Stroke Classification: Critical Role of Unusually Large von Willebrand Factor Multimers and Tissue Factor on Clinical Phenotypes Based on Novel “Two-Path Unifying Theory” of Hemostasis

Unsolved issues in hemostasis

In addition to incompletely understood hemostatic mechanism, clinicians and laboratory hematologists as well as coagulation scientists have used inconsistent terminology in the lexicon of coagulation. These include “coagulation” and “thrombosis,” “blood clots” and “fibrin clots,” “hematoma” and “thrombus,” and “macrothrombus” and “microthrombi.” Further, sometimes the terms “coagulation,” “hemostasis,” and “thrombogenesis” have been used ambiguously to mean the same or similar hemostatic processes in different situations. Also, the term “coagulation” in the laboratory in vitro and that in a disease in vivo convey very different meanings.

Contemporary dogmatic theory on hemostasis has solidified the proposition that all the thromboses must be the same or similar in their character as a result of the final product of hemostasis triggered by TF release and ending with formation of “fibrin clots” via extrinsic cascade/cell-based coagulation. This current concept of the blood clots (thrombus) defines that every thrombus in arterial thrombosis and DVT are made of “fibrin clots,” but perhaps with different NETs. ²⁵ This uniformity thesis has allowed for the clinician to use the same anticoagulation therapy for every thrombotic disorder in absence of understandable thrombogenetic mechanism. ⁴ We know better now that anticoagulation is effective in management of DVT, but it is not effective at all in certain arterial thrombosis such as “DIC” and TTP, ^5,27,31 which are characterized by arteriolar and capillary microthrombosis. Also, AIS caused by fibrin clot with trapped blood cells, cellular components, platelets, and VWF ³⁵ can be prevented effectively with anticoagulation therapy. These clinical and pathological differences among DVT, AIS, and “DIC” have been an enigmatic puzzle to the clinicians and could not be explicated by contemporary hemostatic theories. Clinicians have understood both DVT and DIC are similar disease from the same “blood clots” but occurring in different vasculatures (ie, large vasculature vs microvasculature). Therefore, many clinicians have called DIC is the disorder of microvascular thrombosis instead of that of vascular microthrombosis. If microthrombi of “DIC” are recognized to be different in its character from macrothrombosis of DVT and AIS, the unifying concept of microthrombi and “fibrin clots” perfectly explains the genesis of macrothrombus. Finally, we can affirm that DVT is macrothrombotic disease and “DIC” is microthrombotic disease (ie, VMTD) occurring from separate thrombogenetic paths. ^5,27,31 In retrospect, this ambiguity in terminology has contributed to the delay in identifying the true mechanism of in vivo hemostasis to date. ^5,27

The in vitro based hemostatic theory of in vitro intrinsic and extrinsic coagulation cascades has been very useful in identifying various coagulation disorders due to coagulation factor deficiency utilizing in vitro clotting tests such as prothrombin time, activated partial thromboplastin time, thrombin time, and various coagulation factor assays. ³⁶ However, when comprehensive understanding of in vivo hemostasis is needed, the role of platelets and von Willebrand factors (VWF) could not be assigned into proper places in the mechanism of fibrinogenesis represented by in vitro extrinsic cascade/cell-based model. Additionally, the pathophysiological mechanism producing microthrombosis in VMTD, such as “DIC”, TTP, TTP-like syndrome, HUS, thrombotic microangiopathy, and MODS, including encephalopathy, has not been clearly understood ^5,27 to date. Now, the microblood clots of “DIC” (eg, microthrombi seen in diffuse encephalopathic stroke) are established to be “microthrombi.” ^30,33,37 This reinterpretation of microthrombosis and the concept of microthrombogenesis have helped this author in formulating in vivo hemostasis, which can account for every aspect of hemostasis-related diseases that couldn’t be explained by extrinsic cascade/cell-based coagulation theory, including different stroke phenotypes.

In fact, the role of disseminated VMTD in stroke, which brain phenotype is characterized by encephalopathy presenting with altered mental state, confusion, seizure, and coma rather than localizing neurological symptoms and signs, has been well-known in clinical medicine. ³¹ The support for the concept of microthrombogenesis further comes from increased activity of ULVWF/VWF that has been significantly high in certain stroke syndromes (ie, TIA and AIS). ^{38 –40} However, the issue whether stroke is at high risk due to ULVWF or it causes high ULVWF after stroke is another unresolved issue at this time. This riddle can be easily answered by “two-path unifying theory” of hemostasis.

Novel “two-path unifying theory” of hemostasis

This author has had a long-standing interest on the pathogenesis of “DIC” because this condition is very similar to TTP with the same characteristic of vascular microthrombosis and hematologic features of thrombocytopenia, microangiopathic hemolytic anemia (MAHA) associated with MODS, including the brain and renal dysfunction. ^5,27,31,41 The main difference between “DIC” and TTP has been “DIC” is always associated with critical illnesses and TTP is caused by the deficiency of ULVWF-cleaving protease ADAMTS13, which occurs in hereditary gene mutation-associated TTP (ie, GA-VMTD) or acquired antibody-associated TTP (ie, AA-VMTD). ³¹ In clinical cases, another intriguing feature of VMTD has been there are 2 groups of patients with microthrombosis: one with extremely low ADAMTS13 activity less than 5% of normal and the other with mild to moderately diminished ADAMTS13 activity between 25% and 75%. ³¹ The former group is associated with TTP, but the latter group almost always occurs in critically ill patients with endotheliopathy-associated (EA) diseases such as sepsis and polytrauma. Pathogenetically this latter group is different from immune TTP and hereditary TTP, and has been called as endotheliopathy-associated VMTD (EA-VMTD) or TTP-like syndrome. ³¹ Oftentimes, TTP-like syndrome is associated with coagulopathy. The EA-VMTD-associated coagulopathy has occurred with thrombocytopenia, MAHA, and MODS as seen in TTP; however, unlike TTP, TTP-like syndrome is often characterized by prolonged prothrombin time and activated partial thromboplastin time and hypofibrinogenemia. The coagulopathy in critically ill patients has been termed acute “DIC” when associated with abnormal coagulation parameters, but it is called chronic “DIC” when unassociated with coagulopathy. ^5,27,31 However, it has become apparent that both acute and chronic “DIC” have occurred in association with complement activation, leading to endotheliopathy, molecular changes of exocytosis of ULVWF, and release of inflammatory cytokines. Certainly, this endothelial molecular pathogenesis is identical in both “DIC” and TTP-like syndrome. Thus, “DIC” is concluded to be a clinical phenotype of EA-VMTD as is other TTP-like syndromes. ^5,27,30,31

