International
Journal of
the History
and
Philosophy of Medicine
Int J Hist Philos Med (2011) 1:7-14
HISTORICAL REVIEW
Deep Venous Thrombosis: Hunter, Cruveilhier, Virchow, and
Present-Day Understanding and Clinical Practice
Paul S. Agutter,1∗ P. Colm Malone2
1
Theoretical Medicine and Biology Group, Glossop, Derbyshire, SK13 7RR, UK
2
129 Viceroy Close, Birmingham, B5 7UY, UK
ABSTRACT
During the past 50 years, systematic misunderstandings of Virchow’s contribution to the study of deep venous thrombosis (DVT) have distorted thought and opinion in this field. Virchow’s main concern was with the formation of pulmonary
emboli, not with the generation of venous thrombi per se. In particular, the elements of what has become known as ‘Virchow’s triad’ had nothing to do with the formation of thrombi but with their metastasis to produce emboli. Nevertheless,
Virchow made two particularly important observations about venous thrombogenesis that have been largely overlooked
since the 1950s: thrombi are initiated in valve pockets; and their initiation is associated with a massive local accumulation of leukocytes. He also distinguished clearly between in vivo thrombi and ex vivo blood clots. In this paper we explore the background and motivation to Virchow’s work, which lies in the earlier studies of John Hunter and Jean
Cruveilhier, and the methods he used, particularly his reliance on the recently-improved microscope. We also consider
the work of Virchow’s successors and explore the circumstances leading to the eventual distortion of his legacy. IJHM.
Int. 2010; 1: 7-14. ©2010 Int J Hist Philos Med, Inc.
Key Words: History, Vascular, Vein, Thrombosis
INTRODUCTION
Aspects of the history of medicine help us to appreciate our traditions of thought and practice, and in
some cases they elucidate present-day understanding
of disease pathology, therapy and prophylaxis. This
point is well illustrated by modern perspectives on
deep venous thrombosis (DVT) and its treatment and
management. In the present paper we re-evaluate the
contributions of Hunter, Cruveilhier and Virchow to
this field. Most modern medical textbooks and teaching courses misconstrue Virchow’s seminal work on
thrombosis and embolism and say little or nothing
about his predecessors. Our discussion will explore
the evolution of ideas about the aetiology of DVT,
and will inter alia consider the way in which the
word ‘inflammation’ has been used in relation to
what we now call ‘deep venous thrombi’ and ‘emboli’.
∗Address correspondence to Paul S. Agutter, PhD, Theoretical
Medicine and Biology Group, 26 Castle Hill, Glossop, Derbyshire, SK13 7RR, U.K. Tel: (44)-1457-867679; E-mail:
[email protected]; or P. Colm Malone, MD, 129 Viceroy Close,
Birmingham, B5 7UY, U.K. E-mail: [email protected]
Submitted April 27, 2009; accepted in revised form May 31,
2010.
Advance Access Publication 13 January, 2010 (see www.ijhom.org)
The article is organized as follows. The first substantive section summarizes and compares two modern-day accounts of the aetiology of DVT. The second examines some aspects of Virchow’s writings
about venous thrombi and emboli and corrects a
common misconception. The third considers the origin and fate of the ‘phlebitis’ concept and evaluates
the idea of ‘inflammation’ in relation to DVT. The
fourth reflects on the involvement of venous valves
in thrombogenesis and argues that this involvement
was implicitly recognized by John Hunter. Finally, a
concluding section comments on the implications of
this study for our appreciation of recent medical history and our current understanding and management
of DVT.
The pathological lesion that Hunter termed ‘inflammation of the internal coats of veins’, mistranslated as phlebitis by his successors after his death,
was renamed ‘thrombosis’ by Rudolf Virchow circa
1856. We contend that John Hunter’s theory of life,
and his naked-eye inspection of the lesions, provided
the original framework of reference in terms of which
scientists – including Virchow - sought to understand
the phenomenon for 154 years (until 1947). Those
who prefer mechanical prophylaxis to anticoagulant
treatment for DVT today are still adhering to
Hunter’s (and by implication, Virchow’s) viewpoint,
but they have become a minority.
2 Paul S. Agutter and P. Colm Malone / Int J Hist Philos Med (2011) 1:7-14 THE ETIOLOGY OF DVT
Two accounts of the aetiology of DVT are current.1
The consensus model views DVT as a haematological
disorder and attributes it to some combination, not
well characterized, of ‘stasis’, ‘hypercoagulability’
and ‘vessel wall injury’. The valve cusp hypoxia hypothesis (VCHH) holds thrombogenesis to be initiated in venous valve pockets under conditions of suffocating hypoxaemia; given certain well-defined circumstances, these incipient nidi may develop into
fully-fledged thrombi.
The consensus model
The consensus model is the more widely-held account. It provides a prima facie rationale for anticoagulant and thrombolytic prophylaxis and treatment,
and indeed it has evolved along with these pharmaceutical approaches to clinical practice. Scientifically,
it has motivated research into hereditary and acquired
thrombophilias, and into the molecular responses of
the venous endothelium to hypoxic conditions.2 Research in both these areas has proved fruitful. The
drawback of the consensus model is that it does not
provide a clear aetiological account of DVT. Indeed,
the terms ‘stasis’, ‘hypercoagulability’ and ‘vessel
wall injury’ are not clearly defined or consistently
used.
