10 1038@nrdp20156

PRIMER
Venous thrombosis
Alisa S. Wolberg1,2, Frits R. Rosendaal3,4, Jeffrey I. Weitz5, Iqbal H. Jaffer5,
Giancarlo Agnelli6, Trevor Baglin7 and Nigel Mackman1,2,4,8
Abstract | Venous thromboembolism (VTE) encompasses deep-vein thrombosis (DVT) and pulmonary
embolism. VTE is the leading cause of lost disability-adjusted life years and the third leading cause of
cardiovascular death in the world. DVT leads to post-thrombotic syndrome, whereas pulmonary
embolism can cause chronic pulmonary hypertension, both of which reduce quality of life. Genetic and
acquired risk factors for thrombosis include non‑O blood groups, factor V Leiden mutation, oral
contraceptive use, hormone replacement therapy, advanced age, surgery, hospitalization and long-haul
travel. A combination of blood stasis, plasma hypercoagulability and endothelial dysfunction is thought
to trigger thrombosis, which starts most often in the valve pockets of large veins. Animal studies have
revealed pathogenic roles for leukocytes, platelets, tissue factor-positive microvesicles, neutrophil
extracellular traps and factors XI and XII. Diagnosis of VTE requires testing and exclusion of other
pathologies, and typically involves laboratory measures (such as D‑dimer) and diagnostic imaging. VTE is
treated with anticoagulants and occasionally with thrombolytics to prevent thrombus extension and to
reduce thrombus size. Anticoagulants are also used to reduce recurrence. New therapies with improved
safety profiles are needed to prevent and treat venous thrombosis. For an illustrated summary of this
Primer, visit: http://go.nature.com/8ZyCuY
Venous thromboembolism (VTE) manifests primarily — the so‑called post-thrombotic syndrome (PTS) —
as deep-vein thrombosis (DVT), which typically occurs which reduce quality of life7. VTE is treated with anti-
in the legs, and to a lesser extent as pulmonary embo- coagulants and occasionally with thrombolytics, both of
lism. DVT in arm veins accounts for a low percentage of which are associated with a risk of bleeding. The recent
cases and primarily occurs in patients with cancer and introduction of direct oral anticoagulants (DOACs; also
those with indwelling central vein catheters or ports1. known as new OACs (NOACs)), which directly inhibit
The terms ‘thrombosis’ and ‘embolism’ were coined by either factor Xa or thrombin, has provided clinicians
the German physician Rudolf Virchow, who was the first with new treatment options that have a lower bleeding
to demonstrate a mechanistic link between DVT and risk than conventional therapy.
pulmonary embolism2. However, the so‑called Virchow’s Clues to the operant pathophysiological mechanisms
Triad was attributed to Virchow only 100 years after the are revealed in the morphology of venous thrombi; in
publication of his work in 1856 (REF. 3). Virchow’s Triad contrast to platelet-rich, arterial ‘white’ thrombi, venous
describes three groups of thrombogenic factors: hyper- thrombi contain fewer platelets and are called ‘red clots’
coagulability, changes in blood flow (stasis and turbu- because they have a higher content of red blood cells and
Correspondence to lence) and endothelial dysfunction (FIG. 1). In 2014, the fibrin. Epidemiological studies indicate that both genetic
A.S.W. and N.M. International Society of Thrombosis and Haemostasis (such as the factor V Leiden mutation) and acquired
e-mails: initiated “World Thrombosis Day” (REF. 4) on 13 October (such as surgery) risk factors contribute to VTE. In addi-
[email protected];
— the date of Virchow’s birth. tion, animal studies have revealed roles for leukocytes,
[email protected]
Department of Pathology and
VTE is a multicausal, episodic disorder and a major microvesicles and neutrophil extracellular traps (NETs)
Laboratory Medicine, cause of morbidity and mortality worldwide. It has an in VTE. In this Primer article, we provide an overview of
University of North Carolina annual incidence of 1–2 events per 1,000 person-years the current knowledge of the development and clinical
at Chapel Hill, 819 Brinkhous- and is more common in men than women; the incidence consequences of venous thrombosis and highlight new
Bullitt Building, Chapel Hill,
North Carolina
increases to 1 event per 100 person-years in individu- and emerging treatment options for this condition.
27599–7525, USA. als older than 55 years of age5. Pulmonary embolism is
fatal in ~10% of acute cases, with higher rates in patients Epidemiology
Article number: 15006
doi:10.1038/nrdp.2015.6
with cancer. VTE survivors have a poor prognosis, with Among the cardiovascular causes of death, VTE is
Published online an average one‑year mortality rate of 10%5,6. Up to 50% ranked the third most common in the world, after coro-
7 May 2015 of patients with DVT develop pain, oedema and ulcers nary heart disease and ischaemic stroke, with estimates
NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 1
© 2015 Macmillan Publishers Limited. All rights reserved

PRIMER
Author addresses incidence of VTE than people of Asian descent, whereas

white individuals have an intermediate-level risk13. The
1
Department of Pathology and Laboratory Medicine, University of North Carolina at reasons for these differences are unclear. Acquired risk
Chapel Hill, 819 Brinkhous-Bullitt Building, Chapel Hill, North Carolina 27599–7525, USA. factors include injury of the endothelium by central
2
McAllister Heart Institute, University of North Carolina at Chapel Hill, USA. venous catheters and chemotherapies12. Superficial vein
3
Department of Clinical Epidemiology and Department of Thrombosis and Hemostasis,
thrombosis is associated with an increased risk of VTE
Leiden University Medical Center, The Netherlands.
4
K.G. Jensen Thrombosis Research and Expertise Center, University of Tromsø, Norway. for more than 5 years after the thrombotic event 14,15.
5
Department of Medicine and Department of Biochemistry and Biomedical Sciences, Among the genetic variants, there is a clear inverse
McMaster University, and Thrombosis and Atherosclerosis Research Institute, Hamilton, relationship between prevalence and strength 16.
Ontario, Canada. Heterozygous deficiencies of the natural anticoagulants
6
Division of Internal and Cardiovascular Medicine, Stroke Unit, University of Perugia, Italy. antithrombin, protein C and protein S are rare (<0.02–
7
Department of Haematology, Addenbrooke’s Hospital, Cambridge University Hospitals 0.2%) but increase VTE risk by up to tenfold. Common
NHS Trust, Cambridge, UK. variants with several percent prevalence increase VTE risk
8
Department of Medicine, University of North Carolina at Chapel Hill, USA. by twofold to fivefold. These common variants include
non‑O blood groups, factor V Leiden (rs6025) mutation
of more than 500,000 deaths in the European Union and prothrombin 20210A (rs1799963) mutation. Many
(population 500 million)8 and more than 500,000 deaths common (>20% prevalence) genetic variants (for exam-
in the United States (population 300 million) every year 9. ple, polymorphisms in factors IX17 or XI18–20) have a weak
VTE is reported to be the leading cause of disability- effect (<1.5‑fold increased risk). Some risk factors are not
adjusted life years lost in hospitalized patients9. Despite easily classified, such as hyperthyroidism, use of psycho-
this disease burden, information on VTE incidence in tropic drugs, obesity, alcohol consumption, smoking and
different countries is scarce, and VTE is not listed as part high levels of procoagulant factors. The impact of these
of the “Global burden of diseases, injuries and risk fac- risk factors on overall disease occurrence is a function of
tors study” initiated by the World Health Organization their strength and prevalence; for example, deficiencies in
(WHO) and the World Bank9. protein C, protein S and antithrombin are the strongest
Both genetic and/or acquired risk factors can cause genetic risk factors for thrombosis but, because of their
VTE (BOX 1). For instance, deficiencies in anticoagulants rarity, are responsible for <5% of all thrombotic events
and/or increases in procoagulants can produce a hyper- in the population. BOX 1 lists the risk factors for VTE and
coagulable state, immobility causes blood stasis, and shows that the strongest risk factors are the commonly
surgery can damage veins and/or release procoagulant acquired factors for which thromboprophylaxis with
material into the blood. Importantly, combinations of anticoagulants is usually prescribed.
risk factors seem to be required for VTE initiation10. Patients with VTE have an increased risk of death
VTE risk increases with age, with a sharp rise in inci- for up to 8 years after the initial event. Malignancy is
dence in individuals 55 years of age or older 11. Stasis- the primary cause of death in these patients, followed
related risk factors include surgery, bed rest, pregnancy, by cardiovascular and respiratory diseases5,6. Fewer
plaster cast immobilization and long-haul travel. than half of deaths in the first 30 days after pulmonary
Hypercoagulable states can be caused by malignan- embolism are a direct consequence of the embolism5. In
cies, administration of female hormones used for oral patients without malignancy, there is a twofold increase
contraception, and hormone replacement therapy 12. in mortality compared with controls without VTE, but
Genetic abnormalities that might cause gain of func- this increase is confined to patients with unprovoked
tion in procoagulants (for example, factor V Leiden VTE and those with comorbidities6,21.
or prothrombin G20210A gene variants, or increased
levels of fibrinogen or other clotting factors) or loss of Mechanisms/pathophysiology
function in anticoagulants (for example, deficiencies of Haemostatic clots and thrombotic clots contain many of
antithrombin, protein C or protein S) can also produce the same components but are formed under different cir-
hypercoagulable states12. Ethnicity also influences VTE cumstances. Haemostatic clots form rapidly after vessel
risk; people of African descent have a fivefold higher transection or injury to prevent excessive blood loss. By
contrast, venous thrombi form within the blood vessel
on a largely intact endothelium and develop over several
Box 1 | Risk factors for venous thrombosis by strength days to weeks. A combination of risk factors seem to act
together to trigger VTE. As mentioned above, these fac-
Strong risk factors increase the thrombosis risk by tenfold to 50‑fold, moderate risk tors can be broadly divided into components of Virchow’s
factors by twofold to tenfold, and weak risk factors by less than twofold.
Triad, which include increased blood coagulation, altered
• Strong: major surgery, trauma, plaster casts and cancer
blood flow (stasis) and endothelial dysfunction2,22 (FIG. 1).
• Moderate (genetic): anticoagulant deficiencies, factor V Leiden mutation, These different triggers are discussed below.
non‑O blood groups, prothrombin G20210A mutation
• Moderate (acquired): lifestyle factors, including obesity, smoking and alcohol Animal models of thrombosis
consumption (protective effect for the latter), long-haul travel, oral contraceptives,
Most animal models of venous thrombosis involve
hormone replacement therapy, and medical conditions such as heart failure,
changes in blood flow such as stasis or turbulence.
inflammatory bowel disease and major non-malignant haematological conditions
Arguably, the most physiologically relevant animal
• Weak: all other known genetic variants (prevalence >20%)
model is the baboon model developed by the Wakefield
2 | 2015 | VOLUME 1 www.nature.com/nrdp

PRIMER
a b
Vein Thrombomodulin Tissue factor
Thrombus
KLF2 PAI1
Red blood
EPCR Endothelial cell Endothelial
VCAM1
expression of dysfunction
Fibrin
antithrombotic
factors P-selectin
TFPI
E-selectin
Heparan sulfate
Blood
Valve TPA hypercoagulability
Nitric oxide Leukocyte
activation
Prostacyclin
Tissue factor-positive
microvesicles
CD39 Hypoxia
in valve
Anticoagulants pocket
in blood Neutrophil
Laminar Proteins NETs
flow C and S
Antithrombin
Abnormal Platelet
flow/stasis activation
Endothelium
Figure 1 | Mechanisms that protect from or promote thrombogenesis. a | In healthy veins, Nature Reviews
the | Disease
endothelium Primers
senses
blood flow and provides an antithrombotic surface to prevent clot formation, for example, through Krüppel-like factor 2
(KLF2)-mediated production of thrombomodulin. Other endothelial antithrombotic factors are tissue factor pathway
inhibitor (TFPI), heparan sulfate, tissue plasminogen activator (TPA), nitric oxide, prostacyclin and CD39. In valve pockets,
which have non-laminar blood flow and are prone to hypoxia, increased expression of thrombomodulin and the
endothelial protein C receptor (EPCR) limit thrombin generation. Anticoagulants in the blood, including protein C,
protein S and antithrombin, limit thrombin generation and inhibit locally generated thrombin. b | According to Virchow’s
Triad, a combination of hypercoagulability, haemodynamic changes and endothelial dysfunction underlie thrombus
formation. Stasis induces endothelial dysfunction and expression of cell adhesion molecules such as vascular cell
adhesion protein 1 (VCAM1), P‑selectin and E‑selectin; recruitment of leukocytes to the vessel wall; and initiation of
procoagulant activity, for example, through endothelial downregulation of antithrombotic factors and upregulation
of antifibrinolytic and procoagulant factors such as plasminogen activator inhibitor 1 (PAI1) and tissue factor or through
the release of neutrophil extracellular traps (NETs) and microvesicles. Ultimately, fibrin and red blood cells accumulate,
occluding the vein and causing symptomatic deep-vein thrombosis.
group23. This is partly because the baboon is an upright to the stenosis site and endothelial activation seem to be
animal and its coagulation system is similar to that of the major triggers of thrombosis in this model, although
humans. In this model, stasis is induced for 6 hours in a low level of vessel injury cannot be excluded28. Ferric
the iliac vein using two balloon catheters that enter the chloride and electrolytic injury have also been used to
vessel from opposite directions while the contralateral initiate thrombosis in the inferior vena cava and femoral
iliac vein serves as a control23. Notably, baboons have vein29–35. These techniques induce rapid formation of a
a valve in the affected vessel segment, and this model thrombus but do not involve blood stasis or turbulence
might thus recapitulate some of the pathological events in the initiation phase. All of these animal models have
occurring in humans during thrombus initiation and been used to elucidate the roles of different factors in
allows assessment of valve function24. Most laborato- venous thrombosis (discussed below).
ries use rat or, more commonly, mouse models in which
venous thrombi are provoked by ligature to reduce (ste- Altered blood flow
nosis) or abolish (stasis) flow in the inferior vena cava Abnormal blood flow, such as stasis and turbulence,
for 0–48 hours. Although the murine vena cava does not seems to be an important pathogenic factor for venous
contain valves, the thrombi produced in rodent models thrombosis. For example, a study of 76 patients with
upstream of the ligation site resemble venous thrombi hemiparesis caused by stroke showed that there was a
found in humans and non-human primates, with dis- higher incidence of DVT in the paralysed leg than in
tinct areas composed of platelets and red blood cells25,26. the mobile leg (53% versus 7%, respectively)36. Valve
In the stasis model almost all mice develop thrombi, but pockets of large veins are particularly prone to reduced
complete ligation reduces the delivery of blood compo- blood flow and stasis. Indeed, radiographical and post-
nents, such as platelets, leukocytes and microvesicles, mortem studies found that these pockets frequently
to the site of thrombosis. The stenosis model allows for contain thrombi37. One can speculate that this is partly
delivery of blood components, but thrombus formation caused by activation of the endothelium in the valve
is variable, with incidences in the range of 44–100% pocket due to low blood flow and hypoxia. This notion
48 hours after ligature27. Turbulent blood flow upstream is supported by a study that measured oxygen tension

