- Email: [email protected]
Prognostic implication of initial coagulopathy in out-of-hospital cardiac arrest
Resuscitation 84 (2013) 48–53
Contents lists available at SciVerse ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
Prognostic implication of initial coagulopathy in out-of-hospital cardiac arrest夽 Joonghee Kim, Kyuseok Kim ∗ , Jae Hyuk Lee, You Hwan Jo, Taeyun Kim, Joong Eui Rhee, Kyeong Won Kang Department of Emergency Medicine, Seoul National University Bundang Hospital, 166 Gumi-ro, Bundang-gu, Seongnam-si, Gyeonggi-do, 463-707, Republic of Korea
a r t i c l e
i n f o
Article history: Received 21 May 2012 Received in revised form 24 July 2012 Accepted 3 September 2012
Keywords: Disseminated intravascular coagulation Out-of-hospital cardiac arrest Prognosis Risk factors Cumulative survival rate Coagulopathy
a b s t r a c t Objective: We sought to investigate the prognostic implication of early coagulopathy represented by initial DIC score in out-of-hospital cardiac arrest (OHCA). Methods: OHCA registry was analyzed to identify patients with ROSC without recent use of anticoagulant between 2008 and 2011. Patients were assessed for prehosptial factors, initial laboratory results and therapeutic hypothermia. Outcome variables were survival discharge, 6-month CPC and survival duration within the first week after ROSC. Logistic regression and Cox proportional hazards models were used for both univariable and multivariable analysis. Results: Among 273 eligible patients, initial DIC score was available in 252 (92.3%). Higher DIC score was associated with increased inhospital death (odds ratio [OR], 1.89 per unit; 95% confidence interval [CI], 1.48–2.41) and unfavorable long-term outcome (6-month CPC 3–5; OR, 2.21 per unit; 95% CI, 1.60–3.05). The adjusted ORs for both outcomes were 1.61 (95% CI, 1.17–2.22) and 1.84 (95% CI, 1.26–2.67), respectively. We categorized DIC score in five groups as <3, 3, 4, 5 and >5 and analyzed differential mortality risk using Cox proportional hazards model. Compared with reference group (DIC score < 3), the adjusted HR for early mortality in each remaining group was 1.96 (95% CI, 1.13–3.40), 2.26 (95% CI, 1.27–4.02), 2.77 (95% CI, 1.58–4.85) and 4.29 (95% CI, 2.22–8.30), respectively (p-trend < 0.001). The area under the receiver operating characteristic of DIC score for prediction of unfavorable long-term outcome was 0.79 (95% CI, 0.69–0.88). Conclusion: Increased initial DIC score in OHCA was an independent predictor for poor outcomes and early mortality risk. © 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Post cardiac arrest syndrome (PCAS) is characterized by its high early mortality.1,2 One of the major causes of this early mortality is the systemic ischemic-reperfusion injury previously described as “sepsis-like syndrome” where global activation of coagulation system is commonly found.3,4 In sepsis and trauma, the coagulation activation and following consumptive coagulopathy are well-known phenomenon and their increasing severity is correlated with poor outcome.5–7 The International Society on Thrombosis and Hemostasis (ISTH) disseminated intravascular coagulation (DIC) score is a simple scoring system measuring the severity of the coagulopathy and is used for the diagnosis of DIC.8 Its validity for the diagnosis of DIC was confirmed in a previous study,6 and has been verified to be sensitive to DIC of both infective
and non-infective etiologies.9 In addition, it is strongly correlated with mortality in many conditions including sepsis, trauma and burn.6,10–13 Although significant coagulopathy has been reported also in cardiac arrest patients,14,15 the prognostic implication of DIC score has not been studied in this condition. Considering the significant role of coagulation process in MODS of sepsis or trauma,16,17 it is possible that DIC score has significant prognostic implication in PCAS where MODS is common. We postulated that initial DIC score in resuscitated out-of-hospital cardiac arrest (OHCA) patients is correlated with early mortality rate and thus cardiac arrest outcomes. The objective of this study is to describe the severity of initial coagulopathy in cardiac arrest patients using ISTH DIC score and evaluate its prognostic implication in PCAS.
2. Materials and methods 夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at http://dx.doi.org/10.1016/j.resuscitation.2012.09.003. ∗ Corresponding author. Tel.: +82 31 787 7579; fax: +82 31 787 4055. E-mail address: [email protected] (K. Kim). 0300-9572/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.resuscitation.2012.09.003
We retrospectively studied a prospective registry of all consecutive OHCA victims brought to a single emergency department (ED) from January, 2008 to December, 2011. The institutional review
J. Kim et al. / Resuscitation 84 (2013) 48–53
board at the study facility approved the analysis and issued a waiver of consent. 2.1. Study setting The study facility was a 950-bed tertiary academic hospital in an urban area with a population of 440,000 where about 8.4% of whole population was aged over 65. This area is mostly served by a two-tiered ambulance service provided by fire service run by government. The level of care provided by emergency medical technicians (EMTs) is comparable to that of EMT intermediate in United States. The management of cardiac arrest was based on the recommendations of the 2005 American heart association cardiopulmonary resuscitation guideline until when 2010 guideline from the same organization was published and adopted. Because of routine blood sampling protocol for OHCA, patients underwent blood sampling instantly on ED arrival.18 All laboratory variables considered in this study were part of the initial routine order set for OHCA and thus were measured concurrently from the initial blood samples. 2.2. Participants and data collection ED OHCA registry extending from January, 2008 to December, 2011 was analyzed by researchers. Target population was all ED OHCA patients who achieved ROSC. Exclusion criteria included age <18 years and documented recent use of warfarin, heparin, low molecular weight heparin and intravenous thrombolytics. Prehospital factors including sex, age, arrest location, presence of witness, bystander CPR, time to basic life support (BLS), initial rhythm, prehospital ROSC and cause of arrest were extracted from the registry and chart review. The detailed definition of prehospital factors and other terms are described in supplementary Table 1 (Supplementary materials are available online at RESUSCITATION journal website). Laboratory variables including white blood cell, hematocrit, and platelet count, urea nitrogen, creatinine, bicarbonate, albumin, total bilirubin, C-reactive protein, troponin I, fibrinogen and D-dimer concentration and PT (seconds) were retrieved from chart review. The ISTH DIC score was calculated from PT (seconds), platelet count, serum fibrinogen level and D-dimer concentration (supplementary Table 2). The cutoff values of D-dimer were based on previous studies.6,7 Outcome variables including survival discharge, 6-month cerebral performance category (CPC) score and exact duration of survival within 7 days post-ROSC were obtained by chart review and structured telephone interview. A 6-month CPC score of 1 or 2 was considered a favorable long-term outcome, whereas that of 3–5 was considered an unfavorable one. 2.3. Statistical analysis Analysis of variance (ANOVA), Kruskal–Wallis, Chi-square or Fisher’s exact test was performed as appropriate for comparison between groups. Pearson’s method was used for correlation analysis if the two variables involved are both interval variables and Spearman’s method was used otherwise. Logistic regression and Cox proportional hazards model analysis were used for univariable analysis. To identify independent prognostic factors, multivariable stepwise logistic regression analyses (backward stepwise method based on likelihood ratio) were performed using the variables with their p-value < 0.1 in univariable analysis and their goodness of fit was assessed with the Hosmer–Lemeshow test. The same backward stepwise method was also used in a multivariable Cox proportional hazard regression analysis. The result of logistic regression and Cox proportional hazard model analysis was presented as odd ratios (ORs) and hazard ratios (HRs), respectively, and their 95% confidence intervals (CIs) were also presented. Survival curves based
49
Table 1 Patient characteristics of study population. Sex, male, no. (%) Age, years, mean (SD) Cardiac arrest in public location, no. (%) Witnessed cardiac arrest, no. (%) Bystander chest compression, no. (%) Time to first BLS, min, median (IQR) Shockable initial rhythm, no. (%) Prehospital ROSC, no. (%) Cause of arrest, no. (%) Cardiac Respiratory Traumatic Food asphyxia Hanging Head and neck injury Multiple blunt or penetrating injury Drowning Intoxication Therapeutic hypothermia, no. (%) Overt DIC (DIC score ≥ 5), no. (%) 1-week survival, no. (%) Survival discharge, no. (%) 6-month CPC 1 or 2, no. (%)
149 (59.1) 64.1 (17.3) 52 (20.6) 205 (81.3) 79 (31.3) 4 (0–11) 69 (27.4) 7 (2.8) 91 (36.1) 111 (44.0) 53 (21.0) 18 (7.1) 14 (5.6) 9 (3.6) 5 (2.0) 3 (1.2) 1 (0.4) 81 (32.1) 82 (32.5) 79 (31.3) 54 (21.4) 28 (11.1)
SD, standard deviation; IQR, interquartile range; BLS, basic life support; ROSC, return of spontaneous circulation; CPC, cerebral performance category.
