Thrombelastography and tromboelastometry in assessing coagulopathy in trauma

On admission, 25-35% of trauma patients present with coagulopathy, which is associated with a several-fold increase in morbidity and mortality [1, 2]. Although the management of traumatic coagulopathy differs worldwide [3, 4], the recent introduction of haemostatic control resuscitation [5‐7] and the emerging understanding of acute post-traumatic coagulopathy [1, 2, 8, 9], emphasize the urgent need for adequate haemostatic assays to guide therapy. Classically, coagulopathy is often monitored by plasma-based routine coagulation tests (RCoT) such as activated partial thromboplastin time (APTT) and prothrombin time (PT). These assays were developed half a century ago to monitor haemophilia and anticoagulation therapy, but have unfortunately never been validated for the prediction of haemorrhage in a clinical setting [10, 11]. It should be noted that although abnormal PT and APTT are highly correlated with mortality in trauma patients, the cause of death in these patients is not identified as excessive bleeding [12‐14]. This lack of correlation with clinically relevant coagulopathies can be explained by the fact that plasma-based assays reflect only the small amount of thrombin formed during initiation of coagulation [15, 16]. Consequently, recent reviews have concluded that the plasma-based assays are inappropriate for monitoring coagulopathy or guide transfusion therapy, calling for new tests to monitor these complex patients [17, 18].

In 1994, the classical clotting cascade of haemostasis [19, 20] was challenged by the introduction of a cell-based model of haemostasis emphasizing the importance of tissue factor (TF) as the initiator of coagulation and the pivotal role of platelets for intact haemostasis [21]. The poor correlation between RCoT and clinical bleeding in e.g. trauma and surgery [12‐15, 22‐25] is, hence, explained by this new understanding of haemostasis.

The necessity to analyze whole blood to accurately identify relevant coagulopathies, has led to a revival of the interest in viscoelastic point-of-care haemostatic assays (VHA) such as Thromboelastography (TEG^®) and Rotation Thromboelastometry (ROTEM^®).

The objective of this article is to review the basic principles of VHA, the correlation between the result of VHA and clot formation in accordance with the cell-based model of haemostasis, the current use of VHA-guided therapy in trauma and massive transfusion (haemostatic control resuscitation), limitations of VHA and future perspectives of application of this assay.

According to the cell-based model, haemostasis is described in three phases [34‐36]: Initiation, amplification and propagation. During initiation, circulating activated coagulation factor (F) VII (FVIIa) forms a complex with exposed TF on injured endothelium, which in the amplification stage generates a small amount of thrombin that mainly activates the platelets. In the propagation phase the coagulation factors assemble on the activated platelets generating large amounts of thrombin ("thrombin burst"). The rate and peak of thrombin generation influences the clot structure and stability [37], by activating FXIII to FXIIIa, which cross links fibrinogen and further stabilizes the clot [38]. Furthermore, thrombin activates TAFI to TAFIa which prevents lysis of the fibrin clot [39].

The three different phases of cell-based haemostasis resulting in clot formation are reflected by the VHA. The structural changes in the clot along the VHA trace was recently investigated by scanning electron microscopy demonstrating that the R (TEG)/CT (ROTEM) corresponds to the initiation phase whereas K (TEG)/CFT (ROTEM) reflects the amplification phase [40]. Our group and others have demonstrated that the thrombin burst is reflected by the α-angle (TEG/ROTEM), and determines the clot strength and stability [41, 42]. The ability of VHA to reflect thrombin generation has profound clinical utility because coagulation factor deficiencies secondary to e.g. massive bleeding, dilution, consumption and thrombocytopenia/pathy result in impaired thrombin generation and impaired clot strength (MA (TEG), MCF (ROTEM)) [30, 43]. The whole blood based VHA, therefore, reveals the contribution of all circulating plasmatic and cellular components, in their actual concentrations, to clot formation [44]. Importantly, enhanced fibrinolysis contributes significantly to bleeding in trauma patients as well as patients undergoing cardiac and liver surgery and patients with obstetric complications, and this condition is readily identified by VHA (Ly (TEG), CL (ROTEM)) [45]. In addition, VHA in vitro studies have evaluated the effects of hypothermia [46], acidosis [47], different crystalloids and colloids [48], pro-haemostatic [49] and anti-fibrinolytic drugs [50], with results being highly relevant for the clinical setting.

