Venous thromboembolism (VTE), that consists of the interrelated diseases of pulmonary embolism and deep-vein thrombosis (DVT), is among the top five most common vascular diseases in most countries1. The lifetime risk of VTE is estimated to be 8% overall among US adults2. Approximately 20% of individuals die within 1 year of a VTE diagnosis, sometimes from VTE but often from conditions that provoked the event3.

VTE is preventable, and the risk of VTE is a marker of adverse conditions at the levels of the individual, society and health systems4,5. At the individual level, obesity6,7,8,9 and lack of physical exercise10,11 increase the risk of VTE. These factors are often caused by adverse societal conditions, such as a lack of access to high-quality nutrition or safe places to exercise12. At the health systems level, opportunities are missed to provide effective preventive measures during hospitalization or after selected interventions, or in individuals with a personal or family history of VTE. The pathophysiology of VTE is also distinct among vascular diseases because several established risk factors for atherosclerotic cardiovascular disease, such as hyperlipidaemia, hypertension and diabetes mellitus, are not associated with the risk of VTE independently of established VTE risk factors, such as age and obesity13.

The risk of VTE is multifactorial14 (Table 1 and Fig. 1). In epidemiological studies, VTE is often defined as provoked or unprovoked. Provoked events occur after triggering factors in the previous 3 months, such as immobilization, trauma, surgery, cancer or hospitalization, whereas unprovoked events occur in the absence of these conditions8,15,16,17. Although the classification of provoked and unprovoked can be useful both epidemiologically and clinically, it is controversial18. For instance, some studies might classify pregnancy and hormone-related VTE as provoked. Given that VTE is multifactorial, it can be challenging (or impossible) to identify a single characteristic that led to VTE or to consider as unprovoked an array of smaller thrombotic challenges that led to a VTE event. The 2019 European Society of Cardiology/European Respiratory Society guidelines for the diagnosis and management of acute pulmonary embolism avoided the terminology of ‘provoked’ and ‘unprovoked’, instead focusing on the estimated long-term risk of VTE recurrence (<3% per year being low risk, 3–8% per year being intermediate risk and >8% per year being high risk) on the basis of underlying factors19. However, given the predominance of the provoked and unprovoked terminology in the published literature, we use this terminology in this Review.

Table 1 Risk factors for venous thromboembolism
Fig. 1: Factors with protective or adverse effects on thrombosis potential and risk of VTE.
figure 1

The risk of venous thromboembolism (VTE) is multifactorial. Factors that are thought to be protective and associated with a lower thrombosis potential and VTE risk are shown in blue. Factors that are considered to be harmful and associated with greater thrombosis potential and VTE risk are shown in red. Blood lipid levels, hypertension and diabetes mellitus are not shown in the figure because they are not independently associated with the risk of VTE after accounting for obesity and other risk factors.

VTE is treated with anticoagulant therapy. VTE treatment is divided into three phases: initial stabilization (days to weeks), primary treatment (3–6 months) and secondary prevention (after completion of primary treatment)20. The designation of provoked versus unprovoked VTE also guides treatment duration, particularly whether or not to use secondary prevention strategies, which are recommended for individuals with unprovoked VTE and can include surveillance, anticoagulation therapy during high-risk periods or continued anticoagulation therapy21. However, the distinction between provoked and unprovoked is inconsistently applied in clinical practice and is often not reflective of management decisions. People with incurable cancer and VTE (a provoked event) are offered indefinite anticoagulation therapy, and women with hormone-associated VTE (an unprovoked event) are offered a finite course of treatment21. This inconsistency in VTE management is further confused because minor triggers, such as travel or minor orthopaedic injuries, are often thought of as provoking factors clinically but not in epidemiological studies. The lack of unified definitions and practice recommendations leads to heterogeneity in treatment patterns, prevention and, potentially, outcomes.

VTE is a major cause of morbidity and mortality. Understanding the burden of this condition, as well as its risk factors, is essential to prevent VTE and reduce disparities in VTE incidence and outcomes at the societal level. In this Review, we provide contemporary information on the epidemiology of VTE, including incidence, risk factors and complications. Additionally, we summarize strategies for the primordial, primary and secondary prevention of VTE, and highlight the importance of reducing disparities in VTE incidence and management.