PRIMER
New Traditional anticoagulants induced by hypoxia, inflammation and/or altered blood

drug flow. For example, stasis and turbulent flow are associated
Direct oral anticoagulants
with a reduced expression of KLF2 and activation of the
Drugs in preclinical development
Factor New transcription factor nuclear factor-κB (NF‑κB). NF-κB
XIIa drug
Vitamin K antagonists then induces the expression of various genes that increase
(warfarin) the adhesiveness of the endothelium and that promote
Factor New thrombosis, including cell adhesion molecules (for exam-
XIa drug ple, vascular cell adhesion protein 1 and E‑selectin), the
antifibrinolytic protein plasminogen activator inhibitor
Factor Tissue factor– type 1 (PAI1) and possibly the procoagulant molecule tis-
IXa factor VIIa
sue factor 39–41,43 (FIG. 1). Each of these molecules has been
implicated in thrombogenesis in animal models. In addi-
Factor Xa inhibitors tion, activation of the endothelium results in the trans
(rivaroxaban, apixaban Factor Synthetic pentasaccharide
and edoxaban) Xa (fondaparinux) location of P‑selectin from Weibel–Palade bodies inside
the cell to the cell surface, where it facilitates the recruit-
Thrombin inhibitors ment of both leukocytes and leukocyte-derived micro
(dabigatran)
Thrombin Unfractionated heparin vesicles (small plasma membrane vesicles shed from many
Low-molecular-weight heparin cell types; see below) to the endothelium via its interaction
with P‑selectin glycoprotein ligand 1 (PSGL1)26. Indeed,
Fibrinogen Fibrin inhibiting either P‑selectin or PSGL1 reduces thrombosis
Factor XIIIa in both baboon and mouse models44–46. One study of a
rat model of venous thrombosis suggested that activated
Platelets Red blood cells
endothelium expresses tissue factor 43. However, whereas
some studies have shown that vessel wall tissue factor is
Venous thrombus
the major driver of thrombus formation in the murine sta-
sis model, other studies have shown that leukocyte tissue
Figure 2 | The coagulation cascade and existing and emerging anticoagulant
Nature Reviews | Disease Primers factor contributes to thrombosis in the stenosis model26,47.
drugs. The coagulation cascade is a multistep pathway that leads to the
thrombin-mediated generation of fibrin from fibrinogen, which together with platelets Plasma hypercoagulability
and red blood cells constitute the thrombus. The cascade can be activated by tissue
The coagulation cascade comprises a series of serine
factor–factor VIIa (in the extrinsic pathway) or factor XIIa (in the intrinsic pathway).
Traditional methods of anticoagulation involve vitamin K antagonists and heparin (grey proteases and cofactors that culminate in thrombin gen-
boxes). Low-molecular-weight heparin can also inhibit thrombin generation. Newer eration and conversion of fibrinogen to fibrin48 (FIG. 2).
therapeutics currently in clinical use are enzyme-specific and target either factor Xa or The primary physiological activator of the coagulation
thrombin (pink boxes). Potential targets of drugs that are currently in preclinical cascade is the tissue factor–factor VIIa complex, which
development (green boxes) include factors XIIa, XIa and IXa. triggers the extrinsic pathway. In addition, the cascade
can be activated via factor XIIa (the intrinsic or contact
pathway). Aberrant or excessive activation of coagula-
in the valve pocket of large veins in dogs; a declining tion within vessels leads to thrombin generation, intra-
oxygen gradient was evident from the top to the bottom vascular fibrin deposition and formation of thrombi,
of the pocket after 2 hours of stasis38 (illustrated by blue which then impair blood flow.
areas in the valve pocket in FIG. 1). Because of the ease of collecting and analysing
blood in humans, the best characterized pathophysio
Endothelial dysfunction logical mechanism contributing to VTE in humans
The endothelium of healthy vasculature provides an is plasma hypercoagulability. In vitro studies using
antithrombotic and profibrinolytic surface that pre- spiked or deficient whole blood, plasma or reconsti-
vents clot formation. Blood flow and the laminar tuted reactions that model human plasma show that
shear stress contribute to the maintenance of the vaso plasma hypercoagulability leads to increased throm-
protective phenotype of the endothelium by stimulat- bin generation49–51. These findings are consistent with
ing expression of the transcription factor Krüppel-like studies showing that increased plasma thrombin gen-
factor 2 (KLF2). This master regulator positively regu- eration is a risk factor for both primary and recurrent
lates expression of the anticoagulant protein thrombo venous thrombosis52–61. Mouse models have been used
modulin39–41 (FIG. 1). Valve pockets, which are exposed to directly test the contribution of blood hypercoagula-
to turbulent flow and low flow rather than laminar flow, bility to thrombosis in vivo. For example, mice infused
are programmed to resist thrombosis; the endothelium with prothrombin to mimic the hyperprothrombin
lining expresses higher levels of thrombomodulin and aemia associated with the G20210A mutation have
the endothelial protein C receptor, and lower levels of larger thrombi in the vena cava stasis model and accel-
the procoagulant von Willebrand factor (vWF), than erated fibrin deposition in the electrolytic injury model
adjacent non-valvular endothelium42. of the saphenous vein compared with control mice62.
Under pathological conditions, the endothelium is Elevated levels of factor VIII and fibrinogen also
converted from an anticoagulant to a procoagulant sur- shorten the time to vessel occlusion following ferric
face owing to injury and/or changes in protein expression chloride treatment of the saphenous vein34,35. Moreover,

PRIMER
factor XII-deficient mice have smaller thrombi in the contributes to thrombogenesis in animal models. In a
vena cava stenosis model, suggesting that the intrinsic rat vena cava stasis model, leukocytes express tissue
pathway contributes to venous thrombosis26. Finally, factor 43, and a deficiency of tissue factor in myeloid
the association of VTE risk in humans with impaired cells is associated with reduced thrombosis in the
fibrinolysis of plasma clots formed in vitro63–65 suggests mouse vena cava stenosis model 26. Taken together,
that fibrinolysis limits thrombogenesis. these results support the notion that leukocytes
contribute to venous thrombogenesis.
Platelets, leukocytes and microvesicles Interestingly, the cholesterol-lowering drug rosuva
Platelets. Platelets have traditionally not been thought statin reduces thrombosis in the vena cava ligation
to have a major role in venous thrombosis. However, model and inhibits neutrophil recruitment to the vein
accumulating evidence indicates that platelets contrib- walls and sites of thrombosis86. Statins have also been
ute to thrombosis in mice and humans, and might regu- shown to inhibit monocyte tissue factor expression
late effector functions of innate immune cells recruited in vitro and in vivo87. The JUPITER clinical trial found
to the thrombus66. Thrombocytosis is a risk factor for that rosuvastatin reduces the incidence of VTE by 43%
VTE in patients with cancer or other diseases67–71. In in individuals with elevated levels of C‑reactive protein88.
mice, platelet depletion reduces thrombosis incidence26, However, a meta-analysis of 22 statin trials concluded
and vWF-deficient mice have small thrombi, probably that “the findings from this meta-analysis do not sup-
owing to reduced platelet and leukocyte recruitment 25. port the previous suggestion of a large protective effect
Importantly, a meta-analysis of two clinical trials of statins on venous thromboembolic events” (REF. 89).
(WARFASA and ASPIRE) indicated that low-dose aspi-
rin reduces recurrent VTE after patients completed a Microvesicles. Levels of circulating microvesicles are
6–18 month period of anticoagulation72, and a separate increased in several pathological conditions associ-
meta-analysis in 2014 estimated a 42% risk reduction of ated with venous thrombosis90. For example, leukocyte
recurrent VTE with aspirin therapy 73. Taken together, and platelet microvesicles are increased in the murine
these data suggest that platelets contribute to venous vena cava ligation model, and injection of these micro
thrombogenesis. However, a recent review of the lit- vesicles increases thrombus weight91. However, there is
erature concluded that “while platelet inhibitory agents no convincing clinical evidence to indicate that elevated
might be considered in the prevention of VTE, there levels of total microvesicles increases the risk of VTE92.
is no obvious indication for platelet inhibition in the By contrast, elevated levels of tissue factor-positive
treatment of VTE” (REF. 74). microvesicles might trigger venous thrombosis in some
patients. For example, in patients with pancreatic cancer,
Leukocytes. Activation of the clotting system by leuko- elevated levels of tumour-derived tissue factor-positive
cytes in response to pathogens is part of host defences. microvesicles are associated with an increased risk of
The clotting system serves to immobilize patho- VTE93,94. In a mouse model, tumour-derived micro
gens and recruit immune cells to destroy invading vesicles accumulate at sites of vascular injury, and mice
pathogens. By contrast, in the absence of pathogens, with tumours have larger clots in the mesenteric vascula-
venous thrombosis has been proposed to result from ture than control mice95. Injection of exogenous tumour-
activation of coagulation by leukocytes and crosstalk derived tissue factor-positive microvesicles also enhances
between coagulation and innate immune cells — thrombosis in the mouse stenosis model96. These find-
so‑called immunot hrombosis66. Recent clinical and ings suggest that tissue factor-positive microvesicles may
animal model studies support this premise (FIG. 1). For contribute to venous thrombogenesis10.
example, there is compelling evidence to indicate that
leukocytes contribute to venous thrombosis, particu- Red blood cells
larly in animal models26,75. In patients with cancer, pre- Venous thrombi are called red clots because of their
cancer and pre-chemotherapy, leukoc ytosis is a risk prominent red blood cell content. Traditionally, red
factor for VTE76,77. It has been proposed that neutro blood cells have been thought to be only passively
phils contribute to venous thrombosis by releasing trapped components of the thrombus. However, grow-
NETs, which are extracellular networks consisting of ing evidence suggests that they actively contribute to
DNA, histones and antimicrobial proteins. NETs have coagulation. For example, red blood cell transfusion
been shown to capture platelets and red blood cells alleviates bleeding in anaemic patients97 and is associ-
in vitro 78. Notably, DNA was detected in thrombi in ated with thrombosis in patients with cancer and in
the baboon model78 and in venous thrombi harvested hospitalized patients98–100. The mechanisms by which red
from humans 79,80. DNA and RNA can also activate blood cells contribute to venous thrombosis are poorly
factor XII, which might partly explain the attenuated understood. Red blood cells aggregate in the low shear
thrombosis observed in factor XII-deficient mice26,81. of the venous circulation and increase blood viscosity,
Importantly, neutropenic mice and mice with an which might promote blood stasis101,102. In addition, red
inability to form NETs produce fewer and/or smaller blood cells and microvesicles derived from them can
venous thrombi 26,82. Neutrophils can also enhance have exposed phosphatidylserine — a procoagulant
thrombosis by releasing proteases, such as elastase, phospholipid — and in vitro studies show that both the
that cleave and inactivate the anticoagulant tissue fac- cells and the microvesicles can support thrombin gen-
tor pathway inhibitor 83–85. Leukocyte tissue factor also eration and activate the contact pathway 103–107. Red blood

PRIMER
cells can also activate platelets via release of ADP, stabi- Finally, PAI1-deficient mice have smaller thrombi but
lize clots by decreasing clot permeability 108,109 and delay increased fibrosis, suggesting an interesting relationship
fibrinolysis110. The relative impact of these functions on between thrombosis and fibrosis119,120. These findings
thrombogenesis in vivo has not yet been determined. are particularly relevant in the context of the long-term
Aleman et al.111 reported that red blood cell retention sequelae associated with DVT (PTS) and pulmonary
in venous thrombi requires transglutaminase activity embolism (chronic thromboembolic pulmonary hyper
of factor XIII. Notably, they found that factor XIII- tension). These pathologies are associated with per-
deficient mice have small venous thrombi because the sistent venous obstruction, increased pressure within
thrombi contain significantly fewer red blood cells than vessels and inflammation, which result in valvular
those from wild-type mice. This effect is recapitulated damage and increased vessel wall permeability and stiff-
in human whole-blood clots formed in the presence of ness (scarring). As rapid lysis and resolution of venous
factor XIII inhibitors111. These findings suggest that fac- thrombi have been proposed to reduce PTS and chronic
tor XIIIa contributes to red blood cell retention in clots thromboembolic pulmonary hypertension, efforts to
and, therefore, venous thrombus size. better understand the relationship between thrombosis
and fibrosis are essential.
Resolution of venous thrombi
In addition to their prothrombotic activity, leukocytes Diagnosis, screening and prevention
also seem to be crucial for venous thrombus resolution. Prevention and screening
In a rat model of stasis-induced thrombosis, neutro Hospitalization is a major risk factor for VTE, and at
penia is associated with larger thrombi and increased least half of the outpatients with newly diagnosed VTE
fibrosis, as well as decreased levels of urokinase plas- have a history of hospitalization in the past 3 months.
minogen activator and matrix metalloproteinase 9 Many of these patients received no or inadequate
(MMP9)112. These observations suggest that neutrophil- thromboprophylaxis during their hospital stay. In
derived urokinase plasminogen activator and MMPs fact, because of inadequate thromboprophylaxis, pul-
are involved in resolving thrombi113–115. Monocytes and monary embolism is considered the principal cause of
macrophages are also recruited to venous thrombi, preventable death in hospitalized patients121. To reduce
with high numbers of macrophages present 6–8 days this burden, the risk of VTE should be assessed in all
after thrombus initiation75,116. Monocyte recruitment is patients on admission to hospital, and appropriate
mediated by cysteine–cysteine chemokines and mono- thromboprophylaxis should be prescribed.
cyte chemotactic protein 1 (REFS 117,118). Among other There are several options for thromboprophylaxis122.
activities, monocytes and macrophages contribute to Parenteral anticoagulants — such as heparin, low-
MMP expression in thrombi117,118. In both stasis and molecular-weight heparin (LMWH) or fondaparinux
stenosis murine thrombosis models, deletion of the (a synthetic pentasaccharide, which contains the
gene encoding cysteine–cysteine chemokine recep- antithrombin binding site of heparin) (FIG. 2) — are
tor 2 is associated with reduced monocyte recruitment administered subcutaneously in prophylactic doses.
to venous thrombi, reduced MMP2 and MMP9 activ- Oral anticoagulants, such as vitamin K antagonists
ity in thrombi and impaired thrombus resolution117,118. (for example, warfarin) or DOACs (FIG. 2), are used
for extended prophylaxis after major orthopaedic sur-
gery 123. These approaches are effective at reducing VTE
incidence but are associated with a risk of bleeding.
* * Mechanical methods, such as pneumatic compression
devices with or without elastic stockings, are used in
patients whose risk for bleeding precludes the admin-
istration of anticoagulants, even in prophylactic doses
(for example, patients with intracerebral haemorrhage
or those with very low platelet counts).
Thrombophilia testing to screen for increased VTE
risks includes evaluation for hereditary disorders, such
as the factor V Leiden or prothrombin gene mutations;
deficiency of antithrombin, protein C or protein S; and
acquired disorders, such as antiphospholipid syndrome.
Although the factor V Leiden and prothrombin gene var-
iants place women at higher risk of VTE when they are
prescribed oral contraceptives or hormone replacement
therapy, the absolute risk remains low and screening
Nature of
Figure 3 | Clinical presentation Reviews | Disease Primers
deep-vein
asymptomatic women for these defects is not cost-
thrombosis. These images show patients with extensive effective124. Similarly, thrombophilia screening should
deep-vein thrombosis that included the proximal veins of not be performed in patients with provoked VTE because
the legs. Note the pronounced swelling in the calves and these patients are at low risk of recurrent thrombosis
thighs (indicated by asterisks). Images courtesy of S. Moll, when anticoagulation therapy is stopped125. In those
University of North Carolina, Chapel Hill, USA. with unprovoked VTE, thrombophilia testing should be

