protection Agency. Written informed consent was obtained from patients’ next of kin and general practitioner in all cases and/or from patients regaining consciousness after cardiac arrest. In brief, inclusion criteria were adult patients (≥18 years) resuscitated from OHCA of presumed cardiac cause, who remained unconscious (Glasgow Coma Score (GCS) <8) for more than 20 minutes after sustained return of spontaneous circulation (ROSC). The most important exclusion criteria for the present analysis included unwitnessed asystole, a body temperature of less than 30°C and severe shock at time of admission to hospital defined as sustained systolic blood pressure less than 80 mmHg despite administration of fluids,

vasopressors, inotropes and/or treatment with intra-aortic balloon pump or left ventricular assist device. In this post-hoc study, we analysed data from a cohort of patients with available blood samples from previously published studies regarding inflammation and endothelial biomarkers.[3, 21] Among the available biomarkers, we decided to investigate inflammatory biomarkers, which have previously been described to have potential vasoactive and hemodynamic effects. [5, 22]

Patient management: All patients were admitted to the intensive care unit (ICU) for advanced post-cardiac arrest care and were sedated, intubated and mechanically ventilated throughout the 36-hour intervention period according to the study protocol.[23] All patients were actively cooled with the use of a surface cooling device (Thermowrap® with Allon® unit, MTRE, Israel). Crystalloid fluids were administered in all patients with general treatment goals for central venous pressure (CVP) of 10–15 mmHg to optimize right heart filling pressure, mean arterial pressure (MAP) ≥ 65 mmHg to secure adequate organ perfusion and urine output > 1.5 mL/kg/hour. Patients’ hemodynamics were monitored as soon as possible after ICU admission with heart rate monitoring and an arterial pressure catheter in the radial artery and insertion of a 7.5-F triple lumen Swan-Ganz thermistor and balloon-tipped catheter (Edwards Lifesciences, Irvine, CA). To assess the impact of systemic inflammation and endothelial dysfunction in the early phase of PCAS, we analyzed available baseline data from previously reported invasive hemodynamic measurements (n=154, 94%) and transthoracic echocardiography (n=139, 85%).[24] Inotropics/vasopressors were used to achieve adequate organ perfusion primary directed by clinical parameters as MAP, CVP and urine output to achieve the pre-defined hemodynamic treatment goals. Need of vasopressor support was reported by the cumulative vasopressor index

(CVI), allowing for standardization of equivalent doses of commonly used vasopressor agents and describes the overall degree of vasopressor support.[25] The CVI-score reported was the highest registered dose of vasopressor agents at time of randomization and at each pre-defined time points ± 1 h within 72 h after OHCA.

Blood samples and enzyme linked immunosorbent assay (ELISA) analyses Blood was sampled from an arterial line with first blood samples (T0= baseline) obtained approximately 5 minutes after inclusion and randomization in the TTM-study in the majority of patients. The median time from ROSC to randomization was 114 min (79-149 min). The following serial blood samples were collected at 24, 48 and 72 hours after OHCA and processed as previously described.[3] Activation of the inflammatory system was assessed in EDTA plasma by measurement of the inflammatory cytokines IL-6, IL-10 and TNF-α in duplicates at similar time points by use of a magnetic Bio-Plex Pro Assay (Bio-Rad, Hercules, California, USA).[3] Procalcitonin (PCT) analyses were performed using an immunoluminometric assay (Brahms PCT Kryptor, Brahms, Aktiengesellschaft, Hennigsdorf, Germany) with a detection limit of 0.02 µg/L. Biomarkers of endothelial glycocalyx damage (syndecan-1)[26], endothelial cell activation (sE-selectin)[27, 28], endothelial cell injury (soluble thrombomodulin)[27-29] and endothelial junction disruption (sVEcadherin)[30] were measured in serum in uniplicate by commercially available immunoassays according to the manufactures recommendations with detection limits as previously reported.[2] Statistical analysis The population was stratified by the median value of IL-6 at time of randomization and baseline demographic variables and hemodynamic parameters were compared using Student’s t-test, χ2test and Wilcoxon rank sum test, as appropriate. Data are presented as mean ± standard deviation

(SD) or proportions (%) and for variables with a non-normal distribution, data are presented as median and lower and upper quartile (Q1-Q3). Univariate correlations between variables were assessed by the non-parametric Spearman’s rho (r) correlation coefficient. To investigate associations between CVI-score and the variables contribution to the variation in CVI-score at each time point, multivariable linear regression analysis were performed including covariates with p<0.10 in univariate analysis. In separate models, baseline values of endothelial and inflammatory biomarkers were included to investigate the impact of temporal differences. The models capability of explaining the variation in CVI-score was assessed by R2-values. Multivariable linear regression models were applied after checking for underlying assumptions of linearity, homoscedasticity and collinearity. The serial design of the present study allowed for analysing associations between overall CVI-score and inflammatory and endothelial biomarkers after logarithmic transformation of original data with repeated measures mixed models with an unstructured covariance structure. Associations between CVI-score and inflammatory and endothelial biomarkers were investigated for possible interaction with target temperature by including the interaction term of TTM-group with biomarker as fixed effect in the models. All tests were two-sided and statistical significance was defined as p<0.05 unless other specified. All statistical analyses were performed using the SAS statistical software, version 9.2 (SAS Institute, Cary, NC).