The consensus model was first articulated during
the early 1960s, coinciding with the acceptance of
anticoagulants for clinical use against DVT and embolism.3,4 Its roots lie in the scientific study of blood
coagulation, which dates from the early 19th century
but underwent rapid development during the mid 20th
century. At its outset, this development was largely
driven by the identification of haemophilia in the
British royal family and the need to determine the
cause of the condition and find an effective treatment.
A review by Eagle in 1938 summarised what was
then known of the coagulation mechanism.5 Strikingly, although Eagles mentioned Andral’s isolated
description (1843) of an excessive tendency to coagulate,6 he made no suggestion that haematological research could have any significance for our understanding of DVT.
Immediately after the Second World War, however,
the picture changed. The burgeoning of haematology
as a distinct discipline, fostered by methodological
and conceptual advances in biochemistry, heralded
the birth of the consensus model. During the 1930s,
Dr Paul D White, General Eisenhower’s physician
and a founder of the American Heart Institute, had
been surprised to realise that DVT killed many of his
heart failure patients as well as surgical and decubitus
patients.7 His reaction may have contributed to the
quest for pharmaceutical rather than mechanical approaches to prophylaxis and treatment and thus, later,
to the consensus model. The mechanical and pharmaceutical paradigms associated with the two accounts
of the aetiology of DVT may thus be viewed as re-
flections, respectively, of surgical and medical ideologies and approaches. But the pharmaceutical paradigm soon became pre-eminent.
A 1947 article by Pulvertaft revealed the bewilderment and anger that this development inspired
among followers of the Hunterian tradition.8 In his
introduction, Pulvertaft wrote:
‘The presence of large and well-organized clots of
blood in the pulmonary arteries has long been recognized, but in spite of Harvey's demonstration of the
mechanics of the circulation of the blood, it was left
to Virchow to explain their true nature a hundred
years ago. Since then an enormous literature has
developed around the subjects of thrombosis and embolism… They have perhaps attracted an attention
greater than their clinical importance warrants, but
elementary students of physiology are sometimes
puzzled to hear so many of their masters weaving a
tortuous and highly individual path through the
polysyllabic mazes of blood coagulation… Yet the
penultimate word on the question was written by
Welch in Albutt's System of Medicine in 1897,9 and
much that has been written since has been a bungaloid growth rather than an architectural synthesis; on
re-reading Welch’s excellent study, one feels that
today’s students would be well advised to remember
that there were great men before the Agamemnons of
modern medicine.
Welch’s seminal study of thrombi (actually published in 1899)9 was decidedly in the tradition of Virchow and was justly celebrated. Yet within a decade
of Pulvertaft’s diatribe against the incursion of haematology into the study of DVT, Virchow’s name
had become associated with the so-called ‘triad’ of
stasis + hypercoagulability + vessel wall injury, the
‘Agamemnons of modern medicine’ had begun to
articulate the consensus model, and Welch’s contribution had been more or less forgotten.
The valve cusp hypoxia hypothesis
The VCHH1 may be less familiar to many readers
than the consensus model, so a more detailed description is needed here. It holds that DVT may occur if
there is sustained non-pulsatile (streamline) venous
blood flow, so that the valve cusp leaflets are not
‘flapped’ and the blood in the valve pocket is therefore not exchanged with that in the vein lumen. This
leads to suffocating hypoxaemia in the valve pockets,
resulting in hypoxic injury to the inner (parietalis)
endothelium of the cusp leaflets. Such prolonged
oxygen insufficiency of the endothelium stimulates
the elk-1/egr-1 pathway, which initiates most of the
constellation of responses of endothelial cells to hypoxia and activates a number of chemoattractant and
procoagulant factors.2 When normal pulsatile blood
flow is restored, even transiently, leukocytes and
platelets (attracted by these factors) marginate and
sequestrate at the site of injury to ‘cover and heal’ the
now moribund endothelial layer. These cells selfattach to the necrotic epithelium and constitute the
Int J Hist Philos Med (2011) 1:7-14 / Venous thrombosis 3
first layer of what may become an incipient venous
thrombus.
A frank thrombus may subsequently develop if
regular contractions fail again for a prolonged period,
so the valve pocket is not emptied and refilled with
adequately oxygenated blood. Each further period of
non-pulsatile flow kills all accumulated blood cells in
such a valve pocket. Cells oxygen-starved by each
subsequent immobilization, and expelled back into
the vein lumen, are repetitively replaced by more live
cells charged with oxygen, which are fated to die
when immobilised on the growing sequence of layers
that build up to form the core of the nascent thrombus. If periods of non-pulsatile and pulsatile flow
continue to alternate, serial deposition of white cells
and fibrin may ensue, sequentially or at distinct intervals, and this results in the characteristic ‘Lines of
Zahn’ morphology of a venous thrombus. Only the
blood cells on the outermost layer of a thrombus are
still living.
The VCHH thus re-explains many of the recognized causes of (risk factors for) DVT (e.g. prolonged
inactivity, especially a motionless state while undergoing muscle-relaxant anaesthesia) and it accounts
for the morphology of thrombi. It also predicts that a
venous thrombus will readily embolise, because the
area of endothelium to which it is anchored, the valve
cusp parietalis on which it has grown, is necrotic (by
definition) and is therefore readily detached by the
restoration of pulsatile blood-flow past the obstruction. It provides a prima facie rationale for accepted
non-pharmaceutical approaches to prophylaxis (e.g.
early ambulation, flexion and extension of the ankles,
support stockings and intermittent pneumatic pressure leg devices).
