Detecting Traumatic Brain Injury-Induced Coagulopathy: What We Are Testing and What We Are Not

Abstract

Coagulopathy after traumatic brain injury (TBI) is common and has been closely associated with poor clinical outcomes for the affected patients. TBI-induced coagulopathy (TBI-IC) is consumptive in nature and evolves rapidly from an injury-induced hypercoagulable state. TBI-IC defined by laboratory tests is significantly more frequent than clinical coagulopathy, which often manifests as secondary, recurrent, or delayed intracranial or intracerebral hemorrhage. This disparity between laboratory and clinical coagulopathies has hindered progress in understanding the pathogenesis of TBI-IC and developing more accurate and predictive tests for this severe TBI complication. In this review, we discuss laboratory tests used in clinical and research studies to define TBI-IC, with specific emphasis on what the tests detect and what they do not. We also offer perspective on developing more accurate and predictive tests for this severe TBI complication.

Level of evidence: published reports

Uncontrolled hemorrhage accounts for 30–50% of all trauma fatalities^1,2 and is caused by direct injury to the vasculature and by secondary coagulopathy. Trauma-induced coagulopathy can occur minutes after injury in a prehospital setting and progress rapidly, significantly increasing the risk of secondary bleeding, disability, and death.^3,4 Laboratory and clinical studies have consistently shown that coagulopathy is also common in patients with isolated traumatic brain injury (TBI). TBI-induced coagulopathy (TBI-IC) is a bleeding diathesis that can result in secondary, delayed, or recurrent intracranial or intracerebral hemorrhage. While the clinical manifestations of coagulopathy can be definitively diagnosed by computerized tomography (CT) or magnetic resonance imaging (MRI) scans, laboratory tests used to define hemostatic dysregulations of TBI vary among studies, detect individual aspects of the multistep hemostasis system, and have not been standardized. The finding that the incidence of laboratory-defined TBI-IC is significantly higher than that of clinical coagulopathy suggests that hemostatic dysregulation detected by these tests may not always result in significant clinical coagulopathy. The disparity can be addressed by understanding the nature of TBI-IC and by determining what these tests detect and what they do not.

THE NATURE OF TBI-IC

Common causes of coagulopathy after severe trauma to the body and limbs include significant blood loss (hemorrhagic shock), hemodilution and hypothermia due to fluid resuscitation, systemic metabolic acidosis due to tissue ischemia, dysfunctional platelets and coagulation, and hyperfibrinolysis.^5,6 However, patients with isolated TBI do not suffer significant blood loss, are restricted in fluid resuscitation to prevent cerebral edema and high intracranial pressure, rarely develop hypothermia, and are unlikely to develop systemic metabolic acidosis, suggesting that TBI-IC is mechanistically distinct from deficient and dilutional coagulopathy arising after extracranial trauma and hemorrhagic shock.

When blood samples longitudinally collected in the early hours of TBI are analyzed for coagulation, the fibrinolytic product D-dimer and fibrinogen degradation products (FDPs) are detected within minutes of injury and followed first by a profound depletion of fibrinogen and then by prolonged prothrombin and partial thromboplastin times,^7,8 indicating that TBI-IC is consumptive in nature and is developed through a rapid transition from a trauma-induced hypercoagulable state to a hypocoagulable state.^9,10 This consumptive TBI-IC has been reproduced in mice subjected to lateral fluid percussion injury.^11,12 A critical question is how a localized brain injury is rapidly disseminated to alter the systemic hemostasis system, as all laboratory tests for TBI-IC measure changes in the peripheral blood samples from TBI patients. Brain-derived extracellular vesicles (BDEVs) have been identified as a key disseminating and causal factor for TBI-IC.^11,13 These BDEVs express tissue factor (TF) and abundant anionic phospholipids such as phosphatidylserine (PS).^11,57 Upon release into the circulation, TF- and PS-expressing BDEVs serve as circulating platforms on which extrinsic coagulation is initiated and propagated, thus consuming the coagulation factors.¹⁴ BDEVs, especially metabolically active extracellular mitochondria, which are classified as a subtype of EVs, also activate platelets and endothelial cells to express PS and to release procoagulant EVs.^12,15 The findings from these mouse studies are consistent with clinical reports,^16–18 but controlled clinical trials are needed to determine whether these EVs can serve as predictive and diagnostic markers for TBI-IC. Figure 1 summarizes changes in hemostasis that contribute to the initiation and propagation of TBI-IC.

