Endothelial cells (ECs) are in tight contact with all organs, so that EC activation and damage has been implicated in the development of multiple organ failure (MOF) 1. Among the proposed mechanisms, altered fibrinolysis may promote fibrin deposition and thereby contribute to microvascular alterations 2.

Fibrinolysis and inflammation may be intertwined. Elevated concentrations of C-reactive protein (CRP), especially when they persist over time, are correlated with the risk of MOF and death 3. CRP may inhibit fibrinolysis by inducing release of plasminogen activator inhibitor-1 (PAI-1) from human aortic ECs 4. In addition, the administration of recombinant CRP to volunteers increases circulating PAI-1 levels 5. We previously showed that hypofibrinolysis assessed by the Euglobulin Clot Lysis Time (ECLT) was strongly correlated with CRP concentrations in critically ill patients 6.

ECLT is the test most commonly used to estimate plasma fibrinolytic capacity. The ECLT result represents the balance between tissue plasminogen activator (t-PA) and PAI-1 activities. We sought to define the relationship between ECLT, as a marker of fibrinolysis, von Willebrand factor (vWF), a marker of endothelial dysfunction 7, and the severity of the clinical syndrome in critically ill patients. Organ dysfunction was estimated by the Sequential Organ Failure Assessment (SOFA) score 8 and the risk of hospital mortality by the simplified acute physiologic score (SAPS II) 9.

Of the 49 patients, 18 had sepsis (11 pneumonia, 2 peritonitis, 2 skin infection, 1 urinary tract infection, 1 endocarditis and 1 mediastinitis); the 31 other patients had exacerbated chronic obstructive pulmonary disease (COPD) without evidence of sepsis (n = 21), postanoxic coma or stroke (n = 6), or heart failure (n = 4). Mortality was higher in septic than in non-septic patients (55% vs. 35.4%).

The biological characteristics of the patients are shown in Table 1. Inflammatory variables, such as fibrinogen and CRP levels, were higher in the septic than in the non-septic patients. ECLT was also higher in the septic than in the non-septic patients. There was no significant difference in vWF between septic and non-septic patients.

In the whole population, there was a significant correlation between ECLT and CRP levels (n = 49, r = 0.678, p < 0.001, fig 1A) and between ECLT and SOFA score (r = 0.39, p = 0.009, fig 1B). There was a weak correlation between vWF and CRP levels (r = 0.29; p = 0.04), but not between vWF and either ECLT (r = -0.06, p = 0.65) or the SOFA score (r = -0.02, p = 0.88).

Table 2 shows the correlations between SOFA score, SAPS II score, vWF levels and ECLT in non-septic and septic patients. The relationship between SOFA score and ECLT remained significant in the non-septic patients (r = 0.45, p = 0.01) and was almost significant in the septic patients (r = 0.39, p = 0.11). Other correlations were not significant.

A multiple logistic regression model was used to evaluate the ICU death event (Table 3). The SAPS II (p = 0.048) and ECLT (p = 0.049) were retained in the multiple logistic regression as explanatory variables for ICU mortality.

Several studies have reported a relationship between coagulation abnormalities, organ dysfunction and mortality in critically ill patients 1213141516, but the role of fibrinolysis has not been well studied. Fibrinolytic activity is primarily determined by the balance between t-PA and PAI-1 levels. ECs are responsible for the production and release of t-PA and contribute to the release of PAI-1. Multiple factors, including lipoproteins, cytokines and inflammatory markers, modulate EC production of t-PA and PAI-1 17. There are several reasons why decreased fibrinolysis could be considered as a surrogate marker of EC dysfunction, organ failure and mortality 181920. Indeed, the deposition of fibrin on confluent ECs causes EC aggregation and disorganization of the monolayer, thus increasing permeability 2122. In addition, fibrin is also a potent stimulus for EC production and release of IL-8, a leucocyte chemotactic factor 23.