Results Demographics and hemodynamic data from randomization and within 72 hour after OHCA are presented in Table 1 stratified by low (<87 ng/ml) vs. high (≥87 ng/ml) IL-6 levels at randomization. High levels of IL-6 at randomization were associated with a higher dose of administered adrenaline

during resuscitation attempt (4 mg (2-6 mg) vs. 2 mg (1-3 mg)), increased time to ROSC (26 mins (17-37 mins) vs. 17 mins (12-25 mins)) and higher levels of lactate (8.8 mmol/L (4.8-13 mmol/L) vs. 5.0 mmol/L (3.2-7.4 mmol/L)) compared to low IL-6 levels, all p<0.0001. At time of randomization, 42 (26%) patients received vasopressor support with a median CVI of 2 (1-2). At this time point, level of IL-6 correlated negatively with MAP (r= -0.19, p=0.03) and positively with HR (r= 0.29, p=0.0002), see Table 2. TNF-α did not correlate with MAP (r= 0.04, p=0.66) or HR (r= 0.02, p=0.83). At randomization, biomarkers reflecting endothelial damage, thrombomodulin (r= 0.18, p=0.02) and glycocalyx damage, syndecan-1 (r= 0.18, p=0.03) correlated with HR, but not with MAP or CVI. At time of randomization, median LVEF was 35% (30%-45%) and we found that levels of E-selectin (r= 0.23, p=0.007) correlated positively with LVEF and IL-10 (r= 0.21, p=0.02) and syndecan-1 (r= 0.22, p=0.02) correlated with s’ a marker of systolic function of the left ventricle. None of the other biomarkers showed any correlation with echocardiographic or invasively measured hemodynamic parameters (Table 3). At all time points between 24h-72h after OHCA, the inflammatory marker IL-6 showed a positive correlation with CVI and were negatively correlated with MAP at 48h and 72h (Table 2). The strength of the correlation between IL-6 and CVI-score increased within 72 h, and stratified analysis showed that the correlation coefficient was higher in the TTM33 group compared to the TTM36 group (Table 2 and Figure 1). Endothelial damage reflected by thrombomodulin levels correlated positively with CVI at 48h (r= 0.23, p=0.004) and 72h(r= 0.25, p=0.002), whereas no correlations were found between thrombomodulin and other hemodynamic parameters (Table 2). The correlation between thrombomodulin and CVI-score reported was caused by a significant correlation in the TTM33

group (Figure 1). At 72 h, the majority of inflammatory and endothelial markers correlated with CVI, see Table 2. Linear regression analysis confirmed that IL-6 and thrombomodulin correlated with CVI at corresponding time points, but also revealed that time to ROSC and lactate at randomization were associated with CVI (Table 4). In multivariable analysis, increased time to ROSC and higher IL-6 levels were independently associated with a higher CVI score at 24h, whereas levels of IL-6 and level of target temperature at 48h and 72h were independently associated with higher CVI score, see Table 4. To explore the impact of dynamic differences in biomarkers, we analysed similar models as presented in Table 4 adjusted for baseline levels of biomarkers, which did not change the association with CVI score at 24 h, 48 h or at 72 h. To assess the impact of the temporal changes of inflammatory and endothelial markers on CVI score within 72 h after OHCA and the effect of level of target temperature, we applied repeated measures mixed models to take advantage of the serial design of the study. In individual models, overall levels of PCT β= 0.07 ((95% CI: 0.003 - 0.14), p=0.008)), IL-6 β= 0.2 ((95% CI: 0.1-0.3), p<0.0001)), IL-10 β= -0.08 ((95% CI: -0.17- -0.002), p=0.04)), syndecan β= 0.2 ((95% CI: 0.03 - 0.3), p=0.02)) and thrombomodulin β= 0.2 ((95% CI: 0.1-0.4), p=0.004)) were associated with CVI-score within 72 h after OHCA. In a multivariable model, a twofold overall increase in levels of IL-6 β= 0.2 ((95% CI: 0.1-0.3), p<0.0001)) and IL-10 β= -0.2 ((95% CI: -0.3- -0.06), p=0.005)) within 72 h after OHCA was significantly associated with a higher CVI score, when adjusting for baseline levels of inflammatory and endothelial markers. Level of target temperature significantly modified the association between overall effect of levels of IL-6 on vasopressor support assessed by CVI-score

(TTM33 β=-0.2, (95% CL: -0.3 – -0.04), pinteraction =0.008) whereas no temperature effect was found between CVI-score and IL-10 (pinteraction =0.23). This suggests that a twofold increase in overall IL-6 levels was associated with an overall lower CVI-score in the TTM33 group. In similar models including initial lactate levels and time to ROSC, the interaction with target temperature between levels of IL-6 (pinteraction =0.007) and overall CVI-score was consistent.

