Introduction

Trauma, with the resulting hemorrhage and shock, can impact the patient’s physiology on many levels. One of the most dangerous consequences of trauma is the blood-coagulation system malfunction — a complex pathological condition termed trauma-induced coagulopathy (TIC) [1••, 2–4]. Affecting ~ 25% of all trauma patients, TIC is prevalent worldwide in both civilian and military settings [1••]. Some of the factors causing TIC are intrinsically pathophysiologic, while others are iatrogenic — e.g., dilutional coagulopathy caused by the infusion of crystalloid solutions [2]. TIC may lead to severe bleeding (that often cannot be surgically controlled), or it may cause thromboembolic complications; sometimes, both types of effects are present during different temporal phases of TIC [1••]. This creates an extra layer of complexity in trauma surgery and critical care, necessitating both basic and clinical research to understand the mechanistic aspects of TIC and investigate promising pharmacological intervention strategies.

Blood coagulation in mammals is controlled by a very complex system comprising interactions between dozens of biochemical species, platelets, red blood cells (RBCs), endothelial cells, and other components, all occurring in the context of blood flow, as well as blood-vessel and wound geometry [5–7]. These different components can exist in non-activated or activated states, and the conversions between activation levels add to the combinatorial complexity of the interaction network. A central event in the blood-coagulation process is the formation of a hemostatic thrombus (“clot”) at the site of vascular injury (Fig. 1A). The main components of a thrombus are platelets and the polymeric protein fibrin [5, 6, 8]. Fibrin gives the clot the necessary mechanical stability and hydraulic resistance to perform its hemostatic function [9]. As the clot grows, fibrin is generated from its monomeric precursor, fibrinogen, and this conversion (as well as subsequent fibrin polymerization) is catalyzed by the protein thrombin [10–12]. Thrombin itself is generated from its inactive precursor — prothrombin — in response to injury, and performs multiple functions in the coagulation system (including, e.g., platelet activation and factor XIII activation for fibrin crosslinking) [5, 12]. Thrombin is often regarded as the central enzymatic component in the blood-coagulation system [5, 12], and thrombin generation (TG) impairment can play a significant role in traumatic and surgical coagulopathy [13•, 14, 15].

Fig. 1
figure 1

Thrombin generation: essential concepts. A Main thrombin generation (TG) steps. As a result of bleeding, plasma is exposed to tissue factor, which triggers the enzymatic blood-coagulation network. The reactions in the network lead to the cleavage of prothrombin and generation of small amounts of thrombin, which activates platelets and thereby initiates a TG burst on the platelet surfaces. Ultimately, the generated thrombin converts fibrinogen to fibrin, which becomes part of the body of the growing thrombus. B Schematic representation of TG in vitro. Phases 1, 2, and 3 (marked by colored rectangles) are the TG stages: initiation, propagation, and termination, respectively. The typical one-peaked shape of the TG curve is characterized by five quantitative parameters: lag time (LT, time to the onset of TG), time to thrombin peak (ttP), velocity index (VI, which is the maximum slope of the tangent line), thrombin peak height (PH), and ETP (endogenous thrombin potential, which is the area under the thrombin curve) [18•]. The first three parameters may be regarded as the timing parameters, and the remaining two are the amount parameters. C TG modulation by pathological states and therapeutic modalities. Coagulopathic conditions can shift the normal TG curve (blue) toward insufficient or excessive TG (dark red). Balanced therapeutic intervention acts on both types of coagulopathic TG curves and restores approximately normal TG kinetics (green). In B and C, time 0 corresponds to the TG-initiation time; square brackets denote thrombin concentration

Full size image

Thrombin-generation analysis can help researchers and clinicians to obtain mechanistic insights into TIC and potentially diagnose it [12, 16••]. Furthermore, biochemical components of the TG network could serve as pharmacologic agents or targets for intervention. Here, we review the recent literature on the role of thrombin and related mechanisms in traumatic bleeding, TIC, and their mitigation. Besides TIC, we touch upon the related phenomenon of surgical coagulopathy. Performing surgery on trauma patients is a frequent challenge for anesthesiologists [3]; moreover, the surgery itself may be considered a special case of (controlled) trauma.