In the last 25 years, more than 20 clinical studies reporting on the benefit of using VHA when compared to RCoT to identify coagulopathy and guide transfusion therapy have been published (Table 2). The studies include three randomized clinical trials and involve more than 4,500 patients undergoing major surgery. The majority of studies have been performed in patients undergoing liver or cardiac surgery [51‐57], all reporting of the benefit of using VHA when compared to RCoT, evidenced by reductions in transfusion requirements and need for re-exploration and improved ability to predict the need for blood transfusion in patients with VHA-guided therapy. Importantly, no study have to date reported a benefit of employing plasma-based RCoT to predict bleeding or guide transfusion therapy, when compared to VHA, supporting the scientific rationale of whole blood viscoelastic assays in this setting.

A conserved physiologic haemostatic response, characterized by immediate activation of coagulation and fibrinolysis followed by subsequent fibrinolytic shutdown and later reactivation is often observed in trauma patients [61]. Post-traumatic coagulopathy is classically described as dysfunction and/or consumption of coagulation factors and platelets due to dilution, hypothermia and acidosis i.e., "the bloody viscious cycle" and the ability of VHA to identify these conditions has been extensively reported [22, 23, 58, 62‐70]. At present, 10 studies including more than 700 patients have evaluated VHA in trauma patients (Table 3).

Kaufmann et al. found that in 69 patients with blunt trauma, 65% displayed hypercoagulability upon arrival at the emergency department (ED) whereas only 10% were hypocoagulable. Interestingly, a hypocoagulable TEG was associated with increased ISS and only ISS and VHA, not RCoT, was predictive for early transfusion [62]. Schreiber and colleagues [63] also found that hypercoagulability, as evaluated by VHA, was frequent (62%) in trauma patients (n = 65) upon arrival at the ED, and that this correlated with increased thrombin-antithrombin (TAT) complex generation. APTT, PT and platelet count where within normal limits and could, hence, not identify a hypercoagulable state. Rugeri and colleagues [64] investigated 88 trauma patients and compared their VHA results with that of healthy subjects. They found that trauma patients demonstrated evidence of hypocoagulability, and that this was restricted to those trauma patients also being coagulopathic with RCoT.

Recently, Carroll and colleagues [65] addressed the acute post-traumatic coagulopathy, reported by Brohi et al., [2, 8, 9] by VHA analyses of samples obtained at the scene of accident and upon arrival in the ED in 161 trauma patients. Interestingly, they found that that the clot forming parameters demonstrated hypocoagulability and correlated with fatality, whereas none of the RCoT demonstrated such a correlation. This indicates that VHA is more sensitive in reflecting clinically relevant coagulopathies than RCoT. This has important implications, since the VHA result is available within a short time frame as interventions aiming at normalising the VHA profile and hence the coagulopathy, can be instituted early during resuscitation. The VHA results of acute post-traumatic coagulopathy presented by Carroll et al. do not, however, corroborate the frequency of hyperfibrinolysis reported by Brohi et al. [8]. Only three patients (2%) demonstrated evidence of increased fibrinolysis compared to the hyperfibrinolysis described in the cohort of Brohi, using D-dimer as a marker of fibrinolysis. Levrat and colleagues [66] reported in a cohort of 89 trauma patients that 5 (6%) showed evidence of increased fibrinolysis and that this correlated with euglobuline lysis time. In both the study of Carroll et al. and Levrat et al., hyperfibrinolysis was identified in the most severely injured patients and was associated with increased mortality rate confirming that although rare, this is a very serious condition. A unique feature of VHA is its ability to identify patients with increased fibrinolysis. This enables initiation of specific anti-fibrinolytic therapy, which is associated with decreased blood loss and/or transfusion requirements in non-trauma settings [71]. The role of this therapy in trauma patients [72] is currently under clinical evaluation http://​www.​crash2.​lshtm.​ac.​uk/​.

