Inflammatory response after out‐of‐hospital cardiac arrest—Impact on outcome and organ failure development

Post‐cardiac arrest syndrome that occurs in out‐of‐hospital cardiac arrest (OHCA) patients is characterized by inflammatory response. We conducted a scoping review of current evidence regarding several inflammatory markers' usefulness for assessment of patient outcome and illness severity. We also discuss the proposed underlying mechanisms leading to inflammatory response after OHCA.


Editorial Comment
Use of in inflammatory biomarkers, including CRP, PCT and IL-6 may prove useful in the prognostification of out-of-hospital cardiac arrest along with existing neurological assessments.

| BACKGROUND
Patients successfully resuscitated after out-of-hospital cardiac arrest (OHCA) often develop a pathological state called post-cardiac arrest syndrome (PCAS). 1 PCAS is a complex syndrome that includes various degrees of hypoxic brain injury, myocardial dysfunction, vasodilatation and injury to the kidney and liver.The pathophysiological process behind PCAS is complex but is precipitated by whole-body ischemiareperfusion injury, 1 damaging tissue to various degrees depending on its resilience to ischemia and leading to systemic inflammation, which shares many similarities with sepsis. 2 Increased levels of inflammatory markers, such as procalcitonin (PCT) and interleukin-6 (IL-6), are detected in circulation during PCAS, and their levels have been studied for associations with organ dysfunction and OHCA outcome. 3,4e rationale of these studies is to find an inflammatory marker that would reliably reflect the severity of PCAS and predict evolving organ failures in the short term and mortality and neurological outcomes in the long term.Due to the central nervous system (CNS)'s vulnerability to ischemia and secondary neuroinflammation and its importance for overall recovery, neurological organ failure is in focus after OHCA.However, extracerebral organ failures during post-arrest care also impact on outcome and can be measured with the Sequential Organ Failure Assessment (SOFA) score. 5The SOFA score is widely used for the assessment of organ failure in intensive care and grades the failures of six organ systems (CNS, respiratory, cardiovascular, coagulation, hepatic and renal), based on readily available clinical and laboratory variables. 6e aim of this scoping review is to assess the reported accuracy of inflammatory markers to predict patient outcome and organ failure, with a special focus on the use of methods to adjust for confounding variables in the publications; evaluate the impact of the overall event rate on the predictive accuracy of inflammatory markers; provide an overview of the biological background and possible causality linking inflammation with patient outcome and organ failure after outof-hospital cardiac arrest.

| METHODS
We reviewed the published findings on the prognostic value of inflammatory markers after OHCA, focusing on mortality and neurological outcome up to 12-month follow-up, and the severity of organ failure.Poor outcome was considered as Cerebral Performance Category (CPC) score 3-5, Glasgow Outcome Scale (GOS) 1-3 or nonsurvivor status.As a secondary aim, we report whether the accuracy of these inflammatory markers is different based on the overall event rate of the included studies.
We searched the MEDLINE, PubMed Central, Cochrane CEN-TRAL and Web of Science Core Collection databases from inception to 15 March 2023 for articles on inflammatory response after OHCA in adult patients using the search terms ("inflammation" OR "cytokines") AND "out-of-hospital cardiac arrest."We performed additional searches combining "out-of-hospital cardiac arrest" with the individual inflammatory marker search terms found in the initial search.The exact search terms are provided in an additional file (File S1).Initial screening of the articles was performed by two authors (AMJS, PTP) and the eligibility of the final included studies was confirmed by the same authors.The inclusion and exclusion criteria were decided a priori.The search results and suggested decision of inclusion/ exclusion with reasoning were independently collected in table format by the two authors during screening.Possible disagreements on inclusion were discussed and resolved in unison.Central characteristics of the selected articles were then collected to Table 1 by AMJS.Specific forms for data charting were not used.For synthesis of results, it was decided a priori that AUC would be retrieved from the articles as a simple indicator of association, and that the use of a multivariable model to adjust for confounding variables would be considered necessary for claiming predictive value.
The inclusion criteria for this scoping review were: original article; adult (≥18 years) OHCA patients; the concentration of at least one circulating inflammatory marker had to be measured within 72 h after return of spontaneous circulation (ROSC); organ failure during intensive care or mortality/neurological outcome within the first year after ROSC had to be assessed; statistical significance of the association of the inflammatory marker(s) with outcome or organ failure had to be assessed.Studies on pediatric patients and on animal models were excluded.The study was conducted according to the PRISMA for Scoping Review guidelines (File S3).The study was not registered to PROSPERO.To assess the risk of bias in eligible studies we used Quality In Prognosis Studies tool (File S2). 7

