Behçet’s disease: new insight into the relationship between procoagulant state, endothelial activation/damage and disease activity

Behçet’s disease (BD), also known as Adamantiades–Behçet's disease, is a rare multisystem, inflammatory disease of unknown aetiology [1] and is associated with thrombogenicity [2]. Gender distribution and clinical expression of BD varies with ethnic origin and seems to depend on the geographic area of the BD patients’ residence [3].

The diagnosis of BD is primarily based on clinical manifestations. New international criteria [3] have recently been published as diagnostic criteria for BD and include vascular manifestations, which is of great importance due to the frequency of serious vascular and large vessel involvement as has been recently reported [4]. However, patients who were diagnosed according to the International Study Group Criteria (1990) [5] also meet the new International Criteria (2013).

The pathogenesis of the prothrombotic state in BD is not known, although associated vascular damage/activation seems to be an important key factor [6, 7]. Increased levels of several procoagulant markers [6, 8], defective fibrinolysis [8‐11], and altered platelet function [12, 13] are some of the numerous findings that have been obtained to date; however, these variables have been investigated individually using various methods and, in some cases, have led to conflicting results.

In recent years, there has been growing interest in the use of global coagulation tests in the evaluation of bleeding diathesis and the hypercoagulable state. Rotational thromboelastometry (ROTEM) is a viscoelastometric clotting test that measures the kinetics of clot formation and fibrinolysis, providing global information on the cellular and soluble procoagulant/anticoagulant protein interactions. Calibrated automated thrombogram (CAT) is a thrombin generation test developed by Hemker et al.[14] that quantifies thrombin generation beyond the end points of traditional coagulation tests. Both techniques have been applied successfully to the study of the procoagulant state associated with various conditions, including inflammatory states [15‐17].

Considering that haemostasis emerges from the interplay between different cells, coagulation factors and inhibitors, these global tests may reveal a better understanding of the hypercoagulable condition in BD than has been obtained so far through subrogated and isolated procoagulant markers. Therefore, we investigated the potential utility of ROTEM and CAT in the characterisation of the procoagulant state in BD and assessed whether these tests offer a new insight into the physiopathological mechanisms of the disease and its procoagulant profile.

Of the 33 unrelated BD patients interviewed, 23 were included and compared with 33 age and gender matched healthy subjects. Ten patients were excluded because they did not fulfil inclusion criteria. None of the interviewed patients had signs or symptoms of current thrombosis. The clinical and treatment characteristics of the patients are summarised in Table 1.

As far as we know, this is the first report on the use of ROTEM and CAT in the study of the procoagulant state in BD. The present study has shown that thrombin generation and blood clotting capacity were increased in BD patients, even in the absence of thrombosis. Additionally, we found significant correlations between the ROTEM parameters, plasma ES levels and DA, which suggests links between the procoagulant state, endothelial inflammation and severity of symptoms of the disease.

We found increased levels of fibrinogen in our group of patients with BD. This result is consistent with those observed in previous studies [19‐21] and is not surprising considering that 78% of our patients were in an active state and that fibrinogen is an acute phase reactant [22]. The procoagulant effect of fibrinogen on ROTEM parameters has been reported previously [23], and our results are in agreement with this effect. We observed that fibrinogen levels were significantly correlated with INTEM-MCF in both patients and controls, which suggests that increased levels of fibrinogen in the BD group might be involved in the group’s hypercoagulable profile, as determined by the ROTEM test. Surprisingly, although INTEM-MCF correlated significantly with the fibrinogen levels and DA, we did not observe any dependence between the fibrinogen levels and DA. This issue, which was also observed by Hampton et al.[19], is intriguing but may be the result of the high variability of the fibrinogen synthesis route due to alternative mRNA processing, post-translational modifications and proteolytic degradation, which may lead to different levels of fibrinogen in response to similar inflammatory states [24]. Fibrinogen may also affect thrombin generation. Dielis et al. have shown that fibrinogen may produce not only an anticoagulant effect by increasing LT but also a procoagulant effect by heightening ETP and PH values in normal populations [25]. This dual anticoagulant/procoagulant effect can be explained by the spatial distribution of the thrombin’s binding sites and the kinetics of interaction with its multiple substrates. Increased fibrinogen levels may prolong LT because of this molecule’s ability to bind to thrombin through exosite II, which is needed for the thrombin-mediated FVIII activation. This fibrinogen binding leads to an anticoagulant effect in the initiation phase at low tissue factor concentrations [26, 27]. This evidence has also been supported by Hemker et al. who compared thrombin generation in full and defibrinated plasma [28]. It is interesting to mention that a prolonged LT despite an increased thrombotic risk is observed in patients with antiphospholipid syndrome [29] which suggests that an extended LT does not rule out the presence of a prothrombotic state. Fibrinogen may also increase thrombin generation due to fibrin’s ability to protect thrombin from inhibition by antithrombin III [28, 30]. Our results agree with these hypotheses, given that we found a significant positive correlation between fibrinogen levels and ETP in patients with BD. In summary, although fibrinogen could induce an anticoagulant effect by increasing the LT for generating thrombin, the amount of thrombin generated (ETP) and the strength of the clot (INTEM-MCF) were increased by higher levels of fibrinogen in the BD group. This condition could be responsible, at least in part, for the procoagulant pattern observed in the BD patients by the two global tests.

