a b s t r a c t

Purpose: We aimed to examine the serial changes in coagulofibrinolytic markers that occurred after the restora- tion of spontaneous circulation (ROSC) in cardiac arrest patients, who were treated with Targeted temperature management (TTM). We also evaluated the association between the disseminated intravascular coagulation score and clinical outcomes.

Methods: This was a single-centre, retrospective observational study that included cardiac arrest patients who were treated with TTM from May 2012 to December 2015. The prothrombin time (PT) and partial thromboplas- tin time (PTT), along with the levels of fibrinogen, Fibrin degradation products (FDP), and D-dimer were obtained after ROSC and on day 1, 2, and 3. The DIC score was calculated after ROSC. The primary outcome was the neuro- logic outcome at discharge and the secondary outcome was the 6-month mortality.

Results: This study included 317 patients. Of these, 222 (70.0%) and 194 (61.2%) patients had a Poor neurologic outcome at discharge and 6-month mortality, respectively. The PT, PTT, and Fibrinogen level significantly in- creased over time, while the FDP and D-dimer levels decreased during first three days after ROSC. Multivariate logistic analyses revealed that the DIC score remained a significant predictor for poor neurologic outcome (odds ratio [OR], 1.800; 95% confidence interval [CI], 1.323-2.451) and 6-month mortality (OR, 1.773; 95% CI,

1.307-2.405).

Conclusion: The activity of coagulation and fibrinolysis decreased over time. An increased DIC score was an inde- pendent prognostic factor for poor neurologic outcome and 6-month mortality.

(C) 2017

Introduction

Post-cardiac arrest syndrome mainly results from Ischemic injury during cardiac arrest and reperfusion injury after the restoration of spontaneous circulation (ROSC). ischemic reperfusion injury induces endothelial activation, which is a critical step in endothelial injury and organ damage [1]. In turn, endothelial activation triggers the expression of tissue factor, followed by thrombin generation, which results in coag- ulation. Moreover, an increase in plasminogen activator inhibitor (PAI) activity and a decrease in anticoagulation factors produce a coagulofibrinolytic imbalance after cardiac arrest [2-6]. Coagulofibrinolytic imbalance likely contributes to the development of post-cardiac arrest syndrome. Therefore, coagulofibrinolytic markers can be used as Surrogate markers of outcome in cardiac arrest patients. Recent studies reported that the levels of fibrin degradation products (FDP) and D-dimer have the potential to serve as predictors of

* Corresponding author.

E-mail address: [email protected] (B.K. Lee).

neurologic outcome or mortality in out-of-hospital cardiac arrest (OHCA) patients [7,8]. The activities of enzymes and other factors that participate in coagulation and fibrinolysis change over time after cardiac arrest [9]. However, previous studies analyzed only the initial laboratory data and focused mainly on the association between coagulofibrinolytic factors and the Short-term outcome.

disseminated intravascular coagulation , which is character- ized by increased coagulation and inadequate endogenous fibrinolysis, results in intravascular fibrin formation and microvascular thrombosis of multiple organs and eventually enhances multi-organ dysfunction syndrome in cardiac arrest [3,4,9,10]. DIC frequently occurs after cardiac arrest, with the prevalence of overt DIC development after cardiac arrest reportedly ranging from 42 to 54% [9,11,12]. In previous studies, both coagulofibrinolytic markers and the DIC score were associated with the outcomes of cardiac arrest survivors [7,9,11,12]. Higher DIC scores were associated with worse neurologic outcomes or higher mortality [11,12]. Although targeted temperature management is the stan- dard therapy for cardiac arrest survivors, the abovementioned studies included cardiac arrest patients irrespective of whether they were

http://dx.doi.org/10.1016/j.ajem.2017.04.077

0735-6757/(C) 2017

Fig. 1. Schematic diagram showing the number of patients included in the present study. DNAR, do not attempt resuscitation; TTM, targeted temperature management; ECMO, extracorporeal membrane oxygenation.

treated with TTM [7,9,11,13,14]. The prognostic performance of coagulofibrinolytic factors and DIC score can be changed according to improved outcomes after implementation of TTM in cardiac arrest pa- tients. Therefore, the association of DIC score with outcome in cardiac arrest patients should be reevaluated.

The aims of the present study were to examine the serial changes in coagulofibrinolytic markers that occur after ROSC and to compare the prognostic performance of these factors in cardiac arrest patients who were treated with TTM. We also evaluated the association between the DIC score and neurologic outcome or long-term mortality.

Methods

Study design and population

This was a retrospective observational study of adult comatose car- diac arrest survivors, treated with TTM at Chonnam National University Hospital from May 2012 to December 2015. The Institutional Review Board of Chonnam National University Hospital approved this study (CNUH-2016-104).

Cardiac arrest patients over 18 years of age who underwent TTM were included. Patients were excluded if (1) TTM was interrupted owing to a transfer to another facility, (2) they were taking warfarin,

(3) a target temperature other than 33 ?C was used (e.g., 32 or 36 ?C), and/or (4) extracorporeal membrane oxygenation was applied during post-cardiac arrest care.

