Monocyte subsets predict mortality after cardiac arrest

Abstract After successful cardiopulmonary resuscitation (CPR), many patients show signs of an overactive immune activation. Monocytes are a heterogeneous cell population that can be distinguished into 3 subsets by flow cytometry (classical monocytes [CM: CD14++CD16‐], intermediate monocytes [IM: CD14++CD16+CCR2+] and non‐classical monocytes [NCM: CD14+CD16++CCR2‐]). Fifty‐three patients admitted to the medical intensive care unit (ICU) after cardiac arrest were included. Blood was taken on admission and after 72 h. The primary endpoint of this study was survival at 6 months and the secondary endpoint was neurological outcome as determined by cerebral performance category (CPC)‐score at 6 months. Median age was 64.5 (49.8‐74.3) years and 75.5% were male. Six‐month mortality was 50.9% and survival with good neurological outcome was 37.7%. Monocyte subset distribution upon admission to the ICU did not differ according to survival. Seventy‐two hours after admission, patients who died within 6 months showed a higher percentage of the pro‐inflammatory subset of IM (8.3% [3.8‐14.6]% vs. 4.1% [1.5–8.2]%; P = 0.025), and a lower percentage of CM (87.5% [79.9–89.0]% vs. 90.8% [85.9–92.7]%; P = 0.036) as compared to survivors. In addition, IM were predictive of outcome independent of time to ROSC and witnessed cardiac arrest, and correlated with CPC‐score at 6 months (R = 0.32; P = 0.043). These findings suggest a possible role of the innate immune system in the pathophysiology of post cardiac arrest syndrome.


INTRODUCTION
Out of hospital cardiac arrest (OHCA) is one of the leading causes of death in the Western world affecting ∼86-98 per 100,000 inhabitants per year with only 6-10% of patients surviving to hospital discharge. 1,2 The underlying pathophysiological processes occurring after CA have been termed post cardiac arrest syndrome (PCAS). 3 The main PCAS characteristics include brain injury, myocardial dysfunction, a Abbreviations: AMI, acute myocardial infarction; APACHE II, Acute physiology and chronic health evaulation; APC-H7, allophycocyanin; BMI, body mass index; CA, cardiac arrest; CCR-2, C-C chemokine receptor-2; CD, cluster of differentiation; CI, confidence interval; CM, classical monocytes; CPC, cerebral performance category; CPR, cardiopulmonary resuscitation; CRP, C-reactive protein; ERC, European Resucitation Council; FACS, fluorescence activated cell sorting; HR, hazard ratio; ICU, intensive care unit; IM, intermediate monocytes; IQR, interquartile range; NCM, non-classical monocytes; OHCA, out of hospital cardiac arrest; PCAS, post cardiac arrest sysndrome; PCT, procalcitonin; PE, phycoerythrin; PerCP, peridinin chlorphyll protein; ROS, reactive oxygen species; ROSC, return of spontaneous circulation; RPM, revolutions per minute; TTM, therapeutic temperature management. reactive oxygen species (ROS). 6 Oxidative stress is a strong innate immune system inducer. This systemic inflammatory response may contribute to patient deterioration and ultimately irreversible multiorgan damage and death. Still, a major cause of the high mortality in patients after CA is the irreversible brain injury. 3 Monocytes are circulating leukocytes that migrate into tissues and differentiate into M s. As members of the innate immune system, they have an important physiologic role in the first line of defense against pathogens. However, they have also been causally involved in the pathophysiology of chronic and acute inflammatory conditions. As already described more than 3 decades ago, monocytes are a heterogeneous cell population that can be distinguished into at least 3 subsets according to their CD14 and CD16 surface expression pattern. 7,8 Functions of classical monocytes (CM) include functions such as phagocytosis, while intermediate monocytes (IM) exhibit distinct features including a strong inflammatory reaction upon stimulation. 9 The subset of non-classical monocytes (NCM) on the other hand was described as exhibiting a specific "crawling" behavior at the endothelium during physiologic conditions. 10 Several larger studies have described monocyte subsets as prognostic markers in stable diseases such as chronic kidney disease 11 and coronary artery disease. 12 However, there is only limited data regarding the behavior of monocyte subsets in critically ill patients, while data on monocyte subsets after CA is completely lacking. [13][14][15] Therefore, the aim of the current study was to evaluate the dynamics of circulating monocyte subsets in the pathophysiology and prognosis in survivors of CA admitted to a medical intensive care unit (ICU). Recorded pre-and in-hospital parameters and patient characteristics including Utstein variables are listed in Table 1. The

Subjects and study design
Glasgow-Pittsburgh cerebral performance categories (CPC) were used for neurologic status categorization at 6 months after CA.
CPC 1-2 was defined as favorable neurological outcome while CPC 3-5 was considered poor neurological outcome. 19 The main study endpoint was all-cause mortality at 6 months.