on 1-week survival data were designed using the Kaplan–Meier method, and comparisons were made using the log-rank test. The discriminatory performance of DIC score for long-term outcome prediction was evaluated using the area under the receiver operating characteristic (ROC) curve. The comparison of independent ROC curves was conducted using the method of Hanley and McNeil.19 We performed a sensitivity analysis excluding cases of non-cardiac etiology to see the prognostic implication of DIC score when the potential influences from underlying conditions capable of changing coagulation profile were minimized. P-values < 0.05 were considered significant. All analyses were performed using SPSS 15.0 (SPSS Inc., IL), STATA 10.1 (StataCorp 113 LP, TX) and MedCalc 10.2 (MedCalc Software, Belgium). 3. Results 3.1. Baseline characteristics A total of 547 OHCA patients were brought to the study ED during the study period and 292 patients achieved ROSC (supplementary Fig. 1). After excluding patients with age <18 years or any documented recent use of warfarin, heparin-based products or intravenous thrombolytics, there were 273 eligible patients. The study population finally included 252 (92.3%) patients whose initial laboratory variables required for calculation of initial DIC score were available. The baseline characteristics of the study population are summarized in Table 1. Eighty two (32.5%) showed overt DIC feature with their DIC score equal to or more than 5. Survival discharge to home or nursing home was possible in 54 (21.4%) patients. Favorable long-term outcome (6-month CPC 1 or 2) was achieved in 28 (11.1%) patients. Patient characteristics of the study population were stratified according to initial DIC score (Table 2). Only three (2.3%) patients whose DIC score was more than 3 (N = 130) achieved favorable longterm outcome. Prehospital factors that significantly differed across the stratified groups included age, presence of witness, prehospital ROSC and presumed cardiac etiology. For laboratory variables, all except WBC differed significantly. Significant correlation was found between DIC score and many laboratory variables including hematocrit, bicarbonate, BUN, albumin, total bilirubin and C-reactive
50
Table 2 Patient characteristics stratified according to DIC score. DIC score 0–2 n = 55 39 (70.9) 55.9 (18.2) 12 (21.8) 38 (69.1) 20 (36.4) 2 (0–7) 20 (37.0) 5 (9.1) 27 (49.1)
39 (58.2) 63.8 (14.8) 21 (31.3) 57 (85.1) 25 (37.3) 6 (1–12) 18 (26.9) 1 (1.5) 29 (43.3)
DIC score 4 n = 48 26 (54.2) 73.3 (13.1) 8 (16.7) 41 (85.4) 18 (37.5) 4 (0–12) 9 (18.8) 0 (0.0) 13 (27.1)
DIC score 5 n = 59 36 (61.0) 66.9 (17.1) 7 (11.9) 47 (79.7) 12 (20.3) 5 (0–12) 17 (29.3) 1 (1.7) 18 (30.5)
222 (175–293) 14 (13–15) 305 (260–391) 2.04 (0.68–3.02)
196 (159–244) 15 (15–17) 380 (278–535) 7.32 (3.14–18.77)
165 (137–248) 17 (17–19) 315 (268–438) 15.48 (4.58–20.00a )
149 (106–218) 21 (19–24) 337 (205–484) 15.04 (7.46–20.00a )
11.5 (10.2–13.3) 42.1 (6.0) 15 (13–19) 1.2 (1.0–1.4) 16.7 (4.0) 4.1 (0.5) 0.7 (0.5–0.9) 0.20 (0.20–0.30) 0.04 (0.02–0.10) 26 (47.3)
12.8 (10.4–15.4) 38.0 (7.6) 19 (15–30) 1.3 (1.0–1.9) 14.6 (4.4) 3.6 (0.7) 0.6 (0.4–0.8) 0.34 (0.20–4.70) 0.05 (0.04–0.15) 28 (41.8)
12.6 (9.1–16.1) 36.1 (8.8) 19 (15–43) 1.3 (0.9–2.7) 14.2 (4.6) 3.3 (0.8) 0.8 (0.5–1.0) 0.80 (0.20–5.26) 0.05 (0.03–0.19) 12 (25.5)
13.8 (9.3–18.2) 36.0 (9.0) 23 (14–45) 1.5 (1.1–2.6) 12.0 (5.3) 2.9 (0.7) 1.0 (0.7–1.7) 3.04 (0.48–11.36) 0.14 (0.06–0.36) 14 (23.7)
35 (63.6) 25 (45.5) 17 (31.5)
23 (34.3) 17 (25.4) 8 (11.9)
10 (21.3) 5 (10.4) 0 (0.0)
10 (16.9) 6 (10.2) 2 (3.4)
DIC score 6–8 n = 23 9 (39.1) 58.3 (19.6) 4 (17.4) 22 (95.7) 4 (17.4) 3 (0–12) 5 (21.7) 0 (0) 4 (17.4) 65 (27–92) 23 (20–31) 135 (50–234) 20.00a (20.00a –20.00a ) 11.2 (6.0–13.7) 33.8 (13.5) 19 (11–33) 1.2 (0.9–1.6) 12.7 (5.3) 2.7 (1.0) 0.7 (0.4–1.6) 0.32 (0.20–2.67) 0.27 (0.04–0.56) 2 (8.7) 1 (4.3) 1 (4.3) 1 (4.3)
p
0.109 <0.001 0.087 0.044 0.092 0.380 0.310 0.029 0.021 <0.001 <0.001 <0.001 <0.001 0.127 <0.001 0.002 0.042 <0.001 <0.001 <0.001 <0.001 <0.001 0.002 <0.001 <0.001 <0.001
SD, standard deviation; IQR, interquartile range; BLS, basic life support; ROSC, return of spontaneous circulation, PT, prothrombin time; WBC, white blood cell; BUN, blood urea nitrogen; CRP, C-reactive protein; CPC, cerebral perfomance category. a D-dimer level higher than 20 ug/mL was reported as “>20 ug/mL” in study hospital.
J. Kim et al. / Resuscitation 84 (2013) 48–53
Prehospital and demographic variables Sex, male, no. (%) Age, years, mean (SD) Cardiac arrest in public location, no. (%) Witnessed cardiac arrest, no. (%) Bystander chest compression, no. (%) Time to first BLS, min, median (IQR) Shockable initial rhythm, no. (%) Prehospital ROSC, no. (%) Presumed cardiac etiology, no. (%) DIC score components Platelet, 109 /L, median (IQR) PT, seconds, median (IQR) Fibrinogen, mg/dL, median (IQR) D-dimer, ug/mL, median (IQR) Other laboratory variables for adjustment WBC count, 109 /L, median (IQR) Hematocrit, %, mean (SD) BUN, mg/dL, median (IQR) Creatinine, mg/dL, median (IQR) Bicarbonate, mEq/L, mean (SD) Albumin, g/dL, mean (SD) Bilirubin total, mg/dL, median (IQR) CRP, mg/dL, median (IQR) Troponin I, ng/mL (IQR) Therapeutic hypothermia, no. (%) Cardiac arrest outcomes, no. (%) 1-week survival, no. (%) Survival discharge, no. (%) 6-month CPC 1 or 2, no (%)
DIC score 3 n = 67
J. Kim et al. / Resuscitation 84 (2013) 48–53 Table 3 Stepwise multivariatelogistic regression model analysis of the relationship between poor outcome and various potential prognostic factors. Odds ratio Inhospital death Sex (male) Age (per 1 year) Time to first BLS (per 1 min) Shockable initial rhythm Bicarbonate (per 1 mEq/L) Therapeutic hypothermia DIC score (per 1 unit) Poor long-term outcome (6-month CPC ≥ 3) Age (per 1 year) Time to first BLS (per 1 min) Shockable initial rhythm Presumed cardiac etiology DIC score (per 1 unit)
Table 4 Stepwise multivariateCox proportional hazard model analysis ofthe associations between early mortality risk and various potential prognostic factors.