In recent publications, this author was able to affirm that the concept of “DIC” attributed to activated TF-FVIIa path leading to fibrin clots was ill-founded. “DIC” should be correctly redefined as DIT, which clinical disease is VMTD occurring as a result of activated ULVWF path of hemostasis. More importantly, both “DIC” and DIT with hematologic features of TTP and TTP-like syndrome are characterized by the microthrombosis expression. ^{4,5,7,27,30,31,41} If “DIC” were not the result of TF-FVIIa path-initiated coagulopathy because microthrombi are composed of platelet-ULVWF complexes, it is logical to infer that there has to be another hemostatic path exists via activation of ULVWF and platelets, which contributes to the formation of microthrombi in “DIC.” To this author, this path perfectly fits and works well in in vivo hemostasis framework because the exocytosis of ULVWF interacting with platelets is the first event of in vivo hemostasis in external bodily injury. Now, 2 molecular events are identified; one is well-known TF role activating FVII (TF path) in vitro and in vivo hemostasis and the other is the role of ULVWF recruiting platelets (ULVWF path). Both are essential processes in in vivo hemostasis. Thus, TF path and ULVWF path can be put together and construct a novel in vivo hemostatic theory. This proposition from recognition of false concept of “DIC” has led this author to formulate “two-path unifying theory” of hemostasis as illustrated in Figure 2 and “three different mechanisms of thrombogenesis” as annotated and shaded in the Figure. ^4,5,7,27 Finally, a novel well-defined physiological mechanism of in vivo hemostasis is identified.

As articulated in the schematic illustration in Figure 1, the level (depth) of vascular damage is the critical determinant that produces the different character and components of blood clots as well as severity of stroke syndrome. Following vascular injury involving the damage of ECs and SET/EVT, 2 hemostatic subpaths must be activated. One is microthrombotic (ie, ULVWF) path, and the other fibrinogenetic (ie, TF) path. The third is combined path producing macrothrombosis from unifying of ULVWF and TF paths. In ULVWF path, ULVWF with FVIII are released from Weibel-Palade bodies of ECs and become anchored to the membrane of injured ECs as long elongated strings and recruit platelets to form platelet-ULVWF complexes ^{42 –45} through microthrombogenesis. The platelet-ULVWF complexes become “microthrombi” strings. But, in TF path, TF is released from damaged SET/EVT and activates FVII triggering extrinsic cascade/cell-based coagulation, promoting the sequential activation of coagulation factors and generating thrombin that converts fibrinogen to “fibrin meshes.” Following the activation of these 2 subpaths, final phase leads to the third path called macrothrombotic path in which microthrombi strings and fibrin meshes become comingled. Most likely, at this stage, neutrophils including other blood cells and molecules passively integrate into macrothrombotic path via NETosis. ⁴⁶ This unifying process of microthrombi, fibrin clots, and NETs together at the vascular injury site produces the final “blood clots” (hemostatic plug) to stop bleeding at external bodily injury site, but also forms “blood clots” (macrothrombus) leading to thrombosis at intravascular injury site. These mechanisms are the essence of in vivo blood coagulation and thrombogenesis integrated into the updated framework of “two-path unifying theory” of hemostasis shown in Figure 2. ⁵ In addition, collagen and various adhesion molecules partake actively in primary event of hemostasis, ^47,48 but extracellular traps promoting NETosis ^{25,46,49 –51} are likely secondary cellular components and molecules that participate passively at the unifying stage.

This framework of in vivo hemostatic mechanism has been derived from reinterpretation of “DIC” ^4,7,27,31 that affirms the critical role of ULVWF and platelets in microthrombogenesis. ^5,27,31 This hemostatic mechanism in intravascular injury identifies 3 paths of thrombogenesis, which are microthrombogenesis, fibrinogenesis, and macrothrombogenesis, producing 3 different kinds of blood clots: (1) “microthrombi” for microthrombosis, ⁵² (2) “fibrin clots” for fibrin clot disease and hematoma, and (3) “macrothrombus” for macrothrombosis. In stroke, based on 3 different levels of intravascular wall injury initiating hemorrhage and thrombosis, 5 different phenotypes of stroke are identified. The role of the different levels of vascular damage and resultant phenotypes are summarized in Table 2 and Table 4.

OPEN IN VIEWER

Table 4.

Hemostatic Characteristics of Stroke Syndrome.