Table 1

Baseline characteristics stratified according to neurologic outcome at discharge and 6-month mortality.

	Total	Good	Poor	p	Survivors	Non-survivors	p
	(N = 317)	(n = 95)	(n = 222)		(n = 123)	(n = 194)
Age, years	61.0 (51.0-71.0)	53.0 (43.0-61.0)	65.0 (53.8-73.0)	0.001	54.0 (43.0-62.0)	66.0 (55.0-74.0)	b0.001
Male sex	203 (64.0)	69 (72.6)	134 (60.4)	0.037	86 (69.9)	117 (60.3)	0.082
Comorbidities
Coronary artery disease	50 (15.8)	18 (18.9)	32 (14.4)	0.310	21 (17.1)	29 (14.9)	0.613
Heart failure	23 (7.3)	6 (6.3)	17 (7.7)	0.673	6 (4.9)	17 (8.8)	0.194
Hypertension	130 (41.0)	31 (32.6)	99 (44.6)	0.047	44 (35.8)	86 (44.3)	0.131
Diabetes	84 (26.5)	13 (13.7)	71 (32.0)	0.001	16 (13.0)	68 (35.1)	b0.001
Pulmonary disease	13 (4.1)	1 (1.1)	12 (5.4)	0.118	3 (2.4)	10 (5.2)	0.235
Renal disease	40 (12.6)	7 (7.4)	33 (14.9)	0.066	8 (6.5)	32 (16.5)	0.009
Cerebrovascular accident	18 (5.7)	3 (3.2)	15 (6.8)	0.205	5 (4.1)	13 (6.7)	0.323
Hepatic disease	5 (1.6)	0 (0.0)	5 (2.3)	0.327	1 (0.8)	4 (2.1)	0.652
First monitored rhythm				b 0.001			b0.001
VF/pulseless VT	103 (32.5)	62 (65.3)	41 (18.5)		68 (55.3)	35 (18.0)
PEA	78 (24.6)	20 (21.1)	58 (26.1)		28 (22.8)	50 (25.8)
Asystole	133 (42.0)	12 (12.6)	121 (54.5)		25 (20.3)	108 (55.7)
Unknown	3 (0.9)	1 (1.1)	2 (0.9)		2 (1.6)	1 (0.5)
Etiology				b 0.001			b0.001
Cardiac	174 (54.9)	77 (81.1)	97 (43.7)		89 (72.4)	85 (43.8)
Other medical	79 (24.9)	12 (12.6)	67 (30.2)		15 (12.2)	64 (33.0)
Asphyxia	42 (13.2)	3 (3.2)	39 (17.6)		11 (8.9)	31 (16.0)
Drug overdose	20 (6.3)	3 (3.2)	17 (7.7)		8 (6.5)	12 (6.2)
Drowning	2 (0.6)	0 (0.0)	2 (0.9)		0 (0.0)	2 (1.0)
Witnessed	232 (73.2)	80 (84.2)	152 (68.5)	0.004	103 (83.7)	129 (66.5)	0.001
Bystander CPR	195 (61.5)	62 (65.3)	133 (59.9)	0.370	83 (67.5)	112 (57.7)	0.082
Time to ROSC, min	28.0 (15.0-40.0)	21.0 (15.0-30.0)	31.0 (15.0-41.0)	0.001	22.0 (15.0-30.0)	31.0 (17.8-41.0)	b0.001
Glasgow Coma Scale	3 (3-3)	3 (3-6)	3 (3-3)	b 0.001	3 (3-5)	3 (3-3)	b0.001
Serum lactate, mmol/L	7.3 (4.2-10.5)	6.1 (3.8-8.9)	7.9 (4.9-11.6)	0.001	6.3 (3.8-8.8)	8.2 (5.1-11.7)	b0.001
Glucose after ROSC, mg/dL	226 (168-293)	212 (155-276)	235 (176-302)	0.028	216 (159-283)	234 (178-304)	0.124
PaO2, mm Hg	131 (78-207)	107 (77-189)	141 (80-218)	0.041	121 (77-190)	136 (79-215)	0.171
PaCO2, mm Hg	37.0 (30.0-46.0)	35.8 (31.0-42.1)	38.0 (30.0-47.6)	0.113	35.8 (30.8-42.3)	38.0 (30.0-48.0)	0.103
Initial temperature, ?C	35.9 (34.8-36.7)	36.4 (35.7-37.0)	35.7 (34.6-36.4)	b 0.001	36.3 (35.7-37.0)	35.6 (34.6-36.4)	b0.001
Pre-induction time, min	206 (147-290)	195 (130-252)	215 (153-301)	0.036	196 (137-283)	215 (155-301)	0.116
Induction duration, h	2.3 (1.3-3.3), 315a	3.0 (2.0-4.5)	2.0 (1.0-3.0), 220a	b 0.001	2.8 (2.0-4.3)	1.8 (1.0-3.0), 192a	b0.001
SOFA score	9 (7-12)	7.0 (5.0-10.0)	11.0 (7.0-14.0)	b 0.001	7 (5-10)	11 (8-14)	b0.001
DIC score	4 (3-4)	3 (2-4)	4 (3-5)	b 0.001	3 (2-4)	4 (3-5)	b0.001

VF, ventricular fibrillation; VT, ventricular tachycardia; PEA, pulseless electrical activity; CPR, cardiopulmonary resuscitation; ROSC, restoration of spontaneous circulation; SOFA, sequen- tial organ failure assessment; DIC, disseminated intravascular coagulation.