p
0.47 (0.21–1.06) 1.03 (1.00–1.06) 1.06 (1.00–1.13) 0.20 (0.09–0.44) 0.91 (0.83–1.00) 0.39 (0.18–0.85) 1.61 (1.17–2.22)
0.068 0.035 0.048 <0.001 0.045 0.018 0.003
1.03 (1.00–1.07) 1.16 (1.03–1.31) 0.14 (0.05–0.40) 0.31 (0.11–0.92) 1.84 (1.26–2.67)
0.060 0.017 <0.001 0.034 0.002
BLS, basic life support; CPC, cerebral performance category. Note: The prediction model for inhospital death and poor long-term outcome correctly classified in 83.5% and 92.6% of cases, respectively.
protein. Among them, albumin (Pearson’s r = −0.550, p < 0.001), bicarbonate (Pearson’s r = −0.310, p < 0.001) and hematocrit (Pearson’s r = −0.300, p < 0.001) were most strongly correlated with DIC score. 3.2. Association between DIC score and overall CPR outcome and prognostic performance of DIC score Increased DIC score was a strong risk factor for both inhospital death and unfavorable long-term outcome (6-month CPC 3–5). In a univariable logistic regression analysis (supplementary Table 3), one-unit increase of DIC score was associated with 1.89 (95% CI, 1.48–2.41)- and 2.21 (95% CI, 1.60–3.05)-fold increase of risk for inhospital death and unfavorable long-term outcome, respectively. Backward stepwise multivariable logistic regression models were constructed using the variables with their p < 0.1 in the univariable analyses (Table 3). After adjustment, DIC score remained a significant predictor for inhospital death and unfavorable longterm outcome, with 1.61 (95% CI, 1.17–2.22)- and 1.84 (95% CI, 1.26–2.67)-fold increased risk per unit, respectively. Other independent predictors for inhospital mortality were age, time to BLS, initial rhythm, serum bicarbonate and application of hypothermia. For unfavorable long-term outcome, time to BLS, initial rhythm and presumed cardiac etiology were independent predictors. The long-term prognostic performance of DIC score was assessed by analysis of ROC curve. The area under the ROC curve (AUC) of DIC score for the prediction of unfavorable long-term outcome was 0.79 (95% CI, 0.69–0.88). The optimal cutoff point of DIC score to predict the poor outcome was found between 3 and 4. Using the cutoff point, the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of DIC score were 57.0 (95% CI, 50.2–63.5), 89.3 (95% CI, 71.8–97.6), 97.7 (95% CI, 93.4–99.5) and 20.7 (95% CI, 13.8–29.0), respectively.
51
Age (per 1 year) Shockable initial rhythm CRP (per 1 mg/dL) Therapeutic hypothermia DIC scoregroups (p-trend < 0.001) 1st group (DIC score 0–2, n = 55) 2nd group (DIC score 3, n = 67) 3rd group (DIC score 4, n = 48) 4th group (DIC score 5, n = 59) 5th group (DIC score 6–8, n = 23)
Hazard ratio
p
1.02 (1.01–1.03) 0.49 (0.33–0.73) 1.02 (1.00–1.04) 0.38 (0.26–0.55)
0.003 <0.001 0.047 <0.001
1.00 (Reference) 1.96 (1.13–3.40) 2.26 (1.27–4.02) 2.77 (1.58–4.85) 4.29 (2.22–8.30)
0.017 0.006 <0.001 <0.001
CRP, C-reactive protein.
group, respectively, compared with those in the first group (DIC score < 3). In a multivariable Cox proportional hazard model analysis using stepwise backward method (Table 4), the risk gradient was still significant (p-trend < 0.001). The adjusted hazard ratio for death during the first week after ROSC was 2.77 (95% CI, 1.58–4.85)and 4.29 (95% CI, 2.22–8.30)-fold higher in the fourth and fifth group, respectively, compared with those in the first group. Other independent predictors in the model were age, initial rhythm, Creactive protein and therapeutic hypothermia. Kaplan–Meier survival curves for the five groups during the first week were constructed (Fig. 1). The survival curves were statistically different according to the log-rank test (p = < 0.001, ptrend < 0.001). 3.4. Subgroup analysis of cardiac arrest cases of presumed cardiac origin We performed a sensitivity analysis excluding non-cardiac etiology cases to see the prognostic implication of DIC score when the potential influences from underlying conditions capable of changing coagulation profile were minimized. The magnitude of increased risk represented by ORs and HRs in cardiac etiology subgroup was not lower compared with that of whole study population (Supplementary Table 5). Fig. 2A shows the percentage of patients who achieved 1-week survival, survival discharge and favorable long-term outcome in each DIC score subgroup. Patients with cardiac etiology usually had better chance of achieving survival discharge (p = 0.002) and favorable long-term outcome (p < 0.001), especially if their DIC score was less than five (p = 0.006, p < 0.001,
3.3. Association between DIC score and mortality risk during the first week after ROSC To assess the association between DIC score and early mortality risk, we categorized DIC score in five groups as <3, 3, 4, 5 and >5 and analyzed their differential survival rate during the first week after ROSC. In univariable analysis using Cox proportional hazard model analysis (supplementary Table 4), there was significant gradient of increasing hazard ratio across the five consecutive groups (ptrend < 0.001). The hazard ratio for death during the first week after ROSC was 3.95 (95% CI, 2.34–6.68)- and 6.13 (95% CI, 3.30–11.38)fold higher in the fourth (DIC score 5) and fifth (DIC score > 5)
Fig. 1. Kaplan–Meier plots for cumulative 1-week survival. In fifth DIC score group (DIC score > 5, n = 23), only one patient survived more than 48 h. The difference was significant according to the log-rank test (p < 0.001, p-trend < 0.001).
52
J. Kim et al. / Resuscitation 84 (2013) 48–53
Fig. 2. (A) CPR outcome differences across DIC score subgroups of non-cardiac and cardiac etiology. (B) Comparison of receiving operating characteristic curves of DIC score for predicting poor long-term outcome (6-month CPC 3–5) in non-cardiac origin (dash dot line) and cardiac origin (dashed line) subgroups. The area under the curves were 0.64 (95% CI, 0.41–0.86; p < 0.173) and 0.84 (95% CI, 0.76–0.93; p < 0.001) respectively. However, their difference was not statistically significant (p = 0.095).