It also suggests a new approach to preventing hospital-acquired venous thromboembolism. Having
indicted the pathogenicity of venous valve pockets
undrained for many hours, it follows that the risk of
DVT can be eliminated by ensuring that all venous
valve pockets in the lower limbs are emptied gravitationally at (minimally) half hourly intervals by slight
head-downward tilting of the bed, alternating every
1–1.5 hours with slight foot-downward tilting to preclude the risk of thrombogenesis in the valve pockets
of the upper limb veins.1
Like the consensus model, the VCHH was also
conceived in the 1960s and received powerful experimental support in the late 1970s and early
1980s,10 but its roots are much older.11 They can be
traced to Virchow, and beyond Virchow to Cruveilhier and Hunter. We shall explore this historical development in what follows.
A seeming paradox
After 1962, the consensus model marginalised and
even eclipsed the pathophysiological tradition of
Hunter and Virchow, at least temporarily; Ockham’s
Razor seems to forbid the coexistence of two explanations for the same phenomenon. However, the
originators of the consensus model attached the label
‘Virchow’s triad’ to the alleged pathogenic circumstances (stasis + hypercoagulability + vessel wall
injury). It may seem surprising that both the VCHH
and the consensus model claim to be rooted in the
work of Virchow, who clearly did not advance two
apparently incompatible accounts of the same process. We need to examine Virchow’s work to resolve
this seeming paradox.
WHAT VIRCHOW WROTE AND DID NOT
WRITE
Virchow’s motivation and approach
Rudolf Virchow (1821-1902) began to work on the
aetiology of pulmonary emboli in 1844, under the
direction of Robert Froriep, who set him the challenge of examining the ‘Doctrine of Cruveilhier’ (i.e.
that ‘phlebitis is the root of all pathology’).12 By
1848, he had shown that the semi-solid masses of
blood that form, often fatally, in pulmonary arteries
are not produced in situ but have metastasised from
the peripheral circulation. The papers he published on
these studies, some based on animal experiments and
some on clinical cases, were brought together in his
Thrombose und Embolie,13 published in the year
(1856) that he took up the post of Professor of Pathological Anatomy created for him in Berlin.
Thrombose und Embolie was a remarkable work,
emphasizing a reliance on scientifically demonstrable
fact rather than conjecture, and introducing nomenclature that still survives. (For example, Virchow
gave ‘thrombus’ and ‘thrombosis’ the definitions we
accept today, and he invented the words ‘embolus’
and ‘embolism’.) It may be argued that the discovery
that all pulmonary emboli originate from peripheral
vein thrombi was not original with Virchow, since
Joy14 had come close to this realization in 1840 and
van Swieten15 had broadly grasped the idea as early
as 1776. However, the overwhelming mass of experimental and clinical evidence in Thrombose und
Embolie was entirely Virchow’s, and it made the
conclusion irrefutable.
What is now called ‘Virchow’s triad’ arises from
a misunderstanding of Thrombose und Embolie
For the purposes of the present article, the important
point is that Thrombose und Embolie addresses the
cause of emboli, not the cause of venous thrombi. In
other words, it is not about the aetiology of DVT at
all. The earliest reference we can find to ‘Virchow’s
triad’ as an alleged account of the aetiology of DVT
(usually interpreted as stasis + hypercoagulability +
vessel wall injury) is the 1957 paper by Anning,16
written almost exactly 100 years after Virchow’s
seminal publication. The ‘Virchow’s triad’ concept
seems to have arisen from a misinterpretation of the
following passage from Thrombose und Embolie (pp.
293-294 of the Matzdorff-Bell translation):-
4 Paul S. Agutter and P. Colm Malone / Int J Hist Philos Med (2011) 1:7-14 ‘In the final assessment, the mechanical as well as
the chemical characteristics of the thrombus [in the
pulmonary artery] were studied. The hard rough surface had caused grave damage to the interior wall of
the lung artery… causing inflammation of the lung at
least equivalent in gravity to the effects of the chemical action resulting from the organic substances introduced by accident into the lung… Accordingly, the
sequence of the possible stages and consequences of
blockage may be classified and studied under three
headings:1. phenomena associated with irritation of the vessel and its vicinity;
2. phenomena of blood-coagulation; and
3. phenomena of interrupted blood-flow.’
This list of putative ‘phenomena’ ostensibly resembles ‘Virchow’s triad’. However, from the context, it
is obvious that Virchow was writing about emboli as
‘bungs’ in pulmonary arteries, not about thrombogenesis. Therefore the quoted passage has nothing
to do with the formation of thrombi, only with the
consequences of their metastasis. Other authors have
recognized this widespread misinterpretation,1,11,17-20
but it still persists.