A crucial aspect of physiologic TG is closely linked to the notion of hemostatic balance [17]. This balance is achieved when blood coagulation at the injury site proceeds normally and leads to the cessation of bleeding (i.e., to hemostasis) without excessive intravascular clot formation. In TIC, however, the balance is disrupted [1••]. Therefore, when attempting pharmacologic correction of blood coagulation, it is essential to shift the coagulation process toward its normal state and avoid pushing it too far “to the other side” (i.e., shifting from bleeding to thrombosis and vice versa). We discuss how to avoid the Scylla of hemorrhage and the Charybdis of thrombosis through balanced TG restoration.

The Fundamentals of Thrombin Generation and Coagulopathy Factors

Thrombin generation is initiated when blood flowing from a broken vessel comes in contact with the protein termed tissue factor, which is expressed outside the vascular space. A leading role in this process is played by the enzymatic network triggered by tissue-factor binding; the network propagates enzyme-activation events until thrombin activation is achieved [5]. The main proteins promoting TG — termed procoagulants — are coagulation factors (F) II (prothrombin), FV, FVII, FVIII, FIX, and FX [5]; FXI and FXII may also play some role [7]. It should be noted that the entire protein network responsible for TG, fibrin formation, and fibrinolysis (i.e., biologically controlled fibrin degradation) is very complex and involves enzyme activation and inhibition, protein binding, cofactor action, and protein degradation and contains positive and negative regulatory feedback loops. As an illustration of this complexity, a state-of-the-art mechanistic model of these processes contains 107 protein species and 120 biochemical reactions [18•].

When investigated via in vitro experiments, TG occurs in three phases: initiation, propagation, and termination [5, 19] (alternatively, initiation, amplification, and propagation [7, 12]) (Fig. 1B). Briefly, during initiation, minute amounts of thrombin are generated in blood plasma via enzymatic reactions. Next, localized TG triggers platelet activation, leading to a TG burst on the platelet surface via a positive feedback loop (i.e., the propagation phase). Finally, in the termination phase, thrombin activity is inhibited by antithrombin (a potent direct thrombin inhibitor) and other inhibition mechanisms. Typically, thrombin concentration changes over time as a one-peaked curve (Fig. 1B). To analyze such curves and their changes in a systematic way, researchers look at quantitative TG parameters [18•]. There are five commonly used parameters (nomenclatures vary between test systems): lag time, time to thrombin peak, thrombin peak height, maximum slope of the curve (velocity index), and the area under the curve (endogenous thrombin potential) (Fig. 1B). They represent both the timing of TG and its amount across the three TG phases.

This review is focused on the generation of thrombin. However, other elements of blood coagulation are also essential, and in vivo clot growth can be modulated by any of the following: platelet count and function, polymerization and degradation of fibrin, endothelial activation, blood-flow geometry, and immune-cell activity [1••, 5, 7]. Multiple mechanisms limit excessive clot growth, but most frequently discussed are three endogenous anticoagulants: antithrombin, tissue factor pathway inhibitor (TFPI), and protein C [5]. The existence of these and other control mechanisms in the mammalian blood-coagulation system alludes to the importance of thrombin-level regulation of hemostasis.

TIC and surgical coagulopathy affect the mechanistic aspects of blood coagulation through various coagulopathy factors (i.e., mechanistic drivers of coagulopathy; not to be confused with coagulation factors). The traditional view of coagulopathy emphasized three coagulopathy factors: hemodilution, hypothermia, and acidosis [2]. However, it has become clear that coagulopathic (specifically, prohemorrhagic) conditions can be detected before the onset of, and independently from, the above three factors. This early coagulopathy is termed acute traumatic coagulopathy [20]. Active research of that condition implicated the involvement of protein C as a primary mechanism, but alternative views of that phenomenon exist [21, 22•, 23, 24]. Increased fibrinolysis plays a significant role in coagulopathy [22•, 25], and so does the depletion of specific coagulation factors, such as FV and FVIII [26, 27]. Platelet dysfunction and deficiency, as well as fibrinogen deficiency, can exacerbate coagulopathy [28]. Besides platelets, endothelial cells actively regulate the clot-formation environment, and their dysfunction (endotheliopathy) can contribute to TIC [29]. Thus, many of these (and other) coagulopathy factors can potentially impact TG. Importantly, different coagulopathy factors can be present simultaneously and be manifested to varying degrees in a given TIC case (multifactorial coagulopathy) [2]. Moreover, traumatic coagulopathy is often present in distinct phenotypes, with distinct coagulopathy factors dominant in different patient groups [30•, 31•, 32, 33]. Ultimately, our ability to mitigate coagulopathy is determined by our knowledge of how various coagulopathy factors and their combinations affect different aspects of blood clotting, including TG.