PRIMER
Suspected deep-vein thrombosis or pulmonary embolism is often asymptomatic127. Symptomatic patients fre-
quently complain of chest pain, particularly when tak-
ing a deep breath, and shortness of breath, particularly
Pretest likelihood with exertion, and they may present with rapid breath-
ing and a rapid heart rate. A bloody cough can occur if
Low Moderate or high the embolism completely obstructs a segmental artery
in the lung, causing distal pulmonary infarction.
D-dimer
Diagnostic pathway. First, the pretest likelihood of
DVT or pulmonary embolism should be determined
Normal Abnormal using validated clinical prediction rules, which take into
account components of the clinical assessment, the pres-
ence or absence of risk factors for VTE and whether an
Compression ultrasonography or computed
tomography pulmonary angiography
alternative diagnosis explains the symptoms and signs128.
Patients with a moderate or high pretest likelihood pro-
ceed directly to testing for DVT or pulmonary embolism,
Positive Negative whereas those with a low pretest likelihood should be fur-
ther screened by D‑dimer assay (see below). A negative
No thrombosis Thrombosis No thrombosis D‑dimer result precludes additional testing (FIG. 4).
or embolism or embolism or embolism D‑dimer is a fibrin degradation product generated by
plasmin-mediated proteolysis and is an indirect marker
Figure 4 | Management algorithm for suspected venous Nature Reviews | DiseaseBefore
thromboembolism. Primers
of activation of the coagulation system. D‑dimer can be
diagnostic tests are conducted in patients with suspected deep-vein thrombosis or measured quantitatively in plasma, and point‑of‑care
pulmonary embolism, the pretest likelihood is assessed on the basis of the patient’s risk
whole-blood tests also are available for rapid assess-
profile, clinical presentation and potential alternative diagnoses. Patients with a
moderate or high pretest likelihood directly proceed to diagnostic testing, which ment. Contemporary D‑dimer tests have up to 98%
includes compression ultrasonography or CT pulmonary angiography, whereas patients sensitivity for the diagnosis of VTE, but their specifici-
with a low pretest likelihood are further screened for the presence of D‑dimer — a fibrin ties are relatively low 129 because D‑dimer levels can be
degradation product and thus a marker of thrombosis. However, D‑dimer can also be elevated with advanced age, infection, chronic inflam-
elevated as a result of other conditions, and diagnostic testing is therefore still needed in mation, cancer and other conditions. Consequently, a
patients with a positive D‑dimer test. positive D‑dimer test does not establish a VTE diagno-
sis. Furthermore, because inpatients are more likely to
have a positive D‑dimer test than outpatients, its utility is
restricted to the few patients in whom the results of such limited in hospitalized patients. The major benefit of the
testing would influence decisions regarding the dura- D‑dimer assay is its high negative predictive value. This
tion of treatment. This could include patients who wish means that a normal D‑dimer test makes a VTE diagno-
to stop anticoagulant therapy but who might change sis unlikely in patients with a low pretest likelihood128,
their mind if antithrombin, protein C or protein S defi- thereby eliminating the need for further diagnostic test-
ciencies were detected. Although patients with VTE ing (FIG. 4). D‑dimer levels increase with age, but the use
might have occult cancer, apart from age-appropriate of higher, age-specific cut-offs for normal D‑dimer levels
testing, there is no justification for extensive screening can reduce the need for further testing in the elderly 130,131.
for malignancy after a thrombotic event. As the clinical features of VTE are nonspecific, objec-
tive testing is required for patients with a low pretest likeli-
Diagnosis hood for VTE and an elevated D‑dimer level and for those
Symptoms and signs. The calf is the most common site with a moderate or high pretest likelihood. Compression
of DVT, although ~20% of calf thrombi extend into the ultrasonography is used for the diagnosis of DVT, whereas
more proximal veins of the leg — those behind the knee CT pulmonary angiography (CTPA) or ventilation–
and in the thigh (popliteal and femoral veins, respec- perfusion (V/Q) lung scanning is used for the diagnosis
tively)125. Thrombi in the more proximal veins are more of pulmonary embolism128,132.
likely to break off and cause pulmonary embolism than Compression ultrasonography is a simple non
those in calf veins. Embolism can also arise from upper- invasive test that has replaced contrast venography as
extremity DVT involving the axillary or subclavian the preferred test for diagnosis of DVT. A positive test
veins, but this is a less common source126. shows non-compressible veins (FIG. 5a). The sensitiv-
The symptoms and signs of DVT include swelling in ity and specificity of compression ultrasonography for
the affected limb accompanied by swelling and tenderness proximal DVT are greater than 95% in patients with
along the course of the deep veins (FIG. 3). There can be symptoms suggestive of DVT128. For detection of an
dilatation of the superficial veins of the leg, the skin may isolated calf DVT, the sensitivity and specificity of com-
be red or dusky blue, and the affected limb may be warm pression ultrasonography are lower (~70–75%)133. The
to touch. In some cases, DVT can be asymptomatic, and most frequently used test for the diagnosis of pulmo-
pulmonary embolism might be the first manifestation. nary embolism is CTPA with a multidetector scanner
At least 40% of patients presenting with proximal (FIG. 5b). CTPA is widely available and results are obtained
DVT will have pulmonary embolism; however, this rapidly. Furthermore, it not only identifies emboli but

PRIMER
can also detect other pathologies to provide an alterna- In the perfusion phase, technetium-labelled albumin
tive diagnosis. CTPA can be used as a stand-alone test macroaggregates are injected intravenously, and blood
because of its high sensitivity and specificity (greater flow to the lungs is assessed once the aggregates lodge
than 85% and 90%, respectively)134. Thus, evidence of in the pulmonary microcirculation. Patients are scanned
a thrombus in segmental or larger pulmonary arteries with a gamma counter; areas of lung affected by pulmo-
on CTPA establishes a diagnosis of pulmonary embo- nary embolism exhibit normal air delivery but do not
lism, whereas a negative CTPA excludes it 128. The sig- light up on the perfusion scan because blood flow to that
nificance of subsegmental thrombi is uncertain; in this region is blocked (FIG. 5c). A positive diagnosis is estab-
case, if compression ultrasonography of both legs is lished if there are segmental or larger perfusion defects
negative, a DVT can be excluded and treatment might that ventilate normally, whereas embolism is excluded
not be needed. CTPA can be combined with CT veno by a normal V/Q pattern128,129. A limitation of this tech-
graphy so that the diagnosis of pulmonary embolism and nique is that ~65% of scans exhibit matched or smaller
DVT can be established with a single test and with only mismatched defects, which are sometimes due to other
one injection of contrast dye. Although combined test- pathologies, and fall within the non-high-probability
ing modestly improves sensitivity, it increases radiation category of VTE. These patients require additional test-
exposure and has little effect on the overall detection ing because emboli can be found in up to 40% of this
rate128. Consequently, compression ultrasonography is cohort 128,129. Additional investigation might include
the preferred technique for DVT diagnosis128. bilateral compression ultrasonography to exclude DVT.
V/Q scanning is preferred over CTPA for patients Patients with extensive DVT or pulmonary embo-
with renal impairment and for those with a history of lism might require advanced medical therapies, such
allergy to angiographic contrast dyes129. This two-part as pharmacomechanical thrombolysis to degrade and
test consists of ventilation and perfusion phases. In the remove thrombi in the proximal deep veins of the leg, or
ventilation phase, patients inhale aerosolized radioactive systemic or catheter-directed thrombolysis for massive
xenon or technetium to assess air delivery to the lungs. pulmonary embolism. Extensive DVT involving the iliac
or more-proximal veins can block outflow of blood from
the leg, resulting in marked swelling and pain. In severe
a b cases, the swelling can compress the arteries supplying
blood to the limb, resulting in blanching of the skin and,
in rare cases, progression to gangrene.
Patients with massive pulmonary embolism present
Artery with hypotension, rapid heart rates and low oxygen satu-
Clot ration on room air. Extensive blockage of blood flow into
the lungs places a strain on the right ventricle of the heart,
which might be visible in an electrocardiogram by a new-
onset right bundle branch block or the classic S1Q3T3 pat-
PA tern (FIG. 5d). The right ventricle can dilate as it attempts
PA to squeeze blood into the blocked pulmonary arteries,
and the pulmonary artery pressure often increases. Right
ventricular dilatation in such patients can be detected
on the CTPA or by transthoracic echocardiography. The
echocardiogram might also reveal diminished pumping of
blood from the right ventricle and increased pulmonary
c d artery pressure. Stretching of the right ventricle can result
in increased plasma levels of brain natriuretic peptide or
I
V its precursor, amino‑terminal pro-brain natriuretic pep-
tide, and right ventricular injury as a result of the strain
II might increase plasma levels of troponin135. Although the
positive predictive value of elevated levels of brain natri
uretic peptide, pro-brain natriuretic peptide and troponin
III
Q is low, patients with evidence of right ventricular dysfunc-
tion and increased levels of these biomarkers have a worse
prognosis than those without these findings135.
Figure 5 | Diagnostic signs of venous thromboembolism. Nature a | In compression

Reviews | Disease Primers Differential diagnosis. The differential diagnosis of DVT
ultrasonography, intraluminal thrombus and non-compressibility of the popliteal includes musculoskeletal disorders (such as muscle strain
vein (arrowheads) can be seen. Arrows indicate the popliteal artery (PA). b | This CT
or tear, a ruptured Baker cyst (a synovial cyst of the knee
pulmonary angiogram shows an intraluminal thrombus in the pulmonary artery.
c | This ventilation–perfusion (V/Q) lung scan shows normal ventilation and multiple joint) or twisting injury), lymphatic disorders (such as
perfusion defects that are consistent with pulmonary embolism. d | This lymphangitis or lymphatic obstruction), cardiovascular
electrocardiogram shows the classic S1Q3T3 pattern with a prominent S‑wave in disorders (such as heart failure, peripheral artery disease,
lead I, and a Q‑wave and T‑wave inversion in lead III, which are consistent with right or arterial or venous aneurysms) and inflammatory or
ventricular strain. infective disorders (such as cellulitis).The differential

PRIMER
a Traditional regimen duration of treatment. Patients with VTE occurring in

the setting of transient and reversible risk factors, such as
Anticoagulant effect
Therapeutic window surgery or hormonal therapy, require shorter treatment

durations than patients with an unprovoked VTE or with
ongoing risk factors, such as advanced-stage cancer.
Traditionally, the initial treatment of VTE includes
both a parenteral anticoagulant (intravenous unfraction-
ated heparin (UFH), subcutaneous LMWH or fonda-
parinux) and a vitamin K antagonist, such as warfarin. As
~5 days 3 months to indefinite warfarin takes several days to be fully effective, parenteral
anticoagulation is continued for at least 5 days until the
b DOAC with initial parenteral anticoagulation warfarin-mediated anticoagulation is in the therapeutic
range, as measured by the international normalized ratio
(INR). At this point, parenteral anticoagulation is stopped,

and warfarin is continued for at least 3 months136 (FIG. 6).
As an alternative to warfarin, patients can be transi-
tioned to a DOAC (FIG. 2). In contrast to warfarin, which
requires frequent INR monitoring and dose adjustments,
the DOACs are given in fixed doses without routine
monitoring. Thus, these drugs are more convenient than
5–10 days 3 months to indefinite warfarin, and they are as effective as warfarin at reducing
the risk of recurrent VTE and are associated with less
c DOAC alone bleeding than warfarin137–139.
Alternatively, patients can be treated directly with
rivaroxaban or apixaban, which are oral factor Xa inhibi-

