Association of hyperfibrinolysis with poor prognosis in refractory circulatory arrest: implications for extracorporeal cardiopulmonary resuscitation

Abstract

Background

Identifying candidates for extracorporeal cardiopulmonary resuscitation (eCPR) is challenging, and novel predictive markers are urgently needed. Hyperfibrinolysis is linked to tissue hypoxia and is associated with poor outcomes in out-of-hospital cardiac arrest (OHCA). Rotational thromboelastometry (ROTEM) can detect or rule out hyperfibrinolysis, and could, therefore, provide decision support for initiation of eCPR. We explored early detection of hyperfibrinolysis in patients with refractory OHCA referred for eCPR.

Methods

We analysed ROTEM results and resuscitation parameters of 57 adult patients with ongoing OHCA who presented to our ICU for eCPR evaluation.

Results

Hyperfibrinolysis, defined as maximum lysis ≥15%, was present in 36 patients (63%) and was associated with higher serum lactate, lower arterial blood pH, and increased low-flow intervals. Of 42 patients who achieved return of circulation, 28 had a poor 30-day outcome. The incidence of hyperfibrinolysis was higher in the poor outcome group compared with patients with good outcomes (75% [21 of 28] vs 7.1% [1 of 14]; P<0.001). The ratio of EXTEM A5 to lactate concentration showed good predictive value in detecting hyperfibrinolysis (AUC of 0.89 [95% confidence interval 0.8–1]).

Conclusions

Hyperfibrinolysis was common in patients with refractory cardiac arrest, and was associated with poor prognosis. The combination of high lactate with early clot firmness values, such as EXTEM A5, appears promising for early detection of hyperfibrinolysis. This finding could facilitate decisions to perform eCPR, particularly for patients with prolonged low-flow duration but lacking hyperfibrinolysis.

Editor's key points.

• •Triaging candidates for extracorporeal cardiopulmonary resuscitation (eCPR) is clinically challenging.
• •Hyperfibrinolysis is linked to tissue hypoxia and associated with poor outcomes in out-of-hospital cardiac arrest (OHCA).
• •The potential use of early detection of hyperfibrinolysis as a predictor of outcome in eCPR was analysed using rotational thromboelastometry in 57 adult patients with ongoing OHCA admitted to ICU for eCPR evaluation.
• •The incidence of hyperfibrinolysis was common after OHCA, and was higher in patients with poor outcome after eCPR.
• •Early detection of hyperfibrinolysis in combination with elevated lactate concentration is promising as an approach to facilitate decisions regarding initiation of eCPR after OHCA.

Despite intensive research, campaigns to improve public engagement, and efforts to increase availability of external defibrillators, the prognosis after out-of-hospital cardiac arrest (OHCA) remains poor.1, 2, 3 Over the past decade, mechanical circulatory support systems such as extracorporeal membrane oxygenation (ECMO) have emerged as promising rescue therapy in patients with refractory cardiac arrest. However, identifying optimal candidates for extracorporeal cardiopulmonary resuscitation (eCPR) remains a major challenge.4, 5, 6 Most experts agree that a witnessed arrest and no or at least short periods of no flow are a prerequisite for considering eCPR.⁷ The duration of conventional CPR is also of prognostic value and might help in deciding which patients are suitable candidates for ECMO.⁸ The impact of the quality of cardiac compressions, oxygenation, and overall circulation during conventional CPR remains an area of uncertainty. Long periods of low flow might still be associated with a favourable prognosis if the circulation during conventional CPR is sufficient to maintain the integrity of vital tissues.

