The aetiology of deep venous thrombosis and chronic venous insufficiency

Preface

 

What we now call ‘deep venous thrombosis’ (DVT) has been studied in diverse ways during the past 200-300 years. Each of these approaches contributes to a full modern understanding of aetiology. Therefore, much of this book is a historical survey of the field. However, our remit is broader than the title might suggest: the evolution of ideas about DVT is typical in many ways of medical biology as a whole. Thus, although the aetiology of DVT may seem a narrow topic for a monograph - it implicitly excludes arterial thrombosis and marginalises prophylaxis, therapy, and even such clinically significant sequelae as pulmonary embolism – we hope to engage the reader in a much more general inquiry.

Our historical investigation reveals a 160-year-old schism between two contrasting philosophies of medical and biological research, a schism that is particularly – but by no means uniquely – relevant to the study of DVT. In principle, these philosophies should be complementary rather than competing. So while we wish to elucidate the aetiology of DVT per se, we are also concerned with a more abstract and wide-ranging issue: the future accommodation or rapprochement between two conceptual and methodological traditions.  

To be more specific: present-day ideas about the occurrence, cause and treatment of DVT are dominated by a ‘consensus model’ that was authoritatively articulated during the early 1960s. This model attributes venous thrombosis to a combination of ‘hypercoagulability’, ‘stasis’ and ‘intimal injury’, and presumes on this basis to share common ground with the framework dubbed ‘Virchow’s triad’. In fact, as historical exegesis reveals, the consensus model originated and developed as a by-product of the century-long haematological/biochemical investigation of bleeding diatheses and in vitro studies of blood coagulation that began with the studies of Buchanan in the 1830s and was carried forward in the work of Schmidt and his successors. Scientifically, it owes little to Virchow; philosophically, it owes even less. An up-to-date re-evaluation of Virchow’s actual contribution to the study of thrombosis and embolism is clearly indicated.

The two contrasting approaches to research are clearly identifiable in this aspect of the history of medicine: the ‘mechanistic’ viewpoint, pioneered by Cartesians such as Hoffmann, La Mettrie and Boerhaave, re-articulated by du Bois Reymond and his associates in the mid-19^th century, and reflected today in (e.g.) the consensus model of DVT; and the ‘pathophysiological’ viewpoint, rooted in Harvey’s work, articulated by (e.g.) Hunter, Virchow and Lister, and nowadays marginalised. We suggest that in view of the significance of the work of Virchow and his successors, the ‘pathophysiological’ or ‘vital-materialist’ approach to DVT – and to other areas of medicine and biomedical[1] research - merits reflection, debate and reconsideration. 

Fundamentally, whereas pathologists sought for a century to find and explain the cause of DVT - and manifestly failed to do so - proponents of the consensus model after the Second World War effectively marginalised ‘cause’ and concentrated on therapy and pharmaceutical prophylaxis. The clinical value of this approach is not in question, but it should not be allowed to conceal the gap that it leaves in our understanding of aetiology. The mechanistic approach has enabled the coagulation (fibrinogenesis) process to be elucidated, and leukocyte and platelet congregation and the responses of cells to hypoxia to be characterised in cell-biological and molecular terms. All these are necessary for understanding the aetiology of DVT, but collectively they are not sufficient. The Virchowian, ‘pathophysiological’ literature shows that the circulatory impairment crucially associated with DVT is not reduced linear blood flow rate but non-pulsatile movement, and that ‘endothelial injury’ as a cause of DVT is essentially confined to the parietalis endothelium of the valve cusp leaflets. The ‘mechanistic’ literature articulating the consensus model ignores these crucial points. As it stands, the consensus model tends to confound ‘cause’ with ‘predisposing factors’, obfuscates the circulatory factors tending towards thrombosis under the misleading term ‘stasis’, and loses sight of the fundamental role of venous valves in the formation of thrombi.

More generally: the consensus model does not wholly satisfy the epistemological criteria for an account of DVT aetiology. We argue a case for reconsidering the ‘valve cusp hypoxia hypothesis’ (VCHH), which appears to meet these criteria more completely. The VCHH is in the pathophysiological or ‘vital-materialist’ mould of Harvey, Hunter, Virchow and Lister; that is, its focus is on malfunction at the physiological level. However, as we endeavour to show, it can be reconciled productively with mechanistic accounts of the blood coagulation mechanism and the molecular biology of the venous endothelium. We believe that this reconciliation leads to a novel, intellectually productive and clinically useful account of DVT: an illustration of the unification of traditions that we consider desirable in biomedical research as a whole.

The epistemological/metaphysical overtones of the history we review, and their significance for biomedical research in general, are summed up in an appendix.