Discussion In survivors after OHCA, severity of PCAS is closely associated with the period of ischemia and reperfusion injury, which also triggers and activates the inflammatory and endothelial systems. This study adds to our current knowledge that, IL-6 was consistently and independently associated with level of vasopressor support at 24-72 h after OHCA. Within 72h, overall levels of the proinflammatory biomarker IL-6 and anti-inflammatory biomarker IL-10 was found to be associated with level of vasopressor support, suggesting that systemic inflammation may play a role in PCAS associated hemodynamic dysfunction. However, early systemic inflammation evaluated by TNF-α and IL-6 did not seem to be associated with myocardial dysfunction assessed by cardiac output and vascular resistance nor by echocardiographic parameters. In later stages of PCAS, both systemic inflammation and endothelial injury were associated with level of vasopressor support. In the early phase of PCAS, the clinical condition is often characterized by myocardial dysfunction resulting in a state with low cardiac output, contributing to sustained hypotension.[19] Post-arrest myocardial dysfunction may be caused by myocardial stunning due to global ischemic and large doses of intravenous adrenaline administered during resuscitation attempt.[31] However, studies in post-cardiac arrest patients investigating the possible underlying pathophysiological causes for development of myocardial dysfunction are very limited. In animal models, infusion of the proinflammatory cytokines IL-1β and TNF-α have been proposed as important factors in development

of reversible cardiac depression in sepsis.[32] Niemann et al. demonstrated in an animal model of cardiac arrest, that TNF-α was released in the very early phase after ROSC and found high levels of systemic TNF-α to correlate with decreased systolic function of the left ventricle.[5] The present study could not confirm any relationship between systemic measured inflammatory biomarkers and decreased systolic function of the left ventricle evaluated by thermodilution and echocardiography. Although both studies were conducted in the early phase of PCAS, some important differences are worth discussing. The first blood sample in the current study was obtained with a median time of 114 mins after ROSC, whereas Niemann et al. collected samples within the first 90 mins. The release of TNF-α post-ROSC may be bi-modal with a very early and extensive release followed by a subsequent clearing phase with relative low or non-detectable levels of TNF-α within 24-72 hours after OHCA as previously demonstrated.[3, 33] Furthermore, differences in the methods of evaluating left ventricle function and the design of study may explain the different results. Niemann et al. reported results from an animal model in a controlled laboratory setting with induced ventricular fibrillation were vasopressors were withheld in the post-cardiac period.[5] Whole body ischemia during resuscitation attempt and the reperfusion injury period is thought to be the main trigger of the endothelial and inflammatory systems. However, the possible underlying signalling mechanisms are considered multifactorial and may thus be difficult to identify in details. Systemic inflammation is a strong stimulus for activating and facilitating the diapedesis of immune cells from the blood stream to the target tissue. This mechanism is regulated by the endothelium system.[34] Pro-inflammatory cytokines as TNF-α and IL-6 peaks in the early phase of PCAS and have been proposed to play an important role in activating the endothelium as early levels of IL-6 was associated with a later increase in biomarkers reflecting

endothelial injury.[2] Presence of inflammatory mediators activates the endothelium and glycoproteins as E-selectin are exposed on the surface. E-selectin mediates the interaction between leukocytes and the endothelium and is a marker of endothelial activation and is not presented on the surface in the resting and normal endothelium. In line with this hypothesis, biomarkers reflecting endothelial activation and endothelial injury seemed to peak within 24-48 h after OHCA, although no causal relationship between systemic inflammation and endothelial dysfunction could be concluded from these data.[2] Damage and degradation of the outer glycocalyx layer of the endothelium reflected by circulating levels of syndecan-1 is a marker of increased paracellular permeability and thus extravasation of fluids.[15] We found that levels of syndecan-1 correlated with increased vasopressor support in the later stages of PCAS, however multivariate analyses revealed that IL-6 seemed to be a more important factor. Laurent et al. reported that development of late hemodynamic instability was common after OHCA and the present study suggests that systemic inflammation and endothelial injury may play a role as high levels of IL-6, IL-10, TNF-α and all endothelial markers were associated with increased need of vasopressor support in the later stages of PCAS.[19] The magnitude of the inflammatory response and endothelial injury has been associated with severity of PCAS, which is a major determinant of outcome.[2, 4, 35] It can therefore be speculated whether interventions aiming to decrease systemic inflammation and/or protecting the vascular endothelium may be able to attenuate severity of PCAS and thus improve outcome. Data from animal models of ventricular fibrillation demonstrated that TNF-α blockade with infliximab infused shortly after ROSC attenuated myocardial dysfunction and hypotension leading to improved short term survival, whereas etanercept, a soluble TNF-α antagonist, had no effect.[36, 37] From human data, we know that survivors after OHCA exhibit a similar early release of TNF-α

