If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Medicine I, Clinical Division of Hematology and Hemostaseology, Medical University of Vienna, Waehringer Guertel 18-20, 1090 Vienna, Austria
D-dimer levels but not the DIC score are useful for early prognostication.
Abstract
Background
Recent Korean data suggest a high prevalence of overt disseminated intravascular coagulation (DIC) and a good predictive performance of the ISTH DIC score in successfully resuscitated out-of-hospital cardiac arrest.
Objectives
We hypothesised that in a European cohort of resuscitated out-of-hospital cardiac arrest patients the prevalence of DIC is substantially lower. Furthermore, the determination of D-dimer levels at admission, but not the DIC score, could improve mortality prediction above traditional predictors.
Patients/methods
Data were extracted from a prospective cardiac arrest registry including patients admitted between 2006 and 2015, who achieved return of spontaneous circulation and had parameters for DIC score calculation available. The primary outcome was the prevalence of overt DIC at admission. Secondary outcomes included the association of overt DIC with 30-day mortality and the contribution of the DIC score and D-dimer levels to 30-day mortality prediction using logistic regression. Three stepwise models were evaluated by receiver-operating-characteristic analysis.
Results
Out of 1179 patients 388 were included in the study. Overt DIC was present in 8% of patients and associated with substantial 30-day mortality (83% vs. 39%). The AUC for model 1, including traditional mortality predictors, was 0.83. The inclusion of D-dimer levels significantly improved prognostication above traditional predictors (model 3, AUC 0.89), whereas the inclusion of the DIC Score had no effect on mortality prediction (model 2, AUC 0.83).
Conclusion
Overt DIC was rare in a European cohort of out-of-hospital cardiac arrest patients. D-dimer levels improved 30-day mortality prediction and provided added value to assess early mortality risk after successful resuscitation.
Out-of-hospital cardiac arrest constitutes a major health problem in Western countries. It has a global incidence of 55 per 100.000 person-years as well as an overall poor survival rate ranging from 3% to 31% in Europe [
]. As early outcome prediction is crucial for both treating physicians and relatives, much effort has been put into the identification of variables providing early and accurate prognostication after cardiac arrest [
Predicting survival with good neurological recovery at hospital admission after successful resuscitation of out-of-hospital cardiac arrest: the OHCA score.
The presence of disseminated intravascular coagulation (DIC), as assessed on an 8-point scale accounting points for prothrombin time, platelet count, fibrinogen and D-dimer levels [
The scoring system of the scientific and standardisation committee on disseminated intravascular coagulation of the international society on thrombosis and Haemostasis: a 5-year overview.
In this context, a recent Korean study of patients with out-of-hospital cardiac arrest found a 33% prevalence of overt DIC, which independently and accurately predicted mortality risk [
]. While these data might apply to Korean or even Asian patients in general, no comparable data are available for a non-Asian population.
This study investigated the prevalence of overt DIC, its association with 30-day mortality, and assessed the added prognostic value of the DIC score vs. D-dimer levels above that of traditional mortality predictors in out-of-hospital cardiac arrest in a European cohort of successfully resuscitated patients.
2. Methods
All patients with out-of-hospital cardiac arrest admitted to the Emergency Department - Intensive Care Unit at the Medical University of Vienna from 2006 to 2015 with on-admission coagulation parameters facilitating DIC score calculation available, were extracted from an ongoing prospective observational cardiac arrest registry. Double-data entry into the registry data file was performed by two independent study fellows with range and consistency checks prior to analysis. In case of differences between double-data entries or failed range and consistency checks, original source data were re-analysed by an investigator not involved in primary data entry. Registry-study related blood sampling was performed as soon as vascular access was available upon arrival.
We only included patients who achieved sustained return of spontaneous circulation and received targeted temperature management (32–34 °C). Exclusion criteria comprised conditions known to interfere with DIC score calculation including advanced liver cirrhosis, oral anticoagulation, thrombolysis therapy, thrombotic microangiopathies, bone marrow pathologies and hereditary coagulopathies (supplemental figure).