Blood sampling
Blood for flow cytometry was drawn within 24 h after admission and 72 h after the first blood sampling from either the arterial or central venous line. After the initial 3 mL of blood had been discarded, blood was drawn into an EDTA-tube for immediate flow cytometry. Standard laboratory parameters were measured at the central laboratory of the General Hospital of Vienna.

Statistical analysis
The primary endpoint of this study was survival at 6 months and the secondary endpoint was neurological outcome as determined by cerebral performance category (CPC)-score at 6 months. Sample size calculation analysis revealed that in a cohort with a mortality rate of 50%,

Baseline characteristics
Baseline characteristics are presented in
Monocyte subset distributions at days 0 and 3 did not correlate with age, gender, or body mass index (BMI). In addition, monocyte sub-

Monocyte subset distribution and 6-month mortality
All-cause mortality at 6 months post CA was 50.9% (27 out of 53 patients). Monocyte subset distribution at admission was not different in non-survivors and survivors after 6 months ( Fig. 2A-C)

DISCUSSION
In the present prospective, observational study including 53 consecutive patients admitted to a medical ICU after ROSC, monocyte subset distribution was associated with 6-month survival. Interestingly, monocyte subsets at ICU admission did not differ according to outcome. However, non-survivors showed a significant decrease of CM and a significant increase of IM between admission and day 3. Subsequently, patients who did not survive the first 6 months were charac- were stratified into tertiles of monocyte subsets, patients in the highest tertile of IM were at a 6.4-fold higher risk of dying irrespective of time to ROSC and whether they had a witnessed CA.
We decided to include a mixed cohort of patients both after surviving an out-of-hospital CA as well as after in-hospital CA. All patients experienced sustained ROSC before admission. Most patients were male, had a witnessed CA, and received bystander CPR. Patients suffering from OHCA and a shockable rhythm were more likely to survive, while those with higher lactate levels at admission and longer time to ROSC were more likely to die within 6 months, as expected from previous experience. 20 The PCAS is a syndrome caused by whole body ischemia and reperfusion and involves ongoing myocardial dysfunction, a persisting cause of the initial arrest as well as severe brain injury and severe reperfusion syndrome. 3 As myocardial failure and brain injury are thought to represent the main causes of death, therapeutic interventions and prognostication efforts mainly focus on those areas. However, inflammatory mechanisms have been implicated in myocardial reperfusion injury and thus might play an important role in whole body reperfusion injury. 21,22 As innate immune effector cells, monocytes may play an important role in the pathophysiology of PCAS. 23 According to their surface expression pattern of CD14 and CD16, monocytes can be distin-guished into at least three subtypes, namely CM (CD14 ++ CD16 + ), IM (CD14 ++ CD16 + CCR2 + ), and NCM (CD14 + CD16 ++ CCR2). 8 Initially, 2 subsets have been described, CD14 + CM and CD16 + NCM. The CD16 + subset accounts for ∼10% of circulating monocytes in healthy individuals and has been initially described as pro-inflammatory. However, soon it became evident that there is considerable heterogeneity among CD16 + monocytes, and a third subset, the IM (CD14 ++ CD16 + ) has been described. 8,9 Within the following years, it was shown that this intermediate subset, despite its name, is more than just a subset "in between." 9 This cell type is characterized by a strong inflammatory reaction upon stimulation and by an increased expression of surface molecules involved in reparative mechanisms. 24,25 Several clinical observations have described elevated IM as biomarkers for outcome and disease severity in critical conditions. 13,15,26 In addition, largescale studies evaluated monocyte subsets as predictors of adverse cardiac outcomes in stable cohorts suggesting such a role for both the classical and intermediate subset. 11,12,27 In our analysis, the proportion of CM at admission was inversely correlated with CRP levels, an unspecific marker of inflammatory activity as well as with PCT, a novel biomarker with potential to differentiate between inflammation and infection. In addition, IM showed a significant correlation with PCT levels at admission. It has to be noted that none of the included patients had clinical signs of sepsis. Previous studies however suggest a high proportion of bacteremia after CA, when specifically testing for it. 28 We observed a strong correlation between the widely used disease severity score APACHE II at admission and the proportion of IM on day three after CA. This is in line with a small post cardiac surgery cohort. 29 Therefore, one can speculate that severe organ dysfunction at admission leads to further activation of the innate immune system at day 3. When the APACHE II score was added to the multivariate regression analysis, the association between circulating IM and survival lost significance. Whether circulating monocytes therefore only represent disease severity upon admission or whether they further add specific detrimental effects cannot be answered from our data.