and other pro-inflammatory cytokines.[3, 14] However, this study could not demonstrate any associations between high levels of TNF-α and cardiovascular dysfunction. In contrast, high levels of IL-6 were associated with early hemodynamic variables and independently associated with vasopressor support within 24-72 h after OHCA. The various properties of IL-6 are thought to arise from both local and systemic effects. IL-6 is an upstream regulator of the inflammatory response and plays a central role in initiating and propagating the downstream inflammatory response. Targeting IL-6 with an IL-6 receptor blocker as tocilizumab in order to decrease systemic inflammation, protecting the endothelium and attenuate organ injury may be of interest. Tocilizumab is a treatment option in patients with rheumatoid arthritis, a chronic inflammatory condition, and acts through selective and competitively binding to soluble and membraneexpressed IL-6 receptors and thus blocking IL-6 signal transduction.[38] Whether interventions with TNF-α or IL-6 receptor blockade may be able to decrease the inflammatory response, improve the hemodynamic status and attenuate severity of PCAS and thus affect outcome in an OHCA population may be the subject of future studies. The interpretation of the results of the present study is subject to some important limitations. This study was a post-hoc study of the TTM-trial including only patients at a single centre with available blood samples for analyses of inflammatory and endothelial biomarkers. The hemodynamic management and the use of vasopressors was guided by hemodynamic target goals including MAP ≥ 65 mmHg, which may weaken the possible associations reported between biomarkers and MAP. Need of vasopressor support evaluated by the CVI-score may be a more clinical relevant indicator of hemodynamic dysfunction and was therefore chosen as the primary hemodynamic measure.

Furthermore, echocardiographic and invasive hemodynamic data were missing in a minor proportion of patients at time of randomization, but we do not believe that this would have influenced the results, as we did not observe any trends towards associations between hemodynamic variables and inflammatory and endothelial biomarkers. The observational nature of the study does not allow for independent evaluation of the proposed cause-and-effect relationship. Thus, the present findings should be interpreted as hypothesis generating only in order to form the basis for future studies.

Conclusion: In comatose survivors after OHCA, biomarkers reflecting increasing systemic inflammation and endothelial injury were associated with need for vasopressor support. IL-6 levels were consistently and independently associated with increased need for vasopressor support within 72 h. Target temperature management at 33°C and 36°C modified the association between IL-6 and vasopressor support, presumable due to increased need for vasopressor support in the 33°C group. The correlations between the other investigated inflammatory and endothelial biomarkers showed a temporal increase within 72 h after OHCA, however the strength of the correlations were only modest. Our findings suggest that endothelial injury and systemic inflammation may play a role in PCAS associated cardiovascular dysfunction after OHCA. Interventions aiming at decreasing inflammation and/or protecting the endothelium may be the target of future studies investigating PCAS associated cardiovascular dysfunction.

Conflict of interest statement All authors have reported that they have no financial relationships relevant to the contents of this paper to disclose.

Acknowledgements Karen Dyeremose and Helena Marie Stjernkvist are thanked for the skilled technical assistance.

Funding This work was funded by Trygfonden. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References [1] Nolan JP, Neumar RW, Adrie C, et al. Post-cardiac arrest syndrome: epidemiology, pathophysiology, treatment, and prognostication. A Scientific Statement from the International Liaison Committee on Resuscitation; the American Heart Association Emergency Cardiovascular Care Committee; the Council on Cardiovascular Surgery and Anesthesia; the Council on Cardiopulmonary, Perioperative, and Critical Care; the Council on Clinical Cardiology; the Council on Stroke. Resuscitation. 2008;79:350-79. [2] Bro-Jeppesen J, Johansson PI, Hassager C, et al. Endothelial activation/injury and associations with severity of post-cardiac arrest syndrome and mortality after out-ofhospital cardiac arrest. Resuscitation. 2016;107:71-9. [3] Bro-Jeppesen J, Kjaergaard J, Wanscher M, et al. The inflammatory response after outof-hospital cardiac arrest is not modified by targeted temperature management at 33 degrees C or 36 degrees C. Resuscitation. 2014;85:1480-7. [4] Annborn M, Dankiewicz J, Erlinge D, et al. Procalcitonin after cardiac arrest - an indicator of severity of illness, ischemia-reperfusion injury and outcome. Resuscitation. 2013;84:7827. [5] Niemann JT, Garner D, Lewis RJ. Tumor necrosis factor-alpha is associated with early postresuscitation myocardial dysfunction. Crit Care Med. 2004;32:1753-8. [6] Hollenberg SM, Cunnion RE, Parrillo JE. The effect of tumor necrosis factor on vascular smooth muscle. In vitro studies using rat aortic rings. Chest. 1991;100:1133-7. [7] Nieuwdorp M, Meuwese MC, Vink H, et al. The endothelial glycocalyx: a potential barrier between health and vascular disease. Current opinion in lipidology. 2005;16:507-11. [8] Gando S, Nanzaki S, Morimoto Y, Kobayashi S, Kemmotsu O. Out-of-hospital cardiac arrest increases soluble vascular endothelial adhesion molecules and neutrophil elastase associated with endothelial injury. Intensive Care Med. 2000;26:38-44. [9] Thiagarajan RR, Winn RK, Harlan JM. The role of leukocyte and endothelial adhesion molecules in ischemia-reperfusion injury. Thrombosis and haemostasis. 1997;78:310-4. [10] Aird WC. Endothelium as an organ system. Crit Care Med. 2004;32:S271-9.