 

Synopsis

During the course of this book, two quite different approaches to the study of deep venous thrombosis (DVT) are evaluated and ultimately reconciled. At the outset, some of the material may be unfamiliar; for instance, the distinction between clot and thrombus is often not recognised, though it is fundamental to an understanding of DVT aetiology. We therefore begin by outlining the plan of the book and indicating the ‘end-point’ reached in chapters 11-13.

            Chapter 1 is a general introduction:

• The frequency and incidence of deep venous thrombosis (DVT) and the major known predisposing factors are summarised, and ‘traveller’s thrombosis’ is mentioned.  
• The main pathological consequences of the condition are noted: pulmonary embolism, the main cause of mortality; and post-thrombotic syndrome (a general label for the morbid sequelae, involving valve incompetence and leukocyte infiltration).
• What we term the ‘consensus model’ of DVT, which attributes the condition to some combination of hypercoagulability, stasis and endothelial injury, is then introduced; some variants of this model are outlined.
• A short critique of the consensus model is given, followed by a conjecture about its lasting prominence in the field, and this is followed by a critical discussion of the phrase ‘Virchow’s triad’. An alternative model of DVT is likewise briefly sketched.
• Because both the consensus model and the alternative viewpoint have deep historical roots, we argue that historical exegesis is necessary to explain their respective origins and development and to establish a rational understanding of the aetiology of DVT.

Most of chapter 2 is devoted to a review of the blood coagulation mechanism as it is understood today. No matter how the aetiology of DVT is considered, blood coagulation is clearly involved at some point in the process:

• The development of the blood coagulation ‘cascade’ concept after the Second World War is reviewed in conjunction with the then-increasing clinical use of anticoagulant and (subsequently) thrombolytic treatments. Those developments accompanied the emergence of the consensus model.
• We believe that the consensus model arose from a research tradition essentially unrelated to thrombosis, and that its articulation (a) effectively ousted the then-current pathophysiological viewpoint dating back to Virchow and (b) marginalised studies of the venous endothelium in relation to DVT.
• During this process, Virchow’s explicit distinction between ‘clot’ and ‘thrombus’ fell out of use. Virchow had likened thrombi to clots, but he realised the limitations of his analogy and insisted that the histology of thrombi proves that they form in flowing blood, not in static blood (as clots do).
• Some additional semantic points are raised for later discussion and development.

 The theme of chapter 3 is ‘hypercoagulability’:

• The common assumption that Virchow envisaged ‘hypercoagulability’ as a cause of venous thrombosis is refuted. Two uses of the word ‘hypercoagulability’ are distinguished: a general sense, which entails circular reasoning, and a specific sense, for which we substitute the synonym ‘thrombophilia’.
• The consensus hypothesis that ‘hypercoagulability causes DVT’ yields three predictions. One of these is shown to be weakly corroborated by early studies on ‘experimental thrombi’.
• The second and third predictions are evaluated through a review of the literature on hereditary and acquired thrombophilias and are shown not to be supported by the available evidence.
• It is concluded that thrombophilias increase the likelihood of DVT but are not to be considered ‘causal’. This is followed by a critical discussion of the clinical value of laboratory tests for thrombotic tendencies.
• The need for an alternative to the consensus model of DVT aetiology is re-emphasised.

 We begin the historical exegesis in Chapter 4:

• Two approaches to biomedical investigation are distinguished: the ‘(patho)physiological’, associated with Harvey, Hunter and Virchow, and the ‘mechanistic’, associated with Boerhaave, du Bois Reymond and the consensus model of DVT.
• The origins of both traditions are traced to the 17^th century, the Scientific Revolution and the emergence of natural philosophy.
• Their developments into the 18^th and early 19^th centuries are outlined.
• We identify the lines of investigation that led on the one hand to the elucidation of the blood coagulation mechanism, and on the other to the seminal contributions of Virchow.
• We take particular care to distinguish both mainstream traditions from the once-misleading notion of ‘vital force’, and we consider other possible sources of semantic confusion.