| RESULTS
We found 665 studies, of which 369 were duplicates.Two hundred ninety-six studies were screened for eligibility based on abstract, after which 93 full-text articles were assessed.Fifty-nine studies met our inclusion criteria and were reviewed (Figure 1).In total, we found 65 different inflammatory markers.Among the reviewed studies, IL-6, PCT and C-reactive protein (CRP) were the most extensively studied inflammatory markers with 15, 14 and 17 studies found, respectively.Risk of bias was low in 26/59 studies and moderate in 28/59 studies.
Five studies were of high risk of bias (File S2).
T A B L E 1 Studies included in the review.(Continues)
Eight studies included IL-6 in a multivariable model.Bro-Jeppesen et al. reported in a multicenter study of 682 patients that IL-6 on days two and three was independently associated with 30-day mortality, and on days one to three with SOFA scores of the corresponding days. 43][14][15] Four studies reported that IL-6 at admission had independent prognostic value for ICU mortality, 30-day mortality, six-month neurological outcome and 12-month neurological outcome. 18,19,24,25 Six studies tested the association of IL-6 levels with outcome without adjusting for confounders.Admission IL-6 was found to differentiate neurological outcome and ICU survival in four studies, 16,22,23,26 and neurological outcome on days two and three in one study. 20In a study of 13 patients, IL-6 level in cerebrospinal fluid (CSF), but not in serum, on day two after ROSC was associated with six-month neurological outcome. 17evated IL-6 levels were not associated with prolonged need for circulatory support during the first 5 days of post-resuscitation care. 27study by Akin et al. showed that IL-6 at admission combined with NSE on day three improved prognostic performance for sixmonth neurological outcomes compared to NSE on day three in a multivariable model.25 Admission IL-6 was also superior to admission NSE for predicting 30-day neurological outcome.21

| Procalcitonin
In healthy individuals, small amounts of procalcitonin (PCT) are produced in the thyroid gland, but levels in circulation are not detectable. 28Under inflammatory conditions, PCT production is induced in parenchyma in response to ischemia, monocyte activation, IL-6, 29,30 lipopolysaccharide and IL-1β. 31The prognostic value of PCT after OHCA has been recently evaluated in a metaanalysis. 32 found 14 individual studies on PCT, 3,[11][12][13]15,20,21,25,[33][34][35][36][37][38][39][40] including a total of 2608 patients. Nine studies repod an AUC value for the association between PCT concentration and poor outcome.In four studies, AUC was highest on day one, with a median of 0.84 and values ranging between 0.61 and 0.98 (Table 1, not all day one values shown).PCT concentrations at intensive care unit (ICU) admission were relatively low, with a clear increase within 24 h in the poor outcome group (Figure 2).
Six studies evaluated the independent prognostic performance of PCT after OHCA in multivariable models.Confounding variables were chosen differently, but age and initial rhythm were included in every study.In a study by Pekkarinen et al., PCT on day four (but not at earlier time points) was an independent predictor of 12-month poor neurological outcome. 3Bro-Jeppesen et al. reported that a two-fold increase in PCT between admission and day one independently predicted 30-day mortality. 13Qi et al. showed that PCT had an independent prognostic value for 28-day mortality and neurological outcome at admission, day one and day three. 34PCT at admission was also found to predict in-hospital mortality and poor neurological outcome at hospital discharge but not independently and also being inferior to NSE. 40 Also, Zelniker et al. found no independent prognostic value of PCT, although an association with mortality was recorded. 11Pekkarinen et al. found no independent prognostic value of PCT when day two neurofilament light (NFL) was included in the multivariable model but a significant fixed effect for PCT was found for poor six-month neurological outcome. 39ven studies assessed PCT's prognostic value for neurological outcome or mortality after OHCA without adjusting for confounding variables.In a multicenter study of 932 patients, PCT on days one to three was significantly higher in patients with poor six-month neurological outcomes. 20In smaller single-center studies, PCT at 12 h and on days one to three was associated with poor six-month neurological outcome, and PCT on day one was associated with poor 14-day neurological outcome. 35,37,38No association with neurological outcome was found at ICU admission. 21In a study by Adib-Conquy et al., significant elevations in PCT levels between admission and day one were seen in nonsurvivors. 36The elevation was greater in those who died of refractory shock rather than neurological failure.Mongardon et al.
and Akin et al. also reported significant differences in PCT levels between survivors and nonsurvivors when measured at admission and on days one to three. 25,33e association between severity of organ failure and PCT was assessed in four studies.SOFA score as an indicator of organ failure correlated with PCT on days one to three in two studies.