The multifactorial nature of TIC makes it difficult to come up with a single intervention or treatment algorithm applicable to a heterogeneous population of patients. Combined with the natural intersubject variability, the variability in coagulopathy-factor combinations can result in a broad range of phenotypic responses and outcomes. In the presence of this variability, injury severity appears to be a common denominator, with increased injury severity typically leading to more pronounced coagulopathic states [1••, 2]. Severity of the injury can be assessed at bedside using hemodynamic parameters and stat laboratory test results (e.g., lactate) along with pre-defined injury scores and the need for massive transfusion. These clinical parameters may not be easily incorporated into laboratory models of TIC research, but the recent studies have adopted computational, in vitro, and in vivo (mouse) models to investigate the effects of blood-vessel injury size on thrombus growth [34, 35, 36•]. While these studies do not address coagulopathy, they provide a conceptual foundation for future mechanistic research on the effects of coagulation-trigger strength under coagulopathic conditions.

Methods of Thrombin-Generation Measurement

Thrombin generation is measured in vitro, typically in samples collected from human subjects or experimental animals [37, 38]. The samples contain whole blood or plasma and are sometimes altered to represent specific pathological conditions. Thrombin generation is usually initiated by supplementing a reagent containing tissue factor (or a recalcifying agent, if tissue factor is already included in the system), which reflects blood exposure to this protein upon initiation of bleeding [38]. An alternative measurement strategy involves measuring prothrombin-cleavage fragments in the subject’s blood samples, which may be interpreted as characteristic of ongoing TG in the subject’s blood due to (unspecified) intrinsic causes [39, 40].

Conceptually, measuring TG in minimally altered whole blood should be preferable because the latter incorporates all cellular and humoral elements [41]. However, the first whole-blood TG measurements in trauma research were cumbersome (used manual subsampling) and measured the thrombin–antithrombin complex rather than thrombin itself [42, 43]. A different methodological solution relied on fluorogenic thrombin substrates in plasma, which allowed for more efficient, automated thrombin-kinetics measurements. That method, named calibrated automated thrombinoscopy (CAT), has become widely spread [44]. Most CAT applications in trauma and anesthesiology research use platelet-poor plasma, wherein platelets are replaced with externally added phospholipids; the latter form micelles supporting the enzymatic reactions leading to TG. Platelet-poor plasma has often been the preferred choice in TG measurements because of simplified assay standardization and data comparisons. Yet, the CAT assay is also applicable to platelet-rich plasma [38]. Recently, the automated assay was developed further to measure TG in whole blood; the use of this assay — including cases of trauma and coagulopathy — is being actively explored [45•, 46, 47, 48•, 49•]. Moreover, in the form of microfluidic devices, experimental blood-flow models have emerged as a new blood-coagulation investigation strategy [50–52].

To what extent do in vitro TG measurements characterize in vivo TIC? Can we diagnose coagulopathy using TG assays? Currently, the use of TG assays, such as the CAT, for diagnostic purposes in trauma and surgery is problematic and precluded for at least two reasons: assay variability and difficulty of use in clinical settings [38, 53]. Thus, more traditional methods are utilized for coagulopathy diagnostics, such as prothrombin time and viscoelastic coagulation tests [4, 54]. Consequently, in trauma and surgery, TG assays are used predominantly as research tools [53]. Yet, the sensitivity of TG suggests that it could eventually become an informative means of clinical diagnostics [38, 53]. This is particularly true for the recently emerged whole-blood TG measurement technology [16••]. The development of TG-based clinical assays is desirable because that would allow one to verify that the TG-targeting therapeutic interventions are genuinely balanced. A recent example of TG use to guide anticoagulant therapy suggests that this goal may be within reach [55•].