Blood lactate, potassium, and pH are important markers of hypoperfusion to identify patients who are unlikely to benefit from eCPR. However, none of these parameters has proven to have sufficient sensitivity and specificity to predict clinical outcomes reliably.⁹ The search for other quickly available parameters to predict good or poor prognosis during evaluation for eCPR remains the focus of ongoing research. Recent studies have shown that patients resuscitated from cardiac arrest who developed disseminated intravascular coagulation (DIC) with a hyperfibrinolytic pattern are at a higher risk of multiple organ dysfunction syndrome and poor neurological outcomes.10, 11, 12 Moreover, Viersen and colleagues¹⁰ reported that severe hyperfibrinolysis detected by rotational thromboelastometry (ROTEM) and conventional markers of hyperfibrinolysis were associated with prolonged cardiopulmonary resuscitation time and elevated lactate concentrations. The authors concluded that profound hypoperfusion and tissue hypoxia in patients with OHCA might induce hyperfibrinolysis without prior trauma.

While the exact mechanisms remain elusive, it is assumed that severe hypoperfusion and hypoxia trigger tissue-type plasminogen activator (t-PA) release from Weibel–Palade bodies in endothelial cells.^13,14 Evidence supporting this hypothesis is derived from previously healthy drowning victims who experience prolonged no-flow intervals and develop severe hyperfibrinolysis after resuscitation.¹¹ Likewise, evidence of hyperfibrinolysis on admission to ICU could reflect insufficient CPR performance and might provide an additional marker of hypoperfusion to predict poor outcome, as reported by Buchtele and colleagues.¹⁵ Incorporating ROTEM as a point-of-care assay to screen for hyperfibrinolysis during evaluation for eCPR might therefore aid in decision-making, particularly in patients with prolonged periods of low flow who have exceeded the therapeutic time window but show no signs of hyperfibrinolysis.

Therefore, the aim of this study was to investigate the feasibility of rapid detection or exclusion of hyperfibrinolysis using early ROTEM values during resuscitation and to evaluate their possible roles in decision-making for performing eCPR.

Methods

The study was approved by the ethics committees of Charité—Universitätsmedizin Berlin (EA2/066/23) on June 28, 2023 and was conducted according to the Declaration of Helsinki. Necessity for informed consent was waived because of the retrospective nature of the study.

Patient selection and data collection

The study included patients >18 yr old who failed to achieve return of spontaneous circulation (ROSC) after guideline-based resuscitation efforts and were admitted to the Charité – Universitätsmedizin Berlin ICU for eCPR evaluation between November 2019 and November 2022. Inclusion was limited to patients with presumed cardiovascular causes of circulatory arrest. Patients who did not receive viscoelastic testing during eCPR evaluation were excluded from the study.

Evaluation for extracorporeal cardiopulmonary resuscitation

Upon arrival to ICU, the decision to perform an eCPR was made according to the inclusion and exclusion criteria of our centre. Patients were excluded from eCPR if they experienced an unwitnessed cardiac arrest, did not receive bystander CPR, had an estimated low-flow time to eCPR >90 min, or had a known malignancy.

All patients underwent targeted body temperature management with a target temperature of 33°C. Cooling was performed using a water-circulating gel pad cooling device (Arctic Sun 5000 Temperature Management System; Medivance, Louisville, KY, USA) or ECMO. Each patient received an oesophageal probe for continuous body temperature monitoring.

Resuscitation-related data were collected and analysed based on the international reporting guidelines for patients with OHCA.¹⁶ Poor clinical outcome was defined as death or unfavourable neurological outcome assessed by the cerebral performance category (CPC) score of 3–5 during the ICU stay or on day 30 after OHCA. The neurological condition of patients transferred for rehabilitation within 30 days after resuscitation was evaluated based on routinely received medical reports from the rehabilitation clinics.