tors (FIG. 2). These drugs can be given immediately or
after 1–2 days of parenteral anticoagulation (FIG. 6). If this
single-drug approach is chosen, initial therapy is more
intense and higher doses of rivaroxaban and apixaban
are used for the first 21 and 7 days, respectively, and the
doses are then lowered140–142. All-oral therapy with these
7–21 days 3 months to indefinite drugs eliminates the need for parenteral anticoagula-
tion and for warfarin. This facilitates the transition of
UFH or LMWH DOAC (dabigatran or edoxaban) care from emergency department or hospital to home
Warfarin DOAC (rivaroxaban or apixaban) and reduces health care costs, which are estimated at
~€1,800–3,200 in the European Union and ~US$3,000–
Figure 6 | Anticoagulation regimens. a | Traditional anticoagulation
Nature Reviews with oral vitamin K
| Disease Primers 9,500 in the United States for initial treatment of DVT
antagonists involves short-term bridging with parenteral unfractionated heparin (UFH)
or low-molecular-weight heparin (LMWH) until the oral anticoagulation has reached
or pulmonary embolism9,143,144.
therapeutic levels. b | Patients can also be treated with direct oral anticoagulants Patients with cancer and VTE are at high risk of
(DOACs) following initial parenteral anticoagulation. c | Alternatively, patients can be thrombus recurrence. LMWH has been shown to be more
treated directly with DOAC alone. effective than vitamin K antagonists in these patients145.
Consequently, many patients with cancer are started and
maintained on LMWH instead of warfarin until cancer
diagnosis of pulmonary embolism includes other lung treatment is completed and the cancer goes into remis-
disorders (such as pneumonia, exacerbation of chronic sion136. Similarly, women who develop VTE during preg-
obstructive lung disease or asthma), other causes of nancy are also maintained on LMWH because warfarin
inflammation of the pleura (such as systemic lupus ery- and DOACs can cross the placenta and might increase
thematosus), cardiac disorders (such as congestive heart the risk of fetal bleeding; warfarin is also a teratogenic136.
failure, acute coronary syndrome or pericarditis) and
musculoskeletal conditions (such as rib fractures). Anticoagulant drugs
Parenteral anticoagulation. UFH and LMWH bind
Management to and activate antithrombin, which then inactivates
Anticoagulation is the main treatment in patients with thrombin and factor Xa (FIG. 2). Although fondaparinux
acute VTE. Anticoagulation includes three phases: initial also binds to and activates antithrombin, it only cata
treatment during the first 5–10 days, long-term treatment lyses the inhibition of factor Xa because it is too short
during the subsequent 3–6 months and extended treat- to bridge antithrombin to thrombin. The anticoagulant
ment, which goes beyond 6 months. The goal of initial effect of heparin is measured by the activated partial
VTE treatment is to limit thrombus extension, whereas thromboplastin time (aPTT), whereas, when necessary,
the goal of long-term and extended anticoagulation is the anticoagulant effects of LMWH and fondaparinux
to prevent VTE recurrence. Classifying patients accord- are measured using an anti-factor Xa assay. Immediate
ing to triggering factors helps to determine the optimal parenteral anticoagulation can be achieved with

PRIMER
intravenous UFH, which is dose-adjusted to achieve a is needed to ensure that the INR remains in the thera-
therapeutic aPTT, or with weight-adjusted subcutane- peutic range of 2.0 to 3.0. Warfarin dosing algorithms
ous LMWH or fondaparinux treatment. LMWH and that incorporate pharmacogenetic and clinical variables
fondaparinux can be given subcutaneously in fixed have been developed. Trials comparing warfarin dosing
doses without monitoring and are therefore preferred using these algorithms versus usual practice without
over UFH. However, LMWH and fondaparinux are these algorithms have shown more rapid achievement
cleared by the kidneys and can accumulate in patients of a therapeutic INR but have yet to demonstrate
with impaired renal function. Consequently, these improved outcomes, such as reduced rates of recurrent
patients are treated with UFH, which is mainly cleared thrombosis or less bleeding 147,148. This might change with
via extrarenal routes. UFH is also used for patients with refinements in the dosing algorithms but, in the interim,
extensive VTE who might require systemic, pharmaco- improving the infrastructure for anticoagulation man-
mechanical or catheter-directed thrombolysis because agement is crucial because there is abundant evidence
the dose can be lowered easily to reduce the risk of that warfarin management in dedicated anticoagulation
bleeding. In addition, the short half-life of intravenous clinics surpasses that in the community 149.
UFH and the capacity to rapidly reverse the antico-
agulant effects of UFH with protamine sulfate can be Direct oral anticoagulants. DOACs directly inhibit
advantageous in this setting if there is bleeding. the enzymatic activity of either factor Xa or thrombin
(FIG. 2). DOACs have been compared with conventional
Vitamin K antagonists. Vitamin K antagonists, such as therapy in patients with acute VTE treatment in six
warfarin, inhibit the vitamin K‑dependent synthesis of Phase III clinical trials (TABLE 1). All of the trials used
coagulation factors, including prothrombin and fac- recurrent fatal or non-fatal VTE as the same primary
tors VII, IX and X (FIG. 2). They have been the ‘gold stand- efficacy end point. The main safety outcome was either
ard’ for oral anticoagulation for more than 50 years. The major bleeding or a combined outcome of major and
effective dose of warfarin differs from person to person clinically relevant non-major bleeding.
and can vary over time within an individual patient. This The RE‑COVER trial compared dabigatran with
variability reflects common polymorphisms that influ- warfarin in 2,539 patients with VTE137. VTE recur-
ence the metabolism and pharmacodynamics of warfa- rence rates were similar with dabigatran and warfarin
rin, variations in dietary vitamin K intake and multiple (TABLE 1). Fewer episodes of any bleeding occurred with
drug–drug interactions146. Thus, frequent monitoring dabigatran, although major bleeding rates were similar.
Table 1 | Summary of Phase III clinical trials evaluating new oral anticoagulants

Trial DOAC Compared Design n Primary efficacy Safety outcome: Refs
with outcome* bleeding
(HR; 95% CI) (HR; 95% CI)
RE-COVER Dabigatran Warfarin Randomized, 2,539 with Non-inferior • Major bleeding (0.82; 137
double-blind, acute VTE (1.10; 0.65–1.84) 0.45–1.48)
double-dummy • M/CRN (0.63; 0.47–0.84)
• Any bleeding (0.71;
0.59–0.85)
RE‑COVER II Dabigatran Warfarin Randomized, 2,589 with Non-inferior • Major bleeding (0.69; 138
double-blind, acute VTE (1.08; 0.64–1.80) 0.36–1.32)
double-dummy • M/CRN (0.62; 0.45–0.84)
• Any bleeding (0.67;
0.56–0.81)
Pooled Dabigatran Warfarin Randomized, 5,107 with Non-inferior • Major bleeding (0.73; 138
RE‑COVER double-blind, acute VTE (1.09; 0.76–1.57) 0.48–1.11)
populations double-dummy • M/CRN (0.62; 0.50–0.76)
• Any bleeding (0.7; 0.61–0.79)
EINSTEIN-DVT Rivaroxaban Enoxaparin Randomized, 3,449 with Non-inferior • First M/CRN (0.97; 0.76–1.22) 140
followed by open-label acute proximal (0.68; 0.44–1.04)
warfarin DVT
EINSTEIN-PE Rivaroxaban Enoxaparin Randomized, 4,832 with Non-inferior • Major bleeding (0.49; 141
followed by open-label acute PE with or (1.12; 0.75–1.68) 0.31–0.79)
warfarin without DVT • M/CRN (0.90; 0.76–1.07)
AMPLIFY Apixaban Enoxaparin Randomized, 5,395 with Non-inferior • Major bleeding (0.31‡; 142
followed by double-blind, acute VTE (0.84‡; 0.60–1.18) 0.17–0.55)
warfarin double-dummy (1,836 with PE) • M/CRN (0.44‡; 0.36–0.55)
Hokusai-VTE Edoxaban Warfarin Randomized, 8,240 with Non-inferior • First M/CRN (0.81; 0.71–0.94) 139
double-blind, acute VTE (0.89; 0.70–1.13) • Major bleeding (0.84;
double-dummy (3,319 with PE) 0.59–1.21)
CI, confidence interval; DOAC, direct oral anticoagulant; DVT, deep-vein thrombosis; HR, hazard ratio; M/CRN, major or clinically relevant non-major bleeding;
PE, pulmonary embolism; VTE, venous thromboembolism. *Primary efficacy end point: recurrent fatal or non-fatal VTE. ‡Relative risk.

PRIMER
These results were confirmed in RE‑COVER II138 as well compromised (hypotension) at presentation 132 .
as in a pooled analysis of both RE‑COVER trials138. Thrombolytic therapy can be given systemically
The EINSTEIN-DVT140 and EINSTEIN-PE141 trials or by catheter infusion directly into the thrombus.
compared all oral rivaroxaban (15 mg twice daily for Thrombolytic therapy (using tissue-type plasminogen
3 weeks, followed by 20 mg once daily) with conven- activators) converts endogenous plasminogen to plas-
tional therapy (enoxaparin followed by warfarin). In the min, which then degrades fibrin in clots; alteplase is
EINSTEIN-DVT study, which enrolled 3,449 patients the most commonly used thrombolytic drug. Surgical
with proximal DVT and no clinical evidence of embo- thrombectomy and thrombus maceration and extrac-
lism, rivaroxaban was non-inferior to conventional tion procedures are less often performed. In patients
therapy in preventing recurrent symptomatic VTE140 without hypotension, the benefits of thrombolysis
(TABLE 1). Rivaroxaban and conventional therapy had are offset by more frequent bleeding, including intra
similar bleeding rates. Using the same study design, the cranial bleeding 151. It is unknown whether the benefit-
EINSTEIN‑PE study enrolled 4,832 patients with pulmo- to‑risk balance of thrombolysis will be improved
nary embolism with or without associated DVT. For the using catheter-based technologies with or without
primary efficacy outcome, rivaroxaban was non-inferior adjunctive ultrasound treatment. Local catheter-
to conventional therapy 141 (TABLE 1). The rates of major or directed and/or pharmacomechanical thrombolysis is
clinically relevant non-major bleeding in the two treat- occasionally used for the treatment of extensive DVT
ment groups were similar; however, the rate of major based on the hypothesis that opening the occluded veins
bleeding alone was lower with rivaroxaban than that preserves valve function and reduces the risk of PTS152.
with conventional therapy (1.1% and 2.2%, respectively).
The AMPLIFY trial compared an all oral regimen of Long-term treatment. The optimal duration of anticoag-
apixaban (10 mg twice daily for 7 days, followed by 5 mg ulation after a first VTE episode is still debated. At least
once daily) with conventional therapy in 5,395 patients 3 months of treatment are recommended for patients
with acute VTE142. For the primary efficacy outcome, with provoked VTE136. Longer treatment should be
apixaban was non-inferior to conventional therapy restricted to patients whose provoking factor has not yet
(TABLE 1), but the rate of major bleeding was significantly been resolved. Similarly, at least 3 months of treatment
lower with apixaban (P < 0.001 for superiority). The com- is recommended for patients with unprovoked VTE.
bined rate of major and clinically relevant non-major The decision to extend treatment is based on the risks
bleeding was also lower with apixaban than with conven- of recurrent VTE and bleeding and on patient prefer-
tional therapy (4.3% and 9.7%, respectively; P < 0.001). ences136. For patients with a first unprovoked proximal
The Hokusai-VTE study compared edoxaban with DVT or pulmonary embolism and a low risk of bleed-
warfarin in 8,240 patients with acute VTE (including, indefinite anticoagulation should be considered. For
ing 3,319 with pulmonary embolism)139. All patients most patients with a second unprovoked DVT or pulmo-
received heparin for at least 5 days and were then ran- nary embolism, lifelong treatment is recommended. In
domly assigned to edoxaban or warfarin treatment. Both patients with cancer and VTE, anticoagulation should be
groups received the study drug for 3–12 months, and all continued until the cancer is cured or enters remission.
patients were followed for 12 months. For the primary Rivaroxaban, apixaban, dabigatran and low-dose
efficacy outcome, edoxaban was non-inferior to warfa- aspirin have been compared with placebo for extended
rin (TABLE 1), but major or clinically relevant non-major VTE treatment81–83,140,142,153. All were superior to placebo
bleeding was less frequent with edoxaban (P = 0.004 for for prevention of recurrent VTE with favourable safety
superiority). Notably, the rate of recurrent VTE was lower profiles in terms of major bleeding. Anticoagulants are
with edoxaban than with warfarin in 938 patients with superior to aspirin; rivaroxaban, apixaban and dabigatran
acute pulmonary embolism who had increased levels of reduced the risk of recurrence by at least 80%, whereas
N‑terminal pro-brain natriuretic peptide (at or above aspirin showed only a 32% risk reduction83. Extended
500 pg ml–1) at presentation (3.3% and 6.2%, respectively). treatment with dabigatran was also shown to be non-
In summary, DOACs are non-inferior to conven- inferior to warfarin for prevention of recurrent VTE,
tional therapy in preventing recurrent VTE. Moreover, and dabigatran produced less bleeding 153. Whether the
DOACs are associated with less major bleeding, practicality of DOACs and their favourable safety profiles
although a recent meta-analysis suggested the DOACs will increase the uptake of extended anticoagulation in
might cause more gastrointestinal bleeding 150. More patients with unprovoked VTE remains to be determined.
research is needed because most patients enrolled in
these trials had normal renal function, and there were Vena cava filters. Vena cava filters are used to prevent
few patients of very high or low body weight and rela- emboli from the deep veins of the leg from reaching the
tively few elderly patients — all for whom fixed dosing lungs. Vena cava filters should only be considered in
might not be optimal. In addition, patients with cancer patients with acute, extensive DVT with or without asso-
and VTE were underrepresented in these trials. ciated pulmonary embolism who have contraindications
to anticoagulation, such as serious bleeding, and in those
Other treatment options with recurrent VTE despite optimal anticoagulation136.
Thrombolytic treatment. Thrombolytic treatment is Retrievable filters are also used to prevent pulmonary
indicated for patients with massive pulmonary embo- embolism in high-risk situations whereby anticoagula-
lism — that is, for patients who are haemodynamically tion is temporarily contraindicated (prior to major cancer