respectively). And no one regained consciousness (CPC 4, 5) in 3rd, 4th and 5th group (DIC score > 3, N = 35) if the cause of arrest was presumed to be cardiac-origin. Fig. 2B depicts the ROC curves of DIC score for prediction of unfavorable long-term outcome (6-month CPC 3–5) in whole study population, non-cardiac etiology subgroup and cardiac etiology subgroup. The AUC of DIC score for the prediction of unfavorable long-term outcome was 0.84 (95% CI, 0.76–0.93) in cardiac-origin subgroup while it was 0.64 (95% CI, 0.41–0.86) in non-cardiac origin subgroup. Although the AUC was most large in cardiac-subgroup, the difference between non-cardiac and cardiac etiology subgroup was not statistically significant (p = 0.095). The optimal cutoff point of DIC score for prediction of unfavorable long-term outcome in cardiac-subgroup was identical (between DIC score 3 and 4) as that of whole study population. Using the cutoff point, the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of DIC score were 48.6 (95% CI, 36.7–60.7), 100.0 (95% CI, 82.2–100.0), 100.0 (95% CI, 89.9–100.0) and 33.9 (95% CI, 21.8–47.8), respectively. 4. Discussion In this study, the increased initial DIC score was an independent risk factor for inhospital death and unfavorable long-term outcome. It was also associated with markedly increased rate of death during early post-resuscitation period. These relationships remained significant in cardiac-cause subgroup where the influence from underlying conditions causing coagulopathy is minimized. The underlying mechanism of this observation would be multifactorial. There is high possibility that patients with higher DIC score were more frequently involved with underlying or precipitating conditions that could had already caused varying degree of coagulopathy before cardiac arrest. Although the frequency of such conditions would be much lower in cardiac-cause subgroup, it is not possible to completely eradicate such confounding effect in a retrospective study design targeting cardiac arrest patients. However, the global coagulation activation of cardiac arrest per se could have significant prognostic implication.15 Previous studies suggested that consumptive coagulopathy plays central role in pathogenesis of MODS in critical conditions such as sepsis or trauma.20 As similar global activation of coagulation pathway is also observed in cardiac arrest patients,14 its role in disease progression of PCAS might be significant. In this study, we found a significant correlation between DIC score and early mortality rate. Considering that MODS including cardiac dysfunction is an
important cause of death during early post-resuscitation period, it might be prudent to infer that this early coagulopathy might play a significant role in the development of MODS and subsequent increased mortality. However the lack of data regarding chronological changes in hemodynamics and organ function did not allow us to prove whether such manifestations were actually differed significantly. Generalized activation of coagulation system is an active ongoing process commonly observed during cardiac arrest.14,21 In our study, blood sampling was done immediately after patients’ ED arrival and thus the laboratory values we analyzed represent coagulation status before or immediately after ROSC except for the patients who achieved prehospital ROSC (N = 7, 2.8%). Even for the patients with prehospital ROSC, the median time delay from ROSC to blood sampling was very short (median 2 minutes; IQR, 2–6). Despite this early timing, 82 (32.5%) of whole study population and 22 (24.2%) of cardiac-cause subpopulation showed high DIC score indicative of fulminant DIC (DIC score 5 or more). However, the pathophysiologic consequences of this early change are not yet fully understood in cardiac arrest. Whether this is just an indicator of ischemic injury from cardiac arrest or an active triggering mechanism of following organ dysfunction or both should be the subject of a future study. This study has several limitations. First, the result of this study is based upon retrospective analysis of ED OHCA registry and medical records and thus is limited by possible biases that are intrinsic in such design. Second, although we tried to reduce the number of variables as many as possible by using only significant variables in univariable analysis and adopting stepwise backward logistic regression, the number of events per variables ratio of the multivariable logistic regression models were lower than the desired ratio of >10 suggested by Peduzzi et al.22 The variables considered in the analyses were carefully chosen and we thought that their biological plausibility of influence on both coagulation activation and clinical outcome cannot be easily ignored. As the main purpose of this study was to look into the prognostic significance of DIC score rather than making a good multivariable model to predict ultimate outcome of OHCA patient, we chose not to further remove clinically relevant variables from the beginning. Third, we did not perform multivariable model analysis in cardiac subgroup as its population size was too small for multivariable analysis using large set of prognostic variables we used. Fourth, chronological evaluation of hemodynamic indices, inflammatory markers and functional status of major organs was not possible. Availability of such data could provide more direct evidence of the association
J. Kim et al. / Resuscitation 84 (2013) 48–53
between DIC score and early pathological progressions including myocardial dysfunction, inflammation and MODS. 5. Conclusions Increased initial DIC score in cardiac arrest was an independent risk factor for both inhospital death and unfavorable long-term outcome. The higher mortality risk during early post-resuscitation period could be one possible explanation behind it. Conflicts of interest statement All authors do not have any conflicts of interest. Source of support This study was supported by grant from the Seoul National University Bundang Hospital Research Fund, No. 02-2012-011. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.resuscitation.2012.09.003. References 1. Neumar RW, Nolan JP, Adrie C, et al. Post-cardiac arrest syndrome: epidemiology, pathophysiology, treatment, and prognostication. A consensus statement from the International Liaison Committee on Resuscitation (American Heart Association, Australian and New Zealand Council on Resuscitation, European Resuscitation Council, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Council of Asia, and the Resuscitation Council of Southern Africa); the American Heart Association Emergency Cardiovascular Care Committee; the Council on Cardiovascular Surgery and Anesthesia; the Council on Cardiopulmonary, Perioperative, and Critical Care; the Council on Clinical Cardiology; and the Stroke Council. Circulation 2008;118:2452–83. 2. Herlitz J, Engdahl J, Svensson L, Angquist K-A, Silfverstolpe J, Holmberg S. Major differences in 1-month survival between hospitals in Sweden among initial survivors of out-of-hospital cardiac arrest. Resuscitation 2006;70:404–9. 3. Adrie C. Successful cardiopulmonary resuscitation after cardiac arrest as a “sepsis-like” syndrome. Circulation 2002;106:562–8.
53
4. Adrie C, Laurent I, Monchi M, Cariou A, Dhainaou J-F, Spaulding C. Postresuscitation disease after cardiac arrest: a sepsis-like syndrome? Curr Opin Crit Care 2004;10:208–12. 5. Levi M. Disseminated intravascular coagulation. Crit Care Med 2007;35:2191–5. 6. Bakhtiari K, Meijers JCM, de Jonge E, Levi M. Prospective validation of the International Society of Thrombosis and Haemostasis scoring system for disseminated intravascular coagulation. Crit Care Med 2004;32:2416–21. 7. Dhainaut JF, Yan S, Joyce D, et al. Treatment effects of drotrecogin alfa (activated) in patients with severe sepsis with or without overt disseminated intravascular coagulation1. J Thromb Haemost 2004;2:1924–33. 8. Taylor FB, Toh CH, Hoots WK, Wada H, Levi M, Scientific Subcommittee on Disseminated Intravascular Coagulation (DIC) of the International Society on Thrombosis and Haemostasis (ISTH). Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost 2001;86:1327–30. 9. Levi M, Toh CH, Thachil J, Watson HG. Guidelines for the diagnosis and management of disseminated intravascular coagulation. British Committee for Standards in Haematology. Br J Haematol 2009;145:24–33. 10. Gando S, Wada H, Asakura H, et al. Evaluation of new Japanese diagnostic criteria for disseminated intravascular coagulation in critically ill patients. Clin Appl Thromb Hemost 2005;11:71–6. 11. Sivula M, Tallgren M, Pettilä V. Modified score for disseminated intravascular coagulation in the critically ill. Intensive Care Med 2005;31:1209–14. 12. Cauchie P, Cauchie C, Boudjeltia KZ, et al. Diagnosis and prognosis of overt disseminated intravascular coagulation in a general hospital – meaning of the ISTH score system, fibrin monomers, and lipoprotein–C-reactive protein complex formation. Am J Hematol 2006;81:414–9. 13. Lavrentieva A, Kontakiotis T, Bitzani M, et al. Early coagulation disorders after severe burn injury: impact on mortality. Intensive Care Med 2008;34:700–6. 14. Böttiger B, Motsch J, Böhrer H, Böker T. Activation of blood coagulation after cardiac arrest is not balanced adequately by activation of endogenous fibrinolysis. Circulation 1995;92:2572–8. 15. Adrie C, Monchi M, Laurent I, et al. Coagulopathy after successful cardiopulmonary resuscitation following cardiac arrest. J Am Coll Cardiol 2005;46:21–8. 16. Levi M, van der Poll T. Inflammation and coagulation. Crit Care Med 2010;38:S26–34. 17. van der Poll T, de Boer JD, Levi M. The effect of inflammation on coagulation and vice versa. Curr Opin Infect Dis 2011;24:273–8. 18. Kim J, Kim K, Lee JH, et al. Red blood cell distribution width as an independent predictor of all-cause mortality in out of hospital cardiac arrest. Resuscitation 2012, doi:10.1016/j.resuscitation.2012.01.038. 19. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29–36. 20. Gando S. Microvascular thrombosis and multiple organ dysfunction syndrome. Crit Care Med 2010;38:S35–42. 21. 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. 22. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996;49:1373–9.