[11] Baudry N, Vicaut E. Role of nitric oxide in effects of tumor necrosis factor-alpha on microcirculation in rat. Journal of applied physiology (Bethesda, Md : 1985). 1993;75:23929. [12] Johns DG, Webb RC. TNF-α-induced endothelium-independent vasodilation: a role for phospholipase A₂-dependent ceramide signaling. American Journal of Physiology - Heart and Circulatory Physiology. 1998;275:H1592-H8. [13] Saini HK, Xu YJ, Zhang M, et al. Role of tumour necrosis factor-alpha and other cytokines in ischemia-reperfusion-induced injury in the heart. Exp Clin Cardiol. 2005;10:21322. [14] Adrie C, Adib-Conquy M, Laurent I, et al. Successful cardiopulmonary resuscitation after cardiac arrest as a "sepsis-like" syndrome. Circulation. 2002;106:562-8. [15] Chelazzi C, Villa G, Mancinelli P, De Gaudio AR, Adembri C. Glycocalyx and sepsisinduced alterations in vascular permeability. Crit Care. 2015;19:26. [16] Annborn M, Bro-Jeppesen J, Nielsen N, et al. The association of targeted temperature management at 33 and 36 degrees C with outcome in patients with moderate shock on admission after out-of-hospital cardiac arrest: a post hoc analysis of the Target Temperature Management trial. Intensive Care Med. 2014;40:1210-9. [17] Kilgannon JH, Roberts BW, Reihl LR, et al. Early arterial hypotension is common in the post-cardiac arrest syndrome and associated with increased in-hospital mortality. Resuscitation. 2008;79:410-6. [18] Bergman R, Braber A, Adriaanse MA, et al. Haemodynamic consequences of mild therapeutic hypothermia after cardiac arrest. Eur J Anaesthesiol. 2010;27:383-7. [19] Laurent I, Monchi M, Chiche JD, et al. Reversible myocardial dysfunction in survivors of out-of-hospital cardiac arrest. J Am Coll Cardiol. 2002;40:2110-6. [20] Nielsen N, Wetterslev J, Cronberg T, et al. Targeted temperature management at 33 degrees C versus 36 degrees C after cardiac arrest. N Engl J Med. 2013;369:2197-206. [21] Johansson PI, Bro-Jeppesen J, Kjaergaard J, et al. Sympathoadrenal activation and endothelial damage are inter correlated and predict increased mortality in patients resuscitated after out-of-hospital cardiac arrest. a post hoc sub-study of patients from the TTM-trial. PLoS One. 2015;10:e0120914. [22] Niemann JT, Rosborough JP, Youngquist S, et al. Cardiac function and the proinflammatory cytokine response after recovery from cardiac arrest in swine. J Interferon Cytokine Res. 2009;29:749-58. [23] Nielsen N, Wetterslev J, al-Subaie N, et al. Target Temperature Management after outof-hospital cardiac arrest--a randomized, parallel-group, assessor-blinded clinical trial-rationale and design. Am Heart J. 2012;163:541-8. [24] Bro-Jeppesen J, Hassager C, Wanscher M, et al. Targeted temperature management at 33 degrees C versus 36 degrees C and impact on systemic vascular resistance and myocardial function after out-of-hospital cardiac arrest: a sub-study of the Target Temperature Management Trial. Circ Cardiovasc Interv. 2014;7:663-72.

[25] Trzeciak S, McCoy JV, Phillip Dellinger R, et al. Early increases in microcirculatory perfusion during protocol-directed resuscitation are associated with reduced multi-organ failure at 24 h in patients with sepsis. Intensive Care Med. 2008;34:2210-7. [26] Rehm M, Bruegger D, Christ F, et al. Shedding of the endothelial glycocalyx in patients undergoing major vascular surgery with global and regional ischemia. Circulation. 2007;116:1896-906. [27] Blann AD. Endothelial cell activation, injury, damage and dysfunction: separate entities or mutual terms? Blood Coagul Fibrinolysis. 2000;11:623-30. [28] Blann A, Seigneur M. Soluble markers of endothelial cell function. Clinical hemorheology and microcirculation. 1997;17:3-11. [29] Ishii H, Uchiyama H, Kazama M. Soluble thrombomodulin antigen in conditioned medium is increased by damage of endothelial cells. Thrombosis and haemostasis. 1991;65:618-23. [30] Vestweber D. VE-cadherin: the major endothelial adhesion molecule controlling cellular junctions and blood vessel formation. Arteriosclerosis, thrombosis, and vascular biology. 2008;28:223-32. [31] Kern KB. Postresuscitation myocardial dysfunction. Cardiol Clin. 2002;20:89-101. [32] Kelly RA, Smith TW. Cytokines and cardiac contractile function. Circulation. 1997;95:778-81. [33] Adrie C, Laurent I, Monchi M, et al. Postresuscitation disease after cardiac arrest: a sepsis-like syndrome? Curr Opin Crit Care. 2004;10:208-12. [34] Aird WC. The role of the endothelium in severe sepsis and multiple organ dysfunction syndrome. Blood. 2003;101:3765-77. [35] Roberts BW, Kilgannon JH, Chansky ME, et al. Multiple organ dysfunction after return of spontaneous circulation in postcardiac arrest syndrome. Crit Care Med. 2013;41:1492-501. [36] Niemann JT, Youngquist ST, Shah AP, Thomas JL, Rosborough JP. TNF-alpha blockade improves early post-resuscitation survival and hemodynamics in a swine model of ischemic ventricular fibrillation. Resuscitation. 2013;84:103-7. [37] Youngquist ST, Niemann JT, Shah AP, Thomas JL, Rosborough JP. A comparison of etanercept vs. infliximab for the treatment of post-arrest myocardial dysfunction in a swine model of ventricular fibrillation. Resuscitation. 2013;84:999-1003. [38] Oldfield V, Dhillon S, Plosker GL. Tocilizumab: a review of its use in the management of rheumatoid arthritis. Drugs. 2009;69:609-32.