Global perspective

The incidence of VTE in Europe and the USA is estimated to be ~1–2 per 1,000 person-years1, but varies widely by age, sex, race and medical conditions. In Asia, the rates of VTE are thought to be lower than in Europe and the USA. For instance, the incidence of VTE in South Korea was estimated to be 0.2 per 1,000 person-years22. Fewer data exist for South America23 and Oceania. A study from Buenos Aires, Argentina, found a VTE incidence of 0.7 per 1,000 person-years24, and a study from Perth, Australia, found a VTE incidence of 0.8 per 1,000 person-years25. Very little is known about VTE incidence in Africa26.

The most robust data on VTE incidence come from the USA and Europe. An American Heart Association report from 2021 estimated that approximately 1,220,000 total cases of VTE occur in the USA annually27. This estimate was based on previously unpublished data from the National Inpatient Sample, and showed ~370,000 cases of pulmonary embolism and ~857,000 cases of DVT in 2016 and assumed 30% of DVTs were treated in the outpatient setting. A modelling study estimated that the annual VTE incidence in six countries in Europe (total population 310.4 million) was 296,000 cases of pulmonary embolism and ~466,000 cases of DVT28.

Methodological considerations

Owing to the lack of surveillance for VTE at the population level, estimates of VTE incidence rely on incomplete data and on assumptions to complete the picture. Interpretation of most sources of data on VTE incidence should be viewed in light of secular trends and data characteristics that lead to higher29,30 or lower31,32,33 estimates of VTE incidence (Box 1). There is a clear need for population-based VTE surveillance, particularly in the developing world and in populations at risk of VTE. These data would guide VTE prevention, screening and management efforts.

Acute triggering risk factors

Surgery, bone fracture and minor injuries

Surgery and bone fracture are established risk factors for VTE. The risk of postoperative VTE varies widely by surgery type, and the best strategies to prevent postoperative VTE are unclear34. Risk factors for VTE after surgery include basal risk factors (such as older age and male sex), comorbid conditions (such as obesity, active cancer and malnutrition) and postoperative complications (such as pneumonia, blood transfusions and myocardial infarction)34. The epidemiology of VTE after bone fracture also reflects the diversity of patient populations and type of bone fracture studied, and varies widely. In general, the risk is thought to be higher with lower-leg long-bone fractures, trauma-related bone fractures and operatively treated bone fractures35.

Minor injuries are often an under-appreciated trigger for VTE. The relationship between minor injuries and VTE is difficult to study systematically owing to recall bias in retrospective studies and because most of these minor injuries do not result in medical care. In an analysis from the MEGA study36, minor injuries were shown to increase the risk of VTE by about fivefold, with a higher risk of VTE with more severe injuries, leg injuries and injuries in the previous 4 weeks. Genetic risk factors for thrombosis have been shown to increase the risk of VTE associated with minor injuries compared with control individuals (for example, a 50-fold increased risk for the factor V Leiden variant in the F5 gene, a ninefold increased risk for the prothrombin G20210A polymorphism and a 12-fold increased risk for family history of VTE), suggesting supra-additive interactions36.


Between 40% and 60% of all VTE events occur during or in the 3 months after a hospitalization, and hospitalization increases the risk of VTE by approximately 100-fold16,37,38. The risk of in-hospital VTE is believed to be a function of both underlying medical conditions and immobility, which often characterizes a hospitalization39,40,41. In-hospital VTE is an important indicator of overall health, with even asymptomatic VTE associated with a threefold increased risk of death among hospitalized, acutely ill patients42. Although VTE might not be the primary cause of death in many of these individuals, both VTE symptoms and treatment complicate patient care.