In a retrospective review of 44 combat patients with penetrating trauma Plotkin et al. [23] reported that VHA was a more accurate indicator of blood product requirements than PT, APTT, and INR. They suggested that VHA aided by platelet count and haematocrit should guide blood transfusion requirements. This is in alignment with Martini and colleagues [22] who demonstrated that VHA was superior than PT, APTT, and Activated Clotting Time in detecting clinically relevant clotting abnormalities after hypothermia, haemorrhagic shock and resuscitation in pigs.

Recently Jaeger and colleagues [67] reported of a modification of the VHA (TEG) where the activator kaolin was substituted with TF (RapidTEG). In patients sustaining major blunt trauma they investigated the time from ED arrival to the results of standard TEG, RapidTEG and RCoT were available. RapidTEG was available significantly faster (19.2 min vs. 29.9 min for kaolin TEG and 34.1 min for RCoT). On average the time until the results were available was reduced by approximately 50% for RapidTEG as compared to standard TEG, which may be of clinical relevance.

Important limitations of the VHA exist and should be taken into consideration when interpreting the results of the analysis. Firstly, though it is possible to adjust the temperature at which the blood sample is analysed, VHA is routinely performed at 37°C and therefore the effect of hypothermia will not be recognised [73, 74]. Secondly, the coagulation activators employed results in thrombin formation, which masks the possible inhibition that antithrombotic agents such as aspirin, NSAID, clopidogrel and eptifibatide may have on the platelets ability to aggregate [75]. Consequently, a normal VHA profile does not rule out clinically significant platelet inhibition. Thirdly, the endothelial contribution to haemostasis is not displayed in VHA and therefore, conditions affecting the endothelium such as von Willebrand disease (vWD, quantitative or qualitative defects in vWF and, hence inability of the platelets to adhere to the endothelium), cannot be investigated. If these causes of abnormal bleeding can be excluded, then a normal VHA trace along with clinically significant bleeding necessitating blood transfusion is suspect of a surgical cause. Thus, our group found 97% predictability by VHA in identifying a surgical cause of bleeding in postoperative non-cardiac patients with ongoing transfusion requirements [68].

Recently, the concept of acute traumatic coagulopathy (ATC) was introduced by Brohi et al. [2, 8, 9, 13] based on the observations that coagulopathy, as evaluated by increased PT, APTT and D-dimer levels, was present in trauma patients already upon arrival to the hospital. ATC was independent of traditional causes of coagulopathy but occurred only in patients with evident hypoperfusion. When evaluating trauma patients upon arrival at ED with VHA characteristic profiles are found that are related to ISS and mortality. In patients with minor trauma/tissue injury a normal VHA trace is seen (Figure 2A) whereas in patients with moderate trauma (ISS between 10-20) hypercoagulability is seen (Figure 2B). In patients with severe injury (ISS 20-35), an increased frequency of hypocoagulability is seen (Figure 2C) whereas patients with massive tissue injury (ISS above 30) hyperfibrinolysis is seen (Figure 2D). The different VHA traces indicate that different treatment strategies may be appropriate and this warrants further investigation.

Death due to trauma is the leading cause of lost life years worldwide, with haemorrhage being responsible for 30-40% of trauma mortality and accounting for almost 50% of the deaths the initial 24 h [5]. There is emerging consensus that plasma-based assays are inappropriate for monitoring coagulopathy and guide transfusion therapy in trauma [17, 18], and the cell-based model of haemostasis [21, 34‐36] provides a reliable explanation for this notion. Clinical studies including more than 5000 surgical and/or trauma patients have reported on the benefit of using VHA when compared to RCoT to identify coagulopathy and guide transfusion therapy. However, at present no VHA guided transfusion therapy has been prospectively and independently validated in trauma patients, which is highly warranted.