Contents lists available at SciVerse ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
Prognostic implication of initial coagulopathy in out-of-hospital cardiac arrest夽 Joonghee Kim, Kyuseok Kim ∗ , Jae Hyuk Lee, You Hwan Jo, Taeyun Kim, Joong Eui Rhee, Kyeong Won Kang Department of Emergency Medicine, Seoul National University Bundang Hospital, 166 Gumi-ro, Bundang-gu, Seongnam-si, Gyeonggi-do, 463-707, Republic of Korea
a r t i c l e
i n f o
Article history: Received 21 May 2012 Received in revised form 24 July 2012 Accepted 3 September 2012
Keywords: Disseminated intravascular coagulation Out-of-hospital cardiac arrest Prognosis Risk factors Cumulative survival rate Coagulopathy
a b s t r a c t Objective: We sought to investigate the prognostic implication of early coagulopathy represented by initial DIC score in out-of-hospital cardiac arrest (OHCA). Methods: OHCA registry was analyzed to identify patients with ROSC without recent use of anticoagulant between 2008 and 2011. Patients were assessed for prehosptial factors, initial laboratory results and therapeutic hypothermia. Outcome variables were survival discharge, 6-month CPC and survival duration within the first week after ROSC. Logistic regression and Cox proportional hazards models were used for both univariable and multivariable analysis. Results: Among 273 eligible patients, initial DIC score was available in 252 (92.3%). Higher DIC score was associated with increased inhospital death (odds ratio [OR], 1.89 per unit; 95% confidence interval [CI], 1.48–2.41) and unfavorable long-term outcome (6-month CPC 3–5; OR, 2.21 per unit; 95% CI, 1.60–3.05). The adjusted ORs for both outcomes were 1.61 (95% CI, 1.17–2.22) and 1.84 (95% CI, 1.26–2.67), respectively. We categorized DIC score in five groups as <3, 3, 4, 5 and >5 and analyzed differential mortality risk using Cox proportional hazards model. Compared with reference group (DIC score < 3), the adjusted HR for early mortality in each remaining group was 1.96 (95% CI, 1.13–3.40), 2.26 (95% CI, 1.27–4.02), 2.77 (95% CI, 1.58–4.85) and 4.29 (95% CI, 2.22–8.30), respectively (p-trend < 0.001). The area under the receiver operating characteristic of DIC score for prediction of unfavorable long-term outcome was 0.79 (95% CI, 0.69–0.88). Conclusion: Increased initial DIC score in OHCA was an independent predictor for poor outcomes and early mortality risk. © 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Post cardiac arrest syndrome (PCAS) is characterized by its high early mortality.1,2 One of the major causes of this early mortality is the systemic ischemic-reperfusion injury previously described as “sepsis-like syndrome” where global activation of coagulation system is commonly found.3,4 In sepsis and trauma, the coagulation activation and following consumptive coagulopathy are well-known phenomenon and their increasing severity is correlated with poor outcome.5–7 The International Society on Thrombosis and Hemostasis (ISTH) disseminated intravascular coagulation (DIC) score is a simple scoring system measuring the severity of the coagulopathy and is used for the diagnosis of DIC.8 Its validity for the diagnosis of DIC was confirmed in a previous study,6 and has been verified to be sensitive to DIC of both infective
and non-infective etiologies.9 In addition, it is strongly correlated with mortality in many conditions including sepsis, trauma and burn.6,10–13 Although significant coagulopathy has been reported also in cardiac arrest patients,14,15 the prognostic implication of DIC score has not been studied in this condition. Considering the significant role of coagulation process in MODS of sepsis or trauma,16,17 it is possible that DIC score has significant prognostic implication in PCAS where MODS is common. We postulated that initial DIC score in resuscitated out-of-hospital cardiac arrest (OHCA) patients is correlated with early mortality rate and thus cardiac arrest outcomes. The objective of this study is to describe the severity of initial coagulopathy in cardiac arrest patients using ISTH DIC score and evaluate its prognostic implication in PCAS.
2. Materials and methods 夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at http://dx.doi.org/10.1016/j.resuscitation.2012.09.003. ∗ Corresponding author. Tel.: +82 31 787 7579; fax: +82 31 787 4055. E-mail address: [email protected] (K. Kim). 0300-9572/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.resuscitation.2012.09.003
We retrospectively studied a prospective registry of all consecutive OHCA victims brought to a single emergency department (ED) from January, 2008 to December, 2011. The institutional review
J. Kim et al. / Resuscitation 84 (2013) 48–53
board at the study facility approved the analysis and issued a waiver of consent. 2.1. Study setting The study facility was a 950-bed tertiary academic hospital in an urban area with a population of 440,000 where about 8.4% of whole population was aged over 65. This area is mostly served by a two-tiered ambulance service provided by fire service run by government. The level of care provided by emergency medical technicians (EMTs) is comparable to that of EMT intermediate in United States. The management of cardiac arrest was based on the recommendations of the 2005 American heart association cardiopulmonary resuscitation guideline until when 2010 guideline from the same organization was published and adopted. Because of routine blood sampling protocol for OHCA, patients underwent blood sampling instantly on ED arrival.18 All laboratory variables considered in this study were part of the initial routine order set for OHCA and thus were measured concurrently from the initial blood samples. 2.2. Participants and data collection ED OHCA registry extending from January, 2008 to December, 2011 was analyzed by researchers. Target population was all ED OHCA patients who achieved ROSC. Exclusion criteria included age <18 years and documented recent use of warfarin, heparin, low molecular weight heparin and intravenous thrombolytics. Prehospital factors including sex, age, arrest location, presence of witness, bystander CPR, time to basic life support (BLS), initial rhythm, prehospital ROSC and cause of arrest were extracted from the registry and chart review. The detailed definition of prehospital factors and other terms are described in supplementary Table 1 (Supplementary materials are available online at RESUSCITATION journal website). Laboratory variables including white blood cell, hematocrit, and platelet count, urea nitrogen, creatinine, bicarbonate, albumin, total bilirubin, C-reactive protein, troponin I, fibrinogen and D-dimer concentration and PT (seconds) were retrieved from chart review. The ISTH DIC score was calculated from PT (seconds), platelet count, serum fibrinogen level and D-dimer concentration (supplementary Table 2). The cutoff values of D-dimer were based on previous studies.6,7 Outcome variables including survival discharge, 6-month cerebral performance category (CPC) score and exact duration of survival within 7 days post-ROSC were obtained by chart review and structured telephone interview. A 6-month CPC score of 1 or 2 was considered a favorable long-term outcome, whereas that of 3–5 was considered an unfavorable one. 2.3. Statistical analysis Analysis of variance (ANOVA), Kruskal–Wallis, Chi-square or Fisher’s exact test was performed as appropriate for comparison between groups. Pearson’s method was used for correlation analysis if the two variables involved are both interval variables and Spearman’s method was used otherwise. Logistic regression and Cox proportional hazards model analysis were used for univariable analysis. To identify independent prognostic factors, multivariable stepwise logistic regression analyses (backward stepwise method based on likelihood ratio) were performed using the variables with their p-value < 0.1 in univariable analysis and their goodness of fit was assessed with the Hosmer–Lemeshow test. The same backward stepwise method was also used in a multivariable Cox proportional hazard regression analysis. The result of logistic regression and Cox proportional hazard model analysis was presented as odd ratios (ORs) and hazard ratios (HRs), respectively, and their 95% confidence intervals (CIs) were also presented. Survival curves based
49
Table 1 Patient characteristics of study population. Sex, male, no. (%) Age, years, mean (SD) Cardiac arrest in public location, no. (%) Witnessed cardiac arrest, no. (%) Bystander chest compression, no. (%) Time to first BLS, min, median (IQR) Shockable initial rhythm, no. (%) Prehospital ROSC, no. (%) Cause of arrest, no. (%) Cardiac Respiratory Traumatic Food asphyxia Hanging Head and neck injury Multiple blunt or penetrating injury Drowning Intoxication Therapeutic hypothermia, no. (%) Overt DIC (DIC score ≥ 5), no. (%) 1-week survival, no. (%) Survival discharge, no. (%) 6-month CPC 1 or 2, no. (%)
149 (59.1) 64.1 (17.3) 52 (20.6) 205 (81.3) 79 (31.3) 4 (0–11) 69 (27.4) 7 (2.8) 91 (36.1) 111 (44.0) 53 (21.0) 18 (7.1) 14 (5.6) 9 (3.6) 5 (2.0) 3 (1.2) 1 (0.4) 81 (32.1) 82 (32.5) 79 (31.3) 54 (21.4) 28 (11.1)
SD, standard deviation; IQR, interquartile range; BLS, basic life support; ROSC, return of spontaneous circulation; CPC, cerebral performance category.