Heparin-induced thrombocytopenia

Heparin-induced thrombocytopenia (HIT) is an idiosyncratic immunothrombotic disorder resulting from exposure to heparins43. The aetiology of HIT is thought to be related to IgG antibodies, which bind to platelet factor 4 (PF4) complexed with heparins on the surface of platelets, leading to platelet activation and thrombosis43. Seroconversion (developing antibodies to the PF4–heparin complex) occurs more commonly than the clinical syndrome43, requiring both clinical context and serology to make a diagnosis of HIT. Up to 30–40% of people with confirmed HIT develop arterial and/or venous thrombosis44. Owing to the ubiquitous exposure of patients to heparin during hospitalization, HIT complicates 1 in 1,500 hospital admissions, but can also occur with outpatient heparin exposure or after discharge from hospital43. Heparin formulations such as low-molecular-weight heparins are associated with a lower risk of HIT than unfractionated heparin45.


Acute infection has also been identified as a trigger for VTE, including in two case–crossover studies that evaluated the risk of VTE after hospitalization for infection46,47. In one of these studies, which used data from the ARIC study, hospitalization for infection in the previous 30 days was associated with a 2.7-fold (95% CI 1.1-fold to 6.4-fold) greater risk of VTE relative to control periods of the previous 1 year and 2 years46. A self-controlled, case series study derived from the MEGA cohort used antibiotic treatment as a surrogate for infection. The risk of VTE was fivefold higher (95% CI 4.0-fold to 6.1-fold) during antibiotic treatment than during periods of no antibiotic use in the same patients48. In general, the available studies lack the granularity required to assign individualized risks to specific types of infection.

COVID-19 and COVID-19 vaccination

Early reports from Wuhan, China, suggested a high incidence of thrombosis associated with hospitalization for coronavirus disease 2019 (COVID-19)49. This observation led to a rapid escalation of anticoagulation therapy in people hospitalized for COVID-19 and eventually to clinical trials to assess the use of anticoagulant therapy in patients with mild (outpatient-treated), moderate (treated in general medical wards) or severe (treated in intensive care) COVID-19 as well as the use of anticoagulation prophylaxis in patients after hospital discharge50. However, over time, the role of therapeutically dosed anticoagulation drugs has evolved from the prevention of thrombosis in people with COVID-19 to the prevention of care escalation in those with moderate COVID-19 (ref.50). Treatment of documented VTE in patients with COVID-19 has received less study, but probably should be treated similarly to any provoked VTE event51.

The risk of thrombosis with adenovirus-based COVID-19 vaccination has raised concerns. The adenovirus vector-based vaccines ChAdOx1 nCoV-19 (AstraZeneca) and Ad26.COV2.S (Johnson & Johnson) are associated rarely with an immune-related syndrome of vaccine-associated thrombosis and thrombocytopenia (VITT)52. At the individual level, the risk of VITT is low, with the highest reported incidence of VITT with ChAdOx1 nCoV-19 vaccination (10 per 1 million vaccinated people)52. VITT resembles HIT; however, the currently hypothesized mechanism is that the general inflammatory reaction to the vaccine triggers the production of antibodies to components of the adenovirus vector for the vaccine and to platelet PF4 complexes53. The unavailability of alternative vaccines for billions of people in resource-poor settings combined with the proven efficacy of the adenovirus-based vaccinations means that strategies to diagnose and treat VITT in these settings must be identified.

Other COVID-19 vaccines that are based on mRNA technology (the COVID-19 vaccines from Pfizer–BioNTech and from Moderna) do not seem to trigger VITT54,55. Immune activation through viral infections or vaccinations can lead to inflammation, and inflammation is linked to an increased risk of VTE (see sections ‘Infection’ and ‘Inflammation’). If VTE is triggered by mRNA-based vaccines, the relative and absolute risks are minuscule, because studies with millions of people who received an mRNA-based COVID-19 vaccine have not detected consistent associations with VITT54,56.

Subacute triggering risk factors


High levels of biomarkers of chronic inflammation, such as serum C-reactive protein and serum albumin, have been consistently associated with a greater risk of VTE57,58,59. Genetic studies that identified coagulation-related gene variants linked to the risk of VTE have also highlighted a role of inflammatory pathways in VTE60. However, the most compelling data linking inflammation to the risk of VTE come from randomized, controlled, clinical trials of statins58. In addition to their anti-hyperlipidaemic properties, statins are also potent anti-inflammatory agents and might also improve the coagulation profile61. In the JUPITER trial62, rosuvastatin therapy reduced the risk of VTE compared with placebo (HR 0.57, 95% CI 0.37–0.86). Importantly, blood lipid levels are not associated with the risk of VTE13, suggesting that the role of statins in reducing the risk of VTE is via their anti-inflammatory effects.