| C-reactive protein
CRP is an acute-phase protein released by hepatocytes, where its transcription is induced mainly by IL-6 and IL-1β. 41Thirteen individual studies with a total of 3005 patients evaluated CRP as a prognostic marker, 12,13,16,20,21,25,27,33,[42][43][44][45][46][47] and four studies with 792 patients assessed high-sensitivity CRP (hsCRP). 19,39,48,49While conventional CRP assays have a detection limit of 3 to 8 mg/L, hsCRP assays can detect CRP levels as low as 0.3 mg/L. 50Seven studies reported AUC values for poor outcome ranging from 0.52 to 0.76 (Table 1).The time point of best AUC varied greatly among studies from admission to day three with highest value of 0.76 on day three (Table 1).At ICU admission, the concentration of CRP was usually low but had an upward trend in later time points in both good and poor outcome groups (Figure 2).
A single-center study of 832 patients by Schriefl et al. showed that CRP measured at admission had an AUC of 0.60 but was nevertheless independently associated with 30-day neurological outcome. 43Three studies demonstrated an association of higher CRP levels at admission with mortality but no independent association in multivariable model (not tested in two and not significant in the third), 25,33,42 and two found no difference between neurological outcome groups. 21,46Gregers et al. reported that lower CRP levels at admission and on day one independently predict poor neurological outcome at hospital discharge in a study where all patients underwent extracorporeal cardiopulmonary resuscitation. 45Studies on later time points reported conflicting results: Kim et al. reported unadjusted association with outcome on days one to three, 44  an independent association on day two, 16,48,49 whereas three studies found no association at any time point. 12,13,19,47Most of these studies also reported negative results for other time points.Pekkarinen et al.
reported significant fixed effect of hsCRP for neurological outcome, but this was not independent when combined with day two NFL in a multivariable model. 39garding organ failure, Bro-Jeppesen et al. and Ristagno et al.
demonstrated an unadjusted association of elevated CRP levels with higher SOFA scores, whereas Vaahersalo et al. reported negative results for similar assessment. 12,13,19,48CRP was also not associated with prolonged need for circulatory support during the first 5 days of post-resuscitation care. 27combination marker, CRP to albumin ratio (CAR), was assessed in two studies and reported to have independent value for predicting in-hospital and six-month mortality.42,44