Figure legend Figure 1: Plots of correlation between corresponding levels of IL-6 (upper panel) and thrombomodulin (lower panel) and CVI-score stratified by level of target temperature management at 24h (left), 48h

(middle) and 72 h (right) after out-of-hospital cardiac arrest. Bars represent median value after logarithmic transformation with error bars representing lower and upper quartile (Q1-Q3).

Abbreviations: IL; interleukin, CVI; cumulative vasopressor index, TTM; target temperature management at 33°C or 36°C.

Table 1: Demographics and hemodynamic parameters, n=163. Low IL-6 n=81 Age, years (mean ± SD) 61±11 Male sex (%) 71 (88) Initial rhythm: Ventricular fibrillation (%) 71 (88) Non-perfusing ventricular tachycardia (%) 1 (1) Asystole (%) 2 (2) Pulseless Electrical Activity (%) 6 (9) ROSC after bystander defibrillation (%) 1 (1) Witnessed arrest (%) 72 (89) Bystander CPR (%) 64 (79) Time to advanced life support, mins (Q1-Q3) 8 (5-11) Number of defibrillations, n (Q1-Q3) 3 (2-4)

High IL-6 n=82 61±11 72 (88) 71 (87) 2 (2) 4 (5) 5 (5) 0 (0) 74 (90) 65 (79) 7 (5-11) 4 (2-7)

p-value 0.91 0.98

0.72

0.78 0.97 0.30 0.003

Dose of adrenaline, n (Q1-Q3) Time to ROSC, mins (Q1-Q3) Lactate, first measured, mmol/L (Q1-Q3) Shock at admission, n (%) LVEF at admission, % (Q1-Q3) Target temperature management (TTM33/TTM36), %/% Hemodynamics: T0: MAP, mmHg HR, min-1 CVI T24: MAP, mmHg HR, min-1 CVI T48: MAP, mmHg HR, min-1 CVI T72: MAP, mmHg HR, min-1 CVI

2 (1-3) 17 (12-25) 5.0 (3.2-7.4) 6 (7) 35 (30-45) 47/53

4 (2-6) 26 (17-37) 8.8 (4.8-13) 11 (13) 35 (25-45) 54/46

<0.0001 <0.0001 <0.0001 0.21 0.18 0.39

75±15 70±17 0 (0-1)

72±12 79±19 0 (0-1)

0.10 0.001 0.83

72±9 67±17 2 (1-2)

74±10 67±19 2 (1-2)

0.35 0.81 0.87

79±14 86±19 1 (0-2)

79±14 87±16 1 (0-2)

0.92 0.80 0.40

80±12 84±18 0 (0-0)

80±17 93±18 0 (0-1)

0.91 0.03 0.43

Data are presented as mean ±SD or median and lower and upper quartiles (Q1-Q3) as appropriate. The p-value represents comparison between groups; Low IL-6 (<87 ng/ml) and High IL-6 (≥ 87 ng/ml) stratified by the median value of IL-6 at randomization. A significance level of p<0.05 was chosen. Abbreviations: CPR, cardio-pulmonary resuscitation; CVI, cumulative vasopressor index; HR, heart rate; LVEF, left ventricular ejection fraction; MAP, mean arterial pressure; ROSC, return of spontaneously circulation; TTM; targeted temperature management.

Table 2: Correlations between inflammatory and endothelial biomarkers and hemodynamic parameters within the first 72 hours after out-of-hospital cardiac arrest. MAP r2