In a large retrospective analysis of 1789 ICU patients, Okabayashi et al 12 reported that the SOFA score was correlated with antithrombin concentrations (r = 0.32, p < 0.05), prothrombin fragment concentration (r = 0.4, p < 0.001) and fibrinolysis as assessed by t-PA-PAI-1 complexes (r = 0.52, p < 0.001). Mavrommatis et al 24 studied the activation of fibrinolysis in septic patients during the first day in the ICU, and observed that t-PA/PAI-1 complexes (4.7 ± 0.6 vs 2.1 ± 0.2 ng/mL) and the percentage of fibrinogen/fibrin degradation products (100 vs 57%) were higher in patients with septic shock than in those without (all p < 0.001). These authors also observed a greater activation of fibrinolysis, as reflected by decreased plasminogen activity (41.6 ± 3.3 vs 87.4 ± 3.2%) and increased t-PA and PAI-1 concentrations 24.

ECLT is a global test for fibrinolysis, representing the balance between t-PA and PAI-1 activities. Previously considered as an imprecise method, we have been able to improve the precision and reproducibility of the test with a new semi-automatic device 11. Moreover, in contrast to that of vWF (12 h) 25, the half-lives of t-PA (3–4 min) and PAI-1 (10 min) are very short, suggesting that ECLT could be an instantaneous marker of clinical status. We previously showed that ECLT was related to the number of cardiovascular risk factors (hypertension, smoking habit, diabetes, history of coronary event or stroke, menopausal status) 26 and suggested that ECLT could be a surrogate marker of endothelial dysfunction.

The present study confirms the previously described 6 relationship between ECLT and CRP concentrations. This relationship could be explained by the simple fact that CRP reflects the severity of sepsis and the risk of organ failure. Alternatively, there may be a more pathophysiologic explanation, as CRP can act directly on the endothelium and modulates fibrinolysis 45. Another mechanism for the prolonged lysis time in septic patients could be related to the effects of systemic inflammation on the liver, causing activation of various cells and release of cytokines influencing the production of fibrinogen and PAI-1 2728.

We did not observe, in our population, a significant correlation between ECLT and SAPS II score, as a severity score, but the number of patients was limited and the mortality was quite high in the non-septic group who had significant co-morbidities.

Circulating plasma vWF is derived exclusively from the endothelium (with some minor expression by megakaryocytes). vWF can be produced by ECs when they are stimulated or damaged by various stimuli, including inflammatory cytokines (e.g., interleukin-1 or tumour necrosis factor), hypoxia, thrombin, histamine and leucocyte elastase 29. Increased levels of vWF are also influenced by substances like adrenaline (epinephrine), vasopressin, and even cyclosporin 29. Plasma levels of vWF have been considered as a marker of endothelial dysfunction and damage, although this is controversial 30 and results from studies that have correlated plasma levels of vWF, severity of inflammation and patient outcome are inconsistent 30. Ware et al 31 reported in 559 patients with acute lung injury and/or with acute respiratory distress syndrome that vWF levels were not different in patients with or without sepsis (p = 0.82). In contrast, Kayal et al 32 observed that vWF plasma levels were significantly higher in patients with severe infection (n = 25) than in non-infected patients (n = 7) (p < 0.001). In the present study, there was no significant difference in vWf antigen between septic and non-septic patients. These data confirm that vWF is a somewhat controversial marker of sepsis.

Although, SAPS II and ECLT were retained as explanatory variables of ICU mortality in the multiple logistic regression model, these results must be interpreted carefully in light of the small sample size.

This study has some limitations, including the small sample size and the single-centre nature. Nevertheless, these are interesting data and should be validated in larger heterogeneous populations.

In conclusion, ECLT measurements could represent a marker of alterations in EC function and of organ dysfunction and may be a useful prognostic factor in critically ill patients.

COPD: Chronic Obstructive Pulmonary Disease; CRP: C-reactive Protein; ECLT: Euglobulin Clot Lysis Time; ECs: Endothelial Cells; ICU: Intensive Care Unit; PAI-1: Plasminogen Activator Inhibitor-1; SAPS II: Simplified Acute Physiology Score II; SOFA: Sequential Organ Failure Assessment; t-PA: tissue Plasminogen Activator; vWF: von Willebrand Factor

The authors declare that they have no competing interests.

KZB: Laboratory analysis, writing of the manuscript and design of the study. SO: patient recruitment. MP: writing and analysis of results. PB: recruitment of patients. PC: laboratory analysis. JLV: analysis of the study. DB: coordination and design. MV: statistical analysis and coordination.

All authors read and approved the final manuscript.