PRIMER
Chronic thromboembolic found in one-third of patients159. However, one-half of

hypertension patients report dyspnoea and poor physical performance
Breathlessness
and/or reduced quality of life 6 months to 3 years after
and reduced exercise adequately treated embolism. Thus, pulmonary hyper-
tolerance tension seems to be the extreme manifestation of a much
more common phenomenon of abnormal pulmonary
Residual perfusion159. Thromboendarterectomy can be life-saving
perfusion
defects and can markedly improve quality of life.
Potentially increased
risk of recurrent Palliative care
pulmonary embolism
Currently, there is no effective treatment for estab-
lished PTS. The SOX trial indicated that application
of 30–40 mm Hg elastic compression stockings applied
within 2 weeks of acute DVT did not prevent PTS160 or
reduce leg pain associated with acute DVT161. However,
PTS
patients were only required to wear stockings for 3 days
a week to be classified as frequent users. Two previous
Pain, swelling and open-label randomized trials of compression stockings
skin changes showed an approximate 50% reduction in PTS162,163.
Residual Consequently, the potential benefit of compression
thrombus stockings remains contentious.
and vein Few studies have focussed on the long-term effects
occlusion
of pulmonary embolism and the high prevalence of
Increased risk of impaired exercise tolerance and reduced quality of life
recurrent DVT
in survivors is not well appreciated. Consequently, there
are no specific management strategies for patients with
post-embolism syndrome.
Figure 7 | Long-term consequences of venous thromboembolism.
Nature ReviewsVenous
| Disease Primers
thromboembolism impairs quality of life as a chronic disease with acute exacerbations.
Deep-vein thrombosis (DVT) might result in residual vein occlusion and post-thrombotic Comorbidity
syndrome (PTS). PTS is a major determinant of DVT recurrence risk, and recurrent DVT The cancer diagnosis rate is increased for 6 months
exacerbates residual vein occlusion and PTS, and leads to recurrent pulmonary after presentation with VTE164. Patients with malig-
embolism. Pulmonary embolism results in residual perfusion defects that can ultimately nancy and VTE have a 5.5‑fold increased mortality
result in chronic thromboembolic pulmonary hypertension. It remains uncertain compared with those without thrombosis. More than
whether residual perfusion defects and chronic hypertension give rise to recurrent half of the patients with malignancy and VTE die within
pulmonary embolism in situ. 12 months of diagnosis, mostly because of the cancer 5,6.
The risk of comorbid arterial cardiovascular events is
increased in patients with VTE; however, the connection
surgery in patients with recently diagnosed extensive is poorly understood. Nonetheless, the majority of stud-
DVT or prior to delivery in pregnant women with exten- ies support the concept that patients with unprovoked
sive DVT), and removed when the risk subsides. VTE are at increased risk of arterial events than patients
with provoked VTE and controls165,166. For example,
Quality of life patients presenting with unprovoked VTE before the age
Patient-reported outcomes of 40 years had a fourfold increased risk of myocardial
VTE is a chronic disease with acute exacerbations (FIG. 7). infarction in the subsequent 10 years167. This observation
Quality of life after VTE depends on the presence and suggests that risk reduction measures in patients with
severity of PTS, persistent altered pulmonary perfu- VTE that include exercise, cessation of smoking and
sion, chronic thromboembolic pulmonary hypertension, other lifestyle modifications can both improve quality of
limitations in lifestyle or occupation as a consequence of life and reduce the risk of arterial cardiovascular events.
anticoagulant therapy (particularly in younger patients
and those requiring lifelong therapy), and psychologi- Outlook
cal consequences, such as post-traumatic stress disorder, Unresolved research questions
in pulmonary embolism survivors154. Both generic and As with many diseases, there is a large gap between ani-
disease-specific quality-of-life scores improve 1–4 months mal models of venous thrombosis and the human dis-
after acute DVT, but patients who develop PTS have ease (BOX 2). A major issue is identifying the best animal
persistently poor scores7,155. Quality of life is reduced by model to use for these studies. Although baboons more
pain, decreased physical functioning and mobility, nega- closely resemble human physiology, mice are the most
tive emotional reactions and social isolation156,157, and it frequently used animal for thrombosis studies. However,
declines as PTS increases in severity 158. variations in murine models make it difficult to com-
Chronic pulmonary hypertension occurs in up to 4% pare findings between studies, and standardization of
of patients who have pulmonary embolism, but incom- these models is needed. Furthermore, should results
plete resolution of perfusion defects on lung scans is in mice be confirmed in models of larger animals such

PRIMER
as the baboon? Current pulmonary embolism models NETs have increased microvascular thrombosis170. Taken
involve intravenous injection of tissue factor or plate- together, these studies suggest that reducing neutro
let activators, or removal of an artificial plug to trigger phil release of NETs or treatment with DNase I might
embolization168, but these models do not recapitulate be novel approaches to reducing the incidence of DVT.
mechanisms that trigger spontaneous embolization in Notably, however, a recent review on the role of NETs in
humans. Continued development of models that show venous thrombosis in patients with cancer concluded
embolism from existing DVT, as well as spontaneous that “the current enthusiasm about NETs in cancer is not
pulmonary embolism, are crucial for understanding the yet supported by clinical data” (REF. 171).
pathophysiological mechanisms of VTE.
Most animal studies have analysed the effect of sin- Clinical needs
gle risk factors on thrombosis. These studies typically Development of new, safer drugs. Full acceptance of the
induce thrombosis using mild‑to‑severe vessel injury in currently available DOACs into clinical practice has been
young animals with otherwise healthy vessels and a nor- slowed by two potential limitations. First, although there
mal coagulation profile. However, clinical observations is no need for routine laboratory monitoring of these
indicate that a combination of two or more risk factors drugs, reliable laboratory tests for measuring levels of
is generally required to trigger DVT in humans, such these agents are not widely available. Second, in contrast
as a hypercoagulable state, advanced age, major surgery, to traditional anticoagulants (such as heparin and vita-
immobilization or cancer. Future studies should ana- min K antagonists), the lack of specific reversal agents
lyse the effect of combinations of risk factors on venous for DOACs increases the risk of bleeding complications
thrombosis. Modelling complex diseases associated with in emergency situations. For example, the anticoagulant
thrombosis, such as cancer, might identify new factors activity of UFH can be reversed by administration of
that contribute to its pathogenesis and that could be used protamine sulfate. Specific DOAC reversal agents are
as therapeutic targets. under development; this includes a modified version of
The contribution of blood cells to thrombogenesis factor Xa that can be used to reverse the anticoagulant
remains unclear. Studies have shown a role for leuko- effects of anti-factor Xa agents and an antibody to the
cytes. However, these cells might enhance thrombosis via thrombin inhibitor dabigatran172,173, but these have not
multiple mechanisms, such as by expressing tissue factor, yet been approved for clinical use.
by increasing inflammation and by releasing NETs. By Substantial efforts are also underway to develop new
contrast, they might promote thrombus resolution by anticoagulants that carry little or no risk of bleeding
releasing fibrinolytic enzymes and MMPs. Dissecting — the ‘holy grail’ of anticoagulation174. Data emerg-
the relative importance of these different mechanisms ing from basic science investigations and epidemio-
may guide us to new therapeutic strategies for the pre- logical observations suggest that coagulation proteins
vention and treatment of DVT. In particular, NETs have other than those currently targeted might be viable
become a hot topic in thrombosis research. Recent stud- targets for future anticoagulants, including factors
ies have shown that patients with DVT have elevated IXa, XIa, XIIa (FIG. 2) and XIIIa111,175–186. For example,
DNA levels in the plasma, and that citrullinated histone a recent study examined the effect of reducing fac-
H3, a marker of NETs, can be detected in human venous tor XI expression using an antisense oligonucleotide
thrombi79,169. Interestingly, one study has reported that (FXI-ASO) in patients undergoing total knee replace-
patients with impaired DNase I‑mediated degradation of ment 186. Two dose levels of FXI-ASO were compared
with the LMWH enoxaparin. The low dose of FXI-ASO
Box 2 | Research and clinical priorities (200 mg) reduced factor XI expression to ~33% and
was non-inferior to enoxaparin with respect to symp-
Research priorities tomatic and asymptomatic VTE incidence (27% versus
• Animal models that reflect human pathogenesis 30%) with similar rates of bleeding (3% versus 8%).
• Standardized mouse models However, the higher dose of FXI-ASO (300 mg), which
• Pulmonary embolism models reduced factor XI levels to ~17%, produced a striking
• Models that include combinations of risk factors and diseases associated with reduction in VTE (4%) without increasing bleeding
thrombosis (3%). These exciting results indicate that factor XI con-
• Contribution of blood cells (for example, leukocytes) to thrombosis tributes to VTE development in patients undergoing
• Role of neutrophil extracellular traps in thrombosis
knee surgery and show that reducing factor XI expres-
sion can reduce VTE risk without increasing bleed-
Clinical needs ing. Inhibiting factor XIa activity would be expected
• Laboratory tests for measuring effect of direct oral anticoagulants (DOACs) to produce similar results. Furthermore, factor XIIa
• Reversal agents for DOACs inhibitors, including the tick anticoagulant proteins
• Safer anticoagulants with reduced risk of bleeding infestin‑4 and Ixodes ricinus contact phase inhibitor,
• Novel molecular targets for anticoagulation strategies reduce thrombosis in animal models180,183,184. The factor
• Improved means for rapid removal of thrombus XIIa inhibitory antibody 3F7 effectively reduces fibrin
• Improved imaging and diagnosis deposition in a rabbit model of extracorporeal mem-
• Identification of sensitive and specific biomarkers brane oxygenation185. Notably, because factor XII con-
tributes to thrombosis but has no role in haemostasis,
• Increased awareness of venous thromboembolism symptoms and management
its inhibition is not expected to increase bleeding.

PRIMER
Proteins outside the coagulation cascade might healthy individuals and might be a biomarker of VTE
also be viable targets for reducing thrombogenesis. risk in certain types of cancers. For example, levels of tis-
The Wakefield group has studied the effects of inhib- sue factor-positive microvesicles in the plasma increase
iting the P‑selectin–PSGL1 complex on DVT in the before VTE develops in patients with pancreatic cancer
baboon model187. These studies show that P‑selectin but not in other types of cancer 93,94. The reasons for this
inhibitors and PSGL1 inhibitors reduce thrombosis and apparent specificity are unclear but might reflect the fact
inflammation as effectively as enoxaparin. It would be that pancreatic cancer is often detected at a later stage
interesting to determine whether these inhibitors can than other cancers. Nevertheless, these studies provide a
prevent VTE in humans. proof of principle that tissue factor-positive microvesicles
might be a useful biomarker for predicting thrombosis
Treatment improvements. As discussed above, rapid in a subset of patients and could be combined with other
removal of thrombus and restoration of flow might biomarkers and risk assessment scoring methods.
preserve valvular function and reduce morbidity fol-
lowing DVT; this concept is also called the open-vein VTE awareness
hypothesis152. Randomized clinical trials to evaluate the The rate of VTE is increasing as the number of peo-
efficacy and safety of pharmacomechanical catheter- ple over the age of 50 increases. A recent study ana-
directed thrombolysis in patients with DVT are lysed rates of VTE in Worcester, Massachusetts, USA,
ongoing and will determine whether catheter-directed between 1985 and 2009 (REF. 193). The study concluded
thrombolysis reduces the risk or severity of PTS. One that “despite advances in identification, prophylaxis,
such study is the ATTRACT study 188 (ClinicalTrials. and treatment between 1985 and 2009, the annual rate
gov identifier: NCT00790335), which is expected to of venous thromboembolism has increased and remains
determine the efficacy, safety and cost-effectiveness of high” (REF. 193). There is a need to increase public aware-
pharmacomechanical catheter-directed thrombolysis ness of the symptoms of DVT and pulmonary embolism
for preventing PTS. and to encourage at‑risk and affected individuals to seek
medical attention. In 2006, the US Surgeon General held
Improvements in imaging technology. The algorithms a workshop on VTE, resulting in the Surgeon General’s
used to diagnose VTE and distinguish newly formed 2008 call to action to prevent DVT and pulmonary
thrombus from residual thrombus are complex and can embolism121. This call was designed to “increase public
lack robust positive predictive value. Improvements in awareness, support the development of evidence-based
imaging technologies that increase safety, sensitivity and practice, and carry out research that can address the
specificity might enable better diagnosis of distal DVT gaps in our knowledge” (REF. 121). In the United States,
and subsegmental pulmonary embolism. March is now designated as “DVT awareness month”.
In 2014, the International Society of Thrombosis and
Biomarkers and VTE prediction. Can biomarkers be Haemostasis established a worldwide effort to increase
used to identify patients at risk of VTE? D‑dimer assays VTE awareness4. Non-profit organizations such as Clot
are routinely used for the diagnosis of VTE and to assess Connect 194 have also been created to educate the public
the risk of recurrent VTE; however, D‑dimer tests have about VTE risks, symptoms and treatment. However,
poor predictive values. Furthermore, D‑dimer tests can- more efforts are needed on this front.
not be used to identify patients with incident VTE128. There is also a need to educate medical profession-
Addition of other biomarkers might enhance the predic- als regarding the management of VTE. For example,
tive value of a risk assessment score for VTE. For exam- thromboprophylaxis is still greatly underused. A recent
ple, elevated levels of soluble P‑selectin have been found study described the main reasons for the lack of adher-
in patients with recurrent VTE compared with con- ence to guidelines; these include a lack of VTE awareness
trols189. Notably, high plasma levels of soluble P‑selectin or assessment of VTE risk factors; difficulties in defin-
are predictive of VTE in patients with cancer 190. Indeed, ing risk factors and thromboprophylaxis indications; and
addition of information on both D‑dimer and soluble guidelines that result in inaccurate and unclear indica-
P‑selectin strengthen the Khorana score for predicting tions for thromboprohylaxis195. These issues might be
VTE in patients with cancer190. Other biomarkers might partly remedied through the use of educational pro-
further refine risk assessment scoring models. grammes, which have been shown to increase use of
Circulating microvesicles might be another use- thromboprophylaxis (from 29% to 52%196).
ful biomarker to identify both cancer and non-cancer
patients at increased risk of VTE. However, owing to Conclusions
the high levels of microvesicles in the plasma of healthy Morbidity and mortality from VTE have a considerable
individuals, which are mostly derived from platelets, effect on the global burden of disease. Awareness of and
it seems unlikely that levels of total microvesicles will reduction in modifiable risks is the first step towards
be useful for predicting or diagnosing VTE. Indeed, reducing incidence. Continued clinical, epidemiological
whereas one retrospective study concluded that “high and laboratory investigations are essential for delineating
levels of circulating microparticles are a possible inde- new, effective approaches for decreasing incidence and,
pendent risk factor for VTE” in patients191, another study therefore, long-term sequelae. Integration of these inter-
found no association192. By contrast, tissue factor-positive disciplinary methods is expected to have a substantial
microvesicles are generally not detectable in the blood of impact on the future of VTE and global health.