on 1-week survival data were designed using the Kaplan–Meier method, and comparisons were made using the log-rank test. The discriminatory performance of DIC score for long-term outcome prediction was evaluated using the area under the receiver operating characteristic (ROC) curve. The comparison of independent ROC curves was conducted using the method of Hanley and McNeil.19 We performed a sensitivity analysis excluding cases of non-cardiac etiology to see the prognostic implication of DIC score when the potential influences from underlying conditions capable of changing coagulation profile were minimized. P-values < 0.05 were considered significant. All analyses were performed using SPSS 15.0 (SPSS Inc., IL), STATA 10.1 (StataCorp 113 LP, TX) and MedCalc 10.2 (MedCalc Software, Belgium). 3. Results 3.1. Baseline characteristics A total of 547 OHCA patients were brought to the study ED during the study period and 292 patients achieved ROSC (supplementary Fig. 1). After excluding patients with age <18 years or any documented recent use of warfarin, heparin-based products or intravenous thrombolytics, there were 273 eligible patients. The study population finally included 252 (92.3%) patients whose initial laboratory variables required for calculation of initial DIC score were available. The baseline characteristics of the study population are summarized in Table 1. Eighty two (32.5%) showed overt DIC feature with their DIC score equal to or more than 5. Survival discharge to home or nursing home was possible in 54 (21.4%) patients. Favorable long-term outcome (6-month CPC 1 or 2) was achieved in 28 (11.1%) patients. Patient characteristics of the study population were stratified according to initial DIC score (Table 2). Only three (2.3%) patients whose DIC score was more than 3 (N = 130) achieved favorable longterm outcome. Prehospital factors that significantly differed across the stratified groups included age, presence of witness, prehospital ROSC and presumed cardiac etiology. For laboratory variables, all except WBC differed significantly. Significant correlation was found between DIC score and many laboratory variables including hematocrit, bicarbonate, BUN, albumin, total bilirubin and C-reactive
50
Table 2 Patient characteristics stratified according to DIC score. DIC score 0–2 n = 55 39 (70.9) 55.9 (18.2) 12 (21.8) 38 (69.1) 20 (36.4) 2 (0–7) 20 (37.0) 5 (9.1) 27 (49.1)
39 (58.2) 63.8 (14.8) 21 (31.3) 57 (85.1) 25 (37.3) 6 (1–12) 18 (26.9) 1 (1.5) 29 (43.3)
DIC score 4 n = 48 26 (54.2) 73.3 (13.1) 8 (16.7) 41 (85.4) 18 (37.5) 4 (0–12) 9 (18.8) 0 (0.0) 13 (27.1)
DIC score 5 n = 59 36 (61.0) 66.9 (17.1) 7 (11.9) 47 (79.7) 12 (20.3) 5 (0–12) 17 (29.3) 1 (1.7) 18 (30.5)
222 (175–293) 14 (13–15) 305 (260–391) 2.04 (0.68–3.02)
196 (159–244) 15 (15–17) 380 (278–535) 7.32 (3.14–18.77)
165 (137–248) 17 (17–19) 315 (268–438) 15.48 (4.58–20.00a )
149 (106–218) 21 (19–24) 337 (205–484) 15.04 (7.46–20.00a )
11.5 (10.2–13.3) 42.1 (6.0) 15 (13–19) 1.2 (1.0–1.4) 16.7 (4.0) 4.1 (0.5) 0.7 (0.5–0.9) 0.20 (0.20–0.30) 0.04 (0.02–0.10) 26 (47.3)
12.8 (10.4–15.4) 38.0 (7.6) 19 (15–30) 1.3 (1.0–1.9) 14.6 (4.4) 3.6 (0.7) 0.6 (0.4–0.8) 0.34 (0.20–4.70) 0.05 (0.04–0.15) 28 (41.8)
12.6 (9.1–16.1) 36.1 (8.8) 19 (15–43) 1.3 (0.9–2.7) 14.2 (4.6) 3.3 (0.8) 0.8 (0.5–1.0) 0.80 (0.20–5.26) 0.05 (0.03–0.19) 12 (25.5)
13.8 (9.3–18.2) 36.0 (9.0) 23 (14–45) 1.5 (1.1–2.6) 12.0 (5.3) 2.9 (0.7) 1.0 (0.7–1.7) 3.04 (0.48–11.36) 0.14 (0.06–0.36) 14 (23.7)
35 (63.6) 25 (45.5) 17 (31.5)
23 (34.3) 17 (25.4) 8 (11.9)
10 (21.3) 5 (10.4) 0 (0.0)
10 (16.9) 6 (10.2) 2 (3.4)
DIC score 6–8 n = 23 9 (39.1) 58.3 (19.6) 4 (17.4) 22 (95.7) 4 (17.4) 3 (0–12) 5 (21.7) 0 (0) 4 (17.4) 65 (27–92) 23 (20–31) 135 (50–234) 20.00a (20.00a –20.00a ) 11.2 (6.0–13.7) 33.8 (13.5) 19 (11–33) 1.2 (0.9–1.6) 12.7 (5.3) 2.7 (1.0) 0.7 (0.4–1.6) 0.32 (0.20–2.67) 0.27 (0.04–0.56) 2 (8.7) 1 (4.3) 1 (4.3) 1 (4.3)
p
0.109 <0.001 0.087 0.044 0.092 0.380 0.310 0.029 0.021 <0.001 <0.001 <0.001 <0.001 0.127 <0.001 0.002 0.042 <0.001 <0.001 <0.001 <0.001 <0.001 0.002 <0.001 <0.001 <0.001
SD, standard deviation; IQR, interquartile range; BLS, basic life support; ROSC, return of spontaneous circulation, PT, prothrombin time; WBC, white blood cell; BUN, blood urea nitrogen; CRP, C-reactive protein; CPC, cerebral perfomance category. a D-dimer level higher than 20 ug/mL was reported as “>20 ug/mL” in study hospital.
J. Kim et al. / Resuscitation 84 (2013) 48–53
Prehospital and demographic variables Sex, male, no. (%) Age, years, mean (SD) Cardiac arrest in public location, no. (%) Witnessed cardiac arrest, no. (%) Bystander chest compression, no. (%) Time to first BLS, min, median (IQR) Shockable initial rhythm, no. (%) Prehospital ROSC, no. (%) Presumed cardiac etiology, no. (%) DIC score components Platelet, 109 /L, median (IQR) PT, seconds, median (IQR) Fibrinogen, mg/dL, median (IQR) D-dimer, ug/mL, median (IQR) Other laboratory variables for adjustment WBC count, 109 /L, median (IQR) Hematocrit, %, mean (SD) BUN, mg/dL, median (IQR) Creatinine, mg/dL, median (IQR) Bicarbonate, mEq/L, mean (SD) Albumin, g/dL, mean (SD) Bilirubin total, mg/dL, median (IQR) CRP, mg/dL, median (IQR) Troponin I, ng/mL (IQR) Therapeutic hypothermia, no. (%) Cardiac arrest outcomes, no. (%) 1-week survival, no. (%) Survival discharge, no. (%) 6-month CPC 1 or 2, no (%)
DIC score 3 n = 67
J. Kim et al. / Resuscitation 84 (2013) 48–53 Table 3 Stepwise multivariatelogistic regression model analysis of the relationship between poor outcome and various potential prognostic factors. Odds ratio Inhospital death Sex (male) Age (per 1 year) Time to first BLS (per 1 min) Shockable initial rhythm Bicarbonate (per 1 mEq/L) Therapeutic hypothermia DIC score (per 1 unit) Poor long-term outcome (6-month CPC ≥ 3) Age (per 1 year) Time to first BLS (per 1 min) Shockable initial rhythm Presumed cardiac etiology DIC score (per 1 unit)
Table 4 Stepwise multivariateCox proportional hazard model analysis ofthe associations between early mortality risk and various potential prognostic factors.