Laboratory methods

Blood samples were collected for laboratory studies and ROTEM analysis as soon as vascular access was established. Viscoelastic tests were performed with citrated blood using a ROTEM Sigma device (Tem International, Munich, Germany).¹⁷

An extrinsically activated assay (tissue factor activation, EXTEM) was performed to assess clot dynamics using: clotting time (s), representing the time until initiation of clotting; clot formation time (s), the time until clot firmness reaches 20 mm; A5 and A10 (in mm) measuring early clot firmness, corresponding to the amplitude at 5 and 10 min, respectively; maximum clot firmness, the maximum amplitude of clot firmness during the runtime; and maximum lysis (ML, %), expressed as a percentage of the clot lysed after 60 min of measurement, reflecting fibrinolytic activity (Supplementary Fig. S1). ML ≥15% was defined as hyperfibrinolysis.¹⁸

Data analysis

Statistical analyses were performed with IBM SPSS Statistics Version 28 (IBM, Armonk, NY, USA) and GraphPad Prism version 9 (GraphPad Software, San Diego, CA, USA). For continuous variables, descriptive statistics are given as median with interquartile range (IQR), while categorical variables are presented as absolute and relative frequencies. To compare differences between two independent groups (hyperfibrinolysis vs no hyperfibrinolysis; good outcome vs poor outcome), the Mann–Whitney U-test was used for continuous variables, and χ² test was used for categorical variables. The Kruskal–Wallis test was used to perform a subgroup analysis of resuscitation parameters among the no-ROSC, eCPR, and ROSC groups. To explore the ability of ML to discriminate between patients with and without good outcomes, receiver operating characteristic (ROC) analysis was carried out, including area under the curve measures (AUC) with 95% confidence intervals (CI). The corresponding sensitivity and specificity for the respective cut-off values are given in percentage and CI. Upon discovering the association between markers of tissue hypoxia and hyperfibrinolysis, we conducted a post hoc analysis to assess the ability of early ROTEM values (A5 and A10 in the EXTEM assay) and lactate concentrations to detect or exclude hyperfibrinolysis at an earlier stage. ROC analyses were used to test the ability of A5 and A10 values, or the combination of A5 and lactate concentrations, to discriminate between patients with and without hyperfibrinolysis, as defined by ML ≥15%. Kaplan–Meier curves are used to describe overall survival between patients with hyperfibrinolysis (ML ≥15%) and those without (ML <15%). The log-rank test was used to test differences in overall survival between these groups. A two-sided significance level of 0.05 was applied without adjustment for multiple comparisons. All P-values constitute exploratory data analyses and do not allow for confirmatory generalisation of results.

Results

Fifty-seven patients (median age 54 yr, [IQR 42–61 yr]; 80.7% male) with ongoing OHCA who were admitted to our ICU were included in the study. On arrival, sustained ROSC was established in nine patients (15.8%) after conventional resuscitation efforts. Thirty-three patients (57.9%) received venoarterial-ECMO (VA-ECMO) therapy because of refractory cardiac arrest. Fifteen patients (26.3%) did not meet the inclusion criteria for eCPR (e.g. prolonged resuscitation, unobserved circulatory arrest, no basic life support), and no sufficient ROSC could be established, despite optimisation of resuscitation efforts.

Pathologic fibrinolysis with ML ≥15% was detected in 36 (63.2%) patients [95% CI 0.46–0.82]. Nearly all patients without ROSC in whom eCPR was not initiated showed signs of hyperfibrinolysis (14 out of 15), whereas patients successfully resuscitated by conventional methods had the lowest incidence of hyperfibrinolysis (one out of nine). In the ECMO group, 21 out of 33 patients had hyperfibrinolysis. Baseline characteristics stratified by the presence or absence of hyperfibrinolysis are presented in Table 1 and Supplementary Figure S2.

Table 1.

Baseline characteristics stratified by the presence or absence of hyperfibrinolysis. eCPR, extracorporeal cardiopulmonary resuscitation; INR, international normalized ratio; IQR, interquartile range; PTT, partial thromboplastin time; ROSC, return of spontaneous circulation. *ROSC, eCPR, and no ROSC are distinct, non-overlapping categories and describe the final resuscitation status after admission. ^†Upper limit value of normal or reference range. ^‡Statistical analyses for D-dimer levels were not performed as several measured values exceeded the local laboratory's upper reference limit of 35 mg L⁻¹.