Virchow versus Cruveilhier
Virchow’s primary objective had been to prove that
what Cruveilhier had illustrated as spontaneous/in
situ pulmonary ‘phlebitis’ did not arise de novo in the
(narrowing) pulmonary artery, but came from a peripheral site; in other words, he had set out to refute
the ‘Doctrine of Cruveilhier’. This objective entailed
his replacement of the word ‘phlebitis’ with ‘thrombosis’. His determination to refute Cruveilhier may
have been influenced by the mutual antagonism between German and French scientists during the 19th
century, fostered by the deliberately anti-French
stance adopted by the influential philosopher Herder,
who exhorted his fellow-Germans to ‘spit the green
slime of the Seine from their mouths’.21 There are
other instances of Virchow’s refusal to recognize
French discoveries.11
Virchow’s dismissal of his great French predecessor had unfortunate consequences: posterity has
failed to accord Cruveilhier the respect he merits.22,23
Cruveilhier held the first chair of Pathological Anatomy in history, and among his many achievements he
correctly explained the cause of gastric ulcers, gave
the first description of multiple sclerosis and recognized congenital liver cirrhosis. We shall say more in
the next section about his contribution to our understanding of DVT.
Thrombi resemble blood clots but are not the
same as blood clots
Nowhere is Virchow’s emphasis on the tangible and
demonstrable better illustrated than in his comparison
of thrombi with blood clots.24 Everyone has seen
blood clotting ex vivo. Virchow’s microscopic studies
of thrombi (which his readers had not seen) led him
to say that in certain ways they resemble clots (which
they had seen). He criticized Cruveilhier for refusing
to use the microscope (Cruveilhier was 35 years old
before Lister’s achromatic lens made the microscope
into a reliable instrument, so his lifelong doubts about
the instrument are understandable) and was himself a
committed microscopist. However, he was careful to
point out some important morphological differences
between in vivo venous thrombi and ex vivo clots:‘1. the structure of the thrombus is in layers [i.e. the
lines of Zahn];
2. the fibrin content is denser;
3. the population of colourless corpuscles is denser,
and to a striking degree; these corpuscles were present in the blood from the beginning, and were separated from it with the fibrin… Thus there is evidence
for spontaneous coagulation of the blood within the
vessels, visible where the continuity of the vessels is
interrupted, e.g. by a wound. This large segregation
of colourless corpuscles is clearly associated with the
retardation of the circulating blood.’
In other words, a thrombus is not a clot, or even
like a clot; the resemblance is superficial. (We might
say that this tripartite distinction between ‘clot’ and
‘thrombus’ has a better claim to bear the epithet ‘Virchow’s triad’ than the modern-day consensus-model
mantra ‘stasis + hypercoagulability + vessel wall injury’.) The first point of distinction, the layered structure of the venous thrombus, had been recognized by
Lobstein25 and Bristowe26 but not described so informatively. There was no more detailed exploration
of thrombus structure until 1876, when the newlyimproved Zeiss microscope became available to
Zahn.27 Virchow’s precise drawings of thrombi13,28
make his skill and care as a microscopist evident.
The third point of distinction between thrombus
and clot states that the ‘colourless corpuscles’ forming much of the ‘root’ of the thrombus originate from
the circulating blood. This is ‘obvious’ to our minds,
so why was Virchow so emphatic about it? We shall
return to that question in the following section.
The third point of distinction also indicates the association of thrombi with injured areas of the vessel
wall. In this, Virchow tacitly echoed Cruveilhier,22,23
who believed that what we now call ‘thrombi’ were
always caused by vessel wall injury. The principal
questions are: how may a vein wall be injured without outside (experimental) intervention, and which
specific parts of the vein wall are at risk of such injury?
Venous thrombi form in valves
Virchow’s careful drawings of thrombi depicted them
more or less in the centre of the vein lumen, with
streaks of white material running throughout their
length. Although no valves are shown in most of
these illustrations, Virchow nevertheless stated that
‘the thrombi are seated on the valves’. Notwithstand-
Int J Hist Philos Med (2011) 1:7-14 / Venous thrombosis 5
ing the fact that Thrombose und Embolie is about the
pathogenesis of emboli, not DVT per se, this brief
remark was of the greatest importance for elucidating
the aetiology of DVT and was enough to establish
Virchow’s pre-eminence in the history of thrombosis
research. In his 1858 collection of lectures, Cellular
Pathology, there is a clear illustration of a thrombus
anchored to a valve.28
Virchow and the two accounts of DVT aetiology
We can now propose an answer to the question posed
earlier: why do the two different present-day accounts of the aetiology of DVT, the consensus model
and the VCHH, both claim Virchow as precedent? In
the consensus model, the ‘stasis + hypercoagulability
+ vessel wall injury’ concept is wrongly attributed to
Virchow and seems to have arisen from a misreading
of a passage in Thrombose und Embolie. As for the
VCHH: although Virchow said little about the causes
of DVT, he demonstrated (a) the involvement of leukocytes in the pathogenesis of venous thrombi and
(b) the fact that thrombi are formed on valves. These
two crucial discoveries will be explored further in the
following sections.
White cells and venous thrombogenesis: Hunter
and the concept of ‘inflammation’
The term ‘inflammation’ is at least 2000 years old.
For around 1700 years the leading authority on the
subject was De Medicina, compiled by the 1st century
Roman writer Celsus,29,30 in which the four traditional
signs of information - rubor, calor, tumor and dolor
(redness, heat, swelling and pain) – were identified.
These signs describe manifestation, not cause. What
we recognize nowadays as typical concomitants of
the inflammatory response, such as phagocyte invasion and the release of various cytokines and complement factors, are not signs appreciable by examining physicians; they are detectable only by specific
laboratory investigations.