Stroke Character	Microthrombosis		Macrothrombosis		Macrothrombosis with hemorrhage	Hemorrhage/hematoma
	Focal VMTD	Systemic VMTD	Thrombotic	Embolic
Stroke syndromes	TIA	Encephalopathic stroke	AIS		THS	AHS
Causes
Vascular injury	Intravascular: focal endotheliopathy (eg, atherosclerotic plaque detachment)	Intravascular: disseminated endotheliopathy (eg, EA-VMTD)	Intravascular: local vascular injury/embolism (eg, vascular anomaly; trauma; AF (?); cardiac injury; LAA thrombosis)		Intravascular: local vascular injury beyond vessel wall (eg, uncontrolled hypertension)	Extravascular (EVT): rupture of vascular anomaly; obtuse head/brain injury (ie, AHS)
Level	ECs	ECs	ECs + SET		ECs + SET + EVT	EVT
Nonvascular disease/injurya	NA	Disseminated (eg, GA-VMTD, AA-VMTD)	NA		NA	NA
Hemostatic path	Focal lone ULVWF path activation	Disseminated lone ULVWF path activation	Combined activation of ULVWF and TF paths		Combined activation of ULVWF and TF paths	Lone activation of TF path
Mechanism of stroke genesis	Microthrombogenesis	Microthrombogenesis	Macrothrombogenesis		Macrothrombosis/fibrin clots	Fibrin clots with hematoma
Character of blood clots	Platelet-ULVWF strings	Platelet-ULVWF strings	Combined platelet-ULVWF strings, fibrin meshes and trapped blood cells, DNAs, and histones		Combined platelet-ULVWF strings, fibrin meshes and trapped blood cells, DNAs, and histones; hematoma	Hematoma
Examples	Focal EA-VMTD	Disseminated EA-VMTD, AA-VMTD, GA-VMTD	Cerebral arterial thrombosis; cerebral embolic thrombosis		Thrombo-hemorrhage	ICH, SAH, IVH, SDH, EDH

Abbreviations: AIS, acute ischemic stroke; AF, atrial fibrillation; EC, endothelial cells; EDH, epidural hematoma; EVT, extravascular tissue; ICH, intracerebral hemorrhage; IVH, intraventricular hemorrhage; LAA, left atrial appendage; NA, not applicable; SAH, subarachnoidal hemorrhage; SDH, subdural hematoma; SET, subendothelial tissue; TF, tissue factor; TIA, transient ischemic attack; TTP, thrombotic thrombocytopenic purpura; ULVWF, unusually large von Willebrand factor multimers; VMTD, vascular microthrombotic disease; AA-VMTD, antibody-associated VMTD; EA-VMTD, endotheliopathy-associated VMTD; GA-VMTD, gene mutation-associated VMTD.

^a Nonvascular disease producing microthrombi occurs in TTP (ie, AA-VMTD and GA-VMTD), but still is aberrant hemostasis since ULVWF path is activated and participated in encephalopathic stroke.

Finally, the complete hemostatic mechanism utilizing the essential vascular wall components and coagulation factors is identified and thrombogenetic mechanisms constructed. Four major hemostatic components are ULVWF, platelet, TF, and various coagulation factors, and the involved vascular physiologic participants are ECs and SET/EVT. The application of this hemostatic and thrombogenetic mechanism in the interpretation of stroke provides several clearly defined phenotypes of stroke syndrome, which allows instructive and practical approaches not only in making the diagnosis but also for identifying rational therapy and prognostication.

Encephalopathy as a stroke entity

No doubt, microthrombotic encephalopathy is a stroke phenotype occurring in critically ill patients. Encephalopathy is often provoked as a part of generalized MODS associated with VMTD (ie, TTP and TTP-like syndrome) and is characterized by diffuse intracranial intravascular microthrombosis. ³¹ Transient ischemic attack, which microthrombosis occurs as a result of level 1 damage in focal endotheliopathy, is self-limited and reversible stroke. ^4,7,40 Diffuse encephalopathic stroke is also readily reversible if treated with appropriate treatment in early stage of VMTD even though it is caused by disseminated intracranial microthrombosis because microthrombi strings are composed of platelet-ULVWF complexes and ULVWF can be readily cleaved by antimicrothrombotic agents/regimen such as therapeutic plasma exchange. Diffuse encephalopathic stroke, occurring in hepatic encephalopathy, ^53,54 cardiac encephalopathy, ⁵⁵ pulmonary encephalopathy, ⁵⁶ uremic encephalopathy, ^{57 –59} hypertensive encephalopathy, ⁶⁰ and septic encephalopathy, ^{61 –64} represents the phenotype of stroke involving the brain among MODS associated with generalized VMTD. If masked and undiagnosed in early stage of VMTD and untreated with therapeutic plasma exchange, diffuse encephalopathic stroke could evolve rapidly into MODS resulting in the demise of the patient.

To date, however, because of the lack of understanding on the pathogenesis and yet not recognized as a clinical disease entity for encephalopathy associated with MODS, diffuse encephalopathic stroke of the brain syndrome has not been included in stroke syndrome. Obviously, it is a disseminated form of TIA caused by activated ULVWF path. Diffuse encephalopathic stroke does not respond to TF-based anticoagulation therapy because their pathogenetic mechanism is via microthrombogenesis occurring due to lone activation of ULVWF path, which leads to platelet-ULVWF complexes. It should be emphasized that diffuse encephalopathic stroke in sepsis is usually not due to brain tissue invasion of the pathogen through blood–brain barrier, but is the result of disseminated microthrombi within the microvasculature of the brain. Thus, diffuse encephalopathic stroke is reversible disease if treated in early stage with antimicrothrombotic therapy.