| Leukocyte markers
White blood cell count (WBC) is routinely available in the ICU setting.
We found no studies focusing primarily on WBC as a marker after OHCA.Three studies reported independent prognostic value; lower WBC at admission was associated with poor neurological outcome at hospital discharge in a study where all patients underwent extracorporeal cardiopulmonary resuscitation, 45 and higher WBC and neutrophils on day one were associated with poor six-month neurological outcome in another study. 46Higher WBC on day one independently predicted poor neurological outcome at hospital discharge. 51ven studies reported no differences in WBC between outcome groups 16,21,22,40,42,49,52 and one reported higher WBC (unadjusted) at admission in 30-day nonsurvivors. 53e prognostic value of the neutrophil-to-lymphocyte ratio (NLR) 5][56] Weiser et al. showed that an admission NLR ≥ 6 was independently associated with one-year mortality in a study of 1188 patients. 54Two smaller studies reported independent associations of admission NLR 42 or an increase of NLR during therapeutic hypothermia 55 with in-hospital mortality and one study found that NLR at 72 h independently predicted neurological outcome. 49Also NLR measured after completing targeted temperature management was an independent predictor for in-hospital mortality. 56e delta neutrophil index (DNI) measured at admission was independently associated with both 30-day mortality and neurological outcome in one study. 53The immature/total granulocyte (I/T-G) ratio at admission was associated with ICU mortality in one study without adjustment. 52Significantly lower regulatory T-cell count on day one in 28-day nonsurvivors was reported without adjustment 57 but T-helper cell types 1, 2 and 17 were not different between outcome groups. 58

| Other inflammatory markers
[61][62] Resistin is a marker of neutrophil activation. 63Two studies, one released as two papers, reported an unadjusted association of resistin with 30-day mortality. 11,64,65When measured within 24 h of ROSC, resistin was more strongly associated with 30-day mortality than NSE. 65One study reported no association with six-month neurological outcome. 39GB1 is a nuclear protein of human cells released into circulation from damaged tissue or activated monocytes and macrophages. 66GB1 was assessed in two unadjusted studies with limited sample sizes.One reported an association of plasma HMGB1 on day one with neurological outcome at 30 days. 23Another study found association of intrathecal HMGB1 measured at 48 h (the only time point of intrathecal measurement) with six-month neurological outcome, but not of HMGB1 measured from serum at 6, 24 or 48 h. 67esepsin is the soluble form of CD14, a TLR4 coreceptor binding endotoxin (LPS). 68It was evaluated as a prognostic marker in three studies. 3,34,47In a study by Pekkarinen et al., presepsin at ICU admission was associated with 12-month poor neurologic outcome and ICU death, but the association was no longer significant in a multivariable model. 3Another study, on the contrary, reported admission presepsin to be an independent prognostic marker for a 28-day outcome. 34Presepsin measured within 12 h of ROSC was not associated with 30-day mortality. 47mission IL-8 was associated with in-hospital and one-week mortality in two separate unadjusted studies, but two other studies found no significant differences between the outcome groups. 11,18,69,70Intrathecal but not plasma IL-8 on day two was significantly higher in patients with poor six-month neurological outcomes. 17teopontin (OPN) is a chemoattractant for macrophages and T cells. 71Significant fixed effect during the first 3 days after ROSC was found between OPN and poor six-month neurological outcome. 39terleukin-1 receptor antagonist (IL-1Ra), IL-10, thioredoxin (TRX), pentraxin 3 (PTX3), soluble suppression of tumorigenicity 2 (sST2), kynurenine (KYN), kynurenine to tryptophan ratio (KYN/TRP), kynurenic acid (KYNA) and 3-hydroxyanthranilic acid (3-HAA) measured at ICU admission were associated with ICU F I G U R E 3 Spearman's rank correlation coefficient between the AUC and the overall event rate.AUC and overall event rates for studies in which both were available (Table 1).The study sample size is represented by the dot size, but Spearman rank correlation was not adjusted by study sample size.AUC, area under receiver operating characteristic curve; CRP, C-reactive protein; hsCRP, high-sensitivity C-reactive protein; IL-6, interleukin-6; PCT, procalcitonin.
GDF-15 at admission was also independently associated with increased six-month mortality. 75Perforin-2 at admission was independently associated with both neurological outcome at hospital discharge and 90-day mortality. 76There was a significant but unadjusted increase in sTREM-1 levels between admission and day one in OHCA patients who died of shock but not in survivors. 36Semaphorin 3F on day three was significantly higher in patients who did not survive to hospital discharge but no adjusted model was applied. 77IL-1α, IL-4Ra and TNF-α on day two were associated with 30-day outcome but not independently. 11viewed inflammatory markers that were unable to differentiate outcome groups were macrophage migration inhibitory factor (MIF) 78 and, as part of larger panels, interferon gamma, IL-1β, IL-2, IL-4, IL-5, IL-9, IL-12, IL-13, IL-17A, IL-21, IL-23, eotaxin, granzyme B, CXCL10 (IP-10), CCL2 (MCP1), MIP-1a, MIP-1b, myeloperoxidase and TNFSF12/TWEAK. 11,13,39