^a Number of patients included in the analysis.

Data collection

The following data were obtained from hospital records: age, sex, co- morbidities, first monitored rhythm, etiology of cardiac arrest, presence of a witness on collapse, bystander cardiopulmonary resuscitation (CPR), time to ROSC, Glasgow Coma Scale score after ROSC, serum lactate after ROSC, glucose after ROSC, PaO₂, PaCO₂, initial core temperature, pre-induction time, induction time, coagulofibrinolytic markers, DIC score after ROSC, Sequential Organ Failure Assessment score within the first 24 h after admission [15], neurologic out- come at hospital discharge, and 6-month mortality. The DIC score was calculated using the methods suggested by the International Society of Thromobosis and Haemostasis [16]. The prothrombin time (PT) and partial thromboplastin time (PTT), along with the levels of fibrinogen, FDP, and D-dimer were obtained at after ROSC and day 1, 2, and 3. The antithrombin III level was obtained on days 1, 2, and 3 after ROSC. Neu- rologic outcome was assessed using the Glasgow-Pittsburgh Cerebral Performance Categories (CPC) scale at discharge and recorded as CPC 1 (good performance), CPC 2 (moderate disability), CPC 3 (severe dis- ability), CPC 4 (vegetative state), or CPC 5 (brain death or death) [17]. The primary outcome was a poor neurological outcome (CPC 3-5) at hospital discharge. The secondary outcome was the 6-month mortality.

Statistical analysis

Continuous variables are reported as the median and interquartile range according to the results of a normality test. Comparisons of con- tinuous variables were conducted using Mann-Whitney U tests, as ap- propriate. Categorical variables are presented as frequencies and percentages. Comparisons of categorical variables were performed using ?² or Fisher’s exact tests, as indicated. Receiver operating charac- teristic (ROC) analyses were performed to examine the prognostic per- formance of the coagulofibrinolytic markers and DIC score. Comparisons of the dependent ROC curves were performed using the method described by DeLong et al. [18]. Independent predictors for poor neurologic outcome and 6-month mortality were determined using logistic regression analyses after adjusting for relevant covariates. The DIC score and subgroup of DIC score were modeled using different models to avoid confounding errors. The DIC score was divided into five groups (0-2, 3, 4, 5, >= 6) according to a previous report [11]. To com- pare the 6-month mortality according to the DIC scores, Kaplan-Meier analyses were applied and a log-rank test was used to compare the rate estimates. A Cox proportional hazards regression for survival time was also conducted. The results of the regression analyses are reported as the odds ratio (OR) or hazards ratio (HR) with 95% confidence inter- vals (CI). Variables found to be statistically significant at a level of p b

0.20 in the univariate analyses were included in the multivariate analy-

sis. Backward selection was used to obtain the final model. Linear mixed model analyses were conducted to assess the changes in the coagulofibrinolytic markers over time. A post hoc analysis was per- formed at each time point using the pair-wise Mann-Whitney U test with Bonferroni correction. Data were analyzed using the PASW/SPSS

(TM) software, version 18 (IBM Inc., Chicago, IL, USA), and MedCalc, ver- sion 16.1 (MedCalc Software bvba, Ostend, Belgium). The significance level (two-tailed) was set at 0.05.

Results

Patient selection and characteristics

In total, 399 adult cardiac arrest patients were treated with TTM dur- ing the study period. Of these, 82 patients were excluded, as shown in Fig. 1. Thus, 317 patients were included in the final analysis.

Of the 317 included patients, 222 (70.0%) and 194 (61.2%) had a

poor neurologic outcome at discharge and 6-month mortality, respec- tively. The patients’ clinical characteristics stratified according to

outcome are shown in Table 1. Patients with a Good neurologic outcome were younger and predominantly male. Moreover, patients with a good neurologic outcome had a lower incidence of comorbidities (hyperten- sion and diabetes) and were more likely to have a shockable rhythm, cardiac etiology, higher incidence of witnessed collapse, shorter time to ROSC, higher GCS score after ROSC, lower serum lactate, lower glu- cose, lower PaO₂, higher initial temperature, shorter pre-induction time, longer induction time, lower SOFA score, and lower DIC score compared to patients with a poor neurologic outcome.