p
0.47 (0.21–1.06) 1.03 (1.00–1.06) 1.06 (1.00–1.13) 0.20 (0.09–0.44) 0.91 (0.83–1.00) 0.39 (0.18–0.85) 1.61 (1.17–2.22)
0.068 0.035 0.048 <0.001 0.045 0.018 0.003
1.03 (1.00–1.07) 1.16 (1.03–1.31) 0.14 (0.05–0.40) 0.31 (0.11–0.92) 1.84 (1.26–2.67)
0.060 0.017 <0.001 0.034 0.002
BLS, basic life support; CPC, cerebral performance category. Note: The prediction model for inhospital death and poor long-term outcome correctly classified in 83.5% and 92.6% of cases, respectively.
protein. Among them, albumin (Pearson’s r = −0.550, p < 0.001), bicarbonate (Pearson’s r = −0.310, p < 0.001) and hematocrit (Pearson’s r = −0.300, p < 0.001) were most strongly correlated with DIC score. 3.2. Association between DIC score and overall CPR outcome and prognostic performance of DIC score Increased DIC score was a strong risk factor for both inhospital death and unfavorable long-term outcome (6-month CPC 3–5). In a univariable logistic regression analysis (supplementary Table 3), one-unit increase of DIC score was associated with 1.89 (95% CI, 1.48–2.41)- and 2.21 (95% CI, 1.60–3.05)-fold increase of risk for inhospital death and unfavorable long-term outcome, respectively. Backward stepwise multivariable logistic regression models were constructed using the variables with their p < 0.1 in the univariable analyses (Table 3). After adjustment, DIC score remained a significant predictor for inhospital death and unfavorable longterm outcome, with 1.61 (95% CI, 1.17–2.22)- and 1.84 (95% CI, 1.26–2.67)-fold increased risk per unit, respectively. Other independent predictors for inhospital mortality were age, time to BLS, initial rhythm, serum bicarbonate and application of hypothermia. For unfavorable long-term outcome, time to BLS, initial rhythm and presumed cardiac etiology were independent predictors. The long-term prognostic performance of DIC score was assessed by analysis of ROC curve. The area under the ROC curve (AUC) of DIC score for the prediction of unfavorable long-term outcome was 0.79 (95% CI, 0.69–0.88). The optimal cutoff point of DIC score to predict the poor outcome was found between 3 and 4. Using the cutoff point, the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of DIC score were 57.0 (95% CI, 50.2–63.5), 89.3 (95% CI, 71.8–97.6), 97.7 (95% CI, 93.4–99.5) and 20.7 (95% CI, 13.8–29.0), respectively.
51
Age (per 1 year) Shockable initial rhythm CRP (per 1 mg/dL) Therapeutic hypothermia DIC scoregroups (p-trend < 0.001) 1st group (DIC score 0–2, n = 55) 2nd group (DIC score 3, n = 67) 3rd group (DIC score 4, n = 48) 4th group (DIC score 5, n = 59) 5th group (DIC score 6–8, n = 23)
Hazard ratio
p
1.02 (1.01–1.03) 0.49 (0.33–0.73) 1.02 (1.00–1.04) 0.38 (0.26–0.55)
0.003 <0.001 0.047 <0.001
1.00 (Reference) 1.96 (1.13–3.40) 2.26 (1.27–4.02) 2.77 (1.58–4.85) 4.29 (2.22–8.30)
0.017 0.006 <0.001 <0.001
CRP, C-reactive protein.
group, respectively, compared with those in the first group (DIC score < 3). In a multivariable Cox proportional hazard model analysis using stepwise backward method (Table 4), the risk gradient was still significant (p-trend < 0.001). The adjusted hazard ratio for death during the first week after ROSC was 2.77 (95% CI, 1.58–4.85)and 4.29 (95% CI, 2.22–8.30)-fold higher in the fourth and fifth group, respectively, compared with those in the first group. Other independent predictors in the model were age, initial rhythm, Creactive protein and therapeutic hypothermia. Kaplan–Meier survival curves for the five groups during the first week were constructed (Fig. 1). The survival curves were statistically different according to the log-rank test (p = < 0.001, ptrend < 0.001). 3.4. Subgroup analysis of cardiac arrest cases of presumed cardiac origin We performed a sensitivity analysis excluding non-cardiac etiology cases to see the prognostic implication of DIC score when the potential influences from underlying conditions capable of changing coagulation profile were minimized. The magnitude of increased risk represented by ORs and HRs in cardiac etiology subgroup was not lower compared with that of whole study population (Supplementary Table 5). Fig. 2A shows the percentage of patients who achieved 1-week survival, survival discharge and favorable long-term outcome in each DIC score subgroup. Patients with cardiac etiology usually had better chance of achieving survival discharge (p = 0.002) and favorable long-term outcome (p < 0.001), especially if their DIC score was less than five (p = 0.006, p < 0.001,
3.3. Association between DIC score and mortality risk during the first week after ROSC To assess the association between DIC score and early mortality risk, we categorized DIC score in five groups as <3, 3, 4, 5 and >5 and analyzed their differential survival rate during the first week after ROSC. In univariable analysis using Cox proportional hazard model analysis (supplementary Table 4), there was significant gradient of increasing hazard ratio across the five consecutive groups (ptrend < 0.001). The hazard ratio for death during the first week after ROSC was 3.95 (95% CI, 2.34–6.68)- and 6.13 (95% CI, 3.30–11.38)fold higher in the fourth (DIC score 5) and fifth (DIC score > 5)
Fig. 1. Kaplan–Meier plots for cumulative 1-week survival. In fifth DIC score group (DIC score > 5, n = 23), only one patient survived more than 48 h. The difference was significant according to the log-rank test (p < 0.001, p-trend < 0.001).
52
J. Kim et al. / Resuscitation 84 (2013) 48–53
Fig. 2. (A) CPR outcome differences across DIC score subgroups of non-cardiac and cardiac etiology. (B) Comparison of receiving operating characteristic curves of DIC score for predicting poor long-term outcome (6-month CPC 3–5) in non-cardiac origin (dash dot line) and cardiac origin (dashed line) subgroups. The area under the curves were 0.64 (95% CI, 0.41–0.86; p < 0.173) and 0.84 (95% CI, 0.76–0.93; p < 0.001) respectively. However, their difference was not statistically significant (p = 0.095).