Although anticoagulation therapy is ineffective for diffuse encephalopathic stroke associated with VMTD, ⁶⁵ therapeutic plasma exchange as a surrogate of antimicrothrombotic treatment has been attempted in sepsis and MODS with promising potential if utilized in early stage of MODS. ^{66 –70} The pathogenesis of this stroke is an important concept derived from hemostasis, microthrombogenesis, as well as molecular function of endotheliopathy in EA-VMTD. ²⁷

Role of ADAMTS13 and ULVWF in stroke

The key role of ULVWF in stroke has been found to be significant in TIA and AIS. ^{38 –40,71} In the study of chimeric platelet derived and ECs derived VWF mice, endothelial derived VWF was found to be the major determinant that mediates VWF-dependent ischemic stroke promoting postischemic thrombo-inflammation. ³⁸ Also, a case–controlled study showing of the high level of VWF antigen and activity obtained from the post-stroke blood samples was associated with the occurrence of AIS. This relation was unaffected by the severity of the acute-phase response or by genetic variation or degradation. ³⁹ In prospective observational analytical case–control study, ADAMTS13 activity was reduced and VWF antigen expression increased within 4 weeks of TIA or ischemic stroke onset. ⁴⁰ The study concluded increased ULVWF caused by reduced ADAMTS13 had promoted enhanced platelet adhesion and aggregation contributing to stroke in response to stimulation with collagen and adenosine diphosphate via VWF-mediated pathways.

The implication of increased VWF and decreased ADAMTS13 in post-stroke patients is an unresolved issue. An explanation was decreased protease ADAMTS13 might be caused by genetic or acquired conditions, and the excess of ULVWF due to the decreased protease was thrombogenic, promoting stroke in both TIA and AIS. ³⁹ This has implied that elevated ULVWF was a risk factor for stroke. However, to understand the vascular biology of ULVWF, we have to go back to in vivo mechanism of hemostasis. As shown in Figure 2, and Table 1, thrombosis mediating stroke cannot occur without intravascular injury first. This preconception supports that increased ULVWF/VWF activity, which promotes the activation of microthrombotic path leading to microthrombogenesis and generating microthrombi strings, ^4,7,31 is the result of exocytosis of ULVWF from endothelial injury. Focal endothelial damage releases ULVWF and triggers microthrombosis of TIA (level 1 damage), and local endothelial and SET damage releases ULVWF/TF and triggers macrothrombosis of AIS (level 2 damage). Generalized endotheliopathy releases an extensive amount of ULVWF and promotes disseminated cerebral microthrombosis of diffuse encephalopathic stroke. ²⁷ Thus, increased VWF and ULVWF are not the cause of stroke but are the result of stroke occurring due to ECs damage. This recognition of the vascular biology of ULVWF in stroke may have significant implication in the management and secondary prevention.

Potential role of circulating TF-bearing microvesicles as a diagnostic marker of AIS

Another serious controversy in stroke syndrome is what is the difference between TIA and AIS. Their initial stroke features may be similar, but their subsequent clinical courses are different. How are their pathogenetic mechanisms different between 2 stroke phenotypes? The answers may come from the TF-bearing microvesicles (TF-positive MV) and perhaps other procoagulant-positive MV which are released by apoptotic and damaged cells into circulation, but TF is an integral membrane protein, normally separated from the blood by the vascular endothelium, which plays a key role in promoting blood coagulation. ⁷² With a perforating vascular injury, TF becomes exposed to blood as a result of SET damage and binds plasma FVIIa to activate TF path and initiate extrinsic cascade/cell-based coagulation. ⁷³

The most fundamental aspect of blood coagulation is “hemostasis leading to thrombogenesis can be activated only by vascular injury” as shown in Table 1-(1) and Table 3. This means the thrombogenesis is a natural process in intravascular injury. Procoagulant ULVWF and VWF-positive MV are released/formed from ECs damage in TIA as a result of level 1 injury, but ULVWF, VWF-positive MV, TF, and TF-positive MV are released/formed from ECs and SET damage in AIS as a result of level 2 injury as shown in Figure 1 and Table 4. Therefore, it is natural to expect the expression of ULVWF, VWF-positive MV, TF, TF-positive MV, and also other MV in vascular injury. In stroke, ULVWF and TF would be utilized in the formation of microthrombosis or macrothrombosis, but some MV would be strayed in circulation as noted in clinical investigations. ^{74 –76} Although MV were suspected to be potential regulators of cardiovascular diseases, contributors of pathogenesis of thrombosis, and/or emerging marker and therapeutic targets of stroke, ^{77 –79} it is this author’s opinion that VWF-positive MV and TF-positive MV are not the regulator or cause of thrombogenesis, but simply represent the leftover “microparticles” of thrombogenesis from damaged ECs and SET following intravascular injury in stroke. This thesis is very much consistent with “two-path unifying theory” of hemostasis. However, luckily this TF-positive MV may be the best laboratory marker yet that can differentiate between TIA and AIS. Thus, theoretically strong expression of TF-positive MV supports the diagnosis of AIS due to level 2 damage of the vessel wall, which was consistently demonstrated in recently published articles. ^{74 –76}

Until now, the pathogenetic difference between TIA and AIS has not been clearly understood in coagulation community even though their clinical course is very different and prognostic diversity is intriguing. These also have created anxiety for the patient when the treatment decision was discussed and decided for stroke, and at times some tension occurred between the patient and the clinician. Additionally, clinical trials often had lumped the patients with TIA and AIS together, which might have precluded fair interpretation of therapeutic results without theory-based medicine. Hopefully, this issue separating the different phenotypes between TIA and AIS in stroke could be resolved by the use of TF-positive MV. Additional investigations are needed.

Mechanisms of THS vs AHS

Among 5 major phenotypes of stroke, the pathogeneses of TIA, AIS, and diffuse encephalopathic stroke are simple and straightforward to recognize and be understand. However, 2 hemorrhagic strokes THS and AHS need special explanation from the point view of in vivo hemostasis and vascular physiology.