Compared to non-survivors, survivors were younger and had a lower incidence of comorbidities (diabetes and renal disease); they were also more likely to have a shockable rhythm, cardiac etiology, and higher incidence of witnessed collapse (Table 1). Furthermore, sur- vivors had a shorter time to ROSC, higher GCS score, lower serum lac- tate, higher initial temperature, longer induction time, lower SOFA score, and lower DIC score compared to non-survivors.

Prognostic performance and changes in the coagulation profiles and DIC score over time

The fibrinogen level after ROSC was not a significant predictor of a poor neurologic outcome at discharge (area under the curve [AUC] value, 0.503; 95% CI, 0.446-0.560; p = 0.926). The other

coagulofibrinolytic markers obtained after ROSC and the DIC score showed significant performance to predict a poor neurologic outcome (Fig. 2). The AUC values of the coagulation profiles and DIC score ranged from 0.624 to 0.711.

Fig. 3 shows the changes in the coagulofibrinolytic markers over time. The PT, PTT, and fibrinogen level significantly increased over time, while the FDP, D-dimer, and antithrombin III levels significantly decreased over time (Fig. 3). The PT and PTT, as well as the FDP and D-dimer levels were significantly different between patients with good and poor neurologic outcomes at all time points (Fig. 3).

Fig. 2. Comparison of the area under curve (AUCs) for predicting poor neurologic outcome at discharge between the coagulofibrinolytic profiles and disseminated intravascular coagulopathy score. The prothrombin time, partial thromboplastin time, fibrin degradation product, D-dimer, antithrombin III, and DIC score have AUCs of 0.686 (95% confidence interval [CI], 0.629-0.740), 0.694 (95% CI, 0.636-0.747), 0.675 (95% CI,

0.618-0.730), 0.681 (95% CI, 0.624-0.735), 0.624 (95% CI, 0.565-0.681), and 0.711 (95%

CI, 0.654-0.763), respectively.

Associations between the DIC score and the neurologic outcome at dis- charge or 6-month mortality

The DIC score was included in the multivariate regression analysis, since the DIC score had the highest AUC value. Table 2 shows the results of the multivariate logistic regression analysis for poor neurologic out- come at discharge. A higher DIC score was independently associated with a poor neurologic outcome (OR, 1.800; 95% CI, 1.323-2.451) and all of the DIC score groups compared to the group with a DIC score of 0-2 showed significantly higher ORs of poor neurologic outcome (Table 2).

Table 3 shows the results of the multivariate logistic regression anal- ysis for 6-month mortality. A higher DIC score was independently asso- ciated with the 6-month mortality (OR, 1.773; 95% CI, 1.307-2.405) and all of the DIC score groups compared to the group with a DIC score of 0- 2 showed significantly higher ORs of 6-month mortality (Table 3). A Kaplan-Meier survival analysis revealed that more patients with high DIC scores died during the 6-month period after cardiac arrest (Fig. 4). Table 4 shows the results of the Cox regression analysis of 6-month mortality. A higher DIC score was associated with a higher HR of 6- month mortality (HR, 1.261; 95% CI, 1.089-1.461) and there was a sig- nificant gradient of increasing HRs across the five consecutive DIC score groups (p for trend = 0.003). All of the DIC score groups had higher HRs of 6-month mortality than did the group with a DIC score

of 0-2 (Table 4 and Fig. 4).

Discussion

In the present study, we found that the DIC score and all of the coagulofibrinolytic markers except fibrinogen were able to predict the neurologic outcome at discharge. The PT, PTT, and fibrinogen level in- creased over the 3 days after cardiac arrest, while the D-dimer, FDP, and antithrombin III levels decreased over time. Moreover, all of the markers except fibrinogen were significantly different between patients with good and poor neurologic outcomes. DIC score was independently associated with the neurologic outcome and 6-month mortality, and the patients with higher DIC scores had a higher risk for 6-month mortality. Endothelial activation induced by ischemic reperfusion injury in pa- tients with cardiac arrest has been shown to trigger coagulation by ele- vating the amount of von Willebrand factor that is released from the endothelium [1]. The PT and PTT were shortest at immediately after ROSC and gradually increased over time, while the antithrombin III level was the highest on day 1 after ROSC and decreased over time. It is likely that the coagulation activity was the highest immediately after cardiac arrest and decreased over time owing to the consumption

of coagulation factors and antithrombin III.

Corresponding to coagulation activation, fibrinolytic activity is typi- cally enhanced. Consistent with this, we found that the levels of FDP and D-dimer, which are fibrinolytic factors, were the highest immediately after ROSC and then decreased with time, while the fibrinogen level in- creased over time. In addition, hyperfibrinolysis was prominent early

Fig. 3. Changes in the coagulation profile over the first 3 days after the restoration of spontaneous circulation. The prothrombin time (PT) and partial thromboplastin time (PTT) significantly increased over time (p b 0.001) and the poor neurologic outcome group had significantly longer PTs and PTTs at all time points compared to the good neurologic outcome group (A and B). The fibrinogen level significantly increased over time (p b 0.001), but was only different between patients with good and poor neurologic outcomes at day 3 after ROSC (C). The fibrin degradation product (FDP) and D-dimer levels significantly decreased over time, and the poor neurologic outcome group had significantly higher FDP or D-dimer levels at all time points compared to the good neurologic outcome group (D and E). The level of antithrombin III significantly decreased over time, and the poor neurologic outcome group had a lower antithrombin III level at day 1 and day 2 after cardiac arrest compared to the good neurologic outcome group (F).