| Association between AUC and overall event rate
The overall mortality rates and distributions in neurological outcome categories varied greatly (Table 1).Since patients in a more severe condition are expected to have a more pronounced inflammatory response, we determined Spearman rank correlation coefficients between the inflammatory markers' AUC and the proportion of patients with poor outcome in studies where both were available (Table 1).We limited the assessment to IL-6, PCT and CRP/hsCRP as the most extensively studied markers.Spearman's Rho values for IL-6, PCT and CRP/hsCRP were 0.33, 0.41 and À0.20, respectively (Figure 3).studies.Procalcitonin at 12-24 h or its increase from admission to 24 h was also associated with outcome in several reports.Elevated (hs)CRP levels were associated with outcome and organ failure, but its prognostic value is hampered by delayed kinetics and large overlap between outcome groups (Figure 2).
We observed a trend of higher AUC values of PCT and IL-6 being reported in studies with higher overall poor outcome event rate (Figure 3).While this is in line with inflammation being detrimental after OHCA, the differences in overall event rates between studies impede the generalizability of the results.If the statistical significance for the association of a marker with outcome stems from high values in clearly moribund cases, measuring the marker in cases where the odds are more even may not bring reasonable value to the clinical decision-making.The use of multivariable models (37 out of 59 studies, Table 1) to adjust for confounding factors partly overcame this issue, but the selection of confounding variables was inconsistent.
To be clinically relevant at ICU admission or during the first 24 h after ROSC, the marker should have independent predictive value after adjusting for Utstein-style variables reflecting demographics, comorbidities and the resuscitation event. 79]18 At later time points, the marker should bring additional value to the recommended multimodal predictive assessment 72 h after the OHCA. 80The minimum requirement for a clinically relevant inflammatory marker measured at 48 or 72 h is superiority or nonredundancy to NSE measured at the same time point.Only two studies tested the predictive value of an inflammatory marker in combination with NSE; in one study IL-6 at admission combined with NSE on day three improved prognostication for neurological outcome while and one found no additional value when PCT was combined with NSE. 25,35ree studies compared markers prognostic capabilities with NSE at admission; IL-6 and resistin were superior to NSE in predicting outcome while PCT was not. 21,40,65 In any case, the stage is set for systemic inflammation after OHCA.Immunologically active lung tissue is damaged by compressions and ventilation with 100% oxygen.Aspiration pneumonitis is common, 81 the ischemic gut leaks PAMPs such as endotoxin into the lymph and portal circulation, 81 and ischemic damage occurs in the heart and the CNS, attracting neutrophils to infiltrate tissue to cause further damage. 82The brain, lungs and gut may even form a vicious circle of inflammation (Figure 2). 81These observations support the idea that inflammation causes secondary damage after OHCA.
The most studied inflammatory markers, IL-6, PCT, CRP and leukocyte count-based variables, are all linked, since IL-6 induces the production of PCT and CRP and promotes leukocyte recruitment. 30,41,83IL-6 is an endogenous pyrogen produced by macrophages and stromal cells in response to TLR stimulation by DAMPs and PAMPs. 9In OHCA, liver macrophages reacting to endotoxins released from the ischemic gut are a plausible source for IL-6, but DAMPs released from ischemic tissue may induce IL-6 production by tissue-resident macrophages throughout the body.In circulating monocytes of OHCA patients, NLRP3 inflammasome activation has been reported, 84 85 where it functions as a myokine and has a role in energy metabolism.Muscle tissue releases IL-6 during exercise in a lactate-dependent manner. 10There may not be enough time for de novo protein synthesis in macrophages to explain the rapid elevation of IL-6 already at ICU admission, and striated muscle as a potential source for IL-6 after OHCA warrants further study.A role for muscle tissue and its numerous oxygen-dependent mitochondria in PCAS is suggested by the association of GDF-15 and fibroblast growth factor 21 (FGF-21), markers of mitochondrial stress, with OHCA outcome. 74,86Maybe the systemic response after cardiac arrest is not initially a "sepsis-like syndrome," 2 but rather resembles extremely exhaustive physical exercise?
The mechanisms by which excess inflammation may impair outcome include direct inflammatory damage to brain tissue and systemic circulatory failure combined with impaired microcirculation of multiple organs due to local inflammation, edema and immunothrombosis. 81tracerebral organ failure is associated with poor outcomes after OHCA, 5