The primary outcome was the prevalence of overt DIC at hospital admission. Secondary outcomes were the association of overt DIC with 30-day mortality, as well as the predictive value of the DIC score and of D-dimer levels at admission above that of traditional mortality predictors.
The ISTH DIC 8-point score was applied for DIC calculation [
New disseminated intravascular coagulation score: a useful tool to predict mortality in comparison with acute physiology and chronic health evaluation II and logistic organ dysfunction scores.
](Table 1). A score ≥5 is compatible with overt DIC. The ISTH DIC score includes D-dimer (μg/ml; normal range < 0.5 μg/ml fibrinogen equivalent units), fibrinogen (mg/dl; normal range 200-400 mg/dl), prothrombin time (%; normal range 75–140%) and platelet count (G/l, normal range 150–350G/l). D-dimer levels were quantitatively measured by latex agglutination with the STA-LIA D-Di plus kit with a measuring inaccuracy of <10% [
]. A value below 0.5 μg/ml was used as the upper limit of normal. Results are given as μg/ml fibrinogen equivalent units (normal range < 0.5 μg/ml). For fibrinogen determination the Clauss method was used [
Sustained return of spontaneous circulation was defined as recovery of spontaneous circulation for longer than 20 min. No- and low-flow intervals were defined as the time from collapse to initiation of cardiopulmonary resuscitation and the duration of cardiopulmonary resuscitation respectively. Resuscitation-related parameters were analysed and reported according to Utstein recommendations [
Cardiac arrest and cardiopulmonary resuscitation outcome reports: update and simplification of the Utstein templates for resuscitation registries. A statement for healthcare professionals from a task force of the international liaison committee on resuscitation.
63. (American Heart Association European Resuscitation Council, Australian Resuscitation Council, New Zealand Resuscitation Council, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Council of Southern Africa). Resuscitation,
2004: 233-249
]. Pre-hospital variables are documented by trained Study Nurses and Emergency Department Physicians in special Case Report Forms.
The cause of death was reported as shock (with or without subsequent multi organ failure), when patients required continuous catecholamine support >6 h after the return of spontaneous circulation. Death related to neurological injury was defined as the decision taken by the treating physicians to withdraw life-sustaining treatment due to hypoxic brain damage or manifest intracranial bleeding or the decision to release for organ donation. Neurological prognosis was individually assessed by CT-scans, neuron specific enolase and S-100 levels, as well as neurological examination [
This study was approved by the Ethics Committee of the Medical University of Vienna and carried out in compliance with the Helsinki Declaration.
2.1 Statistical analysis
The study patients were analysed according to 30-day mortality (end of follow-up) and stratified according to survival. The cohort of patients with overt DIC was analysed separately in a second step. Categorical data are presented as numbers (n) and frequencies (%). Medians and inter quartile ranges (IQR) are provided for continuous variables. The primary outcome, the prevalence of overt DIC, is given as a proportion with a 95% confidence interval. The Kaplan-Meier method was used to test the effect of overt DIC on the probability of death at day 30, and the log rank test was used for between-group comparisons of cumulative event rates.
We performed multiple logistic regression across all patients to estimate the effect on 30-day mortality of several candidate predictors, which were judged to be clinically plausible. These included the continuous variables age, no-flow interval, low-flow interval, lactate and creatinine levels and the binary variables sex, initial rhythm (non-shockable vs. shockable), witnessed status bystander cardiopulmonary resuscitation and anaemia (model 1) plus the DIC score (binary; model 2) or plus D-dimer levels (continuous; model 3). No imputation for missing data was performed. Collinearity of independent variables was tested through calculation of variance inflation factors before running the multivariable regression analysis (supplemental table 1). We used a backwards stepwise elimination approach removing all independent variables without a strong association with the outcome variable according to Wald test statistic step-by-step. The performance of the three models was assessed through receiver operating characteristic analysis. Results are given as area under the curve (AUC) with a 95% CI and p-value. Areas under the curve were compared using the method by DeLong et al. [
]. A two-sided p-value < 0.05 was considered statistically significant. The statistical analysis was performed using IBM SPSS version 22.0. Figures were drawn with GraphPad Prism version 7.00.