Figure 1:

A schematic illustration of underlying mechanisms that contribute to the pathogenesis of TBI-induced systemic consumptive coagulopathy. We specifically highlight the contribution of brain-derived extracellular vesicles (BDEVs) and potential soluble factors released from injured brain to TBI-IC by inducing endothelial injury, platelet activation, and a hypercoagulable state (TBI: traumatic brain injury, CBF: cerebral blood flow, ICP: intracranial pressure, and TBI-IC: traumatic brain injury-induced coagulopathy).

TESTS FOR TBI-IC

The prevalence of laboratory-defined TBI-IC varies considerably in published reports, ranging from 10% to 97.2%.^7,19,20 Several factors contribute to this large variation. First, TBI is a collective term for several types of injury, each of which is associated with a differential risk for TBI-IC. For example, coagulopathy is more common in patients with penetrating TBI than in those with blunt injury.^20,21 Patients with cerebral parenchymal injuries (e.g., contusions) are more likely to develop TBI-IC than those with compressing injuries such as epidural hematoma. Second, TBI-IC may occur in more than 80% of patients with severe TBI but is found in less than 1% of patients with mild TBI.^22,23 Third, TBI-IC was defined in these reports by different and non-standardized laboratory tests.^7,19,20 These tests and the timing with which they are performed will define abnormalities in different components of the hemostasis.⁷

The human trauma hemostasis is composed of five distinct but interdependent components: the endothelium, platelets and adhesive ligands, coagulation, fibrinolysis, and vascular tone (e.g., vasoconstriction). Trauma hemostasis is initiated when platelets adhere to the subendothelium exposed at the site of vascular injury (endothelial injury) and become activated and aggregated to seal the wound. This platelet plug is rapidly stabilized by cross-linked fibrin derived from fibrinogen that is cleaved by thrombin, which is the final product of coagulation. When primary hemostasis is achieved, fibrinolysis is initiated to reestablish blood flow. Vasoconstriction is also an integral part of hemostasis by reducing blood flow to promote platelet adhesion at the site of vascular injury. The involvement of individual hemostatic components is sequential but partially overlapping. Laboratory tests often detect changes in the specific component of hemostasis that is dominant at the time of testing. These time-dependent changes in individual hemostatic components explain why the prevalence of TBI-IC defined by these tests varies significantly among studies. Here, we discuss common laboratory tests related to the five components of hemostasis, with emphasis on the targets and limitations of these tests in defining laboratory vs. clinical TBI-IC.

Tests for endothelial injury

In its resting state, the endothelium provides an anti-coagulant and platelet-repellent surface; however, when activated or damaged by traumatic, ischemic, or inflammatory insults, the endothelium becomes highly procoagulant to attract platelets and leukocytes. TBI-induced endotheliopathy has been extensively reported as the cause of vascular permeability and tissue edema. More importantly, it is increasingly recognized that some of the biomarkers identified for endothelial injury also regulate coagulation and propagate coagulopathy. For example, plasma levels of thrombomodulin²⁴ and endothelial protein C receptor indicate the shedding of these molecules from the endothelial surface and also the resultant loss of the endothelial anti-coagulant capability. The shedding of syndecan-1 indicates the depletion of protective glycocalyx, also making the endothelium more procoagulant.²⁵ A study found that patients with TBI-IC had lower plasma levels of soluble thrombomodulin and higher levels of soluble syndecan-1, as compared to patients without coagulopathy.²⁶ EV released from injured endothelial cells are also increasingly used as a marker for plasma procoagulant activity because they are associated with endothelial injury and also express procoagulant activity.^16,27,28 The question is whether these markers for endothelial injury also predict TBI-IC or define its severity.