Table 2 Multivariate analysis of poor neurologic outcome.
	OR (95% CI)	p	OR (95% CI)	p
Age, year	1.045 (1.018-1.073)	0.001	1.048 (1.020-1.077)	0.001
Shockable rhythm	0.161 (0.067-0.388)	b 0.001	0.164 (0.066-0.403)	b0.001
Cardiac etiology	0.386 (0.158-0.945)	0.037	0.357 (0.144-0.885)	0.026
Time to ROSC, min	1.038 (1.013-1.064)	0.003	1.039 (1.013-1.066)	0.003
GCS after ROSC	0.734 (0.563-0.956)	0.022	0.755 (0.578-0.985)	0.038
Glucose after ROSC	1.005 (1.001-1.009)	0.022	1.006 (1.001-1.010)	0.011
PaO2 after ROSC, min	1.010 (1.005-1.015)	b 0.001	1.010 (1.005-1.015)	b0.001
Initial body temperature, ?C	0.694 (0.488-0.985)	0.041	0.694 (0.487-0.989)	0.043
Pre-induction time, min	1.004 (1.000-1.007)	0.035	1.004 (1.000-1.007)	0.002
DIC score	1.800 (1.323-2.451)	b 0.001
DIC score 0-2			Reference
DIC score 3			3.040 (1.119-8.259)	0.029
DIC score 4			2.599 (1.033-6.542)	0.043
DIC score 5			16.841 (3.183-89.098)	0.001
DIC score >= 6			22.778 (2.058-252.138)	0.011

OR, odds ratio; CI, confidence interval; ROSC, restoration of spontaneous circulation; GCS, Glasgow Coma Scale; DIC, disseminated intravascular coagulation.

after ROSC and resolved gradually. This finding was consistent with a previous study showing that DIC with the fibrinolytic phenotype is dominant during the early phase after ROSC and associated with out- come in OHCA patients [12,19]. Fibrinolysis coincides with coagulation after endothelial activation, since Endothelial cells have a considerable amount of tissue plasminogen activator [20]. A study of 63 OHCA pa- tients by Gando et al. reported that tissue plasminogen activity was markedly elevated upon hospital arrival but undetectable at 24 h after admission, while PAI-1 activity was markedly increased at 24 h after ad- mission [4]. Wada et al. also showed that the activity of tissue PAI com- plex decreased over time in cardiac arrest patients [9]. In the present study, DIC with the fibrinolytic phenotype appeared to change into DIC with the balanced fibrinolytic phenotype over time. Suppression of PAI seems to play a major role in the serial changes that occur in fibri- nolysis after cardiac arrest. Here, we found that FDP and D-dimer de- creased markedly between the first and second day after cardiac arrest in the present study and that PAI decreased at 24 h after admis- sion in the previous studies [4,9]. Hence, the activity of PAI should de- crease at around 24 h after ROSC in cardiac arrest patients.

Previous studies have shown that various coagulofibrinolytic markers are associated with neurologic outcome or mortality in OHCA patients [7,8,11,12]. The Coagulation markers acquired during CPR have even been linked to the achievement of ROSC [21]. As such, coagulofibrinolytic markers can be used as surrogate markers of out- come in cardiac arrest patients. Prolonged PTs and PTTs, along with lower antithrombin III levels, were associated with increased poor neu- rologic outcome in the present study, which is in line with the findings of a previous study [7]. Additionally, a study of 182 OHCA patients by Szymanski et al. demonstrated that an elevated D-dimer level was inde- pendently associated with increased 30-day mortality [8]. Ono et al.

analyzed the prognostic performance of various coagulofibrinolytic markers in 315 OHCA patients and showed that the AUCs of all of the ex- amined factors except fibrinogen for poor neurologic outcome ranged from 0.70 to 0.79 [7]. They also reported that FDP had the largest AUC value [7]. The prognostic performance of coagulofibrinolytic markers in the present study showed somewhat lower AUCs (0.62 to 0.69) than those in the study by Ono et al., though we similarly found that fi- brinogen was not associated with neurologic outcome [7]. We assumed that the difference in prognostic performance of the coagulofibrinolytic makers between our study and the study by Ono et al. was related to the difference in the proportion of patients with a poor neurologic outcome (70% vs. 84%). In the previous study, only about 30% of patients were treated with therapeutic hypothermia after OHCA, thereby causing a higher proportion of poor neurologic outcomes [7]. It is known that hyperfibrinolysis, represented by increases in the D-dimer and FDP levels, is associated with hypoperfusion [22]. Furthermore, OHCA pa- tients with hyperfibrinolysis reportedly have a lower pH, base excess, and cerebral Tissue oxygenation and higher Serum lactate level [22, 23]. Increases in the amount of FDP indicate the severity of the hyperco- agulable state, organ hypoperfusion, and thrombus fibrinolysis. There- fore, the increased FDP and D-dimer levels we observed in cardiac arrest patients likely reflect the multi-organ ischemia and explain the reasonable prognostic performance of these factors for poor neurologic outcome or mortality.