 Chapter 5 pursues the first of these two lines of investigation, which led to the elucidation of blood coagulation:

• Studies of blood coagulation had originated among the ‘animal chemists’ of the 18^th and early 19^th centuries, but were placed on a firmer footing in the 1830s by Andral’s pioneering compilation of medical knowledge and by Buchanan’s experimental studies.
• From that time, the history of blood coagulation research can be divided into four arbitrary phases. This chapter concerns the first two phases; the third and fourth phases, covering the period between the Second World War and the present, were explored in chapter 2.
• Phase 1 extended from Buchanan’s publications to the maturation of Schmidt’s ‘classical hypothesis’, when the action of thrombin was identified and the existence of prothrombin and antithrombins was inferred.
• These advances were underpinned by a philosophical movement known as ‘mechanistic materialism’, which (inter alia) marginalised the role of cells in blood function by focusing attention on the soluble components of the plasma.
• Phase 2, covering roughly the first half of the 20^th century, saw the identification of prothrombin and of an ever-increasing number of coagulation factors, and the discoveries of heparin, vitamin K and dicoumarol. By the end of this period, the mechanistic-materialist character of coagulation research was becoming entrenched.  
• Finally, we discuss the origin of this philosophical movement in a schism that took place around 1847. This schism primarily involved Emil du Bois Reymond, the leading protagonist of mechanistic materialism, and Rudolf Virchow, the pioneer of cellular pathology and an exponent of the ‘pathophysiological’ approach to biomedicine.

 Chapter 6 pursues the second of the two lines of investigation mentioned in chapter 4. The focus is on Virchow’s studies of thrombosis and embolism:

• The ‘doctrine of Cruveilhier’ is discussed in order to evaluate the then-customary use of the word ‘phlebitis’ in relation to what we now call ‘thrombosis’.
• Virchow’s life, times and philosophy are reviewed, and we emphasise his complete familiarity with the work of Boerhaave and Hunter as well as that of Cruveilhier and the early 19^th century microscopists who followed him.
• Virchow synthesised those contributions to knowledge by demonstrating that pulmonary emboli originate by metastasis (the word he used in this context) of peripheral venous thrombi to the pulmonary artery. He proved that thrombus ‘embolia’ form in moving, not static, blood; illustrated thrombi apparently associated with the cusps of venous valves (both parietal and ostial);  distinguished clearly between thrombi and clots; and surmised that oxygen is required for thrombosis.  
• Many authors have queried the concept termed ‘Virchow’s triad’. We concur that it perhaps originated in a misapplication of his work
• We discuss Virchow’s opposition to Cruveilhier and his ‘pre-microscopy’ philosophy of biology and medicine.

 Later developments in the Virchowian tradition of investigation are reviewed in chapter 7:

• The words ‘inflammation’, ‘phlebitis’ and ‘pus’ are immensely problematical. They have denoted different entities, and carried markedly different connotations, in different historical eras. These shifts of meaning make 21^st century interpretation of 18^th or 19^th century publications very difficult. Failure to take account of this difficulty has led, and will continue to lead, to persistent misunderstandings.
• The prime casualty of these semantic problems is the once-prominent association of ‘phlebitis’ with venous thrombosis. Indeed, resolution of that misleading association is a key outcome of our historical exegesis.
• The second interpretation problem to be resurrected is the involvement of leukocytes (alongside platelets) in the aetiology of DVT, re-focusing attention on the origin of venous thrombi on the central surfaces of valve cusps. The approach to DVT aetiology implicit in this discussion will be developed progressively in the following chapters.

 The themes of chapter 8 are the meaning of ‘stasis’, the variables of venous blood flow, and the significance and functioning of venous valves in DVT:

• The notion that ‘blood stasis’ is a causal or contributory factor in DVT has Galenic connotations and connections, which are intellectually problematical.
• In the literal English sense (i.e. absolute cessation of movement), ‘stasis’ entails local or organismic death. It has been shown experimentally to be ‘anti-coagulatory’ and must therefore presumably be ‘anti-thrombogenic’.
• The use of ‘stasis’ to denote ‘retarded flow’ or ‘interrupted flow’ is confusing not only for semantic and historical reasons, but also because it focuses attention on mean blood velocity in veins rather than on the pulsatility of blood movement.
• We contend that the (temporary) cessation of pulsatility in venous blood movement is instrumental in DVT.
• The history of the discovery of venous valves, and Harvey’s work in establishing their function, is the background to our reassessment of the evolution and proper interpretation of the ‘stasis’ concept.
• Our contention is that sustained non-pulsatile (streamline) blood flow in veins results in local hypoxaemia in affected venous valve pockets (VVP), the refilling of which furthers the initiation of a potentially thrombogenic sequence of events.