] on European cardiac arrest patients suggests a substantially lower rate of overt DIC and thus a limited applicability of the ISTH DIC score to predict mortality after successful resuscitation. We hypothesised, that, in contrast, D-dimer levels at admission might be useful to improve prognostication.
D-dimer is not only a component of the DIC score but importantly increases dramatically with hypoxia [
] and represents the extent of hyperfibrinolysis occurring during hypoperfusion in cardiac arrest. The time from cardiac arrest to start of resuscitation efforts (no-flow time) is a main determinant of prognosis [
]. Its reliability, however, is often limited because it is commonly estimated or unknown. D-dimer levels on admission may provide a more objective measure for the extent of hypoperfusion and thus may increase the performance of a prediction model above that of traditional risk factors, such as initial rhythm, no- and low-flow time and lactate levels [
Association of initial rhythm with neurologically favorable survival in non-shockable out-of-hospital cardiac arrest without a bystander witness or bystander cardiopulmonary resuscitation.
]. The prognostic value of D-dimer levels might, therefore, not be limited to a small number of cardiac arrest patients like the DIC score, but might be applicable to all successfully resuscitated patients.
3. Results
Out of 1179 patients with out-of-hospital cardiac arrest 388 patients (72% male; 60 years) could be included in the study. 66% of cardiac arrest events (n = 254) were of cardiac origin. Thirty-day mortality was 43% (n = 166). 33% died as a result of shock (n = 54). In 67% of non-survivors a decision to withdraw life-sustaining treatment (n = 108) or release for organ donation was made (n = 4).
Overt DIC was present in 8% (95%CI 5–11%) of patients (n = 30). Non-survivors (43%, n = 166) were older (64 vs. 57 years), had poorer resuscitation characteristics, lower hemoglobin levels (12,4 vs. 13,3 g/dl), higher lactate (10 vs. 7 mmol/l) and D-dimer levels (15 vs. 2 μg/ml) and a higher rate of overt DIC (15 vs. 2%). Platelet counts, prothrombin time and fibrinogen levels were similar between survivors and non-survivors (Table 2).
Table 2Baseline characteristics.
Demographic data
Total n = 388
Survivors n = 222
Non-survivors n = 166
Male sex
278 (72)
170 (77)
108 (65)
Age, years
60 (48–70)
57 (48–67)
64 (51–72)
Medical history
Hypertension
147 (38)
79 (36)
68 (41)
Coronary artery disease
84 (22)
46 (21)
38 (23)
Chronic heart failure
53 (14)
33 (15)
20 (12)
Nicotine abuse
123(32)
71 (32)
52 (31)
COPD
58 (15)
25 (11)
33 (20)
Diabetes mellitus
64 (17)
35 (16)
29 (18)
Resuscitation characteristics
Cardiac arrest at home
166 (43)
80 (36)
86 (52)
Witnessed
320 (83)
194 (87)
126 (76)
Bystander CPR
155 (40)
104 (47)
51 (31)
Initial shockable rhythm
221 (57)
156 (73)
65 (39)
No-flow interval, min
2 (0–6)
1.5 (0–4)
2 (0–9)
Low-flow interval, min
17 (10−30)
15 (8–22)
23 (15–40)
Number of shocks
3 (2–5)
3 (2–4)
3 (1–7)
Cumulative dose of Epinephrine, mg
3 (1–4)
2 (1–4)
3 (2–6)
Cardiac etiology of cardiac arrest
254 (66)
161 (73)
93 (56)
Pre-hospital sROSC
335 (86)
200 (90)
135 (81)
Laboratory values at admission
pH
7.2 (7.0–7.3)
7.2 (7.1–7.3)
7.1 (7.0–7.2)
Lactate, mmol/l
8 (6–11)
7 (4–9)
10 (8–14)
Creatinine, mg/dl
1.15 (0.9–1.6)
1.0 (0.8–1.3)
1.30 (1.0–1.7)
Troponin T, ng/ml
0.15 (0.05–0.78)
0.17 (0.06–0.83)
0.12 (0.04–0.62)
D-dimer, μg/ml
4 (2–14)
2 (1–5)
15 (4–40)
Platelets, G/l
181 (133–241)
187 (141–245)
176 (118–234)
Fibrinogen, mg/dl
349 (271–454)
357 (285–453)
345 (240–454)
Prothrombin time, %
79 (61–93)
80 (63–91)
77 (57–98)
Hemoglobin, g/dl
12.9 (11.3–14.3)
13.3 (12.2–14.5)
12.4 (10.7–13.8)
Anaemia
166 (43)
89 (40)
77 (46)
Overt DIC, n (%; 95%CI)
30 (8; 5–11)
5 (2; 1–5)
25 (15; 10–21)
Cause of death
Shock
54 (14)
N/A
54 (33)
Withdrawal of life-sustaining treatment
108 (28)
N/A
108 (65)
Release for organ transplantation
4 (1)
N/A
4 (2)
Data are presented as n (%) and median (IQR). N/A, not applicable.