Tests for platelets and adhesive ligands

Platelets have moderately low counts, enhanced activation and aggregation, and high procoagulant activity in TBI patients and in animal models of TBI.^29,30 However, how platelets are activated or become hyperreactive has not been studied mechanistically. A not fully explained observation is that platelets from TBI patients respond poorly to adenosine diphosphate and arachidonic acid independent of platelet counts, hemorrhagic shock, and tissue hypoperfusion.^29–32 This selective hypo-response does not appear to be caused by granule depletion of activated platelets,³³ suggesting a new TBI-induced platelet phenotype that remains poorly defined. More importantly, the predictive and diagnostic values of platelet counts and activation states for clinical TBI-IC remain largely unknown.

Despite extensive evidence of platelet activation, the role of adhesive ligands that activate platelets in acute TBI-IC is only beginning to emerge.²⁷ For example, the adhesive ligand von Willebrand factor (VWF) is secreted in a significant amount in patients with trauma including TBI and is associated with poor outcomes.^34,35 More importantly, VWF increases not only its antigen, but also its adhesive activity during acute TBI, as demonstrated by its active conformation³⁶ and enhanced binding to platelets.²⁷ This VWF hyperadhesive activity is attributed to the kinetic deficiency of the metalloprotease ADAMTS-13 (A Disintegrin And Metalloproteinase with a ThromboSpondin type 1 motif, member 13) because TBI patients often have a drastically increased VWF and moderately reduced ADAMTS-13,^37–39 resulting in a drastically reduced ADAMTS-13-to-VWF ratio. This kinetic ADAMTS-13 deficiency slows the rate of VWF cleavage by the metalloprotease, which reduce the adhesive activity of hyperadhesive ultra-large VWF multimers freshly released from activated or injured endothelial cells. The TBI-induced systemic oxidative stress further slows the cleavage by oxidizing VWF and ADAMTS-13, making the former resistant to cleavage⁴⁰ and the latter less active in cleaving VWF.⁴⁰ Consistent with this notion, low ADAMTS-13 activity is associated with coagulopathy; endothelial damage; and mortality in patients with severe trauma,³⁴ including TBI.⁴¹ In mice subjected to severe TBI, exogenous ADAMTS-13 reduces TBI-IC and improves neurological function and survival without impairing the baseline hemostasis.²⁷ However, changes in VWF and ADAMTS-13 have not been integrated into a predictive and diagnostic model for TBI-IC.

Tests for coagulation

The current laboratory tests for TBI-IC are overwhelmingly focused on coagulation. Among these tests, prothrombin time (PT) and its derivative, the international normalization ratio (INR), are widely used in clinical settings with defined threshold levels for coagulopathy (INR≥1.2.−1.3). PT is relevant to vascular injury-induced hemostasis because it measures coagulation initiated through the extrinsic pathway, which is initiated by the integral membrane lipoprotein TF. Tests related to PT but less used in defining TBI-IC are partial thromboplastin time (PTT) and activated partial thromboplastin time (aPTT), which primarily measure the state of the intrinsic pathway (contact pathway). While TBI triggers primarily the extrinsic pathway (tissue factor pathway), the intrinsic pathway is often also activated in trauma and TBI settings to propagate or sustain the level of coagulation initiated by the extrinsic pathway. However, these assays were originally developed to detect bleeding diathesis caused by deficiency in individual coagulation factors; they are therefore less accurate in identifying consumptive coagulopathy and are limited in predicting when TBI-IC will occur and what its severity will be.

Fibrinogen has also been widely used as a marker of TBI-IC, often in combination with PT/INR. This adhesive ligand binds and crosslinks platelets through the activated integrin αIIbβ3 to form a platelet plug, which is stabilized after the platelet-bound fibrinogen is cleaved by thrombin to form a fibrin network.⁴² Plasma levels of fibrinogen are reduced to various extents in TBI patients, especially those with coagulopathy,^43–45 consistent with consumptive coagulopathy. In contrast, changes in FDPs are rarely reported in patients with TBI-IC.⁴⁶ This is surprising because FDPs are closely associated with the rate of fibrinogen cleavage by thrombin and fibrin formation. Nevertheless, levels of plasma fibrinogen and FDPs may accurately measure the rate of intravascular coagulation, but neither has been validated for predicting TBI-IC.