respectively). And no one regained consciousness (CPC 4, 5) in 3rd, 4th and 5th group (DIC score > 3, N = 35) if the cause of arrest was presumed to be cardiac-origin. Fig. 2B depicts the ROC curves of DIC score for prediction of unfavorable long-term outcome (6-month CPC 3–5) in whole study population, non-cardiac etiology subgroup and cardiac etiology subgroup. The AUC of DIC score for the prediction of unfavorable long-term outcome was 0.84 (95% CI, 0.76–0.93) in cardiac-origin subgroup while it was 0.64 (95% CI, 0.41–0.86) in non-cardiac origin subgroup. Although the AUC was most large in cardiac-subgroup, the difference between non-cardiac and cardiac etiology subgroup was not statistically significant (p = 0.095). The optimal cutoff point of DIC score for prediction of unfavorable long-term outcome in cardiac-subgroup was identical (between DIC score 3 and 4) as that of whole study population. Using the cutoff point, the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of DIC score were 48.6 (95% CI, 36.7–60.7), 100.0 (95% CI, 82.2–100.0), 100.0 (95% CI, 89.9–100.0) and 33.9 (95% CI, 21.8–47.8), respectively. 4. Discussion In this study, the increased initial DIC score was an independent risk factor for inhospital death and unfavorable long-term outcome. It was also associated with markedly increased rate of death during early post-resuscitation period. These relationships remained significant in cardiac-cause subgroup where the influence from underlying conditions causing coagulopathy is minimized. The underlying mechanism of this observation would be multifactorial. There is high possibility that patients with higher DIC score were more frequently involved with underlying or precipitating conditions that could had already caused varying degree of coagulopathy before cardiac arrest. Although the frequency of such conditions would be much lower in cardiac-cause subgroup, it is not possible to completely eradicate such confounding effect in a retrospective study design targeting cardiac arrest patients. However, the global coagulation activation of cardiac arrest per se could have significant prognostic implication.15 Previous studies suggested that consumptive coagulopathy plays central role in pathogenesis of MODS in critical conditions such as sepsis or trauma.20 As similar global activation of coagulation pathway is also observed in cardiac arrest patients,14 its role in disease progression of PCAS might be significant. In this study, we found a significant correlation between DIC score and early mortality rate. Considering that MODS including cardiac dysfunction is an
important cause of death during early post-resuscitation period, it might be prudent to infer that this early coagulopathy might play a significant role in the development of MODS and subsequent increased mortality. However the lack of data regarding chronological changes in hemodynamics and organ function did not allow us to prove whether such manifestations were actually differed significantly. Generalized activation of coagulation system is an active ongoing process commonly observed during cardiac arrest.14,21 In our study, blood sampling was done immediately after patients’ ED arrival and thus the laboratory values we analyzed represent coagulation status before or immediately after ROSC except for the patients who achieved prehospital ROSC (N = 7, 2.8%). Even for the patients with prehospital ROSC, the median time delay from ROSC to blood sampling was very short (median 2 minutes; IQR, 2–6). Despite this early timing, 82 (32.5%) of whole study population and 22 (24.2%) of cardiac-cause subpopulation showed high DIC score indicative of fulminant DIC (DIC score 5 or more). However, the pathophysiologic consequences of this early change are not yet fully understood in cardiac arrest. Whether this is just an indicator of ischemic injury from cardiac arrest or an active triggering mechanism of following organ dysfunction or both should be the subject of a future study. This study has several limitations. First, the result of this study is based upon retrospective analysis of ED OHCA registry and medical records and thus is limited by possible biases that are intrinsic in such design. Second, although we tried to reduce the number of variables as many as possible by using only significant variables in univariable analysis and adopting stepwise backward logistic regression, the number of events per variables ratio of the multivariable logistic regression models were lower than the desired ratio of >10 suggested by Peduzzi et al.22 The variables considered in the analyses were carefully chosen and we thought that their biological plausibility of influence on both coagulation activation and clinical outcome cannot be easily ignored. As the main purpose of this study was to look into the prognostic significance of DIC score rather than making a good multivariable model to predict ultimate outcome of OHCA patient, we chose not to further remove clinically relevant variables from the beginning. Third, we did not perform multivariable model analysis in cardiac subgroup as its population size was too small for multivariable analysis using large set of prognostic variables we used. Fourth, chronological evaluation of hemodynamic indices, inflammatory markers and functional status of major organs was not possible. Availability of such data could provide more direct evidence of the association
J. Kim et al. / Resuscitation 84 (2013) 48–53
between DIC score and early pathological progressions including myocardial dysfunction, inflammation and MODS. 5. Conclusions Increased initial DIC score in cardiac arrest was an independent risk factor for both inhospital death and unfavorable long-term outcome. The higher mortality risk during early post-resuscitation period could be one possible explanation behind it. Conflicts of interest statement All authors do not have any conflicts of interest. Source of support This study was supported by grant from the Seoul National University Bundang Hospital Research Fund, No. 02-2012-011. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.resuscitation.2012.09.003. References 1. Neumar RW, Nolan JP, Adrie C, et al. Post-cardiac arrest syndrome: epidemiology, pathophysiology, treatment, and prognostication. A consensus statement from the International Liaison Committee on Resuscitation (American Heart Association, Australian and New Zealand Council on Resuscitation, European Resuscitation Council, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Council of Asia, and the Resuscitation Council of Southern Africa); the American Heart Association Emergency Cardiovascular Care Committee; the Council on Cardiovascular Surgery and Anesthesia; the Council on Cardiopulmonary, Perioperative, and Critical Care; the Council on Clinical Cardiology; and the Stroke Council. Circulation 2008;118:2452–83. 2. Herlitz J, Engdahl J, Svensson L, Angquist K-A, Silfverstolpe J, Holmberg S. Major differences in 1-month survival between hospitals in Sweden among initial survivors of out-of-hospital cardiac arrest. Resuscitation 2006;70:404–9. 3. Adrie C. Successful cardiopulmonary resuscitation after cardiac arrest as a “sepsis-like” syndrome. Circulation 2002;106:562–8.
53
4. Adrie C, Laurent I, Monchi M, Cariou A, Dhainaou J-F, Spaulding C. Postresuscitation disease after cardiac arrest: a sepsis-like syndrome? Curr Opin Crit Care 2004;10:208–12. 5. Levi M. Disseminated intravascular coagulation. Crit Care Med 2007;35:2191–5. 6. Bakhtiari K, Meijers JCM, de Jonge E, Levi M. Prospective validation of the International Society of Thrombosis and Haemostasis scoring system for disseminated intravascular coagulation. Crit Care Med 2004;32:2416–21. 7. Dhainaut JF, Yan S, Joyce D, et al. Treatment effects of drotrecogin alfa (activated) in patients with severe sepsis with or without overt disseminated intravascular coagulation1. J Thromb Haemost 2004;2:1924–33. 8. Taylor FB, Toh CH, Hoots WK, Wada H, Levi M, Scientific Subcommittee on Disseminated Intravascular Coagulation (DIC) of the International Society on Thrombosis and Haemostasis (ISTH). Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost 2001;86:1327–30. 9. Levi M, Toh CH, Thachil J, Watson HG. Guidelines for the diagnosis and management of disseminated intravascular coagulation. British Committee for Standards in Haematology. Br J Haematol 2009;145:24–33. 10. Gando S, Wada H, Asakura H, et al. Evaluation of new Japanese diagnostic criteria for disseminated intravascular coagulation in critically ill patients. Clin Appl Thromb Hemost 2005;11:71–6. 11. Sivula M, Tallgren M, Pettilä V. Modified score for disseminated intravascular coagulation in the critically ill. Intensive Care Med 2005;31:1209–14. 12. Cauchie P, Cauchie C, Boudjeltia KZ, et al. Diagnosis and prognosis of overt disseminated intravascular coagulation in a general hospital – meaning of the ISTH score system, fibrin monomers, and lipoprotein–C-reactive protein complex formation. Am J Hematol 2006;81:414–9. 13. Lavrentieva A, Kontakiotis T, Bitzani M, et al. Early coagulation disorders after severe burn injury: impact on mortality. Intensive Care Med 2008;34:700–6. 14. Böttiger B, Motsch J, Böhrer H, Böker T. Activation of blood coagulation after cardiac arrest is not balanced adequately by activation of endogenous fibrinolysis. Circulation 1995;92:2572–8. 15. Adrie C, Monchi M, Laurent I, et al. Coagulopathy after successful cardiopulmonary resuscitation following cardiac arrest. J Am Coll Cardiol 2005;46:21–8. 16. Levi M, van der Poll T. Inflammation and coagulation. Crit Care Med 2010;38:S26–34. 17. van der Poll T, de Boer JD, Levi M. The effect of inflammation on coagulation and vice versa. Curr Opin Infect Dis 2011;24:273–8. 18. Kim J, Kim K, Lee JH, et al. Red blood cell distribution width as an independent predictor of all-cause mortality in out of hospital cardiac arrest. Resuscitation 2012, doi:10.1016/j.resuscitation.2012.01.038. 19. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29–36. 20. Gando S. Microvascular thrombosis and multiple organ dysfunction syndrome. Crit Care Med 2010;38:S35–42. 21. 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. 22. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996;49:1373–9.