Two challenges complicate the potential of TG as a clinical diagnostic assay. One of them is the intersubject TG variability [18•], which might complicate the diagnostics of coagulopathy based on TG parameters alone. Future development of personalized medicine may provide access to normal TG parameters for each patient, facilitating informative diagnostic comparisons with the subject-specific baseline. Yet, this future currently appears to be remote. In the meantime, TG may need to be used in conjunction with other assays to ascertain reliable diagnostic results. The second challenge is related to TG being an in vitro test. For example, a recent study showed that TG could be preserved even if the plasma is 95% diluted [56]. However, the applicability of this conclusion to clinical situations is doubtful. Deep research is needed to assess the full potential of TG as a point-of-care assay.

Computational Modeling

Complementary to the traditional experimental models of thrombin and fibrin generation in vitro, mechanistic computational modeling has become an essential part of the systems biology concept in blood-coagulation and coagulopathy research [41, 57, 58]. Unlike the statistical models ubiquitous in biomedical literature, these computational models typically consist of time-dependent equations that reflect the production, depletion, binding, activation, and interconversion between the molecular and cellular components represented in the model. Thus, such models can provide a direct link between an observed biological response and its mechanistic underpinnings. The need for computational modeling arises due to the intrinsic complexity of the blood coagulation system, which makes it challenging to generate hypotheses about the system behavior using intuition alone. Computational models facilitate rapid, quantitative hypothesis generation via “virtual screenings” and guide laboratory experimentation and perhaps even clinical studies. Most of such computational models are based on nonlinear ordinary differential equations [18•, 59•, 60••, 61••, 62•, 63, 64, 65•], but algebraic equations [66, 67] and computational fluid dynamics [50, 57, 68•, 69] have also been used. Computational modeling plays a key role in several of the studies cited in this review.

Trauma-induced coagulopathy is not fully understood, and different (often competing) theories exist about its predominant mechanisms [1••, 2]. Experimental research and computational modeling, including TG studies, provide opportunities for focused studies of single or multiple known coagulopathy factors, such as plasma dilution, depletion of specific coagulation proteins, and hypothermia. As mechanistic information accumulates, the models become more accurate, more comprehensive, and increasingly useful in hypothesis generation. These developments hold the promise of providing tools that will ultimately help to unravel the complete mechanistic picture of TIC.

Most traditional computational-modeling approaches focus on TG biochemistry, sometimes combined with blood flow. However, besides the enzymatic reactions, TG (and blood coagulation in general) depends on other factors, such as platelet-activation kinetics and endothelial cells. To reflect this, the new generation of computational models includes additional features, such as platelet action [57] and (hyper)fibrinolysis [68•, 70]. The simplifying assumptions used in model development can be regarded as the necessary first step toward more realistic representations of biology. However, the clinical reality is much more complex, especially in patients with comorbidities and those taking medications (such as antiplatelet drugs) [2]. In such situations, the effects of coagulopathy per se on TG may be particularly difficult to tease out. In any case, the models are mainly intended to streamline laboratory research and are not directly applicable to clinical practice. Therefore, their main benefit is comparable to the discovery and preclinical phases of the typical drug-development process.

Impact of Trauma-Induced Coagulopathy on Thrombin Generation

There are two distinct approaches to investigate the impact of coagulopathy on TG in human blood. One involves using blood (or plasma) samples taken directly from trauma patients. Thrombin generation measured in such samples can be ascribed to TIC superimposed with the individual TG profile of a given patient. However, because the patient’s TG is not measured pre-trauma, there is no baseline to tease out the patient-specific contributions of coagulopathy to the post-trauma TG. (If coagulopathy is expected to result from a planned surgery, then pre-surgery baseline measurements are feasible and should be made.) Moreover, because TIC is often multifactorial, and the information about the severity of each contributing coagulopathy factor may not be readily available, it can be challenging to derive mechanistic conclusions from such measurements.

Another approach is creating coagulopathic states in the test tube — e.g., diluting, cooling, and acidifying plasma or whole blood collected from healthy volunteers. In this analytic approach, we know the contributing coagulopathy factors (and their severities), and subject-specific baseline data are available. However, it can hardly be claimed that such “controlled coagulopathy” fully captures the impact of trauma on blood and its coagulation. The two approaches are complementary, and ultimately both are needed. A third approach, where controlled coagulopathy is induced in an animal, has also been explored [43, 71••, 72]. However, data obtained from animal models cannot be directly used for clinical management in humans due to species-specificity considerations [73, 74].