 Chapter 9 focuses on the valves and valve pockets:

• Our key presumption is that the first stage in the aetiology of DVT is sustained underperfusion of the VVP during extended periods of non-pulsatile flow.  Therefore, the structure, function and pathology of venous valves are discussed. The questions to be answered are: why and how does underperfusion of a VVP potentially lead to thrombosis; and where in the valve - and how - is thrombus formation initiated?
• Valve morphology and function, and certain findings at post-mortem examination, provide the key evidence that VVP are indeed the sites of venous thrombogenesis. The pathological circumstance arises when long-underperfused VVP result in endothelial hypoxia, specifically of the inner (parietalis) surface of the valve cusp(s), and that repetition and persistence of such special conditions may initiate DVT.
• On the presumption that this hypothesis is broadly valid, we calculate the approximate time needed for non-pulsatile venous flow to become pathogenic.
• Hypoxic death of the valve cusp parietalis endothelium during phases of streamline local venous blood flow is treated as instrumental in thrombogenesis.
• This hypothesis accords with some mid-20^th century literature on the pathogenic effects of endothelial hypoxia.
• Published micrographs of venous thrombi are reviewed to elaborate this hypothesis and to establish the stage-by-stage process of thrombogenesis.
• Our account also relates to the induction of DVT by non-fatal carbon monoxide poisoning, to the margination of leukocytes on hypoxic endothelia, to Aschoff’s autopsies of First World War victims, and to recent conflicting evidence about ‘traveller’s thrombosis’.

The idea that valve cusp endothelial hypoxia is instrumental in the aetiology of DVT is extended in chapter 10:

• Hypoxic death of the valve cusp parietalis endothelium during phases of streamline local venous blood flow is treated as instrumental in thrombogenesis.

• This hypothesis accords with some mid-20^th century literature on the pathogenic effects of endothelial hypoxia.
• Published micrographs of venous thrombi are reviewed to elaborate this hypothesis and to establish the stage-by-stage process of thrombogenesis.
• Our account also relates to the induction of DVT by non-fatal carbon monoxide poisoning, to the margination of leukocytes on hypoxic endothelia, to Aschoff’s autopsies of First World War victims, and to recent conflicting evidence about ‘traveller’s thrombosis’.

The valve cusp hypoxia hypothesis is stated in full in chapter 11:

• The VCHH describes the aetiology of DVT in terms of:
  
  • venous valve cusp form and function,
  • hypoxic necrosis of the VVP parietalis endothelium,
  • the active response of viable leukocytes and platelets to the hypoxic or dead endothelium.

It emphasises the gradual, sequential nature of venous thrombogenesis and focuses particularly on why and how blood that entered a VVP ‘alive’ may become ‘dead’ if it is not expelled until some hours later.

• Experimental confirmation of the premises and predictions of the VCHH is reviewed.
• We discuss the sharp distinctions between the VCHH and the consensus (haematological) model in scientific character, medical implications and philosophical connotations.
• DVT ‘risk factors’ are explained in the light of the VCHH
• The prophylactic applications of the hypothesis are discussed.

 In chapter 12, the VCHH is supplemented and enriched by recent discoveries in the molecular biology of endothelial cells and their responses to hypoxia:

• Gene expression patterns in endothelial cells (EC), presumably including those of the valve cusp parietalis, change in response to hypoxia and other challenges.
• These changes, the signalling networks involved and the consequences for cell phenotype have been elucidated in considerable detail, providing a mechanistic underpinning for the VCHH. In particular, they define mechanisms for:
  • the increased congregation and anchoring of leukocytes and platelets on the hypoxic area,
  • the effects of activated neutrophils on the injured vascular endothelium,
  • enhanced blood coagulation in the immediate neighbourhood.

• A significant part of the molecular-biological literature in this field concerns the effects of hypoxia on vasodynamics. We presume that this has limited relevance to the aetiology of DVT, though it must be pertinent to perfusion of venous vessel walls via the vasa venarum.
• We hope that this assimilation of ‘mechanistic’ findings into the ‘pathophysiological’ VCHH exemplifies how biomedicine might adopt a re-unified approach to problems, as outlined in the preface and discussed further in chapter 4.
  

Chapter 13 illustrates the value of an integrated, a priori account of the value of DVT by explaining the variable condition of post-mortem blood:

• Post-mortem blood may either consist of semi-solidified reddish masses (resembling either clots, thrombi or their combinations), or be entirely ‘white’ thrombi; or, in occasional cases, be wholly liquid and incoagulable. Such variations have confused pathologist observers, and we suggest an alternative explanation in terms of the VCHH (not an original explanation, as will be seen).
• The central question is whether it can ever be rightly imagined that intravascular blood coagulates after death. This question was intensely debated in the early 20^th century but was then resolved ad hoc.
• To attempt a resolution, we re-explore the First World War controversy, reconsider the forensic and judicial interpretations/ implications of the states of post-mortem blood, and re-draw some clinical inferences.
• This leads to the conclusion that so-called ‘post-mortem’ thrombi or clots - when they are not wholly absent - are invariably agonal thrombi.

The Appendix reflects on the philosophical underpinnings of the study of DVT throughout history and in the present day.

 

Members

Log in