PRIMER
1. Grant, J. D. et al. Diagnosis and management of upper 25. Brill, A. et al. von Willebrand factor-mediated platelet 49. Butenas, S., van‘t Veer, C., Cawthern, K.,
extremity deep-vein thrombosis in adults. Thromb. adhesion is critical for deep vein thrombosis in mouse Brummel, K. E. & Mann, K. G. Models of blood
Haemost. 108, 1097–1108 (2012). models. Blood 117, 1400–1407 (2011). coagulation. Blood Coag. Fibrinol. 11 (Suppl. 1),
2. Virchow, R. Gesammalte Abhandlungen zur 26. von Bruhl, M. L. et al. Monocytes, neutrophils, and S9–S13 (2000).
Wissenschaftlichen Medizin. 219–732 (in German) platelets cooperate to initiate and propagate venous 50. Allen, G. A. et al. Impact of procoagulant concentration
(Medinger Sohn & Co., Frankfurt, 1856). thrombosis in mice in vivo. J. Exp. Med. 209, on rate, peak and total thrombin generation in a model
3. Bagot, C. N. & Arya, R. Virchow and his triad: 819–835 (2012). system. J. Thromb. Haemost. 2, 402–413 (2004).
a question of attribution. Br. J. Haematol. 143, This study demonstrates the roles of leukocyte 51. Machlus, K. R. et al. Effects of tissue factor,
180–190 (2008). tissue factor, factor XIIa, neutrophils and platelets thrombomodulin, and elevated clotting factor levels on
4. International Society on Thrombosis and Haemostasis. in a mouse model of thrombosis. thrombin generation in the calibrated automated
About World Thrombosis Day [online], http://www. 27. Geddings, J. et al. Strengths and weaknesses of a new thrombogram. Thromb. Haemost. 102, 936–944
worldthrombosisday.org/about/ (2014). mouse model of thrombosis induced by inferior vena (2009).
5. Naess, I. A. et al. Incidence and mortality of venous cava stenosis: communication from the SSC of the 52. Wichers, I. M. et al. Assessment of coagulation and
thrombosis: a population-based study. J. Thromb. ISTH. J. Thromb. Haemost. 12, 571–573 (2014). fibrinolysis in families with unexplained thrombophilia.
Haemost. 5, 692–699 (2007). 28. Myers, D. Jr. et al. Selectins influence thrombosis in a Thromb. Haemost. 101, 465–470 (2009).
6. Flinterman, L. E., van Hylckama Vlieg, A., mouse model of experimental deep venous 53. Brandts, A., van Hylckama Vlieg, A., Rosing, J.,
Cannegieter, S. C. & Rosendaal, F. R. Long-term thrombosis. J. Surg. Res. 108, 212–221 (2002). Baglin, T. P. & Rosendaal, F. R. The risk of venous
survival in a large cohort of patients with venous 29. Diaz, J. A. et al. Thrombogenesis with continuous thrombosis associated with a high endogenous
thrombosis: incidence and predictors. PLoS Med. 9, blood flow in the inferior vena cava. A novel mouse thrombin potential in the absence and presence of
e1001155 (2012). model. Thromb. Haemost. 104, 366–375 (2010). activated protein C. J. Thromb. Haemost. 5, 416–418
This study describes the long-term mortality in a 30. Cooley, B. C., Szema, L., Chen, C. Y., Schwab, J. P. & (2007).
large cohort of Dutch patients with DVT. Schmeling, G. A murine model of deep vein 54. Dargaud, Y., Trzeciak, M. C., Bordet, J. C., Ninet, J. &
7. Kahn, S. R. et al. The postthrombotic syndrome: thrombosis: characterization and validation in Negrier, C. Use of calibrated automated
evidence-based prevention, diagnosis, and treatment transgenic mice. Thromb. Haemost. 94, 498–503 thrombinography +/– thrombomodulin to recognise
strategies: a scientific statement from the American (2005). the prothrombotic phenotype. Thromb. Haemost. 96,
Heart Association. Circulation 130, 1636–1661 31. Wang, X., Smith, P. L., Hsu, M. Y., Ogletree, M. L. & 562–567 (2006).
(2014). Schumacher, W. A. Murine model of ferric chloride- 55. van Hylckama Vlieg, A. et al. Elevated endogenous
8. Cohen, A. T. et al. Venous thromboembolism (VTE) in induced vena cava thrombosis: evidence for effect of thrombin potential is associated with an increased risk
Europe. The number of VTE events and associated potato carboxypeptidase inhibitor. J. Thromb. of a first deep venous thrombosis but not with the risk
morbidity and mortality. Thromb. Haemost. 98, Haemost. 4, 403–410 (2006). of recurrence. Br. J. Haematol. 138, 769–774 (2007).
756–764 (2007). 32. Aghourian, M. N., Lemarie, C. A. & Blostein, M. D. 56. Tripodi, A. et al. The endogenous thrombin potential
9. ISTH Steering Committee for World Thrombosis Day. In vivo monitoring of venous thrombosis in mice. and the risk of venous thromboembolism. Thromb.
Thrombosis: a major contributor to the global disease J. Thromb. Haemost. 10, 447–452 (2012). Res. 121, 353–359 (2007).
burden. J. Thromb. Haemost. 12, 1580–1590 33. Cardenas, J. C. et al. Overexpression of the cell cycle 57. ten Cate-Hoek, A. J. et al. Thrombin generation in
(2014). inhibitor p16INK4a promotes a prothrombotic patients after acute deep-vein thrombosis. Thromb.
This systematic review of the literature shows that phenotype following vascular injury in mice. Haemost. 100, 240–245 (2008).
VTE is a major burden of disease across Arterioscler. Thromb. Vasc. Biol. 31, 827–833 (2011). 58. Hron, G., Kollars, M., Binder, B. R., Eichinger, S. &
low-income, middle-income and high-income 34. Machlus, K. R., Cardenas, J. C., Church, F. C. & Kyrle, P. A. Identification of patients at low risk for
countries. Wolberg, A. S. Causal relationship between recurrent venous thromboembolism by measuring
10. Mackman, N. New insights into the mechanisms of hyperfibrinogenemia, thrombosis, and resistance to thrombin generation. JAMA 296, 397–402 (2006).
venous thrombosis. J. Clin. Invest. 122, 2331–2336 thrombolysis in mice. Blood 117, 4953–4963 (2011). 59. Eichinger, S., Hron, G., Kollars, M. & Kyrle, P. A.
(2012). 35. Machlus, K. R., Lin, F.‑C. & Wolberg, A. S. Prediction of recurrent venous thromboembolism by
11. Silverstein, M. D. et al. Trends in the incidence of deep Procoagulant activity induced by vascular injury endogenous thrombin potential and D‑dimer. Clin.
vein thrombosis and pulmonary embolism: a 25‑year determines contribution of elevated factor VIII to Chem. 54, 2042–2048 (2008).
population-based study. Arch. Intern. Med. 158, thrombosis and thrombus stability in mice. Blood 118, 60. Besser, M., Baglin, C., Luddington, R., van Hylckama
585–593 (1998). 390–398 (2011). Vlieg, A. & Baglin, T. High rate of unprovoked recurrent
12. Reitsma, P. H., Versteeg, H. H. & Middeldorp, S. 36. Warlow, C., Ogston, D. & Douglas, A. S. Deep venous venous thrombosis is associated with high thrombin-
Mechanistic view of risk factors for venous thrombosis of the legs after strokes. Part I — generating potential in a prospective cohort study.
thromboembolism. Arterioscler. Thromb. Vasc. Biol. incidence and predisposing factors. BMJ 1, J. Thromb. Haemost. 6, 1720–1725 (2008).
32, 563–568 (2012). 1178–1181 (1976). 61. Tripodi, A. et al. High thrombin generation measured
13. Zakai, N. A. & McClure, L. A. Racial differences in 37. Sevitt, S. The structure and growth of valve-pocket in the presence of thrombomodulin is associated with
venous thromboembolism. J. Thromb. Haemost. 9, thrombi in femoral veins. J. Clin. Pathol. 27, 517–528 an increased risk of recurrent venous
1877–1882 (2011). (1974). thromboembolism. J. Thromb. Haemost. 6,
14. Roach, R. E. et al. The risk of venous thrombosis in 38. Hamer, J. D., Malone, P. C. & Silver, I. A. The PO2 in 1327–1333 (2008).
individuals with a history of superficial vein thrombosis venous valve pockets: its possible bearing on 62. Aleman, M. M. et al. Elevated prothrombin promotes
and acquired venous thrombotic risk factors. Blood thrombogenesis. Br. J. Surg. 68, 166–170 (1981). venous, but not arterial, thrombosis in mice.
122, 4264–4269 (2013). 39. Lin, Z. et al. Krüppel-like factor 2 (KLF2) regulates Arterioscler. Thromb. Vasc. Biol. 33, 1829–1836
15. Cannegieter, S. C. et al. Risk of venous and arterial endothelial thrombotic function. Circ. Res. 96, (2013).
thrombotic events in patients diagnosed with e48–e57 (2005). 63. Meltzer, M. E., Lisman, T., Doggen, C. J.,
superficial vein thrombosis: a nationwide cohort study. 40. Lin, Z., Hamik, A., Jain, R., Kumar, A. & Jain, M. K. de Groot, P. G. & Rosendaal, F. R. Synergistic effects of
Blood 125, 229–235 (2015). Krüppel-like factor 2 inhibits protease activated hypofibrinolysis and genetic and acquired risk factors
16. Rosendaal, F. R. & Reitsma, P. H. Genetics of venous receptor‑1 expression and thrombin-mediated on the risk of a first venous thrombosis. PLoS Med.
thrombosis. J. Thromb. Haemost. 7 (Suppl. 1), endothelial activation. Arterioscler. Thromb. Vasc. 5, e97 (2008).
301–304 (2009). Biol. 26, 1185–1189 (2006). 64. Meltzer, M. E. et al. Venous thrombosis risk associated
17. Bezemer, I. D. et al. F9 Malmo, factor IX and deep vein 41. Atkins, G. B. & Jain, M. K. Role of Krüppel-like with plasma hypofibrinolysis is explained by elevated
thrombosis. Haematologica 94, 693–699 (2009). transcription factors in endothelial biology. Circ. Res. plasma levels of TAFI and PAI‑1. Blood 116, 113–121
18. Bezemer, I. D. et al. Gene variants associated with 100, 1686–1695 (2007). (2010).
deep vein thrombosis. JAMA 299, 1306–1314 42. Brooks, E. G. et al. Valves of the deep venous system: 65. Fay, W. P., Parker, A. C., Condrey, L. R. & Shapiro, A. D.
(2008). an overlooked risk factor. Blood 114, 1276–1279 Human plasminogen activator inhibitor‑1 (PAI‑1)
19. Li, Y. et al. Genetic variants associated with deep vein (2009). deficiency: characterization of a large kindred with a
thrombosis: the F11 locus. J. Thromb. Haemost. 7, 43. Zhou, J., May, L., Liao, P., Gross, P. L. & Weitz, J. I. null mutation in the PAI‑1 gene. Blood 90, 204–208
1802–1808 (2009). Inferior vena cava ligation rapidly induces tissue factor (1997).
20. Austin, H. et al. New gene variants associated with expression and venous thrombosis in rats. 66. Engelmann, B. & Massberg, S. Thrombosis as an
venous thrombosis: a replication study in White and Arterioscler. Thromb. Vasc. Biol. 29, 863–869 intravascular effector of innate immunity. Nat. Rev.
Black Americans. J. Thromb. Haemost. 9, 489–495 (2009). Immunol. 13, 34–45 (2013).
(2011). 44. Wakefield, T. W. et al. P‑selectin and TNF inhibition 67. Connolly, G. C., Phipps, R. P. & Francis, C. W. Platelets
21. Andresen, M. S. et al. Mortality and recurrence after reduce venous thrombosis inflammation. J. Surg. Res. and cancer-associated thrombosis. Seminars Oncol.
treatment of VTE: long term follow‑up of patients with 64, 26–31 (1996). 41, 302–310 (2014).
good life-expectancy. Thromb. Res. 127, 540–546 45. Furie, B. & Furie, B. C. Role of platelet P‑selectin and 68. Jensvoll, H., Blix, K., Braekkan, S. K. & Hansen, J. B.
(2011). microparticle PSGL‑1 in thrombus formation. Trends Platelet count measured prior to cancer development
22. Wolberg, A. S., Aleman, M. M., Leiderman, K. & Mol. Med. 10, 171–178 (2004). is a risk factor for future symptomatic venous
Machlus, K. R. Procoagulant activity in hemostasis 46. Myers, D. D. Jr. et al. P‑selectin antagonism causes thromboembolism: the Tromso Study. PLoS ONE 9,
and thrombosis: Virchow’s Triad revisited. Anesth. dose-dependent venous thrombosis inhibition. e92011 (2014).
Analg. 114, 275–285 (2012). Thromb. Haemost. 85, 423–429 (2001). 69. Zakai, N. A., Wright, J. & Cushman, M. Risk factors for
23. Meier, T. R. et al. Prophylactic P‑selectin inhibition 47. Day, S. M. et al. Macrovascular thrombosis is driven venous thrombosis in medical inpatients: validation of
with PSI‑421 promotes resolution of venous by tissue factor derived primarily from the blood a thrombosis risk score. J. Thromb. Haemost. 2,
thrombosis without anticoagulation. Thromb. vessel wall. Blood 105, 192–198 (2005). 2156–2161 (2004).
Haemost. 99, 343–351 (2008). 48. Mackman, N., Tilley, R. E. & Key, N. S. Role of the 70. Khorana, A. A., Francis, C. W., Culakova, E. &
24. Myers, D. D. Jr. Nonhuman primate models of extrinsic pathway of blood coagulation in hemostasis Lyman, G. H. Risk factors for chemotherapy-associated
thrombosis. Thromb. Res. 129 (Suppl. 2), S65–S69 and thrombosis. Arterioscler. Thromb. Vasc. Biol. 27, venous thromboembolism in a prospective
(2012). 1687–1693 (2007). observational study. Cancer 104, 2822–2829 (2005).