Hormonal states

Oral contraceptive use, hormone replacement therapy, and pregnancy

The risk of VTE is elevated in women receiving oestrogen-based contraceptives, hormone replacement therapy or infertility treatment, as well as during pregnancy63. In a cohort study of ~1.2 million women from four national registries, women taking combined (oestrogen and progestogen) oral contraceptives or progestin-only oral contraceptives were at approximately sixfold and threefold increased risk, respectively, of VTE compared with women who did not take oral contraceptives64. Although the prevalence of exposure to oral contraceptives and the magnitude of the association with VTE is relatively large, owing to the overall low incidence of VTE in this age group, 2,000 women would need to shift from combined oral contraceptives to progestin-only contraceptives to prevent one VTE annually64. Oestrogen therapy intensity seems to be important, with the highest risk of VTE in those individuals receiving continuous oestrogen therapy and a lower risk with increasing break periods in the therapy cycle65. Exposure to high-intensity exogenous hormones in the context of fertility treatments (for example, to simulate egg development and in vitro fertilization) can lead to ovarian hyperstimulation syndrome. Up to 10% of women with severe ovarian hyperstimulation syndrome develop VTE66. Users of hormone replacement therapy have an increased risk of VTE, similar to that observed with oral contraceptives67; however, the increased absolute baseline risk of VTE in this older patient population might change the risk–benefit ratio of hormone replacement therapy in this setting67. Research in transgender women might also provide additional insights into the role of sex hormones in the risk of VTE68.

VTE complicates ~1.2 in 1,000 pregnancies69 and accounts for approximately 9% of pregnancy-associated deaths in the USA70. Pregnancy is a procoagulant state and is also associated with greatly elevated steroid hormone levels, mechanical venous obstruction and venous stasis, which increase the risk of VTE before delivery and during the puerperium71,72. In the postpartum period, the risk of VTE is highest during the first week after delivery (~0.9 per 1,000 deliveries)73. The risk of VTE drops to ~0.25 per 1,000 deliveries in the second week and continues to drop through to the 12th week. Furthermore, among individuals who are pregnant or postpartum, certain obstetric procedures and complications, such as caesarean delivery, pre-eclampsia, haemorrhage and postpartum infection, are thought to be associated with an increased risk of VTE (OR 1.3–6.4)73. Pregnancy-associated VTE differs by race/ethnicity in the USA, but whether this variation is due to differences in risk factors, genetics or access to health care has not been well studied74.


Endogenous testosterone levels are not associated with VTE risk75,76. However, exogenous testosterone therapy might increase the risk of VTE77,78,79. Hypothesized mechanisms include elevated haematocrit levels (which can increase blood viscosity), platelet accumulation and elevated thromboxane A2 production by platelets and increased concentrations in the blood, all of which can increase the risk of VTE77,78,79. The three largest studies on this topic have all relied on administrative databases. One study that used a case–crossover design determined that testosterone therapy is associated with a doubling in VTE risk80. Another study using a cohort design found a 70% greater risk with testosterone therapy81. A third study found no significant association82. Although the available data for VTE risk and testosterone therapy are not entirely conclusive, these findings, together with those suggesting that testosterone therapy is associated with increased risk of atherosclerotic disease83,84,85, suggest that caution should be used when prescribing exogenous testosterone. Research among transgender men might also provide novel insights into the role of testosterone in the risk of VTE68.