p

HR r2

p

CVI r2

p

T0, n=163 PCT IL-6 TNF-α IL-10 E-selectin Thrombomodulin Syndecan-1 VE-cadherin

-0.04 -0.19 0.04 -0.12 -0.01 -0.03 -0.07 0.05

0.68 0.03 0.66 0.15 0.87 0.68 0.36 0.54

0.06 0.29 0.02 0.26 0.05 0.18 0.18 -0.06

0.47 0.0002 0.83 0.001 0.54 0.02 0.03 0.45

0.08 0.10 -0.10 0.08 0.09 0.12 0.07 -0.07

0.35 0.21 0.21 0.35 0.25 0.13 0.39 0.33

T24, n=149 PCT IL-6 TNF-α

-0.12 0.002 0.03

0.18 0.97 0.70

0.22 0.29 -0.09

0.01 0.0003 0.26

0.13 0.19 0.06

0.13 0.02 0.45

IL-10 E-selectin Thrombomodulin Syndecan-1 VE-cadherin

0.10 -0.12 -0.11 -0.11 -0.03

0.24 0.14 0.19 0.21 0.71

-0.02 0.16 0.20 0.19 0.14

0.82 0.06 0.01 0.02 0.10

-0.07 0.09 0.13 0.03 -0.02

0.38 0.25 0.11 0.71 0.77

T48, n=145 PCT IL-6 TNF-α IL-10 E-selectin Thrombomodulin Syndecan-1 VE-cadherin

-0.15 -0.19 0.09 0.02 -0.08 -0.13 -0.45 0.02

0.10 0.03 0.34 0.83 0.38 0.14 0.004 0.86

0.25 0.16 -0.20 -0.07 0.13 0.07 0.007 -0.05

0.004 0.06 0.02 0.41 0.14 0.43 0.93 0.60

0.26 0.31 -0.03 0.04 0.14 0.23 0.27 -0.16

0.002 0.0001 0.74 0.61 0.09 0.004 0.001 0.06

T72, n=145 PCT -0.11 0.37 0.31 0.01 0.27 0.002 IL-6 -0.33 0.005 0.25 0.04 0.39 <0.0001 TNF-α -0.03 0.82 0.11 0.36 -0.26 0.002 IL-10 -0.08 0.54 0.23 0.07 0.05 0.58 E-selectin -0.39 0.0006 0.23 0.31 0.17 0.04 Thrombomodulin -0.21 0.08 0.15 0.20 0.25 0.002 Syndecan-1 -0.09 0.46 0.07 0.57 0.23 0.005 VE-cadherin 0.12 0.35 0.12 0.33 -0.18 0.03 Abbreviations: CVI; cumulative vasopressor index, endothelial cell activation (sE-selectin), endothelial cell injury (thrombomodulin), endothelial junction disruption (sVE-cadherin), glycocalyx damage (syndecan-1), HR; heart rate, interleukin-6 (IL-6), interleukin-10 (IL-10), MAP; mean arterial pressure, PCT; procalcitonin, Tumor Necrosis Factor-α (TNF-α).

Table 3: Correlations between inflammatory and endothelial biomarkers and echocardiographic and invasive hemodynamic parameters at admission after out-of-hospital cardiac arrest.

Echocardiography

Pulmonary arterial catheter

n=139

n=154

LVEF

s’

E/e’

TAPSE

CI

SV I

SV RI

r2

p

r2

p

r2

p

r2

p

r2

p

r2

p

r2

p

PCT

0.0 7

0.4 5

0.0 01

0. 99

0. 13

0. 19

0.0 3

0. 79

0.0 5

0. 57

0. 12

0. 19

0.0 01

0. 99

IL-6

0.0 9

0.3 1

0.0 9

0. 32

0. 05

0. 59

0.0 9

0. 31

0.0 2

0. 83

0. 16

0. 06

0.0 4

0. 65

TNF-α

0.0

0.6

0.0

0.

0.

0.

0.0

0.

0.0

0.

0.

0.

0.0

0.

4

2

2

81

02

78

8

40

2

29

05

59

6

48

IL-10

0.0 2

0.8 3

0.2 1

0. 02

0. 02

0. 85

0.1 0

0. 26

0.0 6

0. 51

0. 06

0. 48

0.0 8

0. 31

sE-selectin

0.2 3

0.0 07

0.0 08

0. 93

0. 10

0. 25

0.1 0

0. 25

0.1 0

0. 22

0. 03

0. 73

0.1 0

0. 22

Thrombom odulin

0.0 3

0.7 2

0.0 7

0. 43

0. 04

0. 65

0.0 2

0. 85

0.0 1

0. 88

0. 08

0. 35

0.0 2

0. 81

Syndecan-1

0.0 07

0.9 4

0.2 2

0. 02

0. 03

0. 70

0.0 02

0. 99

0.1 3

0. 11

0. 21

0. 01

0.1 0

0. 23

sVEcadherin

0.0 2

0.7 9

0.0 8

0. 39

0. 05

0. 58

0.1 0

0. 27

0.0 05

0. 95

0. 01

0. 88

0.0 1

0. 89

Abbreviations: CI; cardiac index, early peak transmitral filling velocity (E), endothelial cell activation (sE-selectin), endothelial cell injury (thrombomodulin), endothelial junction disruption (sVE-cadherin), glycocalyx damage (syndecan-1), interleukin-6 (IL-6), interleukin-10 (IL-10), lateral peak early diastolic tissue velocities (e’), lateral peak systolic (s’) tissue velocities, PCT; procalcitonin, SVI; stroke volume index, SVRI; systemic vascular resistance index, tricuspid annular plane systolic excursion (TAPSE), Tumor Necrosis Factor-α (TNF-α).