| STRENGTHS AND LIMITATIONS OF THE STUDY
A major strength of our review was a systematic approach to retrieve relevant studies from multiple databases, but heterogeneity and usage of unadjusted models in the included studies were limitations.Heterogeneity of the studies and the large number of different markers made a meta-analysis unfeasible.Our study was limited to OHCA patients, limiting the generalizability of the results.In this scoping review, we wanted to give an overview of the numerous different inflammatory markers studied in the context of OHCA.The broad scope precluded a detailed analysis of single markers.In some cases, it was not obvious if a given marker is inflammatory or not.In these cases, we attempted to be inclusive rather than exclusive.

| CONCLUSIONS
An inflammatory response resulting in the release of multiple inflammatory mediators into circulation is induced after OHCA.Therefore, inflammatory markers are potentially useful for early risk stratification after OHCA but thus far no study has been able to use this information in patient management.PCT and IL-6 have moderate prognostic value for outcome during the first 24 h of the ICU stay.The outcome prediction accuracy appears to be associated with the study overall event rate, the association being stronger in studies with higher mortality due to more severely ill patients.Studies on the association of inflammatory marker levels with multiple organ failure have had conflicting results.
Andersson et al. on day one, 20 Chong et al. and Kim et al. on day three and Ristagno et al.

4 | DISCUSSION 4 . 1 |
Prognostic value of inflammatory markers after out-of-hospital cardiac arrestIn this scoping review, we summarized the published data on the association of inflammatory markers with mortality, neurological outcome, and organ failure in adult OHCA patients.The 59 articles matching the inclusion criteria clearly indicate that systemic inflammation is induced after OHCA, since most of the studied markers displayed elevated levels in circulation during the first days of post-resuscitation care.However, the prognostic value and temporal kinetics of the different markers varied greatly.Of the most studied markers, IL-6 and the parameters reflecting the mobilization of neutrophils into circulation (NLR, DNI, I/T-G ratio and crude WBC) measured at ICU admission were associated with outcome or organ failure in multiple and many of the reviewed studies reported an association of inflammatory markers with multiple organ failure, usually measured with the SOFA score.Due to the limitations of the cardiovascular component of the SOFA score, 87 studies focusing on the association of inflammatory markers with cardiovascular failure often used other measures.Still, the association of inflammation with organ failure is less clear than the association with outcome.Further studies are needed to determine if inflammation is a causal factor or a result of neurological damage leading to poor outcomes after OHCA.Studies on inflammatory markers with prognostic value should focus on the early post-ROSC period up to 24 h.This warrants studies on markers that do not require de novo protein synthesis, such as complement activation products and DAMPs like HMGB1.Furthermore, any new biomarker studied for potential prognostic value should be tested in a multivariable model with prognostic variables currently in clinical use.