Compared to the number of studies on TBI-induced coagulation, there are far fewer studies on how intrinsic anticoagulants regulate the development of TBI-IC. For example, there is little information regarding the role of tissue factor pathway inhibitor (TFPI), which is the primary inhibitor of the coagulation initiation, in acute TBI-IC. In a prospective study of 120 patients with severe isolated TBI, plasma levels of TFPI were found to be low in patients with coagulopathy, but this low TFPI level was not associated with the state of coagulopathy or clinical outcomes of the patients.⁴⁷ The serine protease inhibitor antithrombin is a small glycoprotein that forms a complex with thrombin to block the proteolytic conversion of fibrinogen to fibrin. The thrombin-antithrombin complex in peripheral blood was elevated during early injury in a small study of patients with severe isolated TBI, but its causal or regulatory role in the development of TBI-IC was not investigated.⁴⁸ Finally, protein C is a plasma vitamin K-dependent glycoprotein zymogen that becomes activated serine protease upon binding to thrombin.^49,50 Activated protein C proteolytically inactivates FVa and FVIIIa.⁵¹ This anti-coagulant activity is primarily detected on the surface of endothelial cells because activated protein C forms a complex with the endothelial protein C receptor and thrombomodulin.⁵¹ Studies have suggested that a maladaptive protein C pathway in response to brain injury and hemorrhagic shock causes a rapid APC-mediated coagulopathy.⁵² However, the value of these intrinsic anti-coagulation factors for predicting TBI-IC has not been investigated in sufficient numbers of patients.

Tests for fibrinolysis

During the late stage of coagulation, the polymerized fibrin exposes noncompetitive high-affinity sites for tissue-type plasminogen activator (tPA) and plasminogen located in the αC-domain, which are cryptic in fibrinogen.⁵³ The D domain also undergoes conformational changes to expose low-affinity tPA- and plasminogen-binding sites.⁵⁴ Upon binding, tPA activates the colocalized plasminogen to plasmin, which cleaves crosslinked fibrin to dissolve the clot and reestablish the blood flow in an occluded vessel. This fibrinolytic process is kinetically slow during hemostasis but appears to develop early and rapidly in TBI, resulting in an acute hyperfibrinolytic state.⁷ This observation suggests that TBI-induced hyperfibrinolysis may develop independent of fibrin formation. This notion is supported by the finding that tPA deficiency reduces persistent intracerebral hemorrhage but does not prevent systemic coagulopathy in TBI mice.⁵⁵ The underlying mechanism of this fibrin-independent hyperfibrinolysis remains largely unknown. A prospective multicenter observational cohort study conducted at five European trauma centers and including 914 patients with TBI reported that death was associated with a hyperfibrinolytic profile defined by a drastic increase in plasma D-dimer and S100A10 and decreased tPA; however, only S100A10 was negatively correlated with the VHA maximum lysis on thrombelastometry.⁵⁶ One interesting note is that the cytoplasmic protein S100A10 can function as a cell-surface plasminogen receptor,^56,57 which could potentially drive early hyperfibrinolysis during acute TBI, but such a tentative regulatory pathway has not been investigated in TBI. The fibrinolytic product D-dimer has been widely measured in patients with TBI and found to be drastically elevated in these patients.^43,58 However, D-dimer is non-specific and varies significantly among patients, making it less accurate alone for predicting and diagnosing TBI-IC. Although TBI-induced hyperfibrinolysis remains poorly understood, the antifibrinolytic agent tranexamic acid, a synthetic derivative of the amino acid lysine that reversibly blocks lysine binding sites on plasminogen molecules, has been shown to reduce TBI-induced death moderately when given within 3 hours after injury, without increasing the risk of vascular occlusive events, as reported in the CRASH 3 trial.⁵⁹ However, the improvement was primarily achieved in patients with mild-to-moderate TBI (GCS 9–15) but not those with severe head injury (GCS 3–8).