Table 4: Associations between cumulative vasopressor index, demographics and inflammatory and endothelial biomarkers. Cumulative Vasopressor Index 24 h after OHCA Univariable

β (95%CI) Age Sex

per 5 years Male

Initial rhythm

VF/VT

Witnessed arrest

Yes

Bystander CPR

Yes

Time to ROSC Initial lactate

per 5 min mmol L-1

TTM-group

36°C

PCT

2-fold

IL-6

2-fold

TNF-α

2-fold

IL-10

2-fold

sE-selectin

2-fold

Thrombomodulin

2-fold

0.06 (-0.02-0.1) 0.1 (-0.6-0.5) -0.06 (-0.7-0.5) -0.1 (-0.7-0.5) -0.2 (-0.7-0.2) 0.09 (0.04-0.1) 0.05 (0.01-0.09) -0.1 (-0.5-0.2) 0.1 (0.03-0.2) 0.2 (0.05-0.3) 0.07 (-0.08-0.2) -0.05 (-0.2-0.08) 0.2 (-0.20.5) 0.2 (-0.050.5)

pvalue

Multivariable Adj. R2=0.11 β (95%CI)

p-value

0.14 0.88 0.84 0.65 0.34 0.0004

0.07 (0.02-0.1)

0.01

0.003 NS

0.50 0.009 0.004

NS 0.1 (0.01-0.2)

0.02

0.36 0.44 0.32 0.10

NS

48 h after OHCA Univariable

β (95%CI) 0.07 (-0.02-0.2) 0.03 (-0.6-0.7) -0.3 (-1.0-0.5) -0.01 (-0.7-0.7) -0.1 (-0.6-0.4) 0.1 (0.07-0.2) 0.06 (0.02-0.1) -0.5 (-0.9- -0.08) 0.2 (0.06-0.2) 0.2 (0.1-0.3) -0.08 (-0.3-0.08) -0.02 (-0.2-0.1) 0.3 (-0.1-0.8) 0.6 (0.3-0.9)

p-value

72 h after OHCA Univariable

Multivariable Adj. R2=0.22 β (95%CI)

p-value

0.14 0.90 0.49 0.96 0.70 <0.0001

0.11 (0.06-0.2)

0.005 0.02

NS -0.5 (-0.9- -0.1)

0.0006 0.0001

0.0001

0.01 NS

0.2 (0.06-0.3)

0.004

0.32 0.80 0.13 0.0005

NS

β (95%CI) 0.07 (-0.02-0.2) 0.3 (-0.2-0.9) -0.2 (-1.0-0.5) 0.3 (-0.4-1.0) -0.4 (-0.9- -0.01) 0.05 (0.01-0.1) 0.05 (0.01-0.1) -0.6 (-1.0- -0.2) 0.1 (0.02-0.2) 0.3 (0.2-0.4) -0.3 (-0.4- -0.09) 0.02 (-0.1-0.2) 0.1 (-0.3-0.5) 0.6 (0.3-0.8)

p-value

Multivariable Adj. R2=0.33 β (95%CI)

p-value

0.13 0.23 0.57 0.44 0.06

NS

0.08

NS

0.01

NS

0.001

-0.4 (-0.8- -0.1)

0.02 <0.0001 0.002

0.008 NS

0.3 (0.2-0.4) -0.4 (-0.5- -0.2)

<0.0001 <0.0001

0.78 0.52 <0.0001

NS

Syndecan-1

2-fold

sVE-cadherin

2-fold

0.09 (-0.05-0.2) -0.11 (-0.4-0.2)

0.18 0.41

0.3 (0.2-0.5) -0.6 (-1.3-0.03)

0.0002

NS

0.06

NS

0.3 (0.1-0.4) -0.6 (-1.3-0.05)

0.003

NS

0.07

NS

Regression coefficients (β) with 95% confidence intervals (95%CI), p-values are displayed for univariable and multivariable linear models showing the associations between the covariates and cumulative vasopressor index (CVI) after out-of-hospital cardiac arrest (OHCA). Covariates with p≤0.10 in univariate analysis were included in multivariable analysis and adjusted R2 values are reported. Predicted changes in CVI-score associated with one unit increase in the explanatory variables (age (5 years older), sex (male), initial rhythm (ventricular fibrillation/ventricular tachycardia), witnessed arrest (yes), bystander cardiopulmonary resuscitation (CPR) (yes), time to return of spontaneous circulation (ROSC) (per 5 minutes), initial levels of lactate (per mmol/L), level of target temperature management (TTM36 vs. TTM33), procalcitonin (PCT), interleukin-6 (IL-6), Tumor Necrosis Factor- α (TNF-α), interleukin-10 (IL10), endothelial cell activation (sE-selectin), endothelial cell injury (thrombomodulin), glycocalyx damage (syndecan-1), endothelial junction disruption (sVEcadherin) (all log-transformed, 2-fold higher). IL-6, IL-10, PCT, TNF-α, sE-selectin, thrombomodulin, syndecan-1 and sVE-cadherin levels at 24h, 48h and 72h were entered in the models of CVI at 24h, 48h and 72h, respectively. NS, non-significant.