PRIMER
71. Simanek, R. et al. High platelet count associated with mouse xenograft model of human pancreatic cancer. 121. Office of the Surgeon General (US); National Heart,
venous thromboembolism in cancer patients: results Blood 119, 5543–5552 (2012). Lung, and Blood Institute (US). The surgeon general’s
from the Vienna Cancer and Thrombosis Study 97. Ho, C. H. The hemostatic effect of packed red cell call to action to prevent deep vein thrombosis and
(CATS). J. Thromb. Haemost. 8, 114–120 (2010). transfusion in patients with anemia. Transfusion 38, pulmonary embolism. Rockville (MD): Office of the
72. Castellucci, L. A. et al. Efficacy and safety outcomes 1011–1014 (1998). Surgeon General (US) [online], http://www.ncbi.nlm.
of oral anticoagulants and antiplatelet drugs in the 98. Khorana, A. A. et al. Blood transfusions, thrombosis, nih.gov/books/NBK44178/ (2008).
secondary prevention of venous thromboembolism: and mortality in hospitalized patients with cancer. 122. Kahn, S. R. et al. Prevention of VTE in nonsurgical
systematic review and network meta-analysis. BMJ Arch. Intern. Med. 168, 2377–2381 (2008). patients: antithrombotic therapy and prevention of
347, f5133 (2013). 99. Gangireddy, C. et al. Risk factors and clinical impact of thrombosis, 9th ed: American College of Chest
73. Simes, J. et al. Aspirin for the prevention of recurrent postoperative symptomatic venous thromboembolism. Physicians evidence-based clinical practice guidelines.
venous thromboembolism: the INSPIRE J. Vasc. Surg. 45, 335–341; discussion 341–332 Chest 141 (Suppl. 2), e195S–e226S (2012).
Collaboration. Circulation 130, 1062–1071 (2014). (2007). This paper provides evidence-based guidelines for
74. Riedl, J., Pabinger, I. & Ay, C. Platelets in cancer and 100. Kumar, M. A. et al. Red blood cell transfusion VTE prevention in non-surgical patients. These
thrombosis. Hamostaseologie 34, 54–62 (2014). increases the risk of thrombotic events in patients with guidelines are important because pulmonary
75. McGuinness, C. L. et al. Recruitment of labelled subarachnoid hemorrhage. Neurocrit. Care 20, embolism is the principal cause of preventable
monocytes by experimental venous thrombi. Thromb. 84–90 (2014). death in hospitalized patients.
Haemost. 85, 1018–1024 (2001). 101. Andrews, D. A. & Low, P. S. Role of red blood cells in 123. Falck-Ytter, Y. et al. Prevention of VTE in orthopedic
76. Blix, K., Jensvoll, H., Braekkan, S. K. & Hansen, J. B. thrombosis. Curr. Opin. Hematol. 6, 76–82 (1999). surgery patients: antithrombotic therapy and
White blood cell count measured prior to cancer 102. Baskurt, O. K. & Meiselman, H. J. Blood rheology and prevention of thrombosis, 9th ed: American College of
development is associated with future risk of venous hemodynamics. Semin. Thromb. Hemost. 29, Chest Physicians evidence-based clinical practice
thromboembolism — the Tromso study. PLoS ONE 8, 435–450 (2003). guidelines. Chest 141 (Suppl. 2), e278S–e325S
e73447 (2013). 103. Peyrou, V. et al. Contribution of erythrocytes to (2012).
77. Connolly, G. C. et al. Leukocytosis, thrombosis and thrombin generation in whole blood. Thromb. 124. Wu, O. et al. Screening for thrombophilia in high-risk
early mortality in cancer patients initiating Haemost. 81, 400–406 (1999). situations: systematic review and cost-effectiveness
chemotherapy. Thromb. Res. 126, 113–118 (2010). 104. Horne, M. K., 3rd, Cullinane, A. M., Merryman, P. K. & analysis. The Thrombosis: Risk and Economic
78. Fuchs, T. A. et al. Extracellular DNA traps promote Hoddeson, E. K. The effect of red blood cells on Assessment of Thrombophilia Screening (TREATS)
thrombosis. Proc. Natl Acad. Sci. USA 107, thrombin generation. Br. J. Haematol. 133, 403–408 study. Health Technol. Assess. 10, 1–110 (2006).
15880–15885 (2010). (2006). 125. Kearon, C. Natural history of venous
This study shows that NETs capture platelets and 105. Whelihan, M. F., Zachary, V., Orfeo, T. & Mann, K. G. thromboembolism. Circulation 107, I22–I30 (2003).
red blood cells, and enhance fibrin deposition. Prothrombin activation in blood coagulation: the 126. Levy, M. M., Albuquerque, F. & Pfeifer, J. D. Low
79. Savchenko, A. S. et al. Neutrophil extracellular traps erythrocyte contribution to thrombin generation. incidence of pulmonary embolism associated with
form predominantly during the organizing stage of Blood 120, 3837–3845 (2012). upper-extremity deep venous thrombosis. Ann. Vasc.
human venous thromboembolism development. 106. Rubin, O. et al. Red blood cell-derived microparticles Surg. 26, 964–972 (2012).
J. Thromb. Haemost. 12, 860–870 (2014). isolated from blood units initiate and propagate 127. Meignan, M. et al. Systematic lung scans reveal a high
80. Martinod, K. & Wagner, D. D. Thrombosis: tangled thrombin generation. Transfusion 53, 1744–1754 frequency of silent pulmonary embolism in patients
up in NETs. Blood 123, 2768–2776 (2014). (2013). with proximal deep venous thrombosis. Arch. Intern.
81. Kannemeier, C. et al. Extracellular RNA constitutes a 107. Van Der Meijden, P. E. et al. Platelet- and erythrocyte- Med. 160, 159–164 (2000).
natural procoagulant cofactor in blood coagulation. derived microparticles trigger thrombin generation via By performing systematic lung scans in patients
Proc. Natl Acad. Sci. USA 104, 6388–6393 (2007). factor XIIa. J. Thromb. Haemost. 10, 1355–1362 presenting with proximal DVT, the authors show
82. Martinod, K. et al. Neutrophil histone modification (2012). that at least 40% of such patients have
by peptidylarginine deiminase 4 is critical for deep 108. van Gelder, J. M., Nair, C. H. & Dhall, D. P. The asymptomatic emboli.
vein thrombosis in mice. Proc. Natl Acad. Sci. USA significance of red cell surface area to the permeability 128. Bates, S. M. et al. Diagnosis of DVT: antithrombotic
110, 8674–8679 (2013). of fibrin network. Biorheology 31, 259–275 (1994). therapy and prevention of thrombosis, 9th ed:
83. Massberg, S. et al. Reciprocal coupling of coagulation 109. Cines, D. B. et al. Clot contraction: compression of American College of Chest Physicians evidence-based
and innate immunity via neutrophil serine proteases. erythrocytes into tightly packed polyhedra and clinical practice guidelines. Chest 141 (Suppl. 2),
Nat. Med. 16, 887–896 (2010). redistribution of platelets and fibrin. Blood 123, e351S–e418S (2012).
84. Petersen, L. C., Bjorn, S. E. & Nordfang, O. Effect of 1596–1603 (2014). This paper provides an evidence-based approach to
leukocyte proteinases on tissue factor pathway 110. Wohner, N. et al. Lytic resistance of fibrin containing the diagnosis of DVT.
inhibitor. Thromb. Haemost. 67, 537–541 (1992). red blood cells. Arterioscler. Thromb. Vasc. Biol. 31, 129. Huisman, M. V. & Klok, F. A. Diagnostic management
85. Steppich, B. A., Seitz, I., Busch, G., Stein, A. & Ott, I. 2306–2313 (2011). of acute deep vein thrombosis and pulmonary
Modulation of tissue factor and tissue factor pathway 111. Aleman, M. M. et al. Factor XIII activity mediates red embolism. J. Thromb. Haemost. 11, 412–422 (2013).
inhibitor‑1 by neutrophil proteases. Thromb. blood cell retention in venous thrombi. J. Clin. Invest. 130. Linkins, L. A. et al. Selective D‑dimer testing for
Haemost. 100, 1068–1075 (2008). 124, 3590–3600 (2014). diagnosis of a first suspected episode of deep venous
86. Patterson, K. A. et al. Rosuvastatin reduced deep This study shows that factor XIIIa activity is thrombosis: a randomized trial. Ann. Intern. Med.
vein thrombosis in ApoE gene deleted mice with required to retain red blood cells in venous clots 158, 93–100 (2013).
hyperlipidemia through non-lipid lowering effects. and that inhibition of factor XIIIa reduces clot size. 131. Righini, M. et al. Age-adjusted D‑dimer cutoff levels to
Thromb. Res. 131, 268–276 (2013). 112. Varma, M. R. et al. Neutropenia impairs venous rule out pulmonary embolism: the ADJUST‑PE study.
87. Owens, A. P. 3rd et al. Monocyte tissue factor- thrombosis resolution in the rat. J. Vasc. Surg. 38, JAMA 311, 1117–1124 (2014).
dependent activation of coagulation in 1090–1098 (2003). 132. Authors/Task Force Members et al. 2014 ESC
hypercholesterolemic mice and monkeys is inhibited 113. Sood, V. et al. Urokinase plasminogen activator guidelines on the diagnosis and management of acute
by simvastatin. J. Clin. Invest. 122, 558–568 independent early experimental thrombus resolution: pulmonary embolism: the task force for the diagnosis
(2012). MMP2 as an alternative mechanism. Thromb. and management of acute pulmonary embolism of the
88. Glynn, R. J. et al. A randomized trial of rosuvastatin Haemost. 104, 1174–1183 (2010). European Society of Cardiology (ESC) endorsed by the
in the prevention of venous thromboembolism. 114. Singh, I. et al. Failure of thrombus to resolve in European Respiratory Society (ERS). Eur. Heart J. 35,
N. Engl. J. Med. 360, 1851–1861 (2009). urokinase-type plasminogen activator gene-knockout 3033–3073 (2014).
89. Rahimi, K. et al. Effect of statins on venous mice: rescue by normal bone marrow-derived cells. 133. Kearon, C., Ginsberg, J. S. & Hirsh, J. The role of
thromboembolic events: a meta-analysis of published Circulation 107, 869–875 (2003). venous ultrasonography in the diagnosis of suspected
and unpublished evidence from randomised 115. Deatrick, K. B. et al. The effect of matrix deep venous thrombosis and pulmonary embolism.
controlled trials. PLoS Med. 9, e1001310 (2012). metalloproteinase 2 and matrix metalloproteinase 2/9 Ann. Intern. Med. 129, 1044–1049 (1998).
90. Owens, A. P. 3rd & Mackman, N. Microparticles in deletion in experimental post-thrombotic vein wall 134. Klok, F. A., Mos, I. C., Kroft, L. J., de Roos, A. &
hemostasis and thrombosis. Circ. Res. 108, remodeling. J. Vasc. Surg. 58, 1375–1384.e2 Huisman, M. V. Computed tomography pulmonary
1284–1297 (2011). (2013). angiography as a single imaging test to rule out
91. Ramacciotti, E. et al. Leukocyte- and platelet-derived 116. Wakefield, T. W. et al. Venous thrombosis-associated pulmonary embolism. Curr. Opin. Pulmonary Med.
microparticles correlate with thrombus weight and inflammation and attenuation with neutralizing 17, 380–386 (2011).
tissue factor activity in an experimental mouse model antibodies to cytokines and adhesion molecules. 135. Pasha, S. M. et al. NT‑pro-BNP levels in patients with
of venous thrombosis. Thromb. Haemost. 101, Arterioscler. Thromb. Vasc. Biol. 15, 258–268 acute pulmonary embolism are correlated to right but
748–754 (2009). (1995). not left ventricular volume and function. Thromb.
92. Owens, A. P., 3rd & Mackman, N. MP’s and VTE’s: 117. Henke, P. K. et al. Targeted deletion of CCR2 impairs Haemost. 108, 367–372 (2012).
fact or fiction. Thromb. Res. 128, 505–506 (2011). deep vein thombosis resolution in a mouse model. 136. Kearon, C. et al. Antithrombotic therapy for VTE
93. Khorana, A. A. et al. Plasma tissue factor may be J. Immunol. 177, 3388–3397 (2006). disease: antithrombotic therapy and prevention of
predictive of venous thromboembolism in pancreatic 118. Ali, T. et al. Monocyte recruitment in venous thrombus thrombosis, 9th ed: American College of Chest
cancer. J. Thromb. Haemost. 6, 1983–1985 (2008). resolution. J. Vasc. Surg. 43, 601–608 (2006). Physicians evidence-based clinical practice guidelines.
94. Geddings, J. E. & Mackman, N. Tumor-derived tissue 119. Obi, A. T. et al. Plasminogen activator‑1 Chest 141 (Suppl. 2), e419S–e494S (2012).
factor-positive microparticles and venous thrombosis overexpression decreases experimental This paper provides evidence-based guidelines for
in cancer patients. Blood 122, 1873–1880 (2013). postthrombotic vein wall fibrosis by a non-vitronectin- the treatment of VTE.
95. Thomas, G. M. et al. Cancer cell-derived dependent mechanism. J. Thromb. Haemost. 12, 137. Schulman, S. et al. Dabigatran versus warfarin in the
microparticles bearing P‑selectin glycoprotein 1353–1363 (2014). treatment of acute venous thromboembolism. N. Engl.
ligand 1 accelerate thrombus formation in vivo. 120. Baldwin, J. F. et al. The role of urokinase plasminogen J. Med. 361, 2342–2352 (2009).
J. Exp. Med. 206, 1913–1927 (2009). activator and plasmin activator inhibitor‑1 on vein wall 138. Schulman, S. et al. Treatment of acute venous
96. Wang, J. G. et al. Tumor-derived tissue factor remodeling in experimental deep vein thrombosis. thromboembolism with dabigatran or warfarin and
activates coagulation and enhances thrombosis in a J. Vasc. Surg. 56, 1089–1097 (2012). pooled analysis. Circulation 129, 764–772 (2014).