Variables	Cohort (n=57)		Hyperfibrinolysis (n=36)		No hyperfibrinolysis (n=21)		P-value
Age (yr), median [IQR]	54	[42–61]	56	[45–63]	54	[37–59]	0.18
Gender, male (n, %)	46	80.7	29	80.6	17	81	0.97
Resuscitation status*							<0.001
No ROSC* (n, %)	15	26.3	14	38.9	1	4.8
eCPR* (n, %)	33	57.9	21	58.3	12	57.1
ROSC* (n, %)	9	15.8	1	2.8	8	38.1
Resuscitation values
Witnessed (n, %)	47	82.5	31	86.1	16	76.2	0.34
Basic life support (n, %)	49	86.0	31	86.1	18	85.7	0.97
Shockable rhythm (n, %)	35	61.4	20	55.6	15	71.4	0.24
No-flow time (min), median [IQR]	0	[0–5]	0	[0–5]	0	[0–5]	0.98
Low-flow time (min) median [IQR]	70	[60–85]	72.5	[60–82.5]	70	[60–90]	0.73
Cumulative dose of epinephrine (mg) median [IQR]	7	[6–9]	8	[6–10]	6	[4–8]	0.022
Lactate (mmol L−1) (<1.8)†, median [IQR]	13.31	[11–16.5]	14.5	[13.2–18.4]	11.1	[9.5–13.6]	<0.001
pH (7.35–7.45)†, median [IQR]	6.91	[6.8–7.09]	6.84	[6.76–6.96]	7.1	[6.92–7.2]	<0.001
Coagulation values
Platelet count (150–370 nl−1)†, median [IQR]	190	[140–228]	153	[125.5–198]	216	[192–282]	0.002
Haemoglobin (12.5–17.2 g dl−1)†, median [IQR]	13.2	[11.6–14.2]	13.3	[11.2–14.4]	13,0	[11.6–13.9]	0.92
Prothrombin time (70–130%)†, median [IQR]	55	[44–68]	55	[39–68]	64	[46–68]	0.39
INR (0.9–1.25)†, median [IQR]	1.35	[1.18–1.57]	1.37	[1.2–1.7]	1.35	[1.18–1.51]	0.48
PTT (s) (26–40)†, median [IQR]	120.4	[65–170]	126.8	[80.9–170]	120.4	[61.95–170]	0.87
D-dimer‡ (<0.5–35 mg l−1)†, median [IQR]	35	[22.03–35]	35	[35–35]	34.72	[16.1–35]
Fibrinogen (1.6–4 g L−1)†, median [IQR]	1.98	[1.26–2.49]	1.805	[1.15–2.44]	2.08	[1.68–2.94]	0.07
Maximum lysis (%), median [IQR]	89	[2.5–99.5]	99	[92.5–100]	0	[0–3]	<0.001

Overall, circulation was established in 42 patients (73.4%), of whom 28 (66.7%) had a poor 30-day outcome, including 26 non-survivors. While 10 patients died because of severe multiorgan dysfunction syndrome, in 16 patients the decision to discontinue life-sustaining therapy was made after the determination of an unfavourable neurological prognosis. Of these 16 patients, 10 had a CPC score of 5, while the remaining six had a CPC score of 4. Two patients were transferred to a neurological early rehabilitation program but had poor neurological performance at 30 days post-resuscitation, with a CPC score of 3. Among the patients with good neurological outcome, 10 patients were assessed with a CPC score of 1, while the remaining four had a CPC score of 2.

Patients with an unfavourable outcome had increased ML (median 93% [IQR 9–99%] 1% [IQR 0–3%], P<0.001), and the prevalence of hyperfibrinolysis with ML ≥15% was higher compared with the group with better outcomes (75%, or 21 out of 28 patients vs 7.1%, or one out of 14 patients; P<0.001). Other outcome-related parameters are shown in Table 2.

Table 2.