The presence of overt DIC was associated with a two-fold rise in 30-day mortality (83% vs. 39%, p < 0.001, Fig. 1). Patients with overt DIC less frequently had a cardiac cause of cardiac arrest (40 vs. 68%, p = 0.002), had a higher rate of non-shockable rhythm (70 vs. 39%), longer no-flow (5 vs. 2 min, p = 0.003) and low-flow times (23 vs. 17 min, p = 0.04) and higher lactate levels (13 vs. 8 mmol/l, p < 0.001). Sex (57 vs. 73% male, p = 0.06), age (65 vs. 59 years, p = 0.31), rate of bystander cardiopulmonary resuscitation (30 vs. 41%, p = 0.25) and witnessed events (83 vs. 82%, p = 0.90) were similar between patients with and without overt DIC.
Fig. 1Kaplan-Meier plot of estimated mortality to day 30 according to the presence of overt disseminated intravascular coagulation at admission.
Fig. 2 shows the incremental value of overt DIC (model 2) and of D-dimer levels (model 3) above that of traditional risk factors (model 1) for mortality prediction by comparison of ROC curves derived from these models. Unadjusted and adjusted associations between each covariable and 30-day mortality are available with the supplement. The AUC for model 1 was 0.83 (95%CI 0.79–0.88). Inclusion of the DIC score had no added predictive effect (AUC model 2: 0.83; 95%CI 0.79–0.88). Inclusion of D-dimer levels resulted in a significantly increased AUC (AUC model 3: 0.89; 95%CI, 0.86–0.93; p < 0.001).
Fig. 2Receiver Operating Characteristic (ROC) curves of three models to predict 30-day mortality.
Model 1 (dotted line): Inclusion of traditional mortality predictors only (age, sex, initial rhythm, witnessed status, bystander cardiopulmonary resuscitation, no-flow interval, low-flow interval, creatinine, lactate levels and anaemia). 5 variables remained significant (age, initial rhythm, no-flow interval, low-flow interval and lactate levels).
Model 2 (dotted line): Inclusion of traditional mortality predictors plus DIC score. The 5 variables which remained significant were identical to that of Model 1.
Model 3 (solid line): Inclusion of traditional mortality predictors plus D-dimer levels. Variables that remained significant were age, initial rhythm, no-flow interval, low-flow interval and D-dimer levels.
The inclusion of D-dimer into the model increased the AUC (Model 1 vs. Model 3; p < 0.001).
This study assessed the prevalence of overt DIC, its association with 30-day mortality, and its added prognostic value vs. that of D-dimer levels at hospital admission in a selected European cohort of out-of-hospital cardiac arrest patients. Overall, overt DIC was rare. Even though overt DIC was associated with a two-fold rise in mortality, its inclusion in a model did not affect predictability. D-dimer levels, in contrast, improved mortality prediction.
Our study was driven by the hypothesis that elevated D-dimer levels in out-of-hospital cardiac arrest do not generally indicate the presence of overt DIC but reflect the presence of hyperfibrinolysis following hypoxia and hypoperfusion [
]. Accordingly, non-survivors had significantly higher D-dimer levels compared to survivors, but similar platelet counts, fibrinogen levels and prothrombin time.