As shown in Table 2, TIA is the result of focal ECs damage (level 1 damage) from minor endothelial injury (eg, small atherosclerotic plaque detachment) or transient endothelial dysfunction (eg, atrial fibrillation) which leads to microthrombi strings composed of platelet-ULVWF complexes. These microthrombi strings should be resolved by circulating protease ADAMTS13 and the patient recovers after transient ischemic event. Acute ischemic stroke is the result of local endothelial injury (eg, detached atherosclerotic atheroma) causing deeper vascular damage into SET (level 2 damage) that leads to the unifying process of both microthrombi from ULVWF path and fibrin clots from TF path. The damage produces macrothrombus at the site of detached atheroma. However, in addition to level 2 damage, if the outer vascular wall tunica externa and beyond (i.e., EVT), schematically shown in Figure 1, is damaged (eg, severe hypertension), intracranial intravascular hemostasis occurs in the vascular lumen and bleeds into the brain, it leads to level 3 damage. Hemostasis in the lumen creates macrothrombosis from activated ULVWF and sTF paths, and bleeding into EVT (brain tissue) causes hemorrhage/hematoma from activated eTF path. This is called THS.

In contrast, certain intracranial vascular injury associated with vascular anomaly (eg, subarachnoid aneurysm) or obtuse trauma (eg, head/brain injury while on antiplatelet therapy) may produce intracranial extravascular injury (level e damage) leading to almost pure hemorrhagic stroke without chance to produce thrombosis because hemorrhage does not occur into the vascular lumen, but into EVT as noted in Table 1. Acute hemorrhagic stroke develops due to hemorrhage into the brain tissue (eg, ICH) or cavitary space between the coverings of the brain (eg, SAH, IVH, SDH, EDH). These stroke phenotypes collectively are called AHS. This preconception is derived from theoretical thesis based on “two-path unifying theory” of hemostasis. ^4,7,27 Thus, these 2 different pathogeneses between THS and AHS may dictate very different approach in their management.

Origin of embolic stroke: atrial fibrillation versus atrial cardiopathy

Approximately 20% to 40% of stroke is caused by embolic thrombi originated from the heart or other extracranial vasculature. ⁸⁰ According to Framingham study report in 1991, the attributed risk of stroke from all cardiovascular contributors had decreased with age except from atrial fibrillation, for which the risk increased significantly after 34 years follow-up. ⁸¹ Atrial fibrillation has been assigned to be the most common cause of embolic stroke. ^82,83 Current views on the mechanism of thrombogenesis in atrial fibrillation rest on the century-old hypothesis that fibrillation of the atrium produces stasis of blood, which causes thrombus formation and embolism to the brain. ⁸⁴ Recently however, the validity of this prevailing model of atrial fibrillation and thromboembolism has been questioned ^{22 –24,80,85,86} because embolic stroke occurs without atrial fibrillation and straightforward association between atrial fibrillation and stroke does not convincingly demonstrate temporality, specificity, or a biological gradient, and it was not concordant with the totality of the available experimental evidence. ²²

Atrial fibrillation seems to fulfill the criteria of Virchow’s triad, which are necessary for thrombus formation: (1) blood stasis, (2) endothelial dysfunction, and (3) altered hemostasis. Blood stasis is evident in the left atrium of patients where flow velocity is markedly reduced concomitantly with impaired contractility of left atrial appendage. ⁸⁴ It is still not clear, however, how remodeling-related blood stasis can be implicated favoring thrombus formation in atrial fibrillation. ⁸⁷ This hypothesis has been recently challenged by the demonstration that atrial remodeling does not influence clotting activation and thrombus formation in canine models. ⁸⁸ Instead, there is clear association between atrial cardiopathy and ischemic stroke, and left atrial appendage has long been considered to be a prime site for thrombus formation in patients with atrial fibrillation. ²²

Hence, we are confronted with 2 very basic issues in the triangle of (1) atrial fibrillation, (2) atrial cardiopathy, and (3) thrombus formation, which are all involved in embolic stroke one way or another. The first issue is what is the primary event in thrombosis formation? Is this event atrial fibrillation or atrial cardiopathy? The second issue is what is the character of thrombus? Is it microthrombi or macrothrombus? To understand these 2 issues, we have to go back to the basics of hemostasis principle. The first principle in Table 1 is “hemostasis, leading to thrombosis, can be activated only by intravascular injury.” According to this logic, the answer to first issue comes from contentious claim of “intravascular injury” because atrial fibrillation alone does not cause vascular injury (Table 3). The answer to second question is related to the character of thrombus, which is obviously macrothrombus that is the case in embolic thrombi.

Atrial fibrillation, which does not cause intravascular injury that triggers hemostasis, could cause only endothelial dysfunction ⁸⁹ without SET damage. It promotes exocytosis of ULVWF without release of TF, and activates ULVWF path and produces transient “microthrombi” strings causing TIA. ⁸⁹ Thus, atrial fibrillation alone without vascular injury which produces damage to ECs and SET cannot cause macrothrombus that can lead to embolic AIS. It is very likely that “cryptogenic stroke” (ie, microembolic TIA) due to transient microthrombi may be prevalent in atrial fibrillation. Could it be possible that cryptogenic microembolic stroke has been included in embolic stroke (ie, macroembolic AIS) in some of clinical trials? Also, could it be possible that thrombosis within left atrial appendage due to cardiopathy promote both the onset of atrial fibrillation and embolic AIS? It is known the thrombus within left atrial appendage is macrothrombus that produces AIS. ^85,90,91 The hypothesis of atrial fibrillation alone promoting macrothrombus is probably incorrect according to in vivo hemostatic theory. This hemostatic theory also refutes Virchow’s concept that the triad is needed for thrombus formation since the only true cause of thrombosis is “intravascular injury.” Once intravascular injury causing ECs and SET damage occurs, ULVWF and TF paths trigger macrothrombus formation. The role of thrombophilic state or blood stasis is not the initiating cause of thrombus formation, but is the potentiating factor for thrombogenesis only after vascular injury has occurred. ⁷

Hope is this on-going theoretical debate ^{89,91 –93} and clinical trials based on conceptual issues of the pathophysiological mechanism in cardiac thrombus formation in atrial fibrillation and cardiopathy clarify the pathogenesis of embolic stroke and microembolic stroke sooner than later and provide a firmer recommendation for the prevention of this unique stroke syndrome.