Initial applications of the CAT assay indicated increased TG in the trauma-patient plasma compared to the normal control group [19, 75]. Other studies also reported elevated TG in trauma but showed that more severely injured and coagulopathic patients had reduced TG compared to the less severely injured ones [76••, 77]. These results can be understood if we posit that, while the occurrence of injury per se increases the TG potential in patient plasma, the onset of coagulopathy (particularly in severe cases) leads to its relative reduction. This notion is supported by a very recent study using a prototype point-of-care device to measure TG in trauma-patient whole blood [16••]. In that study, TG in trauma patients treated as one group exceeded that in the healthy control group; however, in the subgroup of trauma patients with shock and those requiring early massive transfusion, TG was reduced, with thrombin-peak height falling below the healthy-control level. Another recent study, using a different experimental assay, concluded that TG in trauma patients was adequate despite the presence of acute traumatic coagulopathy [40]. Thus, it is plausible that, under the conditions of the study, the coagulopathy phenotype did not affect TG, which happens in some animal trauma models [72]. Alternatively, the trauma-induced TG increase and the coagulopathy-induced TG decrease could have offset each other.

Cardiac surgery — particularly the use of a cardiopulmonary bypass — typically reduces thrombin generation [14, 59•, 65•, 78]. However, in certain situations, thromboembolic complications (apparently associated with uncontrolled thrombin activity) are possible [79, 80]. Remarkably, post-bypass TG can be delayed but have a higher peak than pre-bypass TG, rendering the bypass effects on TG equivocal [81]. As a different example, analysis of TG during liver transplants shows a time-dependent TG-impairment pattern [82•]. Taken together, coagulopathy and TG impairment are relatively common in certain types of complex surgery, but the diagnostic utility of TG has not been firmly established.

A detailed computational and experimental study of controlled multifactorial coagulopathy induced in vitro demonstrated that dilution with normal saline reduces plasma TG, consistent with many prior in vitro studies [18•]. At the same time, it is known that the quantitative effects of dilution on TG can depend on the diluent [42]. Moreover, under certain conditions, TG remains sufficient despite the dilution [56]. The effects of hypothermia on TG are somewhat ambivalent. When assessed by measuring the thrombin–antithrombin complex in whole blood, hypothermia decreases the TG rate [43, 83]. However, when TG is measured in human plasma using the CAT, hypothermia delays the onset of the thrombin peak (consistent with the whole-blood results) but robustly increases the area under the thrombin curve [18•]. What is needed now is to better understand how these in vitro effects translate to in vivo traumatic coagulopathy. For example, the abovementioned ambivalence of hypothermia effects on in vitro TG may suggest that hypothermia exacerbates both the early prohemorrhagic and late prothrombotic TIC phenotypes [18•].

Computational modeling predicted a significant plasma-TG reduction under acidosis [18•], consistent with earlier whole-blood measurements in a porcine model of coagulopathy [43]. However, the corresponding plasma CAT experiment showed no significant effect of acidosis on TG [18•]. An independent study of a synthetic coagulation proteome showed a moderate acidosis-induced late-TG increase, and this result was correlated with reduced inhibitory activity of antithrombin [84]. Recent laboratory studies underscore the technical challenges of rigorous assessment of acidosis effects in vitro [85•, 86, 87]. In summary, a mechanistic understanding of the effects of acidosis on TG, as well as its clinical consequences, requires further in-depth investigations [88•].

Restoration of Thrombin Generation in Coagulopathy

Because TG can be insufficient or excessive in trauma and surgery, it is necessary to develop robust measures for therapeutic TG improvement. Naturally, this improvement should aim to shift TG closer to its normal state but avoid shifting it too far. In that sense, TG restoration should be balanced [17, 89].