The low 8% rate of overt DIC found in the current study supports our previous preliminary estimates (95% CI, 1–9) [
]. Compared with our cohort, patients included in the study by Kim et al. more rarely had a shockable initial rhythm (27% vs. 57%), had a lower rate of bystander cardiopulmonary resuscitation (31% vs. 40%) and achieved pre-hospital sustained return of spontaneous circulation in <3% of cases (vs. 86%). Targeted temperature management was also only received by 32% (vs. 100%). The high prevalence of overt DIC observed by Kim et al. and the associated unfavourable outcome is likely based on the poor baseline characteristics of the Korean study cohort. This makes a direct comparison of both study populations difficult. However, we applied the same selection criteria on our cardiac arrest registry as Kim and colleagues did and included only out-of-hospital cardiac arrest patients who achieved sustained return of spontaneous circulation and had coagulation markers available. This raises the question as to whether these marked differences between European and Asian out-of-hospital cardiac arrest patients may result from cross-national differences in resuscitation policies including pre-hospital termination-of-resuscitation rules.
Although the rate of overt DIC was higher in non-survivors, platelet counts, fibrinogen levels and prothrombin time did not differ significantly between survivors and non-survivors. Nonetheless, it is unclear whether cardiac arrest per se causes DIC. Disseminated intravascular coagulation is often a tissue factor triggered condition. However, no data are available on the extent of tissue factor expression in cardiac arrest or on autoptic examinations confirming disseminated fibrin deposits in cardiac arrest non-survivors. It is conceivable that hypoxia related endothelial damage with ensuing blood exposition to interstitial tissue factor might provoke systemic thrombin release leading to uncontrolled coagulation activation [
], multiple organ failure. In this context, Adrie and colleagues found a significant association between thrombin levels and the extent of organ dysfunction in successfully resuscitated cardiac arrest patients [
]. However, while these mechanisms of coagulopathy might occur within hours or days after successful resuscitation (as part of the post cardiac arrest syndrome), data from drowning-related cardiac arrest patients, obtained immediately after admission to the hospital, suggest that primary hyperfibrinolysis rather than DIC is the initial pattern of coagulopathy in many cases of prolonged cardiac arrest [
]. Hyperfibrinolysis by hypoxia- and hypoperfusion-related tissue-type plasminogen activator release is independent of widespread or even localised thrombosis. This mechanism could explain the relatively rare prevalence of overt DIC along with a high number of patients with elevated D-dimer levels, as found in our study, because on-admission D-dimer indicates fibrinolysis not overt DIC early after cardiac arrest.
This hypothesis is supported by previous studies reporting the highest fibrinolytic activity in cardiac arrest patients with early death [
Predicting survival with good neurological recovery at hospital admission after successful resuscitation of out-of-hospital cardiac arrest: the OHCA score.
Impact of time to return of spontaneous circulation on neuroprotective effect of targeted temperature management at 33 or 36 degrees in comatose survivors of out-of hospital cardiac arrest.
Target temperature management of 33 degrees C and 36 degrees C in patients with out-of-hospital cardiac arrest with initial non-shockable rhythm - a TTM sub-study.
]. D-dimer may add additional prognostic value by partly correcting for the imprecision inherent to reported no- and low-flow intervals. The inclusion of admission d-dimer levels in our model accordingly improved mortality prediction above established risk factors. In contrast, the DIC score, owing to its low prevalence, had no additional predictive value in this European cohort of cardiac arrest patients.