Antiplatelet agents

An overview of major clinical trials and meta-analyses by Hackam and Spence ¹⁰¹ has concluded that aspirin-dipyridamole is an acceptable antiplatelet therapy for patients with noncardio-embolic ischemic stroke or TIA and probably superior to aspirin alone. However, there are many limitations in the body of randomized trials evaluating antiplatelet therapy for the secondary prevention of ischemic stroke and TIA. And the evidence is insufficient to claim the benefit of aspirin-dipyridamole or clopidogrel monotherapy in long-term prevention. ¹⁰¹

This conclusion is not surprising because TIA (level 1 vascular damage) is the result of lone activation of ULVWF path, which is a benign transient stroke probably with no lasting impact toward acute brain syndrome. Transient ischemic attack is physiologically resolved by naturally occurring protease ADAMTS13 without any specific therapy. Antiplatelet agent would not be beneficial for resolving formed microthrombi, but is expected to be efficacious in preventing recurrence of TIA. However, AIS would not be benefitted with antiplatelet agent in therapeutic intervention, secondary prevention, and prophylaxis because it is due to macrothrombosis with significant brain syndrome that occurs due to combined activation of ULVWF and TF paths (level 2 vascular damage). Besides, antiplatelet agent is ineffective against activated TF path even in secondary prevention.

In the future, clinical trials must compare the effectiveness of antiplatelet agents between TIA and AIS separately since their thrombogenetic mechanisms are different. Theoretically antiplatelet agents would inhibit ULVWF path only and be useful in secondary prevention of TIA, but antiplatelet agent alone would not be effective in secondary prevention of AIS and embolic stroke.

Recombinant ADAMTS13

Recently, decreased ADAMTS13 and increased ULVWF/VWF have been closely correlated to ischemic stroke. ^{38 –40,71} Retrospective studies have concluded that the high level of VWF was associated with increased risk of developing ischemic stroke. ^39,102 However, please note that the blood samples in patients with stroke were obtained after stroke for the study of ULVWF/VWF. Current interpretation is increased VWF/ULVWF is a prothrombotic marker for stroke. ¹⁰³ But, as discussed earlier in the text, “two-path unifying theory” of hemostasis supports transiently increased VWF and ULVWF is not the cause of ischemic stroke, but is the result of increased exocytosis from endothelial damage causing TIA and AIS in vascular injury. It is likely that decreased ADAMTS13 activity could have been related to imbalance due to excessive exocytosis of ULVWF or might have preexisted due to heterozygous mutation of ADAMTS13 gene in some patients, which potentiated TIA and AIS once intravascular injury had occurred. ³¹ Theoretically, recombinant ADAMTS13 (rADAMTS13) would cleave ULVWF present in unresolved microthrombi strings and therapeutically benefit for “severe and prolonged” TIA in level 1 damage and for secondary prevention in a short term, but would not benefit AIS caused by level 2 damage because AIS develops due to macrothrombus from combined activation of ULVWF and TF path and cannot be countered by antimicrothrombotic therapy. Parenteral antimicrothrombotic therapy (eg, rADAMTS13) might be practical for “one” short-term therapeutic use in “severe and prolonged” TIA, and oral antimicrothrombotic therapy (eg, N-acetylcysteine) may be very safe and useful for secondary prevention of recurrent TIA. At this time, oral antiplatelet agent is indicated in secondary prevention of TIA. Oral antimicrothrombotic therapy with N-acetylcysteine needs well-designed clinical trials. Theoretically, antiplatelet agents and antimicrothrombotic agents have no place for the treatment and prevention of AIS. On the other hand, both rADAMTS13 and therapeutic plasma exchange as a surrogate for ADAMTS13 are anticipated to be effective in the treatment of diffuse encephalopathic stroke associated with generalized VMTD in very early stage. ^104,105

Antithrombin agents and antifibrinogenetic agents

For many decades, vitamin K inhibitor (coumadin) and naturally occurring anticoagulant from bovine and porcine intestinal mucosa (unfractionated heparin) have been extensively utilized in clinical practice. Newer anticoagulants, including antithrombin agents such as low-molecular-weight heparin (LMWH), argatroban, bivalirudine, and lepirudin, have been introduced, and recently antifibrinogenetic agents, such as FXa inhibitors and pentasaccharides as well as LMWH, have been added to the armamentaria of anticoagulants for prophylaxis, secondary prevention, and therapeutic effect, particularly among patients with cardio-embolic stroke.