It has been suggested that to achieve balanced restoration of TG — and of thrombus formation in general — the pharmacologic intervention must itself be balanced (Fig. 1C). This conclusion was reached in studies of a simplified in vitro system reflecting the possible use of prothrombin complex concentrates (PCCs), which are combinations of purified pro- and anticoagulant proteins [61••, 90]. Such combinations have recently emerged as potential next-generation therapeutics for traumatic and surgical coagulopathy, and their optimal composition is a research topic [91]. Currently available PCC products contain therapeutic amounts of vitamin K–dependent coagulation factors, but they contain notably below-therapeutic amounts of antithrombin [91]. This is logical by design, because most PCCs are generally intended to be used for defined procoagulant-factor deficiency states, including warfarin therapy and hereditary prothrombin deficiency. In addition, PCCs are increasingly used to treat bleeding conditions associated with direct factor Xa and thrombin inhibitors — such as rivaroxaban and dabigatran — when direct antidotes are unavailable [92]. The main purpose of using PCCs in the current clinical practice is to enhance TG, and thus considerations for regulating TG have been largely omitted [93].

Computational modeling with subsequent in vitro experiments suggested that the combination of the procoagulant factors FII, FIX, and FX with a sufficiently large amount of antithrombin can accurately restore TG in plasma diluted with normal saline [61••, 90]. The benefits of combining procoagulants with antithrombin were also highlighted in an independent computational study, where the diluted-plasma composition data had been taken directly from patients on cardiopulmonary bypass [59•]. Traditional coagulation tests, including prothrombin time and thromboelastography, have focused on how fast clotting takes place, but not on the balance of coagulation. If factor Xa and thrombin are not inhibited by antithrombin in a timely manner, there is an increased risk of intravascular thrombus formation [79, 80]. Thus, in the setting of highly complex multifactorial- coagulopathy reduction [18•], it may be logical to supplement both pro- and anticoagulant proteins.

As a next step, it would be informative to explore the efficacy and safety of the balanced-intervention therapeutic paradigm under more complex coagulopathy scenarios, e.g., including hypothermia and acidosis. Computations using an existing model of controlled multifactorial coagulopathy can guide such analysis in the presence of intersubject variability [18•]. Reassuringly, a recent study using a porcine model of trauma concluded that combining PCCs with additional antithrombin improves therapeutic outcomes and reduces the occurrence of thromboembolic events [71••]. The published retrospective study of patients with TIC using the US national Trauma Quality Improvement Program (TQIP) data set over 2 years suggested that the cohort who received fresh frozen plasma (FFP) plus PCC early (< 60 min of the admission) seemed to have lower transfusion requirements in RBCs and FFP when compared to the propensity-matched cohort who received FFP without PCC [94••]. The authors’ secondary analyses suggested a lower in-hospital mortality was associated with FFP plus PCC compared to FFP alone (17.5% vs. 27.7%, P = 0.01) without increasing the risk of deep-vein thrombosis and pulmonary thromboembolism. However, more clinical data are needed to fully assess the safety and efficacy of PCC, as well as PCC in combination with other agents.

There are some clinical data suggesting rebound hypercoagulability and thrombotic potential following the administration of PCCs in trauma patients [95•, 96••]. The use of antithrombin as a primary anticoagulant in a PCC formulation is theoretically supported by its potency as a natural anticoagulant [61••] in the setting of acquired AT deficiency due to traumatic bleeding and surgery [97•, 98]. This concept is further supported by computational modeling and in vitro measurements [61••, 65•, 90, 99]. Further clinical evaluations of antithrombin supplementation are warranted to assess its efficacy to reduce the incidence of venous thromboembolism in trauma [100].

In the context of PCC use to enhance TG, of note is the interaction between thrombin and fibrin. Clinically, fibrinogen supplementation is often used to mitigate coagulopathy by improving thrombus formation [101••]. At the same time, in vitro assays indicate that fibrin reduces thrombin activity by binding with it [102]. However, in other in vitro assays, the resulting decrease in thrombin activity is temporary and is more than compensated by the subsequent TG increase [103]. It is unknown how this would translate to in vivo situations, but computational analyses suggest that both PCC and fibrinogen should be supplemented for simultaneous normalization of thrombin and fibrin levels in coagulopathy [50, 70].