There are fewer reports on the role of fibrinolysis inhibitors in the development of TBI-IC. In plasma samples collected from rats subjected to polytrauma, the plasmin inhibitor α2 macroglobulin decreased significantly within 2 hours and then increased by 4 hours.⁶⁰ Mice deficient in plasminogen activator inhibitor-1 (PAI-1) developed severe intracerebral hemorrhage after TBI.⁵⁵ A PAI-1 increase was detected along with elevated plasmin activity and high levels of plasma plasminogen and D-dimers in mice subjected to polytrauma,⁶⁰ suggesting that TBI may cause an imbalance between fibrinolysis and its inhibition. This imbalance is evident in studies showing that trauma patients with hyperfibrinolysis had significantly increased levels of tPA, but unchanged or even reduced levels of PAI-1, as compared to patients without hyperfibrinolysis.^61,62

Global tests

To address the limitations of individual tests, viscoelastic devices such as thromboelastometry (ROTEM) and thromboelastography (TEG) are increasingly used to globally evaluate hemostasis in TBI patients. These devices measure the viscoelastic properties of a blood clot to define its formation, stability and strength.^63,64 They detect changes in blood samples from TBI patients but have not functionally linked to traditional hemostatic tests. Furthermore, these devices do not adequately measure endothelial injury or factors regulating vascular tone. The DIC score has also been used to define TBI-IC in some studies.^65,66 This score system integrates measurements of platelets, fibrinogen/fibrin, and PT, but not those of endothelial injury. Its ability to accurately define TBI-IC remains to be evaluated. These global tests improve the detection of coagulopathy and its severity, but their application for predicting TBI-IC remains to be investigated.

Tests for cerebral vascular autoregulation and hemostasis

There is extensive laboratory and clinical evidence that TBI impairs cerebrovascular autoregulation, which maintains a steady CBF over a wide range of perfusion pressures during homeostasis.⁶⁷ Cerebral vasospasm also develops in patients with subarachnoid hemorrhage,⁶⁸ which is reported in 39–65% of TBI cases,⁶⁹ and has been independently associated with poor outcomes.^69,70 However, despite the finding that TBI triggers a catecholamine burst caused by acute sympathoadrenal activation, there is limited information regarding how changes in vascular tone regulate local and systemic hemostasis.^25,71 It has been recently reported that BDEVs drastically reduced CBF of non-injured mice by triggering calcium-dependent signaling in vascular smooth muscle cells.⁷² The CBF reduction in those mice was transient and recovered rapidly in surviving mice but persisted in dying mice and was epinephrine nonresponsive. These findings are consistent with clinical observation that cerebrovascular dysregulation is associated with poor outcomes in TBI patients.^73,74

Summary

Coagulopathy is a frequent and life-threatening complication of TBI. It is consumptive in nature and involves all components of the trauma hemostasis. There is a significant disparity of incidence between laboratory-detected hemostasis dysregulation and clinical coagulopathy, suggesting that the current laboratory tests have not met the clinical needs to accurately predict TBI-IC and define its severity. These laboratory tests are insufficient or inadequate in several key aspects. First, these tests (except the global tests) were originally developed to detect bleeding diathesis caused by deficiencies in individual coagulation factors and may not be accurate for detecting consumptive coagulopathy. Second, they detect specific components of hemostasis but are not sufficient to recapitulate the state of hemostasis involving all key components. Third, the tests diagnose coagulopathy after it occurs but none can predict when or if it will occur. Finally, the tests have proven to define the impact of coagulopathy on outcomes of TBI, but there is no consensus on whether treatments should be considered on the basis of their results to reduce coagulopathy. The comprehensive evaluation and standardization of laboratory tests is therefore needed to identify those that can predict the occurrence and severity of clinical TBI-IC in a timely fashion, ideally using a composite score that integrates laboratory and clinical information. A assay-guided algorithm combined with artificial intelligence could potentially create individualized models for predicting TBI-IC. In a pilot study, Gratz et al. used a thromboelastometric-guided algorithm to identify 21 of 32 patients as having TBI-IC, as compared to 5 identified through conventional coagulation tests.⁷⁵ To achieve this goal, large clinical trials are required to evaluate all available as well as new tests for predicting and diagnosing clinically relevant TBI-IC and to generate large data sets, from which a composite score system can be developed and validated.