In regard to antithrombotic therapy for stroke, experts have expressed a broad spectrum of opinions. Surveys of practitioners have also demonstrated a lack of consensus on the use of anticoagulants for ischemic stroke. This uncertainty has existed largely due to incomplete understanding of the pathogenesis of different phenotypes of stroke and the lack of definitive clinical data based on the “hemostatic evidence of medicine.” A review by the panel of the Stroke Council of the American Heart Association found no strong evidence for effectiveness of anticoagulants in treating AIS. Several clinical trials have suggested that emergent anticoagulation regimens have no significant effect in improving clinical outcomes for patients with AIS. ¹⁰⁶

Currently, there is lack of consensus on the recommendation and benefit for using full-dose anticoagulation for treatment of patients with AIS because of limited efficacy and increased risk of bleeding complications. Latest Cochrane review ¹⁰⁷ on anticoagulation for AIS did not provide any evidence that the early use of anticoagulants is of overall benefit to people with strokes caused by blood clots. More research is needed to find out if there are ways to select the people with stroke who will benefit from anticoagulants without suffering the hemorrhagic complications.

This conclusion from therapeutic clinical trials could have been anticipated because AIS is the result of intravascular damage that is resulting from activation of ULVWF path and TF path of hemostasis in intravascular injury at the area of in situ thrombus. Once a macrothrombus is formed as a result of unifying process of microthrombi strings from ULVWF path and fibrin meshes from TF path, the macrothrombus would unlikely be resolved with any kind of anticoagulation therapy until natural fibrinolysis is activated. Further, preventive anticoagulation is impractical in isolated AIS episode. But the patient with a persistent underlying disease such as atrial cardiopathy with recurring thrombus formation in the left atrial appendage, mechanical valve associated atrial fibrillation ^{85,89,108 –110} that is at high risk for recurrent “intravascular injury” might be benefitted with TF path inhibitors in primary prevention of embolic stroke. According to in vivo hemostatic theory, nonvalvular atrial fibrillation in isolated more likely produces TIA rather than AIS. Therefore, antiplatelet agent might be beneficial for primary and secondary prevention of TIA-in isolated nonvalvular atrial fibrillation.

Tissue plasminogen activators

The modern era of the thrombolytic/fibrinolytic therapy has begun in the early 1990s. ¹¹¹ Among thrombolytic agents such as streptokinase, urokinase and tissue plasminogen activator (tPA), alteplase that is a tPA approved by the US Food and Drug Administration (FDA) for the treatment of AIS has been considered to be the mainstay of thrombolytic therapy. ¹¹² Theoretically, this is a very reasonable therapeutic agent for AIS due to its character of macrothrombus if utilized in the symptom onset of stroke in less than 3 hours. The efficacy of alteplase for AIS is established. ^{112 –114} Although thrombolytic/fibrinolytic agents are theoretically ideal, it is still in early stage of clinical use as safety issue is being addressed. The usage of thrombolytic therapy remains low in clinical practice. ^115,116 We will wait for the results of further controlled clinical trials to provide clinicians with evidence-based medicine that is selected and initiated on the concept of theory-based medicine.

Desmopressin

Desmopressin, a synthetic analog of vasopressin, in addition to its utility in several conditions such as diabetes insipidus, uremic bleeding, and nocturnal enuresis has been approved for use in the treatment of hemophilia A, von Willebrand disease type 1, and congenital and acquired platelet dysfunction. The mechanism of action of desmopressin is its agonist activity at V2 receptors that leads to a rise in intracellular concentrations of cyclic adenosine monophosphate, which in turn induces exocytosis of ULVWF from its storage sites (ie, Weibel–Palade bodies) of ECs into circulation. ¹²⁰

Based on the mechanism of in vivo “two-path unifying theory” of hemostasis, ^5,7,27 it is expected the treatment of desmopressin could initiate and potentiate hemostasis exactly at the active bleeding site of intracranial vascular injury that could be repaired by ULVWF and activated FVII if treated in very early stage of AHS. Desmopressin administration would release highly thrombogenic ULVWF and FVIII at the vascular injury site in large quantity and activate ULVWF path, and recruit platelets to form microthrombi strings. Microthrombi strings would be unified with fibrin meshes formed from activated TF path due to hemorrhage of AHS and may induce efficient macrothrombogenesis producing natural hemostatic “patch” at the extravascular bleeding site. More than 3 decades ago, the use of desmopressin has been advocated in the treatment of hemostatic disorders. ^{121 –123} According to hemostatic theory, desmopressin might be an ideal drug if the diagnosis of AHS can be established in a very early stage.

Desmopressin has been used in both primary ICH in the absence of underlying vascular malformation or coagulopathy and secondary ICH due to underlying vascular anomaly or anticoagulation state. ^117,118 Hematoma expansion is a detrimental outcome of spontaneous ICH, resulting in significant neurological deficits, and enhancement of ICH-induced brain injury. ¹²⁴ Desmopressin stabilized the platelet function in neurosurgical patients with secondary intracranial hemorrhage, ^117,125 and was associated with a decreased likelihood of intracranial hemorrhage/hematoma expansion during the first 24 hours in retrospective assessment of its effectiveness and safety in patients with antiplatelet-associated ICH. ¹²⁶ Theoretically, desmopressin appears to be one drug that has potential for early treatment of ICH and other AHS. It would be an interesting proposal to test the hemostatic effect of desmopressin in ICH, SAH, and other AHS.

rFVIIa

Recombinant activated FVII (rFVIIa) is a potent promoter of TF path, which triggers extrinsic cascade/cell-based coagulation leading to fibrinogenesis to produce thrombin and subsequent fibrin meshes. Although rFVIIa has been approved by the US FDA 2 decades ago for hemophilia A and B with inhibitors, bleeding episodes in hemophilia, and FVII deficiency, ^127,128 its off-label use has been controversial, especially to enhance hemostasis in patients with bleeding who may or may not have an underlying coagulation defect. ¹²⁸ While evidence-based guidelines do exist for rFVIIa treatment in hemophilia, none is available for its off-label use.