Patient characteristics according to 30-day clinical outcome. cCPR, conventional cardiopulmonary resuscitation; eCPR, extracorporeal cardiopulmonary resuscitation; INR, international normalized ratio; IQR, interquartile range; PTT, partial thromboplastin time. *Because of the low prevalence of the disease, statistical analysis was not applicable. ^†Upper limit value of normal or reference range. ^‡Statistical analyses for D-dimer levels were not performed as several measured values exceeded the local laboratory's upper reference limit of 35 mg L⁻¹.

Variables	Cohort (n=42)		Poor outcome (n=28)		Good outcome (n=14)		P-value
Age (yr), median [IQR]	54	[41.3–61.5]	50	[42–62]	55	[32.5–61.5]	0.89
Gender, male (n, %)	32	76.2	19	67.6	13	92.2	0.07
Resuscitation method							0.017
eCPR (n, %)	33	78.6	25	89.3	8	57.1
cCPR (n, %)	9	21.4	3	10.7	6	42.9
Comorbid diseases
Coronary artery disease (n, %)	28	66.7	18	64.3	10	71.4	0.64
Diabetes mellitus (n, %)	5	11.9	2	7.1	3	21.4	0.31
Smoker (n, %)	9	21.4	5	17.9	4	28.6	0.45
Arterial hypertension (n, %)	12	28.6	6	21.4	6	42.9	0.17
Chronic heart failure (n, %)	9	21.4	4	14.1	5	35.7	0.13
Chronic kidney disease* (n, %)	2	4.8	1	3.6	1	7.1	n.a.
Cause of cardiac arrest							0.71
Cardiac (n, %)	38	90.5	25	89.3	13	92.3
Unknown (n, %)	4	9.5	3	10.7	1	7.1
Resuscitation values
Witnessed (n, %)	36	85.7	22	78.6	14	100.0	0.61
Basic life support (n, %)	39	92.9	26	92.9	13	92.9	1.0
Shockable rhythm (n, %)	27	64.3	16	57.61	11	78.6	0.17
No-flow time (min), median [IQR]	0	[0–3]	0	[0–3]	0	[0–1.25]	0.39
Low-flow time (min), median [IQR]	72.5	[55–86.6]	80	[56–90]	60	[31–72]	0.054
Cumulative dose of epinephrine (mg), median [IQR]	7	[5–8]	7	[6–9]	6	[4–7]	0.026
Lactate (mmol L−1) (<1.8)†, median [IQR]	12.9	[10–16]	13.9	[10.9–16.4]	11	[5.5–13.7]	0.036
pH (7.35–7.45)b, median [IQR]	6.95	[6.87–7.15]	6.91	[6.79–7.04]	7.12	[6.94–7.20]	0.008
Coagulation values
Platelet count (150–370 nl−1)†, median [IQR]	194.5	[147–238]	191	[142–240]	211	[161–247]	0.27
Haemoglobin (12.5–17.2 g dl−1)†, median [IQR]	12.9	[11.3–13.9]	11.8	[11–13.7]	13.5	[11.5–14.2]	0.19
Prothrombin time (70–130%)†, median [IQR]	61	[44–68]	55	[39–68]	63	[47–72]	0.55
INR (0.9–1.25)†, median [IQR]	1.34	[1.18–1.57]	1.34	[1.18–1.71]	1.30	[1.15–1.30]	0.54
PTT (s), (26–40)†, median [IQR]	138	[82.8–170]	165	[83.2–170]	120	[50.2–159]	0.23
D-dimer‡ (<0.5–35 mg L−1)†, median [IQR]	35	[20–35]	35	[22.8–35]	34.72	[20–35]	—
Fibrinogen (1.6–4 g L−1)†, median [IQR]	1.9	[1.26–2.6]	1.9	[1.2–2.5]	2.0	[1.6–2.7]	0.40
Maximum lysis (%) (<15%)†, median [IQR]	21.5	[0–99]	93	[9–99]	1	[1–3.5]	<0.001
Maximum lysis ≥15% (n, %)	22	52.4	21	75	1	7.1	<0.001