The early microscopists of the 17th century, van
Leeuwenhoek, Janssen, Kepler, Malpighi and Huygens, studied ‘inflamed’ tissues and challenged the
doctrine of Celsus.31 Their findings enabled John
Hunter (1728-1793) to define ‘inflammation’ more
precisely.32 He wrote: ‘Inflammation is to be considered only as a disturbed state of parts... it is not to be
considered a disease, rather a salutary operation
[‘healing’ in today’s terms] consequent on either violence or disease’. In other words, inflammation is a
response to tissue injury; the term ‘inflammation’ is
descriptive but has no aetiological implication. Hunter
then distinguished two types of ‘inflammation’, ‘adhesive’ and ‘suppurative’. In today’s terminology, these
presumably equate to ‘degenerative inflammation’
(e.g. arthritis) and ‘infection-related inflammation’,
though when Hunter was writing the concept of ‘infection’ lay more than 60 years in the future.
‘Inflammation of the internal coats of veins’, the title
of Hunter’s 1793 publication,32 denotes what we now
(following Virchow) call venous thrombosis. What
Hunter observed were intravenous coagula. Why did
he use the word ‘inflammation’ in this context, and
why did he specifically implicate ‘the internal coats of
veins’? The latter question is challenging because he,
like Cruveilhier and Virchow after him, observed the
white-streak-coagulum in the middle of the vein lumen. The former question invokes what Hunter accepted as a signifier of inflammation, ‘pus’.
The concept of ‘pus’ in the 18th and 19th centuries
In Hunter’s day, it was well known that vein lumens
could become occluded with what he termed ‘coagulable lymph, pus and blood’.33,34 Pus was considered
to betoken inflammation, and inflammation was
manifest as pus. Although ‘pus’ and ‘purulent inflammation’ are irrevocably associated today with
bacterial activity, that association would not be established until the work of Pasteur and Lister in the
1850s.35,36 In Hunter’s day, any white amorphous
material in the body was ‘pus’. It was considered to
enter blood vessels from the surrounding tissues,
which made its appearance in the centre of the vein
lumen surprising.
Therefore, when Hunter observed an accumulation
of white material in intravascular coagula, he identified it as ‘pus’ and inferred ‘inflammation’. Since
‘inflammation’ was traditionally associated with heat,
he devoted considerable time and attention to determining whether blood coagulation liberates heat, but
found it did not. However, in view of the invariable
association of intravascular coagula with ‘pus’, he
continued to describe the formation of such coagula
as ‘inflammation’ of the ‘internal coats of veins’.
It was a few decades later, shortly after cell theory
was articulated, that leukocytes were recognized as
distinct entities. The source of ‘pus’ and its relationship to leukocytes was intensely debated during the
1840s. The English microscopists Addison37,38 and
Waller39 concluded that ‘pus globules’ and ‘leukocytes’ were identical, and that pus was normally
formed by diapedesis of leukocytes from the blood
vessels into the tissues, not the other way round, as
previously believed. Remarkably, Virchow rejected
that inference. He believed that the admixture of white
material in the interstices of red venous thrombi represented a venous coagulum associated with a focus of
pus. His generation of scientists therefore considered
‘coagulated blood inside vessels’ to merit the suffix ‘itis’. Nevertheless, he amassed evidence to show that
the leukocytes constituting the white parts of thrombi
originated from within the blood stream, not within
the surrounding tissues, and he used this discovery to
further his attack on Cruveilhier. This, it seems, is
why he was so emphatic about an observation that we
accept nowadays as a near-truism (see previous section).
6 Paul S. Agutter and P. Colm Malone / Int J Hist Philos Med (2011) 1:7-14 Cruveilhier
Gilbert Brechet (1784-1845) appears to have read
Hunter’s ‘inflammation of the internal coats of veins’
as ‘inflammation of the veins’ and translated it to the
Greek-derived word ‘phlebitis’. He may have used
the word originally in a memoir read to the Paris
Académie des Sciences in 1832.40 Brechet was a
close friend of Jean Cruveilhier (1791-1874) and succeeded him as Professor of Anatomy in 1836.
Cruveilhier adopted his interpretation of Hunter and
made free use of the neologism ‘phlebitis’. Virchow
subsequently attributed the concept of ‘phlebitis’ to
Hunter. Much more recently, Browse et al. attributed
the strange hybrid term ‘phlebothrombosis’ to the
same source.41
Cruveilhier wrote:
‘In a way, phlebitis dominates all pathology… The
first effect of all phlebitis is coagulation of the blood
and its adherence to the walls of the vessel/s… Stagnation of the venous blood, and ‘serosity’ [oedematous
swelling?] in the corresponding parts, result from obstruction of the venous circulation in the inflamed vessel when the collateral veins are not sufficient for circulation…’42 And in relation to what we would now
call a pulmonary embolus, he observed: ‘I divided
these clots and found in the centre of the main clot a
collection of pus with all the characteristics of phlegmonous pus.’
Here and elsewhere in his work, Cruveilhier revealed his capacity for careful and detailed naked-eye
observation and rational inference. If we substitute
‘thrombosis’ for ‘phlebitis’ and ‘leukocyte accumulation’ for ‘pus’, then – allowing for some overstatement
– these passages of his writing are still valid today.