Logically, a balanced therapeutic strategy involving both pro- and anticoagulants could potentially be used for both insufficient and excessive TG (Fig. 1C). While early TIC (hours after trauma) is typically associated with a prohemorrhagic phenotype, late coagulopathy (2–3 days after trauma) may shift the phenotype to a prothrombotic state [1••]. Overall, it is often unclear how the bleeding tendency will develop over time after trauma. Therefore, anticoagulants may need to be administered concurrently with, or following, procoagulant administration. Illustrating this notion, the ability to reduce the risk of both excessive and insufficient TG has been reported for FFP [104••].

FFP has been, and remains, the standard therapeutic strategy to treat traumatic bleeding and coagulopathy, particularly so in the USA [105, 106]. Because FFP is expected to contain all the biochemical blood-coagulation components naturally present in plasma, it may be regarded as an intrinsically balanced intervention to restore normal TG. (Due to its molecular composition, FFP can influence blood coagulation components other than thrombin. For example, it may impact fibrinogen availability and fibrinolysis control.) The same principle is expected to hold for a related product, lyophilized plasma [107]. More generally, the 1:1:1 (plasma:platelets:RBCs) resuscitation paradigm extends the concept of balanced TG to include platelets required for TG enzymes to be fully functional [3, 5].

Despite the perceived advantage of FFP in restoring both procoagulant and anticoagulant proteins, the concentrations and activity status of the coagulation proteins in FFP can be variable; other complications linked to FFP use may also arise [108–110]. To find robust alternatives to FFP, active research is ongoing to compare the efficacy and safety of PCCs with those of FFP in trauma and surgery [111]. Some European centers are starting to use PCC instead of FFP. Indeed, PCC appears to reduce blood-transfusion requirements in coronary surgery [112•], which is consistent with the results of a propensity score–matched cohort study of PCC use in cardiac surgery [113••]. PCCs are now included in the European trauma and surgery recommendations [101••, 114]. Two very recent pilot clinical trials have suggested the feasibility of using PCC as an alternative to FFP [115, 116], and another randomized trial suggested the superiority of PCCs [117••].

A re-emerging paradigm involves using whole blood rather than FFP to mitigate traumatic coagulopathy [118, 119•, 120, 121]. This approach, which has a long history in (mainly military) trauma [120], may be another step in developing the balanced-intervention concept. Published reports indicate similar or improved outcomes when using whole blood compared to blood components [119•, 120, 122]. However, there are also some reservations [1••]. It is possible that, just like the existence of distinct phenotypes in coagulopathy, specific therapeutic modalities (PCCs, FFP, whole blood, etc.) will prove to be advantageous in different traumatic and surgical scenarios [123].

As a final illustration of the concept of balance, here is an example of an unbalanced therapeutic strategy. Recombinant activated FVII (commonly abbreviated as rFVIIa) is a strong procoagulant that was previously considered a promising drug for traumatic bleeding; later its benefits were reconsidered due to accumulating evidence of thromboembolic complications [124]. Both computational and experimental in vitro research on TG in diluted plasma suggested that rFVIIa brings about an intrinsically unbalanced TG modulation. Indeed, it accelerates TG, but the resultant TG profile remains distorted and does not approximate a normal TG time course [61••, 65•, 90]. It is plausible that this distortion is responsible for the high probability of thromboembolism associated with this drug, which came to be regarded only as a means of last resort in traumatic bleeding [125, 126].

Furthermore, it is plausible that a similar type of TG distortion may also characterize PCC formulations lacking anticoagulant components. While it is conceivable that TG may be increased by careful PCC dosing (without supplementing antithrombin) [96••], it may be challenging to achieve a robust balance between insufficient and excessive TG using the dose of procoagulants as the only tunable parameter. Moreover, supplementing antithrombin with PCC may provide a robust solution for the late-phase coagulopathy with a prothrombotic phenotype.

Conclusion

The pivotal roles of TG in hemostasis and pathological thrombosis have been extensively explored through various techniques, including plasma marker measurements, TG assays, and computational modeling. The latter technique is relatively new, but it has provided significant insights into optimal regulation of TG and further development of balanced therapeutic strategies. There are yet many unanswered questions about the applicability of computational TG models to in vivo coagulation. However, computational models are increasingly improved to capture platelet-vascular interactions and platelet-surface TG reactions. These newer models are likely to fill the gap between in silico and in vivo, and they will fuel the future growth of this research field in the years to come.