Previous studies have reported a correlation between platelet count (PTS) and clot formation speed and strength [31]. However, we did not find any differences in PTS between the patients and the controls. Although higher platelet activation and response to stimulus have been reported in BD patients [13, 32], our research group recently published a study of platelet function in the same group of BD patients included in the present study and found no differences in platelet activation markers between the controls and the BD patients, either at baseline conditions or after stimulation with agonists (TRAP and ADP) [33]. After considering this finding and the lack of differences in PTS and platelet contribution to the ROTEM trace between the groups, we conclude that platelets are not the cause of the deviation in the ROTEM results.

A study by Spiezia et al. (2008) suggests that erythrocyte count (ERY) may decrease clot firmness [34]. In the present study, we did not observe any significant differences in the ERY between the patients and controls, and therefore the contribution of ERY to the differences observed in the ROTEM trace between the two groups seems negligible.

It has been reported that abnormal fibrinolysis may contribute to thrombosis, atherosclerosis and vascular stenosis [35]. Fibrinolysis has been previously studied in BD with conflicting results, which was probably due to the high variability of patients included in the studies and the different assays used in each case [8‐10, 19, 36‐38]. In the present study, BD patients showed tPA antigen levels similar to controls but had significantly increased PAI-1 antigen plasma levels, suggesting a possible hypofibrinolitic profile in this group. Systemic inflammation as presented in BD could increase PAI-1 levels [39, 40]. Moreover, it has been reported that platelet stimulation by thrombin induces platelet synthesis and release of active PAI-1 [41, 42] and, in fact, increased platelet activation has been highly correlated to plasma PAI-1 activity in acute stroke patients [43]. Thrombin may also stimulate platelet synthesis and secretion of TGF beta [44‐46]. TGF beta, in turn, may increase the synthesis of PAI-1 in endothelial cells [44‐46]. These mechanisms may explain, at least in part, the increased plasma levels of PAI-1 in BD patients because they show systemic activation of coagulation and increased thrombin production in response to stimulus. Increased levels of PAI-1 can increase the clot formation speed and clot stability [47, 48] due to the rapid and irreversible blockage of the protease activity of tPA, the main plasminogen activator [49]. Our results agree with this observation given that we found a significant correlation between antigenic levels of PAI-1 and INTEM-CFT, INTEM-α and INTEM-MCF, which points to PAI-1 as a key factor in the procoagulant state observed in BD patients by this test. In spite of the fact that an association between levels of PAI-1 and thrombosis in BD has not been reported, relief from vascular events and oral ulcers after treatment with profibrinolytic agents has been observed in these patients [50, 51]. Moreover, we and other groups have observed a positive correlation between PAI-1 levels and DA, suggesting a probable association between the impaired fibrinolysis in BD and the severity of the disease symptoms [19, 52]. Whether this finding reflects a causal relation between BD symptoms and defective fibrinolysis is an issue that needs to be evaluated in further studies with larger numbers of patients.

The procoagulant state observed by the CAT and ROTEM tests in the BD patients was supported by the increase in plasma TAT, a marker of intravascular thrombin formation. However, the TAT level did not correlate to the ROTEM and CAT parameters. A lack of correlation between TAT levels and CAT and ROTEM values has been previously reported [53‐55] which suggests that the TAT level might indicate that activation of coagulation had occurred but does not necessarily reflect the patient’s procoagulant potential at the time of the sampling.

In contrast to previous reports that indicated high DD levels in BD patients with active disease and deep vein thrombosis [11], we did not find any differences in DD between the BD patients and controls. This controversial result may be due to the absence of signs, symptoms or recent history of thrombosis in our patients.

Endothelial damage has been described as a potential key factor involved in the prothrombotic state of BD [7, 8], and ES, a marker of endothelial damage/activation, has been found to be increased in the active state of the disease [56]. Our results were in correspondence with this data as we found higher levels of ES in the BD patients compared with controls that correlate with DA (Figure 1). When analyzing the correlation between ES and ROTEM and CAT parameters, we found a significant correlation between ES levels and the ROTEM parameters but not between ES levels and the CAT parameters. We also failed to obtain any correlation between CAT parameters and DA that by the contrary showed correlation with the procoagulant profile observed by the ROTEM test. One hypothesis to explain this effect may be based on the fact that the CAT test is only able to depict the thrombin generation capacity of the plasma, whereas the ROTEM test describes thrombin generation, clot formation and fibrinolysis. As shown above, these processes may be altered in this disease, and therefore the ROTEM test may be a more appropriate test for describing the associated endothelial and inflammatory pathological condition of the disease.

Our results support the existence of an increased procoagulant state in BD patients, and they raise the question about the usefulness of anticoagulant treatments in these patients. This is a matter of controversy as there is no consensus among rheumatologists from various countries [57] because lifelong anticoagulation treatment might increase the risk of fatal bleeding [58]. Moreover, given that blood vessel wall inflammation is a possible cause of thrombosis in BD patients, the use of immunosuppressive therapy might be a more rational choice [59, 60].

The ROTEM test is a useful tool for studying of hypercoagulable state in BD. Additionally, CAT experiments reveal that plasma from BD patients is able to produce faster and higher thrombin generation. Our data also indicated that endothelial activation/damage is involved in both the clinical manifestation and procoagulant state of this pathology. Moreover, the heightening of fibrinogen and PAI-1 may be important components in the procoagulant condition of the disease, whereas the effect of platelets seemed to be almost negligible. Further studies are warranted to evaluate the relationship between PAI-1 levels and the symptoms of the disease and to determine whether the ROTEM test and ES levels are useful tools/markers for monitoring therapeutic response and disease progression in BD patients.