According to in vivo hemostatic theory and thrombogenetic mechanism, rFVIIa even with pharmacologic dose is expected to be ineffective in AHS because fibrin clots enhanced by rFVIIa alone cannot achieve complete hemostasis without the hemostatic plug, which needs ULVWF and platelets at bleeding intravascular injury site. The patient with AHS already has enough TF available from EVT damage (eg, brain issue area), but sufficient quantity of ULVWF would not be available at bleeding intravascular injury and EVT site. For example, in ICH, excessive bleeding at the site of a small intracranial ECs damage extending to EVT into the brain tissue does not produce sufficient ULVWF due to intravascular pressure gradient. Thus, the sufficient hemostatic patch composed of microthrombi strings does not occur. But TF present in EVT hemorrhage promotes regional fibrinogenesis and produces fibrin meshes leading to cerebral hematoma (Table 1). In this situation, intravenous treatment with pharmacologic dose of rFVIIa would further increase cerebral hematoma. Also, pharmacologic dose of rFVIIa almost surely would promote pathologic “intravascular fibrinogenesis” if continuously administered. If additional vascular injury occurs, intravascular macrothrombogenesis could produce severe macrothrombotic complication. Thomas et al ¹²⁹ appropriately cautioned the administration of FVIIa to patients with arterial injuries and injured mesenteric and cerebral vessels might be especially susceptible to thrombosis. This macrothrombotic complication can be explained by in vivo hemostatic theory.

Well-controlled clinical trials were done utilizing rFVIIa for ICH with an enthusiasm in earlier clinical trial, ¹¹⁹ but at the additional confirmation trials with rFVIIa were proved to be unsuccessful in improving survival for acute ICH. ^{130 –132} In retrospect, as presaged from “two-path unifying theory” of hemostasis, the lack of benefit of rFVIIa for ICH and theoretical complications of serious of disseminated fibrin thrombi and macrothromboembolic complications could have been anticipated because hemorrhage/hematoma from EVT (ie, brain) damage in ICH cannot be corrected by induction of intravascular fibrinogenesis. It can be concluded that rFVIIa is not effective and is contraindicated in the management of AHS according to both in vivo hemostatic theory-based medicine and evidence-based medicine following clinical trials.

In reflecting on the review of past clinical trials, this author would like to caution designing clinical trials in selecting the patients for each stroke syndrome separately. A note of consideration is summarized in Table 6. First, TIA, which is focal microthrombotic disease due to level 1 vascular damage causing lone activation of ULVWF path, should be separated from AIS, perhaps using TF-positive MV marker, because AIS is local macrothrombotic disease due to level 2 vascular damage causing activation of combined ULVWF and TF paths. Transient ischemic attack and AIS occur as a result of 2 different hemostatic thrombogeneses. Second, we know that antiplatelet regimen is expected to be effective for TIA prevention according to theory-based medicine (Table 5) and also was found to be effective by evidence-based medicine. ¹⁰¹ Third, THS due to level 3 vascular damage is also a very different stroke syndrome from AIS, which pathogenetic differences are defined in Table 2 and Table 4. Thus, any suspected THS should not be included in AIS trials and in AHS trials.

OPEN IN VIEWER

Table 6.

Proposed Theoretical Considerations in the Clinical Trial for Different Stroke Phenotypes.

Stroke syndromes caused by ULVWF path (TIA) – Antiplatelet agent is effective in secondary prevention. – Oral antimicrothrombotic agent (NAC?) may be beneficial in secondary prevention. – Parenteral antimicrothrombotic agent (rADAMTS13) is likely beneficial for encephalopathic stroke. Stroke syndromes caused by combined ULVWF and TF paths (AIS) – Anti-FXa, antithrombin or antifibrinogenetic agents are probably ineffective for completed stroke. – Anti-FXa, antithrombin or antifibrinogenetic agents may be beneficial for stroke in progress. – Oral anti-FXa, antithrombin or antifibrinogenetic agents may be beneficial in primary and secondary prevention in conditions such as atrial cardiopathy, and mechanical heart valve because the risk of vascular injury is high. – Theoretically, oral anti-FXa, antithrombin or antifibrinogenetic agents is expected to be ineffective in nonvalvular atrial fibrillation because according to in vivo hemostasis it is more likely to respond to antiplatelet agent. This may need a clinical trial for confirmation. – Thrombolytic agent (recombinant t-PA) is effective in early treatment within 3 hours of the onset if intracranial hemorrhage is definitely excluded. Stroke syndromes caused by combined ULVWF and TF paths (THS) – To be individualized with general care for underlying pathology. Stroke syndromes caused by lone TF path (AHS) – Desmopressin may be beneficial for AHS within few hours if the patient is available on-site. Need a carefully designed clinical trial. – rFVIIa is not only ineffective but also is contraindicated based on in vivo hemostatic theory because of unneeded systemic fibrinogenesis causing fibrin clot disease and potential macrothromboembolic complication.

Abbreviations: AHS, acute hemorrhagic stroke; AIS, acute ischemic stroke; NAC, N-acetyl cysteine; rFVIIa, recombinant activated factor VII; TF, tissue factor; THS, thrombo-hemorrhagic syndrome; TIA, transient ischemic attack; t-PA, tissue plasminogen activator; ULVWF, unusually large von Willebrand factor multimers.

Interestingly, anti-thrombin/antifibrinogenetic therapy of LMWH in AIS caused by large artery occlusion had reduced early neurologic deterioration. ¹³³ Perhaps this finding suggests certain macrothrombosis (eg, AIS) in in situ stroke is not a completed stroke event yet when it is diagnosed, but thrombogenesis might be a continuing process. This proposition may support thrombosis in progression could still be benefited from anticoagulation treatment in early stage of stroke. This factor should be weighed in interpreting the result of clinical trials among different groups of the same stroke phenotype.