Based on the above findings, we evaluated the potential of ML to distinguish between patients with favourable and unfavourable outcomes using ROC analysis (Fig. 1a). ML in EXTEM resulted in an AUC of 0.85 [95% CI 0.7–1.0] for poor outcome. The ML cut-off of ≥15% predicts poor outcomes with 92.9% [95% CI 0.69–1.0] specificity and 75% sensitivity [95% CI 0.57–0.87]. Kaplan–Meier analyses (Fig. 1b) showed a higher 30-day survival rate in the non-hyperfibrinolysis group compared with the hyperfibrinolysis group (log-rank P<0.001). Considering the strong association between hyperfibrinolysis and hypoperfusion markers, including lactate, we evaluated the potential of early available ROTEM values, including EXTEM A5, A10, and lactate concentration to detect fibrinolysis earlier. As displayed in Figure 2a, EXTEM A5 resulted in an AUC of 0.76 [95% CI 0.63–0.89] to predict hyperfibrinolysis, while EXTEM A10 resulted in an AUC of 0.79 [95% CI 0.64–0.9]. Lactate had an AUC of 0.78 [95% CI 0.63–0.89]. Combined analysis revealed that EXTEM A5 to lactate ratio yielded the highest AUC of 0.89 [95% CI 0.8–1]. Using a cut-off value of ≤3.1 for EXTEM A5 to lactate ratio, the corresponding sensitivity and specificity for detecting hyperfibrinolysis were 88.2% [95% CI 0.7–1.0] and 85% [95% CI 0.6–1.0], respectively (Fig. 2b).

Fig 1.

(a) Receiver operating characteristic curve of maximum lysis (ML) discriminates patients with and without poor outcome as defined by poor neurologic outcome or death at day 30. Maximum lysis in EXTEM resulted in an area under the curve (AUC) of 0.85 [95% CI 0.7–1.0] for poor outcome, an ML cut-off of ≥15% predicted poor outcome with 93% [95% CI 0.69–1.0] specificity and 75% sensitivity [95% CI 0.57–0.87]. (b) Probability of survival to day 30 based on the occurrence of hyperfibrinolysis defined as ML ≥15% (P<0.001). Dashed lines represent the 95% CI at each time point, displayed as confidence bands. CI, confidence interval; HF, hyperfibrinolysis.

Fig 2.

(a) Receiver operating characteristic curve of EXTEM A5 (AUC 0.76 [95% CI 0.63–0.89]), EXTEM A10 (AUC 0.79 [95% CI 0.64–0.9]), and lactate concentration (AUC 0.78 [95% CI 0.63–0.89; P=0.001]) to predict hyperfibrinolysis. (b) Receiver operating characteristic curve of the EXTEM to lactate ratio (AUC 0.89 [95% CI 0.8–1]) to predict hyperfibrinolysis.

Discussion

We used ROTEM to examine coagulation profiles in patients with refractory OHCA who were being evaluated for extracorporeal cardiopulmonary resuscitation (eCPR). We describe a rapid approach to identify hyperfibrinolysis by combining early clot firmness (EXTEM A5) with lactate (EXTEM A5 to lactate ratio). Our findings support the hypothesis that absence of hyperfibrinolysis measured by ROTEM strengthens the rationale to initiate ECMO despite a long low-flow time as it supports sufficient tissue perfusion during the conventional resuscitation phase.

Compared with previous studies^10,15,19, which reported incidence of hyperfibrinolysis from 36% to 50% in patients with OHCA, our study cohort showed an incidence of hyperfibrinolysis of 62%. This likely results from persistent circulatory arrest until admission, preceded by prolonged hypoperfusion and tissue hypoxia. Supporting this, we observed elevated lactate values and longer low-flow times in our cohort than in previous studies.^10,15,19 Moreover, nearly all patients (93%) in whom ROSC was not achieved and who did not receive ECMO had evidence of hyperfibrinolysis. In contrast, patients in whom ROSC was established using conventional resuscitation efforts had the lowest rates of hyperfibrinolysis. Comparing resuscitation parameters between groups, we found that the non-ROSC group had longer periods of low flow and higher lactate values, and had received higher cumulative doses of epinephrine, which supports a link between hyperfibrinolysis and tissue hypoxia. Among patients treated with ECMO, a substantial proportion (63%) had evidence of hyperfibrinolysis.