In addition to coagulofibrinolytic markers, the DIC score has also

been suggested as a surrogate marker of outcome in cardiac arrest pa- tients [7,11,12]. The DIC score can express both abnormalities of coagu- lation and fibrinolysis. In a study by Wada et al. that included 388 OHCA patients, an increased DIC score was associated with increased all-cause mortality [12]. Kim et al. demonstrated that the DIC score was an

Table 3

Multivariate analysis of 6-month mortality.

OR (95% CI) p OR (95% CI) p

Age, year 1.050 (1.026-1.074) b0.001 1.053 (1.027-1.080) b0.001

Diabetes 2.734 (1.266-5.903) 0.010 3.234 (1.445-7.239) 0.004

Shockable rhythm 0.195 (0.100-0.384) b0.001 0.172 (0.085-0.346) b0.001

Witness 0.331 (0.149-0.731) 0.006 0.330 (0.147-0.740) 0.007

Time to ROSC, min 1.031 (1.010-1.053) 0.003 1.031 (1.009-1.052) 0.005

PaO₂ after ROSC, mm Hg 1.008 (1.004-1.012) b0.001 1.008 (1.003-1.012) b0.001

SOFA score 1.120 (1.010-1.243) 0.032 1.152 (1.033-1.285) 0.011

DIC score 1.773 (1.307-2.405) b0.001

DIC score 0-2 Reference

DIC score 3 3.626 (1.400-9.389) 0.008

DIC score 4 4.991 (2.023-12.309) b0.001

DIC score 5 6.374 (1.874-21.674) 0.003

DIC score >= 6 6.329 (1.033-38.788) 0.046

OR, odds ratio; CI, confidence interval; ROSC, restoration of spontaneous circulation; SOFA, sequential organ failure assessment; DIC, disseminated intravascular coagulation.

Fig. 4. Kaplan-Meier survival curve and Cox regression survival curve according to the disseminated intravascular coagulopathy (DIC) score. The univariate Kaplan-Meier survival analysis revealed that a higher DIC score was associated with higher 6-month mortality (A). The multivariate Cox regression curve shows that groups with a DIC score of 3, 4, 5, and N 6 had higher Hazard ratios for 6-month mortality compared to the group with a DIC score of b 2 (B).

independent predictor for poor neurologic outcome in OHCA patients and they found that the risk of poor neurologic outcome increased by 84% with a 1 unit increase in the DIC score [11]. In our study, the DIC score rather than the other coagulofibrinolytic markers had the highest AUC value in predicting poor neurologic outcome. We also found that the DIC score was an independent predictor for poor neurologic out- come, and, like the study by Kim et al., we noted that the risk of poor neurologic outcome increased by 80% with a 1 unit increase in the DIC score [11]. The DIC score remained a significant predictor of 6-month mortality even after adjusting for the survival time. Additionally, in the present study, the risk for 6-month mortality increased by 26% with a 1 unit increase in the DIC score. Therefore, like the time to ROSC and shockable rhythm, which are well-known robust predictors of outcome in cardiac arrest patients, the DIC score can also be used as a surrogate marker for predicting the neurologic outcome or long- term mortality in cardiac arrest patients.

This study has several limitations. First, our results were based on retrospective analyses of the records of cardiac arrest patients at a single centre, and thus selection bias could have been present. Our study de- sign limits the ability to identity only association relationships rather

than causation. Second, an increased FDP/D-dimer ratio reflects DIC with the fibrinolytic phenotype, and Wada et al. showed that DIC with the fibrinolytic phenotype was dominant using the FDP/D-dimer ratio in cardiac arrest patients [12,19]. It was not appropriate to show the FDP/D-dimer ratio in the present study, since the maximal D-dimer value is limited to 35.2 mg/L fibrinogen-equivalent units in our labora- tory. Third, although about 55% of our patients had a cardiac etiology, we included all non-Traumatic cardiac arrest patients according to the Utstein template [24]. Therefore, the etiology of cardiac arrest was het- erogeneous in the present study. The coagulation and fibrinolysis activ- ity may be different according to the etiology of cardiac arrest as the asphyxial arrest had more severe ischemic insult than cardiac arrest [25]. Fourth, despite our efforts, there may have been unmeasured con- founding factors that were not included in our multivariate analyses. Fifth, the effects of different target temperatures on both coagulation and fibrinolytic activities remain controversial [26,27]; therefore, only patients who had a target temperature of 33 ?C were included. Finally, we did not take into consideration transfusion, factor replacement, or taking direct oral anticoagulants, which may have affected coagulation and fibrinolysis during post-cardiac arrest care. Hence, we investigated