We were able to show that upon admission, monocyte subsets were not associated with 6-month mortality. However, 72 h after admission, we could observe a remarkable shift toward an increase in IM accompanied by a decrease in CM. This fact is of major importance and scientific interest. One explanation for this observed "delay" in monocyte subset shift might be an impaired immune response in the first hours after CA as analyzed elsewhere. 30 Another hypothesis may be that  Several observations confirm our findings of a strong inflammatory activation in patients after CA, with 1 study analyzing monocyte behavior. [21][22][23] The authors of the latter finding suggested changes in monocyte pattern recognition receptor signaling pathways and inflammasome activation may play a role in PCAS. 23 Monocyte subsets were implicated in the pathophysiology of acute myocardial infarction and correlated with outcome and left ventricular function in those patients. [33][34][35] The detailed mechanisms are not understood; therefore, one could speculate that the observed monocyte subset changes in our study might be due to myocardial injury.
Conversely, the findings in AMI might be caused by reperfusion injury.
Thus, our results might also be explained by a whole body reperfusion injury after ROSC in CA.
Another important aspect of PCAS and a common cause of death is severe brain damage. In our study, we could show a significant correlation between the proportion of IM and the CPC-score at 6 months.
This finding is of major importance, as the main goal for the patient and the health care system is discharge of patients with acceptable neurologic outcome. In this regard, monocyte subsets were described as predictors of defect size and outcome in other situations characterized by an acute and sudden brain injury, such as stroke. 36,37 In those studies, monocyte subsets were not only implicated as surrogate parameters of stroke size but more as reparative cells probably acting as a surrogate for a reparative process underway, which has been supported by experimental findings. 38 Of interest, monocyte subsets have further been implicated in the deterioration of chronic degenerating neurologic diseases such as multiple sclerosis. 39 Whether an increase of IM may reflect the size of neurologic damage or rather the reparative efforts underway cannot be answered by this analysis and warrants further mechanistic studies in experimental settings.
Further studies are strongly warranted to broaden our understanding of the pathophysiologic mechanisms involved in our findings and to design potential therapeutic interventions. Whether monocyte sub-set distribution constitutes a biomarker of an ongoing and escalating inflammatory activation or monocyte subsets have potential reparative effects or damaging effects is currently unknown. The recently published observation that IM exhibit a distinct micro-RNA profile might be a first step in designing such an intervention. 25 Several limitations of the current study need to be mentioned.
First, this was a single-center study including both patients after out-of-hospital and in-hospital CA. Due to the rather small sample size, individual analysis of patient groups was not feasible. Still, for the treating ICU physician, it is of major interest whether a marker can predict outcome independent of case specifics. Further, larger studies are warranted to analyze whether monocyte subset distribution is involved in the pathophysiology of patients after both in-and out-of-hospital CA.
As the present analysis is of observational nature, we can only describe an association between monocyte subset distribution and mortality but cannot draw any causality from our findings. In addition, only circulating monocyte subsets and routinely measured inflammatory markers were assessed, while other soluble mediators reflecting systemic inflammatory activation such as IL-6 or IL-1 were not measured in our population. Furthermore, blood was drawn within the first 24 h after ICU admission, which is why we cannot exclude time-dependent variations within the first 24 h, which may render early monocyte subset measurements feasible for prognostication. Still, our findings suggest a later increase in IM being predictive of outcome.
In conclusion, we provide evidence for the predictive value of monocyte subset distribution for 6-month survival in patients admitted to a medical ICU after surviving a CA independent from time to ROSC and witnessed cardiac arrest. Our results contribute further aspects on the role of the innate immune system in the pathophysiology of PCAS and may pose a therapeutic target.