Regarding patient outcomes, hyperfibrinolysis, defined as ML ≥15%, was associated with poor prognosis after restoring circulation, irrespective of the use of ECMO or conventional methods, which aligns with the results of previous studies.^11,15 Buchtele and colleagues,¹⁵ evaluating the optimal cut-off value of ML for predicting poor outcomes in patients with hyperfibrinolysis, found that a cut-off value ≥20% ML resulted in a specificity of 100% for predicting poor outcomes.

In contrast to our findings, Schöchl and colleagues¹⁹ also utilised ROTEM to investigate the incidence of hyperfibrinolysis in patients with OHCA, but reported no significant difference in lysis parameters between patients who achieved ROSC and those who did not. However, there are notable differences in study design and clinical context between this study and ours. Primarily, the timing of blood sampling differed; Schöchl and colleagues¹⁹ performed blood sampling immediately on-scene, capturing the early phase of cardiac arrest. In contrast, our samples were collected upon ICU admission. These differences could be crucial, as prolonged CPR is likely to increase the risk of developing tissue hypoxia and hyperfibrinolysis.

Detecting or excluding hyperfibrinolysis by ROTEM can take up to 60 min depending on the severity of fibrinolysis, making it unreliable as a decision-making tool under time-critical circumstances, such as initiation of eCPR.^20,21 Several studies have demonstrated the predictive power of rapidly available clot firmness values such as EXTEM A5 and A10 in early diagnosis of hyperfibrinolysis. However, these values showed good sensitivity and specificity only in cases of severe hyperfibrinolysis, where clot was lysed after a short measuring time.^21,22 The combination of EXTEM A5 and lactate (EXTEM A5 to lactate ratio) improved the AUC in ROC analysis to 0.89, resulting in high sensitivity and specificity for detecting hyperfibrinolysis. Low values of A5 and A10 could be the consequence of overt hyperfibrinolysis, which outweighs the procoagulant and antifibrinolytic potential of the clot, increasing consumption of procoagulant and antifibrinolytic factors.^21,23

The trend towards lower fibrinogen concentrations observed in the hyperfibrinolysis group also imply fibrinogenolysis, as excessive plasmin generation not only degrades fibrin but also breaks down fibrinogen, resulting in fibrin/fibrinogen degradation products (FDP).^24,25 Alongside the decreased concentration of fibrinogen, the lower platelet counts in the hyperfibrinolysis group are likely to contribute to the observed low A5 and A10 measurements.

In an observational study, Wada and colleagues¹² revealed that >50% of patients with OHCA presented with DIC on admission, and a subgroup of these patients had DIC with a fibrinolytic phenotype associated with poor outcomes. They further showed that high lactate concentrations as a marker of tissue hypoxia independently predicted occurrence of DIC with a fibrinolytic phenotype. The key features of this fibrinolytic DIC phenotype were prolonged prothrombin time, decreased platelet count, elevated D-dimer and FDP concentrations, low fibrinogen concentrations, and increased FDP/D-dimer ratio suggesting consumption coagulopathy with excessive fibrin and fibrinogenolysis as the result of severe hypoperfusion. Similar observations and conclusions were made by Schwameis and colleagues¹¹ in drowning-induced OHCA. We observed alterations in conventional coagulation tests in our cohort, regardless of their fibrinolytic status in ROTEM, indicating DIC was present in both scenarios. However, the intensity of tissue hypoxia seems to promote progression of DIC towards a hyperfibrinolytic phenotype.^11,12 Another reason for altered coagulation values, particularly increased partial prothrombin time (PTT), might be the initial treatment with unfractionated heparin, commonly given by emergency physicians as a loading dose of 5000 IU i.v. alongside aspirin for suspected cardiac origin of OHCA. Additionally, hypoxia-induced damage to the endothelium can result in disruption of the glycocalyx, resulting in release into the circulation of glycosaminoglycans, which are known for their heparin-like effect and can also prolong the PTT.^11,26,27