Table 4 Cox regression analysis of 6-month mortality.
	HR (95% CI)	p	HR (95% CI)	p
Age, year	1.029 (1.017-1.041)	b 0.001	1.027 (1.015-1.039)	b0.001
Diabetes	1.450 (1.047-2.008)	0.025	1.469 (1.055-2.047)	0.023
Shockable rhythm	0.503 (0.317-0.799)	0.004	0.450 (0.282-0.717)	0.001
Cardiac etiology	0.706 (0.501-0.994)	0.046	0.617 (0.431-0.883)	0.008
Witness	0.687 (0.494-0.954)	0.025	0.724 (0.511-1.025)	0.069
Time to ROSC, min	1.011 (1.003-1.018)	0.006	1.009 (1.001-1.017)	0.022
Bystander CPR	0.809 (0.570-1.148)	0.236	0.715 (0.515-0.993)	0.045
Serum lactate, mmol/L	1.051 (1.013-1.090)	0.008	1.066 (1.027-1.107)	0.001
PaO2 after ROSC, mm Hg	1.001(1.000-1.003)	0.056	1.002 (1.000-1.003)	0.018
SOFA score	1.107 (1.052-1.164)	b 0.001	1.102 (1.047-1.159)	b0.001
DIC score	1.261 (1.089-1.461)	0.002
DIC score 0-2			Reference
DIC score 3			2.465 (1.348-4.507)	0.003
DIC score 4			2.595 (1.472-4.576)	0.001
DIC score 5			2.232 (1.161-4.288)	0.016
DIC score >= 6			4.312 (1.969-9.444)	b0.001

HR, hazards ratio; CI, confidence interval; ROSC, restoration of spontaneous circulation; CPR, cardiopulmonary resuscitation; SOFA, sequential organ failure assessment; DIC, disseminated intravascular coagulation.

the prognostic performance using only the variables obtained after ROSC and included the DIC score calculated after ROSC into the multi- variate regression models.

In conclusion, the DIC score and all of the examined coagulofibrinolytic markers, except fibrinogen, showed reasonable prognostic performance for poor neurologic outcome in cardiac arrest survivors who were treated with TTM. An increased DIC score, which had the highest AUC value, was an independent prognostic factor for poor neurologic outcome and 6-month mortality.

Funding sources/disclosures

This work supported by a grant (CRI16024-1) Chonnam National University Hospital Biomedical Research Institute.

Conflict of interest statement

No authors have any conflicts related to this work.