PRIMER
139. Hokusai, V. T. E. I. et al. Edoxaban versus warfarin for A randomised controlled trial. Thromb. Haemost. 112, 184. Decrem, Y. et al. Ir‑CPI, a coagulation contact phase
the treatment of symptomatic venous 1137–1141 (2014). inhibitor from the tick Ixodes ricinus, inhibits
thromboembolism. N. Engl. J. Med. 369, 1406–1415 162. Brandjes, D. P. et al. Randomised trial of effect of thrombus formation without impairing hemostasis.
(2013). compression stockings in patients with symptomatic J. Exp. Med. 206, 2381–2395 (2009).
140. EINSTEIN Investigators et al. Oral rivaroxaban for proximal-vein thrombosis. Lancet 349, 759–762 185. Larsson, M. et al. A factor XIIa inhibitory antibody
symptomatic venous thromboembolism. N. Engl. (1997). provides thromboprotection in extracorporeal
J. Med. 363, 2499–2510 (2010). 163. Prandoni, P. et al. Below-knee elastic compression circulation without increasing bleeding risk. Sci. Transl
141. The EINSTEIN–PE Investigators. Oral rivaroxaban for stockings to prevent the post-thrombotic syndrome: a Med. 6, 222ra217 (2014).
the treatment of symptomatic pulmonary embolism. randomized, controlled trial. Ann. Intern. Med. 141, This study shows that a factor XIIa-specific
N. Engl. J. Med. 366, 1287–1297 (2012). 249–256 (2004). antibody reduces activation of coagulation in an
142. Agnelli, G. et al. Oral apixaban for the treatment of 164. Sorensen, H. T., Mellemkjaer, L., Steffensen, F. H., extracorporeal membrane oxygenator, suggesting
acute venous thromboembolism. N. Engl. J. Med. Olsen, J. H. & Nielsen, G. L. The risk of a diagnosis of that targeting factor XIIa might reduce thrombosis.
369, 799–808 (2013). cancer after primary deep venous thrombosis or 186. Büller, H. R. et al. Factor XI antisense oligonucleotide
143. Spyropoulos, A. C. & Lin, J. Direct medical costs of pulmonary embolism. N. Engl. J. Med. 338, for prevention of venous thrombosis. N. Engl. J. Med.
venous thromboembolism and subsequent hospital 1169–1173 (1998). 372, 232–240 (2015).
readmission rates: an administrative claims analysis 165. Franchini, M. & Mannucci, P. M. Association between This study shows that antisense therapy to
from 30 managed care organizations. J. Manag. Care venous and arterial thrombosis: clinical implications. decrease factor XI levels reduces VTE in patients
Pharm. 13, 475–486 (2007). Eur. J. Internal Med. 23, 333–337 (2012). undergoing elective primary total knee
144. Ruppert, A., Steinle, T. & Lees, M. Economic burden of 166. Becattini, C., Vedovati, M. C., Ageno, W., Dentali, F. & arthroplasty, without increasing bleeding.
venous thromboembolism: a systematic review. Agnelli, G. Incidence of arterial cardiovascular events 187. Ramacciotti, E. et al. P‑selectin/ PSGL‑1 inhibitors
J. Med. Econom. 14, 65–74 (2011). after venous thromboembolism: a systematic review versus enoxaparin in the resolution of venous
145. Lee, A. Y. et al. Low-molecular-weight heparin versus a and a meta-analysis. J. Thromb. Haemost. 8, 891–897 thrombosis: a meta-analysis. Thromb. Res. 125,
coumarin for the prevention of recurrent venous (2010). e138–e142 (2010).
thromboembolism in patients with cancer. N. Engl. 167. Spencer, F. A., Ginsberg, J. S., Chong, A. & Alter, D. A. 188. Washington University School of Medicine. The
J. Med. 349, 146–153 (2003). The relationship between unprovoked venous Attract Study [online], http://attract.wustl.edu/
146. Jonas, D. E. & McLeod, H. L. Genetic and clinical thromboembolism, age, and acute myocardial (2009).
factors relating to warfarin dosing. Trends Pharmacol. infarction. J. Thromb. Haemost. 6, 1507–1513 (2008). 189. Ay, C. et al. High concentrations of soluble P‑selectin
Sci. 30, 375–386 (2009). 168. Rectenwald, J. E. et al. Experimental pulmonary are associated with risk of venous thromboembolism
147. Pirmohamed, M. et al. A randomized trial of genotype- embolism: effects of the thrombus and attenuation of and the P‑selectin Thr715 variant. Clin. Chem. 53,
guided dosing of warfarin. N. Engl. J. Med. 369, pulmonary artery injury by low-molecular-weight 1235–1243 (2007).
2294–2303 (2013). heparin. J. Vasc. Surg. 43, 800–808 (2006). 190. Ay, C. et al. High plasma levels of soluble P‑selectin
148. Kimmel, S. E. et al. A pharmacogenetic versus a 169. Diaz, J. A. et al. Plasma DNA is elevated in patients are predictive of venous thromboembolism in cancer
clinical algorithm for warfarin dosing. N. Engl. J. Med. with deep vein thrombosis. J. Vasc. Surg. Venous patients: results from the Vienna Cancer and
369, 2283–2293 (2013). Lymphat. Disord. 1, 341–348.e1 (2013). Thrombosis Study (CATS). Blood 112, 2703–2708
149. Mearns, E. S. et al. Quality of vitamin K antagonist 170. Jimenez-Alcazar, M. et al. Impaired DNase1‑mediated (2008).
control and outcomes in atrial fibrillation patients: a degradation of neutrophil extracellular traps is 191. Bucciarelli, P. et al. Circulating microparticles and risk
meta-analysis and meta-regression. Thromb. J. 12, 14 associated with acute thrombotic microangiopathies. of venous thromboembolism. Thromb. Res. 129,
(2014). J. Thromb. Haemost. http://dx.doi.org/10.1111/ 591–597 (2012).
150. van Es, N., Coppens, M., Schulman, S., Middeldorp, S. jth.12796 (2014). 192. Ay, C., Freyssinet, J. M., Sailer, T., Vormittag, R. &
& Buller, H. R. Direct oral anticoagulants compared 171. Pabinger, I. & Posch, F. Flamethrowers: blood cells and Pabinger, I. Circulating procoagulant microparticles in
with vitamin K antagonists for acute venous cancer thrombosis risk. Hematology Am. Soc. Hematol. patients with venous thromboembolism. Thromb. Res.
thromboembolism: evidence from Phase 3 trials. Educ. Program 410–417 (2014). 123, 724–726 (2009).
Blood 124, 1968–1975 (2014). 172. Lu, G. et al. A specific antidote for reversal of 193. Huang, W., Goldberg, R. J., Anderson, F. A.,
151. Meyer, G. et al. Fibrinolysis for patients with anticoagulation by direct and indirect inhibitors of Kiefe, C. I. & Spencer, F. A. Secular trends in
intermediate-risk pulmonary embolism. N. Engl. coagulation factor Xa. Nat. Med. 19, 446–451 (2013). occurrence of acute venous thromboembolism: the
J. Med. 370, 1402–1411 (2014). 173. Honickel, M. et al. Reversal of dabigatran Worcester VTE study Am. J. Med. 127, 829–839.e5
152. Popuri, R. K. & Vedantham, S. The role of thrombolysis anticoagulation ex vivo: porcine study comparing (2014).
in the clinical management of DVT. Arterioscler. prothrombin complex concentrates and idarucizumab. 194. Clot Connect. Venous Thromboembolism Resource.
Thromb. Vasc. Biol. 31, 479–484 (2011). Thromb. Haemost. 113, 728–740 (2015). [online], http://www.clotconnect.org/ (2010).
153. Schulman, S. et al. Extended use of dabigatran, 174. Mackman, N. Triggers, targets and treatments for 195. Kerbauy, M. N., de Moraes, F. Y., Kerbauy, L. N.,
warfarin, or placebo in venous thromboembolism. thrombosis. Nature 451, 914–918 (2008). Conterno Lde, O. & El‑Fakhouri, S. Venous
N. Engl. J. Med. 368, 709–718 (2013). 175. Renne, T. et al. Defective thrombus formation in mice thromboprophylaxis in medical patients: an
154. Noble, S., Lewis, R., Whithers, J., Lewis, S. & lacking coagulation factor XII. J. Exp. Med. 202, application review. Rev. Assoc. Med. Bras. 59,
Bennett, P. Long-term psychological consequences of 271–281 (2005). 258–264 (in Portuguese) (2013).
symptomatic pulmonary embolism: a qualitative study. 176. Kleinschnitz, C. et al. Targeting coagulation factor XII 196. Anderson, F. A. Jr. et al. Changing clinical practice.
BMJ Open 4, e004561 (2014). provides protection from pathological thrombosis in Prospective study of the impact of continuing medical
155. Kahn, S. R. et al. Determinants of health-related cerebral ischemia without interfering with hemostasis. education and quality assurance programs on use of
quality of life during the 2 years following deep vein J. Exp. Med. 203, 513–518 (2006). prophylaxis for venous thromboembolism. Arch.
thrombosis. J. Thromb. Haemost. 6, 1105–1112 177. Howard, E. L., Becker, K. C., Rusconi, C. P. & Intern. Med. 154, 669–677 (1994).
(2008). Becker, R. C. Factor IXa inhibitors as novel
156. Ashrani, A. A. et al. Impact of venous anticoagulants. Arterioscler. Thromb. Vasc. Biol. 27, Acknowledgements
thromboembolism, venous stasis syndrome, venous 722–727 (2007). The authors thank T. Wakefield, P. Henke, D. Myers and
outflow obstruction and venous valvular incompetence 178. Schumacher, W. A., Luettgen, J. M., Quan, M. L. & R. Kasthuri for reading the manuscript, and A. Conrad for
on quality of life and activities of daily living: a nested Seiffert, D. A. Inhibition of factor XIa as a new approach assistance with Figure 1.
case–control study. Vasc. Med. 15, 387–397 (2010). to anticoagulation. Arterioscler. Thromb. Vasc. Biol. 30,
157. van Korlaar, I. et al. Quality of life in venous disease. 388–392 (2010). Author contributions
Thromb. Haemost. 90, 27–35 (2003). 179. Zhang, H. et al. Inhibition of the intrinsic coagulation Introduction (N.M. and A.S.W); Epidemiology (F.R.R.);
158. Kahn, S. R. et al. Prospective evaluation of health- pathway factor XI by antisense oligonucleotides: a novel Mechanisms/pathophysiology (N.M. and A.S.W.); Diagnosis,
related quality of life in patients with deep venous antithrombotic strategy with lowered bleeding risk. screening and prevention (I.H.J. and J.I.W.); Management
thrombosis. Arch. Intern. Med. 165, 1173–1178 Blood 116, 4684–4692 (2010). (G.A.); Quality of life (T.B.); Outlook (N.M. and A.S.W.); and
(2005). 180. Muller, F., Gailani, D. & Renne, T. Factor XI and XII as overview of the Primer (N.M.).
This study confirms the causal association between antithrombotic targets. Curr. Opin. Hematol. 18,
PTS and reduced quality of life. 349–355 (2011). Competing interests
159. Klok, F. A. et al. The post‑PE syndrome: a new concept 181. Povsic, T. J. et al. Use of the REG1 anticoagulation G.A. has received personal fees from Boehringer Ingelheim,
for chronic complications of pulmonary embolism. system in patients with acute coronary syndromes Bayer Healthcare, Daiichi Sankyo, Sanofi, Pfizer and
Blood Rev. 28, 221–226 (2014). undergoing percutaneous coronary intervention: results Bristol–Myers Squibb; all are outside the scope of the sub-
This comprehensive review introduces the concept from the phase II RADAR-PCI study. EuroIntervention. mitted work. T.B. has received honoraria for participating in
of post-pulmonary embolism syndrome as a 10, 431–438 (2014). scientific advisory boards for Boehringer Ingelheim and
common outcome, analogous to PTS after DVT. 182. Vavalle, J. P. et al. The effect of the REG2 Daiichi Sankyo, and has received an unconditional travel
160. Kahn, S. R. et al. Compression stockings to prevent anticoagulation system on thrombin generation grant from Boehringer Ingelheim. N.M. served as consult-
post-thrombotic syndrome: a randomised placebo- kinetics: a pharmacodynamic and pharmacokinetic ant to Merck and Bayer. F.R.R. is listed on several patents
controlled trial. Lancet 383, 880–888 (2014). first‑in‑human study. J. Thromb. Thrombolysis 38, for prothrombotic gene variants. J.I.W. served as a consult-
This placebo-controlled trial shows that early 275–284 (2014). ant to and received honoraria from Boehringer Ingelheim,
application of compression stockings after acute 183. Hagedorn, I. et al. Factor XIIa inhibitor recombinant Bristol–Myers Squibb, Pfizer, Johnson & Johnson, Daiichi
DVT fails to prevent PTS. human albumin infestin‑4 abolishes occlusive arterial Sankyo, ISIS Pharmaceuticals and Portola. A.S.W. served
161. Kahn, S. R. et al. Graduated compression stockings to thrombus formation without affecting bleeding. as a consultant to Merck. I.H.J. has no competing financial
treat acute leg pain associated with proximal DVT. Circulation 121, 1510–1517 (2010). interests to report.

10 1038@nrdp20156

Uploaded by

Copyright:

Available Formats

10 1038@nrdp20156

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10 1038@nrdp20156

Uploaded by

Copyright:

Available Formats

PRIMER

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 1

© 2015 Macmillan Publishers Limited. All rights reserved

Author addresses incidence of VTE than people of Asian descent, whereas

2 | 2015 | VOLUME 1 www.nature.com/nrdp

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 3

© 2015 Macmillan Publishers Limited. All rights reserved

New Traditional anticoagulants induced by hypoxia, inflammation and/or altered blood

4 | 2015 | VOLUME 1 www.nature.com/nrdp

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 5

© 2015 Macmillan Publishers Limited. All rights reserved

6 | 2015 | VOLUME 1 www.nature.com/nrdp

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 7

© 2015 Macmillan Publishers Limited. All rights reserved

Figure 5 | Diagnostic signs of venous thromboembolism. Nature a | In compression

8 | 2015 | VOLUME 1 www.nature.com/nrdp

© 2015 Macmillan Publishers Limited. All rights reserved

a Traditional regimen duration of treatment. Patients with VTE occurring in

Therapeutic window surgery or hormonal therapy, require shorter treatment

(INR). At this point, parenteral anticoagulation is stopped,

rivaroxaban or apixaban, which are oral factor Xa inhibi-

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 9

© 2015 Macmillan Publishers Limited. All rights reserved

Table 1 | Summary of Phase III clinical trials evaluating new oral anticoagulants

10 | 2015 | VOLUME 1 www.nature.com/nrdp

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 11

© 2015 Macmillan Publishers Limited. All rights reserved

Chronic thromboembolic found in one-third of patients159. However, one-half of

12 | 2015 | VOLUME 1 www.nature.com/nrdp

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 13

© 2015 Macmillan Publishers Limited. All rights reserved

14 | 2015 | VOLUME 1 www.nature.com/nrdp

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 15

© 2015 Macmillan Publishers Limited. All rights reserved

16 | 2015 | VOLUME 1 www.nature.com/nrdp

© 2015 Macmillan Publishers Limited. All rights reserved

NATURE REVIEWS | DISEASE PRIMERS VOLUME 1 | 2015 | 17

© 2015 Macmillan Publishers Limited. All rights reserved

You might also like