In lecture 10 of Cellular Pathology, Virchow wrote:28
‘It was imagined by John Hunter that pus… was furnished as a product of secretion by the wall of the vessel. This doctrine, however, presented some difficulty
…Cruveilhier was right... that the so-called pus in the
veins never, in the first instance, lies against the wall
of the vein, but is always seen first in the centre of the
previously existing coagulum which marks the start of
the process. He imagined that the pus was secreted
from the vessel wall, but did not remain there, but by
means of ‘capillary attraction’ made its way into the
centre of the coagulum.28
Having made this admission, however, he proceeded
to attack Cruveilhier’s ‘capillary attraction’ explanation, declaring it to be ‘mystical’. Later in the same
lecture, however, he distinguished the ‘reactive blood
semi-solidification’ of thrombogenesis from the ‘inflammatory state’ of the vein wall that had previously
been alleged to cause it.28 That distinction alone constituted a remarkable advance in understanding.43
‘Pus’, infection and confusion
Virchow made a remarkable statement about the
‘white substance’ (the collected leukocytes) involved
in thrombus formation: he said it was puriform but
not purulent’. The phrase was remarkable because it
preceded Pasteur’s recognition of bacterial action,
and also preceded Lister’s realisation that bacterial
action, infection and pus formation were interrelated.35 Lister recognised that the leukocytes associated with venous thrombi had nothing to do with bacteria; his Croonian lecture clearly distinguished between infection and intravascular coagulation.36 Nevertheless, another 80 years was to pass before it was
generally accepted that bacterial infection was not
involved in the aetiology of DVT.11 The longevity of
the term ‘phlebitis’ to denote DVT testifies to the
persistence of that confusion, notwithstanding the
best efforts of Virchow, Lister and Pasteur.
The President of the Royal College of Surgeons
during Lister’s early years, Sir James Paget, was
critically dismissive of the new notions about ‘pus’
until 1879, when Lister received a standing ovation in
an international meeting in Amsterdam.44,45 This is
not to criticise Paget, but to indicate how difficult it is
for senior generations to assimilate new developments. Anyone seeking to revise ‘received wisdom’
must re-educate their seniors, whose lessons have
ossified. The same phenomenon affected Cruveilhier
– and Semmelweiss - but perhaps in reverse.
Venous thrombi originate in valve pockets
Cruveilhier, like Hunter, was puzzled by the location
of the ‘pus’ in the middle of intravenous coagula. Until
Virchow discovered otherwise, it was held that pus
was ‘exuded’ from the vessel walls and tissues external to them, so it should have been found immediately
adjacent to the vein wall, not in the middle of a coagulated red blood mass. Cruveilhier seems not to have
found a way of explaining this mysterious observation.
However, perhaps Hunter had already done so.
Hunter was cautious about the use of microscopes
and admonished his readers not to mislead themselves by ‘what can be seen through magnifying
glasses … globules etc.’.34 Significantly, he wrote
that warning - in a footnote on page 42 of his treatise
- some 30 years before JJ Lister’s achromatic lens
was introduced, when microscopes were unreliable;
so his dubiety was justified. The consequence, however, was that he – like Cruveilhier - observed intravascular coagula only by naked eye. He never
stated that thrombi are initiated on the valves, which
are impossible to see with the unaided eye when they
are smothered by a thrombus, but it seems highly
likely that he made that interpretation. Harvey had
described the valves as ‘eminences’ on the ‘internal
lining (tunicula intima)’ of the vein.46 Hunter was
familiar with Harvey’s masterpiece. Hence, we suggest, his nomenclature: inflammation (i.e. an accumulation of leukocytes) of the internal coats (the eminences on the internal linings, i.e. the valves) of
veins. As we have seen, Virchow – aided by his
skilled use of the improved microscope – recognized
Int J Hist Philos Med (2011) 1:7-14 / Venous thrombosis 7
that venous thrombi are initiated on the valves (his
illustrations in Cellular Pathology explicitly showed
them in the valve pockets) and that blood-borne leukocytes are involved in their formation. So whilst he
acknowledged the importance of Hunter’s contribution, he was considerably less generous to Cruveilhier.
The significance of leukocytes in thrombogenesis
was acknowledged by subsequent writers such as
Welch9,47 and Aschoff,48 and the role of platelets by
Bizzozero,49 though all this is largely overlooked by
proponents of the modern consensus model of DVT
aetiology. As for the other major discovery of Hunter
and Virchow: it is often acknowledged that thrombi
are initiated at the venous valves, but the aetiological
significance is seldom regarded. In fact, the association of thrombogenesis with the valves was overlooked for almost a century after Virchow, recognized during a period of two decades beginning in the
1950s, and then marginalized again as essentially
irrelevant to the consensus model.
A paper by Hadfield50 published in 1950 contains
line drawings that precisely reflect Aschoff’s ‘Kopfteil, Halsteil and Schwanzteil’ of thrombi.48 Remarkably, however, those drawings show no valves
or valve pockets. Hadfield’s thrombi therefore arose,
wholly imaginatively, from ‘wrinkles’ on the vessel
wall, perhaps implying some unspecified abnormality
of the endothelium. It seems that in Hadfield’s time,
the first half of the 20th century, the presence and significance of valves in veins had come to be overlooked or ignored.
However, only a year after Hadfield’s paper appeared, Paterson and the McLachlin brothers produced a seminal paper on valve pocket thrombi.51
Subsequent papers by the same authors52,53 further
established the valves as the site of thombogenesis, as
Virchow had described a century earlier. Later, an
excellent series of pathological studies by Sevitt
demonstrated that venous thrombi are rooted (and
therefore presumably initiated) on the inner (parietalis) surfaces of the valve cusps.54,55 These observations proved crucial for the development of the
VCHH of DVT aetiology.11 The seminal work of
Paterson, McLachlin and Sevitt is not cited by proponents of the consensus model and is therefore not
well known.
CONCLUSION
This paper illustrates some of the ways in which salient discoveries by eminent researchers can be masked
or distorted by subsequent commentators. Cruveilhier
made a major contribution to understanding the aetiology of DVT and pulmonary embolia, but Virchow’s critique has marginalized him in the eyes of
posterity. Virchow himself was concerned with the
pathogenesis of embolism, not of DVT, yet – paradoxically – he made two crucial discoveries about
venous thrombogenesis: the involvement of leuko-
cytes and the role of venous valves. Yet posterity has
not acknowledged these discoveries; indeed, the consensus model of DVT aetiology has no place for either of them. (We believe that Hunter had implicitly
recognized both, but largely because of the limitations of contemporaneous microscopes he was unable
to articulate either explicitly.) Likewise, Virchow’s
tripartite distinction between in situ thrombi and ex
vivo clots is largely ignored nowadays. Instead, most
modern writers read Virchow’s summary account of
the factors involved in pulmonary embolism as principles of the aetiology of DVT, hence the erroneous
notion of ‘Virchow’s triad’.
Associated with the distortion and misrepresentation of Virchow’s work is the fact that the consensus
model of DVT aetiology remains more widely accepted than the VCHH. A possible factor is the intimate connection of the consensus model with anticoagulant and thrombolytic prophylaxis and therapy,
and of the VCHH with mechanical prophylaxis. The
former approach to clinical practice entails profits for
the pharmaceutical industry; the latter does not.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Malone PC, Agutter PS. The aetiology of deep venous
thrombosis. QJM. 2006;99(9):581-593.
Pinsky DJ, Yan SF, et al. Hypoxia and modification of the
endothelium: implications for regulation of vascular homeostatic properties. Semin Cell Biol. 1995;6(5):283-294.
Wessler S. Thrombosis in the presence of vascular stasis.
Am J Med. 1962;33:648-666.
Mustard JF, Murphy EA, Rowsell HC, Downie HG. Factors influencing thrombus formation in vivo. Am J Med.
1962;33:621-647.
Eagle H. Recent advances in the blood coagulation problem. Medicine 1938;16:95-138.
Andral G. Essai d’Hématologie Pathologique. Paris:
Fortin, Masson et Cie Libraries; 1843.
”Paul Dudley White”. Dictionary of American Biography.
Supplement 9: 1971-1975. New York: Charles Scribner's
Sons; 1994.
Pulvertaft RJ. Post-operative pulmonary embolism. Ann R
Coll Surg Engl. 1947;1(4):181-190.
Welch WH. Thrombosis. In: Allbutt TC, ed. A System of
Medicine. Vol 6. London: Macmillan; 1899:155-165.
Hamer JD, Malone PC. Experimental deep venous thrombogenesis by a non-invasive method. Ann R Coll Surg
Engl. 1984;66(6):416-419.
Malone PC, Agutter PS. The Aetiology of Deep Venous
Thrombosis: A Critical, Historical and Epistemological
Survey. Dordrecht: Springer; 2008.
Nuland SB. The Fundamental Unit of Life: Sick Cells,
Microscopes and Rudolph Virchow. In: Nuland SB, ed.
Doctors: The Biography of Medicine. New York: Random
House: 1988:304-343.
Virchow RLK. Thrombosis and Embolie. Barth JA, ed.
Klassiker der Medizin Herausgegeben von Karl Sudhoff.
Leipzig: 1910. In: Matzdorff AC, Bell WR, eds. Canton:
Science History Publications; 1998.
Joy WB. Phlebitis, or inflammation of veins. In: Tweedie
A, ed. A System of Practical Medicine. Vol IV. London:
Whittaker & Co.; 1840.
Van Swieten G. Commentaries upon Boerhaave’s Aphorisms Concerning the Knowledge and Cure of Diseases.
Vol VIII. Edinburgh: Charles Elliot; 1776.
8 Paul S. Agutter and P. Colm Malone / Int J Hist Philos Med (2011) 1:7-14 16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Anning ST. The historical aspects of venous thrombosis.
Med Hist. 1957;1(1):28-37.
Brinkhous KM. The problem in perspective. In: Sherry S,
ed. Thrombosis. Washington: National Academy of Sciences; 1969:37-48.
Owen CA. A History of Blood Coagulation. Rochester:
Mayo Foundation for Medical Education and Research;
2001.
Brotman DJ, Deitcher SR, Lip GY, Matzdorff AC. Virchow’s triad revisited. South Med J. 2004;97(2):213-214.
Dickson BC. Venous thrombosis: on the history of Virchow’s triad. Univ Toronto Med J. 2004;81(3):166-171.
Barnard FM. Herder's Social and Political Thought: From
Enlightenment to Nationalism. Oxford: Clarendon;
1965:34.
Duckworth D. Phlebitis and thrombosis. Lancet.
1913;181(4668):481-482.
Talbott JH. A Biographical History of Medicine: Excerpts
and Essays on the Men and their Work. New York: Grune
and Stratton; 1970:531-533.
Virchow RKL. Gessamelte Abhandlungen zur Wissenschaftlichen Medizin von Rudolf Virchow. Frankfurt am
Main: Meidinger; 1856:514-515.
Lobstein JF. De l’obstruction et de l’oblitération des
veines. In: Lobstein JF, ed. Traité d’Anatomie
Pathologique II: L’Anatomie Pathologique Spéciale. Paris:
FG Levrault; 1833.
Bristowe JS. Softening clots in the heart in a case of associated renal disease – hypertrophy of the heart – pulmonary and cerebral apoplexy. Trans Path Soc Lond.
1856;7:134-166.
Zahn FW. Untersuchungen über thrombose. Zentralbl Med
Wissenschaften. 1876;10:129-153.
Virchow RLK. Die Cellularpathologie in ihrer Begründung auf physiologische und pathologische Gewebelehre. Berlin: Hirschwald; 1858: lecture 10.
Spencer WG, ed. Celsus: De Medicina. 2nd edition. Cambridge: Harvard University Press; 1971.
Spencer WG. Celsus' De Medicina-A learned and experienced practitioner upon what the art of medicine could
then accomplish. Proc R Soc Med. 1926;19(Sect Hist
Med):129-139.
Locy WA. Malpighi, Swammerdam and Leeuwenhoek.
Pop Sci. 1901;58(6):561-584.
Hunter HJ. Observations on the inflammation of the internal coats of veins. Trans Soc Improv Med Chir Knowledge 1793;1:18-41.
Abernethy J. An essay on the ill consequences sometimes
succeeding to venasection. In: Abernethy J, ed. Surgical and
Physiological Essays. London: James Evans; 1793.
Hunter J. Treatise on the blood, inflammation and gunshot
wounds (1794). Palmer J, ed. London: Longman; 1934.
Lister J. On the early stages of inflammation. Phil Trans R
Soc 1858;148(pt. 2):645-702.
Lister J. Croonian Lecture: On the coagulation of the
blood. Proc Roy Soc. 1863;12:580-611.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
Addison W. On the colourless corpuscles and on the
molecules and cytoblasts in the blood. Lond Med Gaz NS.
1842;30:144-148.
Addison W. Experimental and Practical Researches on
Inflammation. London: Churchill; 1846.
Waller AV. Microscopic observations on the perforation
of the capillaries by the corpuscles of the blood and on the
origin of mucus and pus globules. Phil Mag. 1846;29:271397.
Holman E. Arteriovenous aneurism: Clinical evidence
correlating size of fistula with changes in the heart and
proximal vessels. Ann Surg. 1924;80(6):801-816.
Browse NL, Burnand KG, Lea Thomas M. Aetiology of
DVT. In: Browse NL, ed. Diseases of the Veins. London:
Edward Arnold; 1988:447-460.
Cruveilhier J. Inflammation of the pulmonary artery and
lobular pneumonia followed by phlebitis. In: Long BR, ed.
Selected Readings in Pathology. London: Baillière Tindal
and Cox; 1929.
Lidell JA. On thrombosis and embolism. Am J Med Sci.
1872;63(128):325-365.
The reports of the two referees (Sir) James Paget and John
Goodsir in the possession of the Royal Society: Report on
Mr Lister’s paper on the early stages of inflammation. Br J
Surg. 1932;20(79):462-466.
Newsom SW. Pioneers in infection control - Joseph Lister.
J Hospl Infect. 2003;5(4):246-253.
Harvey W. Exercitatio anatomica de motu cordis et sanguinis in animalibus. Frankfort, 1628. In: Willis R, Eliot
CW, eds. On the Motion of the Heart and Blood in Animals. Harvard Classics Vol. 38. New York: Collier and
Sons; 1910.
Welch WH. The structure of white thrombi. Trans Path
Soc Philadelphia. 1887;13:281-300.
Aschoff L. Lectures in General Pathology. New York:
Hueber; 1924:254.
Bizzozero G. Űber einen neuen Formbestandtheil des
Blutes und desen Rolle bei der Thrombose und der Blutgerinnung. Virchows Arch Pathol Anat. 1882;90(2):261332.
Hadfield G. Thrombosis. Ann R Coll Surg Engl.
1950;6(4):219-234.
McLachlin J, Paterson JC. Some basic observations on
venous thrombosis and pulmonary embolism. Surg Gynecol Obstet. 1951;93(1):1-8.
Paterson JC, McLachlin J. Precipitating factors in venous
thrombosis. Surg Gynecol Obstet. 1954;98(1):96-102.
Paterson JC. The Pathology of Venous Thrombi. In:
Sherry S, ed. Thrombosis. Washington: National Academy
of Sciences; 1969:321-331.
Sevitt S. Organization of valve pocket thrombi and the
anomalies of double thrombi and valve cusp involvement.
Br J Surg. 1974;61(8):641-649.
Sevitt S. The structure and growth of valve pocket thrombi
in femoral veins. J Clin Pathol. 1974;27(7):517-528.