References

1. Geppert A, Zorn G, Delle-Karth G, Koreny M, Siostrzonek P, Heinz G, et al. Plasma concentrations of von Willebrand factor and intracellular adhesion molecule-1 for prediction of outcome after successful cardiopulmonary resuscitation. Crit Care Med 2003;31:805-11.
2. Adrie C, Monchi M, Laurent I, Um S, Yan SB, Thuong M, et al. Coagulopathy after suc- cessful cardiopulmonary resuscitation following cardiac arrest: implication of the protein C anticoagulant pathway. J Am Coll Cardiol 2005;46:21-8.
3. Bottiger BW, Motsch J, Bohrer H, Boker T, Aulmann M, Nawroth PP, et al. Activation of blood coagulation after cardiac arrest is not balanced adequately by activation of endogenous fibrinolysis. Circulation 1995;92:2572-8.
4. Gando S, Kameue T, Nanzaki S, Nakanishi Y. Massive fibrin formation with consecu- tive impairment of fibrinolysis in patients with out-of-hospital cardiac arrest. Thromb Haemost 1997;77:278-82.
5. Gando S, Nanzaki S, Morimoto Y, Kobayashi S, Kemmotsu O. Tissue factor and tissue factor pathway inhibitor levels during and after cardiopulmonary resuscitation. Thromb Res 1999;96:107-13.
6. Geppert A, Zorn G, Delle-Karth G, Heinz G, Maurer G, Siostrzonek P, et al. Plasmino- gen activator inhibitor type 1 and outcome after successful cardiopulmonary resus- citation. Crit Care Med 2001;29:1670-7.
7. Ono Y, Hayakawa M, Maekawa K, Kodate A, Sadamoto Y, Tominaga N, et al. Fibrin/ fibrinogen degradation products (FDP) at hospital admission predict neurological outcomes in out-of-hospital cardiac arrest patients. Resuscitation 2017;111:62-7.
8. Szymanski FM, Karpinski G, Filipiak KJ, Platek AE, Hrynkiewicz-Szymanska A, Kotkowski M, et al. Usefulness of the D-dimer concentration as a predictor of mor- tality in patients with out-of-hospital cardiac arrest. Am J Cardiol 2013;112:467-71.
9. Wada T, Gando S, Mizugaki A, Yanagida Y, Jesmin S, Yokota H, et al. Coagulofibrinolytic changes in patients with disseminated intravascular coagulation associated with post-cardiac arrest syndrome-fibrinolytic shutdown and insuffi- cient activation of fibrinolysis lead to organ dysfunction. Thromb Res 2013;132: e64-9.
10. Gando S. Microvascular thrombosis and Multiple organ dysfunction syndrome. Crit Care Med 2010;38:S35-42.
11. Kim J, Kim K, Lee JH, Jo YH, Kim T, Rhee JE, et al. Prognostic implication of initial co- agulopathy in out-of-hospital cardiac arrest. Resuscitation 2013;84:48-53.
12. Wada T, Gando S, Ono Y, Maekawa K, Katabami K, Hayakawa M, et al. Disseminated intravascular coagulation with the fibrinolytic phenotype predicts the outcome of patients with out-of-hospital cardiac arrest. Thromb J 2016;14:43.
13. Callaway CW, Donnino MW, Fink EL, Geocadin RG, Golan E, Kern KB, et al. Part 8: post-cardiac arrest care: 2015 American Heart Association guidelines update for car- diopulmonary resuscitation and emergency cardiovascular care. Circulation 2015; 132:S465-82.
14. Peberdy MA, Callaway CW, Neumar RW, Geocadin RG, Zimmerman JL, Donnino M, et al. Part 9: post-cardiac arrest care: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122:S768-86.
15. Vincent JL, de Mendonca A, Cantraine F, Moreno R, Takala J, Suter PM, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on “sepsis-re- lated problems” of the European Society of Intensive Care Medicine. Crit Care Med 1998;26:1793-800.
16. Taylor Jr FB, Toh CH, Hoots WK, Wada H, Levi M. Scientific subcommittee on dissem- inated intravascular coagulation of the International Society on thrombosis and haemostasis. Towards definition, clinical and laboratory criteria, and a scoring sys- tem for disseminated intravascular coagulation. Thromb Haemost 2001;86: 1327-30.
17. Booth CM, Boone RH, Tomlinson G, Detsky AS. Is this patient dead, vegetative, or se- verely neurologically impaired? Assessing outcome for comatose survivors of cardi- ac arrest. JAMA 2004;291:870-9.
18. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Bio- metrics 1988;44:837-45.
19. Asakura H. Classifying types of disseminated intravascular coagulation: clinical and animal models. J Intensive Care 2014;2:20.
20. Lowenstein CJ, Morrell CN, Yamakuchi M. Regulation of Weibel-Palade body exocy- tosis. Trends Cardiovasc Med 2005;15:302-8.
21. Schochl H, Cadamuro J, Seidl S, Franz A, Solomon C, Schlimp CJ, et al. Hyperfibrinolysis is common in out-of-hospital cardiac arrest: results from a pro- spective observational thromboelastometry study. Resuscitation 2013;84:454-9.
22. Viersen VA, Greuters S, Korfage AR, Van der Rijst C, Van Bochove V, Nanayakkara PW, et al. Hyperfibrinolysis in out of hospital cardiac arrest is associated with markers of hypoperfusion. Resuscitation 2012;83:1451-5.
23. Duvekot A, Viersen VA, Dekker SE, Geeraedts Jr LM, Schwarte LA, Spoelstra-Man AM, et al. Low cerebral oxygenation levels during resuscitation in out-of-hospital cardiac arrest are associated with hyperfibrinolysis. Anesthesiology 2015;123:820-9.
24. Perkins GD, Jacobs IG, Nadkarni VM, Berg RA, Bhanji F, Biarent D, et al. Cardiac arrest

    and cardiopulmonary resuscitation outcome reports: update of the Utstein resusci- tation registry templates for out-of-hospital cardiac arrest: a statement for healthcare professionals from a task force of the International Liaison Committee on Resuscitation (American Heart Association, European Resuscitation Council, Aus- tralian and New Zealand Council on Resuscitation, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Council of Southern Africa, Resuscitation Council of Asia); and the American Heart Association Emergency Car- diovascular Care Committee and the Council on Cardiopulmonary, Critical Care, Peri- operative and Resuscitation. Circulation 2015;132:1286-300.

    Vaagenes P, Safar P, Moossy J, Rao G, Diven W, Ravi C, et al. Asphyxiation versus ven- tricular fibrillation cardiac arrest in dogs. Differences in cerebral resuscitation effects-a preliminary study. Resuscitation 1997;35:41-52.

25. Jeppesen AN, Kirkegaard H, Ilkjaer S, Hvas AM. Influence on temperature on thromboelastometry and platelet aggregation in cardiac arrest patients undergoing targeted temperature management. Crit Care 2016;20:118.
26. Jacob M, Hassager C, Bro-Jeppesen J, Ostrowski SR, Thomsen JH, Wanscher M, et al. The effect of targeted temperature management on coagulation parameters and bleeding events after out-of-hospital cardiac arrest of presumed cardiac cause. Re- suscitation 2015;96:260-7.