5. Practical implications
D-dimer levels are a known unspecific marker of disease severity and mortality in critical illness [
Our data suggest that D-dimer levels may not only be a global marker for poor outcome in out-of-hospital cardiac arrest, but are also independently associated with short-term mortality after successful resuscitation, in addition to established risk factors. Our findings might be of interest for healthcare professionals who are first to respond to resuscitated patients upon hospital admission, because D-dimer levels may only be prognostically valuable if determined immediately after admission. Due to their short half-life, D-dimer levels likely rapidly decrease after return of spontaneous circulation (and subsequent termination of hyperfibrinolysis) and lose their prognostic value. As early best possible prognostication is critical for both treating physicians and relatives, prospective investigation of the predictive performance of D-dimer levels may be warranted. D-dimer levels suggesting hyperfibrinolysis may not only be used as part of prognostication and resource allocation after cardiac arrest but could also possibly help to guide early post-resuscitation care. While anti-fibrinolytic treatment is an established approach improving outcome in trauma patients with fibrinolytic hemorrhage, no data are yet available for cardiac arrest patients. The current study, however, suggests that an interventional trial investigating the safety and efficacy of anti-fibrinolytic treatment based on d-dimer levels may be warranted in specific bleeding cardiac arrest patients, or those undergoing invasive procedures such as extracorporeal membrane oxygenation.
6. Conclusions
Overt DIC is rare in a European cohort of patients with resuscitated out-of-hospital cardiac arrest, limiting the broad applicability of the DIC score for prognostication. D-dimer levels in contrast might prove useful to early assess mortality risk after successful resuscitation. A prospective investigation may be warranted.
7. Limitations
The registry design of our study has obvious limitations. The availability of D-dimer levels was limited, because D-dimer is not yet routinely assessed in out-of-hospital cardiac arrest patients. Thus, we had to exclude a substantial number of patients prior to analysis. Furthermore, we missed patients who had no blood sampling due to their very poor condition at admission. Our data may therefore underestimate the true prevalence of overt DIC in out-of-hospital cardiac arrest. Although our study cohort is largely comparable to the total population of our cardiac arrest registry in terms of age, sex, witness status, rate of bystander cardiopulmonary resuscitation, initial shockable rhythm, rate of cardiac etiology of cardiac arrest and lactate levels [
Progress in the chain of survival and its impact on outcomes of patients admitted to a specialized high-volume cardiac arrest center during the past two decades.
] the associated bias has to be considered when interpreting our results. The limited outcome rate per predictor variable further affects the validity of the analysis. Furthermore, we used a d-dimer level of <0.5 μg/ml as the upper limit of normal in all patients. A future study, however, should test the effect on mortality of age-adjusted d-dimer cutoffs as age-adjustment may increase prognostic accuracy.
Source of funding
Funded by the Austrian Science Fund FWF_F 5404-B21 (Special Research Program, Cellular Mediators Linking Inflammation and Thrombosis).
Predicting survival with good neurological recovery at hospital admission after successful resuscitation of out-of-hospital cardiac arrest: the OHCA score.
The scoring system of the scientific and standardisation committee on disseminated intravascular coagulation of the international society on thrombosis and Haemostasis: a 5-year overview.
New disseminated intravascular coagulation score: a useful tool to predict mortality in comparison with acute physiology and chronic health evaluation II and logistic organ dysfunction scores.
Cardiac arrest and cardiopulmonary resuscitation outcome reports: update and simplification of the Utstein templates for resuscitation registries. A statement for healthcare professionals from a task force of the international liaison committee on resuscitation.
63. (American Heart Association European Resuscitation Council, Australian Resuscitation Council, New Zealand Resuscitation Council, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Council of Southern Africa). Resuscitation,
2004: 233-249
Association of initial rhythm with neurologically favorable survival in non-shockable out-of-hospital cardiac arrest without a bystander witness or bystander cardiopulmonary resuscitation.
Impact of time to return of spontaneous circulation on neuroprotective effect of targeted temperature management at 33 or 36 degrees in comatose survivors of out-of hospital cardiac arrest.
Target temperature management of 33 degrees C and 36 degrees C in patients with out-of-hospital cardiac arrest with initial non-shockable rhythm - a TTM sub-study.
Progress in the chain of survival and its impact on outcomes of patients admitted to a specialized high-volume cardiac arrest center during the past two decades.
30266-8&id=fx1.jpg){kind=link}
30266-8&id=gr1.jpg){kind=link}
30266-8&id=gr2.jpg){kind=link}