Given the high incidence of hyperfibrinolysis in our cohort, which was associated with poor prognosis, the ratio of EXTEM A5 to lactate is promising for detecting or ruling out hyperfibrinolysis rapidly. The ability to exclude the presence of a hyperfibrinolytic phenotype at an early stage could help inform the decision to commence ECMO therapy in uncertain cases. This would be especially useful in cases with prolonged low-flow times exceeding local inclusion criteria. The absence of hyperfibrinolysis as a marker of effective conventional resuscitation might be an argument for putting the patient on extracorporeal circulation despite a long low-flow time. The quality of low-flow circulation in cardiac arrests during CPR is usually unclear. It is clearly a key factor in neurological outcomes after prolonged cardiac arrest, and has always to be considered when determining whether a patient is a candidate for eCPR. During the short time window when the decision has to be made whether a patient is suitable for eCPR, no single parameter has been proven useful. The combination of EXTEM A5 and lactate can be obtained in a reasonable timeframe to guide decision-making concerning eCPR. More prospective studies are needed to validate the ability of these parameters to reflect CPR quality during prolonged resuscitation.

Some limitations must be taken into account when interpreting these data. Firstly, the relatively small number of patients and the retrospective design limit the robustness and confidence of the predictive model developed for clinical outcomes after OHCA. The relatively low incidence of OHCA cases eligible for eCPR, combined with the stressful nature of resuscitation efforts that hindered ROTEM analysis for every admitted patient limited our ability to assemble a larger cohort. Secondly, study generalisability is constrained by the single-centre design and the potential variety of viscoelastic devices utilised by other healthcare institutions. Thirdly, the current study focused strictly on patients with circulatory arrest, presumed to have a cardiovascular cause, who underwent ROTEM analysis after admission to ICU for eCPR evaluation. The potential of selection bias related to such an exclusive cohort must be considered. Fourthly, because of the laboratory's limitation to provide exact values for D-dimer measurements >35 mg L⁻¹, we were unable to analyse differences between the groups. Consequently, D-dimer concentrations cannot be utilised to assess the severity of hyperfibrinolysis. We also did not assess specific fibrinolysis markers, such as t-PA, plasminogen, and plasminogen activator inhibitor-1. Admittedly, there is no standardised test for detecting hyperfibrinolysis. Therefore, it is essential to acknowledge the potential bias that might arise because of the lack of an international standardised value.

Given these limitations, our findings should be considered as hypothesis generating and, while aligning with previous studies regarding the association between hyperfibrinolysis and poor prognosis, it is necessary to conduct prospective multicentre studies to validate our prediction models regarding the outcome and early detection of hyperfibrinolysis. In particular, the ratio of EXTEM A5 to blood lactate concentration should be evaluated as a potential marker guiding the use of eCPR.

Conclusions

Our findings highlight the high prevalence of hyperfibrinolysis in patients with out-of-hospital cardiac arrest, and confirm previous studies linking hyperfibrinolysis to poor clinical outcomes. The combination of lactate concentration with early clot firmness values is a promising approach to rapid detection or exclusion of hyperfibrinolysis, which could aid in decision-making regarding use of extracorporeal CPR. Further prospective studies are necessary to validate these observations.

Authors’ contributions

Study design: AM, JN, JK, DZ

Acquisition of data: AM

Analysis and interpretation of data: AM, JN, JK, SKP

Initial draft of the manuscript: AM, JK, JN

Manuscript revision: all authors

Declaration of interest

The authors declare that they have no conflicts of interest.

Handling Editor: Hugh C Hemmings Jr

Appendix A. Supplementary data

The following is the Supplementary data to this article: