Investigating Novel and Conventional Biomarkers for Post-resuscitation Prognosis: The Role of Cytokeratin-18, Neuron-specic Enolase, and Lactate

Introduction: We investigated the associations of cell death-related release of cytokeratin-18 (CK-18) and its caspase-cleaved form (CCCK-18) with outcomes, clinical parameters, and neuron-specic enolase (NSE) among cardiac arrest (CA) victims. Methods: We determined CK-18, CCCK-18, and NSE plasma concentrations within 6, at 24, and 72 post-CA hours of 54 successfully resuscitated patients. Baseline characteristics, laboratory, vital parameters, prognostic scores were recorded. The follow-up was 30 days. Results: Neither the absolute values of CK-18, CCCK-18, CCCK-18/CK-18 ratio, or NSE, nor their kinetics showed signicant difference according to survival and neurological outcome. CK-18 decreased in good renal function in contrast to renal failure. Lactate levels were increased in non-survivors and poor neurological outcomes. In receiver operator characteristic (ROC) analyses lactate was comparable with prognostic scores in predicting 30-day mortality and neurological outcome. The ROC curve of initial lactate indicated a cut-off as 4.90 mmol/L to predict 30-day mortality (specicity=80%; sensitivity=74%) and 6.00 mmol/L (specicity=80%; sensitivity=84%) for poor neurological outcome. Conclusion: This study was the rst to investigate the post-resuscitation prognostic value of CK-18 and CCCK-18 with NSE. Although we could not prove their prognostic value, we conrmed the role of initial lactate in predicting mortality and neurological outcome in unselected resuscitated patients. Continuous data are presented as median values with interquartile range [percentiles 25–75] or mean ± standard deviation, categorical data as the number of subjects and percentages. CA: cardiac arrest, CPR: cardiopulmonary resuscitation; SOFA: Sequential Organ Failure Assessment Score; SAPS II: Simplied Acute Physiology Score II; ICU: intensive care unit. arrest; CRP = C-Reactive Protein; INR International Normalized Ratio; non-signicant 0.05)


Introduction
Even after successful cardiopulmonary resuscitation (CPR), the long-term survival of cardiac arrest (CA) victims remains poor despite all efforts 1,2 . Early death within 3 days after out-of-hospital cardiac arrest (OHCA) occurs mostly due to circulatory failure, while later death is mainly related to severe hypoxicischemic encephalopathy and the subsequent withdrawal of life-sustaining therapy following prognostication of poor neurological outcome, which is responsible for approximately half of the deaths 3,4 . Two-thirds of patients surviving admission to intensive care unit (ICU) after OHCA die due to brain injury, while neurological death only occurs in one-fourth of patients with in-hospital cardiac arrest (IHCA), in whom multiple organ failure drives mortality 5,6 .
Predicting the overall survival and neurological function of CA victims is amongst the biggest challenges facing the medical team. Early predictors of outcome that would support clinical decision-making are required to avoid costly and intensive resources in cases of futility, to guide level-of-care decisions and goals-of-care conversations with family members 7 . The current guidelines and studies are mostly dealing with neurological prognostication in patients remaining comatose after CA 8, 9 , while high proportion of patients (especially after IHCA) reach acceptable neurological function but suffer from multiple organ failure, which may lead to death independently of neurological status 10 .
In the past decades, multiple biomarkers have been tested for neuroprognostication after CA (e.g. NSEneuron-speci c enolase; S100 calcium-binding protein B; neuro lament light chain, tau protein, glial brillary acidic protein) 11 , which are released by injury of neuronal or glial cells re ecting the neurological function. Besides, it would be necessary to nd a reliable biomarker in unselected resuscitated population after cardiac arrest which could provide useful information about the general outcome and survival without focusing only on the neurological status.
Ischemia and reperfusion cause intracellular Ca 2+ overload and generation of reactive oxygen species conditions, what predisposes to mitochondrial injury 12 and cytochrome c release 13 resulting in the activation of executioner caspases and cleavage of cytokeratins through complex biochemical pathways 14,15 . Cytochrome c -as a cell death marker -is released into the bloodstream after resuscitation from cardiac arrest in a rat model of ventricular brillation and closed-chest resuscitation 16 . Cytokeratins are proteins of the intermediate lament group and components of intracytoplasmic cytoskeleton in the epithelial and parenchymal cells 17 . Ischemia-reperfusion injury as the consequence of CA and CPR leads to increased systemic apoptotic and necrotic cell death 18 . The activity of caspases during apoptosis leads to the fragmentation of the cytokeratin-18 (CK-18), which is considered to be an apoptosis-speci c cell death biomarker (caspase-cleaved cytokeratin-18 -CCCK-18), while necrotic cell death results in the release of the full-length CK-18 to the circulation. The fragments of the CK-18 cleaved by caspases can be recognised by a monoclonal antibody and in combination with the full CK-18 measurement, the predominant mode of cell death can be determined 19 .
Previous studies found an association between the increased levels of CK-18 and its caspase cleaved fragments and outcome in different disorders. The increased level of CCCK-18 in septic and critically ill patients was associated with mortality in previous studies 20 . CCCK-18 concentrations are elevated in patients with acute myocardial infarction compared to stable or unstable angina patients and the marker is signi cantly increased at the site of coronary occlusion as compared to peripheral blood samples 21 .
More studies investigated the marker in neurological disorders, such as ischemic stroke 22 , intracerebral 23 , and aneurysmal subarachnoid hemorrhage 24 , and traumatic brain injury 25 .
These considerations suggest that CK-18 and CCCK-18 could provide additional information to the prognosis of CA survivors as well since this condition often leads to critical illness and neurological disorder. We hypothesised that the level of systemic cell death should be in association with the mortality and outcome after CA. Therefore, the following study was performed to characterise for the rst time the release pattern of CK-18 and CCCK-18 in a prospective unselected cohort of resuscitated patients to evaluate their prognostic value for mortality, outcome, and their association with previously studied NSE, conventionally used clinical and laboratory parameters, prognostic scoring systems.

Baseline characteristics
Characteristics, etiology of the CA and CPR, vital parameters and comorbidities of the study cohort are summarized in Table 1. 54 patients were enrolled (median age: 67 [61-78] years, 48% male), 72% suffered IHCA, 69% had non-shockable initial rhythm. We did not nd statistically signi cant differences between non-survivors (n = 38; 70%) and survivors on age, gender, and the duration of CPR. Among survivors, the CPC category was signi cantly better, as expected (3.0 [2.3-3.0] vs. 4.5 [3.0-5.0], p < 0.001) and signi cantly favourable SAPS II and SOFA scores were recorded as shown in Table 1. Investigating biomarkers according to 30-day mortality and neurological outcome

Routine laboratory parameters
We recorded and analysed the main conventionally used laboratory parameters regarding 30-day mortality summarised in Table 2. Non-survivors had elevated initial liver function parameters and troponin. Hemoglobin and hematocrit values were lower among survivors at 24 and 72 post-CA hours as shown in Table 2 . 3. a-b.).  (Fig. 4. b.) The ROC curve of initial lactate per se indicated a cut-off as 4.90 mmol/L to predict 30-day mortality (sensitivity: 74%; speci city: 80%) and as 6.00 mmol/L (speci city: 80%, sensitivity: 84%) for poor neurological outcome (CPC4-5).

Discussion
To date, this is the rst study investigating the prognostic value for mortality and neurological outcome of CK-18 and its caspase-cleaved form in an unselected resuscitated population. The levels of the CK-18 referring to total cell death including apoptosis and necrosis were elevated, and the CCCK-18/CK-18 ratio was decreased among resuscitated patients compared to other study populations, that refers to a large extent of cell death dominantly due to necrosis 26 . Contrary to expectations, the investigated biomarkers or their change did not differ based on 30-day mortality. None of these markers showed connection with neurological function either. Despite our initial hypothesis, (i.e. survival depends on the rate of cell death after CA) mortality could rather be determined by damage of a smaller group of cells responsible for critical function and survival, but this signal may vanish in the mass of the total cell death. Survival may also rather depend on the remaining functional capacity and the ability to recover than the extent of damage that the above biomarkers may represent.
More surprisingly, no correlation was found between the biomarkers and the duration of CPR either, though a longer state of arrest is expected to cause a larger extent of cell damage. On the other hand, information about length and severity of prearrest hypoxia, and time between CA and start of effective resuscitation was lacking, though this period may not only be as important as the duration of resuscitation but also interferes with the results. Similarly, no correlation was found between the duration of resuscitation and survival.
We observed similarly negative results when evaluating the relationship between other organ system failure and cell death markers, except renal function. While CK-18 levels show a slowly declining trend over time in patients with normal renal function, it remains high in patients with renal insu ciency. The signi cantly higher 72-hour CK-18 levels in patients with renal failure after CA mostly could be due to the impaired renal elimination and less probably due to increased release from injured renal epithelial cells 27,28 . In a previous study increased serum concentration of total CK-18 was found in patients with chronic kidney disease stages 3-5, while there was no signi cant elevation of CCCK-18 levels similar to our ndings 29 .
We could not prove any connection of the cell death biomarkers with neurological outcome. Practically Consequently, we tested a marker that represents neuronal injury more speci cally. In previous studies, NSE was described as a prognostic marker for poor neurological outcome after cardiac arrest and it has an additional role in neuroprognostication in the current guidelines 11 . In contrast to our expectations, there was no signi cant difference in our study population concerning this marker and we could not con rm the prognostic value for 30-day mortality or neurological outcome. Explanation could be the high heterogeneity of our unselected population, which was mostly composed of various IHCA cases, and even the OHCA group had diverse etiology. Only a few papers published data about the prognostic value of NSE in patients after IHCA 32 . The prognostic value of NSE was higher for OHCA than for IHCA patients in the study of Kaspar et al. 33 . They explained the difference with higher number of confounders and mortality without hypoxic-ischemic encephalopathy in IHCA patients.
Failure of tissue perfusion during cardiac arrest leads to anaerobic metabolism. Lactate is the end product of anaerobic metabolism that can be used as a marker of cellular hypoxia and to predict mortality in critical illness 34 . Previous investigations have proved the utility of lactate as a marker for disturbances of tissue perfusion to predict survival in cardiac arrest patients 35,36 . According to our results, the lactate values per se had moderate prognostic value for 30-day mortality compared to the ones of SOFA and SAPS II. The reliability of these widely used prognostic scoring systems is equivocal because of their moderate discrimination ability and they have been consequently considered as not clinically relevant in post-resuscitation care 37 . In a study made up of OHCA patients treated with therapeutic hypothermia, Acute Physiology And Chronic Health Evaluation II, SAPS II, and OHCA scores had moderate accuracy to distinguish neurological outcomes and mortality, moreover, the SOFA score was found to be a poor predictor in this population 38 . Bisbal et al. reported the SAPS III as less accurate in determining the in-hospital mortality of post-CA patients compared to other general prognostic scores 39 . Signi cant efforts have been made in recent years to develop scoring systems that can be used to estimate the outcome of resuscitated patients 40 , and these often require background information about the patient and about often missing peri-arrest circumstances. Though promising, these scores are not as widely used as the classic general prognostic scores.
It was reported by Grimaldi et al. that lactate levels < 5.1 mmol/L were associated with favourable outcomes at discharge from the intensive care unit after CA 41 . Similar to their results in our population the cut-off was 4.90 mmol/L to predict 30-day mortality and 6.00 mmol/L for poor neurological prognosis. A most recent publication emphasised the need for development of prognostication tool in IHCA 36 . They retrospectively investigated the predictive role of the initial lactate levels among patients who suffered IHCA and required mechanical ventilation. According to their results elevated lactate levels were associated with mortality and the AUC indicated moderate ability to predict mortality similar to our results. This study mainly focused on mortality, while we evaluated the prognostic role of lactate levels concerning the neurological outcome. According to our results elevated lactate levels measured within 6 hours after CA could help the prediction of coma, vegetative state, or brain death.
Although in our recent study we could not prove the prognostic value of novel cell death markers (CCCK-18, CK-18), and previously well-investigated NSE, we con rmed the role of initial lactate levels for prediction of mortality and neurological outcome in unselected resuscitated patients. At a quick glance, the simply obtainable early lactate level may give similarly useful information per se than the SAPS II and SOFA scores, which can be calculated later after admission and requires more time.
The strength of our study was the prospective design with an unselected CA patient cohort as we aimed to nd a reliable independent biomarker predictor that can be used in general resuscitated population. We performed repeated sampling three times within the rst 72 hours after CA instead of only admission sampling to evaluate the change of the investigated biomarkers. The biomarker results (including NSE) were not available to clinicians during patient care and thus could not in uence treatment decisions. A major limitation of the current study is the relatively small sample size. The follow-up period ended on the 30th day after CA, and we have no data about long-term survival or neurological status. Prognostication of neurological outcome after cardiac arrest is based on clinical neurological examination, electrophysiological investigations, neuroimaging, and biomarkers. 9 However, in our study, we only focused on biomarkers in the peripheral blood and we did not evaluate the mentioned examinations or tools for prognostication. Most of our patients obtained the best CPC of 3, 4, or 5, which are mentioned as poor neurological outcomes in most recent studies contrary to our categorisation (CPC 4-5 for poor and 1-3 for good outcome). Given that the follow-up period was relatively short, we assumed that the fate of patients in CPC 3 category is not hopeless, thus we chose the other classi cation method.

Conclusion
In this study we could not prove the prognostic value of the CK-18 and its caspase cleaved form as biomarkers of systemic epithelial cell death, nor the one of NSE, while initial lactate levels were found to be a simple yet moderate predictor for mortality or neurological outcome in our unselected general resuscitated population.

Study population
Adult patients (age >18 years) who suffered in-hospital or out-of-hospital CA were prospectively enrolled after successful cardiopulmonary resuscitation to this observational pilot study conducted at the Before enrolment in the study, written informed consent was obtained from legal representatives or from the alert, oriented patients themselves.
Sample and data collection Data collected included patients' anamnestic information, comorbid conditions, variables that are necessary for calculating prognostic scores (Sequential Organ Failure Assessment score -SOFA; Simpli ed Acute Physiology Score -SAPS II), the circumstances, and presumed cause of the CA, and CPR. We recorded the laboratory parameters (electrolytes; markers of renal and hepatic function; in ammatory parameters; complete blood count, lactate as a marker of anaerobic shift and severe tissue hypoperfusion), blood gas, and vital parameters (heart rate, systolic and diastolic blood pressure, mean arterial pressure, ventilation mode) at the time of blood sampling. We examined the presence of vital organ failure during the post-resuscitation period. Organ system failure was de ned as follows: -circulation: the patient still required vasopressor or inotropic support at 24 hours -respiration: the patient still required positive pressure ventilation support at 24 hours -liver: transaminase elevations exceeding three times the normal value -kidney: decreased urine production (<500 ml/day) or creatinine clearance below 30 ml/min.
The SAPS II and SOFA scores were calculated by addition of the worst parameters of the rst 24 hours after CA. To avoid false pessimistic estimation of neurological function and to get information about patients who died in ICU or before 30 days, the best neurological status was recorded during ICU stay using the cerebral performance category (CPC) score 10 corresponding beta values and 95% con dence intervals. The diagnostic power of the prognostic scores and lactate was assessed using the area under the curve (AUC) of the receiver operator characteristic

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplementmaterial.docx