Basal risk factors

Demographic risk factors

Older age

The risk of VTE rises exponentially with age7,15,86. For example, in the LITE study15, the race-adjusted and sex-adjusted incidence of VTE per 1,000 person-years was 0.72 in those aged 40 to <55 years, 1.58 in those aged 55 to <65 years, 2.47 in those aged 65 to <75 years, 3.12 in those aged 75 to <85 years and 6.96 in those aged ≥85 years. In a more recent analysis, the hazard ratios of VTE per decade of age were 2.67 (95% CI 2.45–2.91) (in the Emerging Risk Factors Collaboration (n = 731,728)) and 1.81 (95% CI 1.71–1.92) (in the UK Biobank (n = 421,537))7. The greater risk of VTE at older ages might be partly a result of the higher prevalence in older individuals of VTE risk factors, such as obesity, cancer, hospitalization or other comorbid conditions. Coagulation factor concentrations also typically increase with age87,88,89,90, which might partly explain the increased risk of VTE in older adults.

Sex and gender

The risk of VTE according to sex varies by life stage. Women of child-bearing age are at increased risk of VTE, probably as a consequence of exposure to oral contraceptives and pregnancy91,92,93,94 (see section on ‘Hormonal states’). At older ages, cis-men are at a modestly greater risk of VTE than cis-women16,95,96. Some of this excess risk might be due to differences in the prevalence of underlying risk factors, particularly height95. VTE risk among transgender people is complex and might partially reflect gender-affirming hormone treatments and surgical procedures97.

Race and ethnicity

The association between race and ethnicity and the risk of VTE is complex98,99 (Fig. 2). In the USA, the incidence of VTE has been shown to differ by race and ethnicity100,101,102. Black American individuals seem to have the highest risk of VTE, followed by white, Latina/Latino and Asian American individuals, respectively100,101,102 (based on US definitions of race and ethnicity99). However, VTE rates are similar in Black American and white American individuals after accounting for differences in the prevalence of underlying risk factors100. Geography also has a role in both VTE incidence and pulmonary embolism mortality in the USA, with the highest incidence in south-eastern USA103,104. Worldwide, differences in the prevalence of risk factors (such as obesity), genetics, overall health and access to health care are all likely to influence the risk of VTE in each region. However, available data are mostly from North America and Europe, with few or no data from Africa, Asia, Australia and South America4,105. Cross-national comparisons are difficult owing to differences in (or the lack of) disease surveillance systems and differences in health-care practices. Disentangling the effects of race and geography that are due to social (such as structural racism) and genetic constructs in epidemiological studies might inform efforts to reduce disparities in VTE incidence. There is also a need to include individuals from diverse populations in clinical trials. Figure 2 shows potential reasons, at least in the context of the USA, for the racial and ethnic differences and disparities in VTE incidence98.

Fig. 2: Racial/ethnic differences in VTE incidence and potential underlying factors.
figure 2

The graph shows the relative risk of venous thrombosis according to racial/ethnic categories. The relative risk is based on a review of the literature on absolute risk of venous thromboembolism (VTE) in various cohort studies (see ref.98 for more details). The figure shows the factors that might lead to a higher or lower risk of VTE in each racial/ethnic group. Adapted with permission from ref.98.


Physical inactivity

Numerous observational studies have found that greater physical activity is associated with lower VTE risk, as has been thoroughly reviewed previously10,11. In a 2020 meta-analysis of 14 prospective cohorts, the relative risk of VTE in individuals with high physical activity was 0.87 (95% CI 0.79–0.95) compared with those with low physical activity11. Greater cardiovascular fitness, objectively measured using a cycle ergonometric test, has also been associated with lower risk of VTE in a study in Swedish men106. Mechanisms linking greater physical activity to lower VTE risk include improved endothelial function, lower concentrations of coagulation factors, lower BMI and increased blood flow11. Conversely, low physical activity through a sedentary lifestyle, habitual sitting at work and more time spent watching television have all been linked to a greater risk of VTE107,108. However, results have not been entirely consistent109. The population-attributable risk of VTE owing to physical inactivity might be as high as 4%110.


Research assessing diet as a risk factor for VTE has generally yielded neutral, inconsistent or modest findings111. This lack of strong evidence might reflect the difficulty in assessing diet as well as the uncertainty about what specific elements of the diet influence the risk of VTE. Diet might modify the risk of VTE through intake of specific nutrients, altering levels of haemostatic and fibrinolytic biomarkers, and/or through being associated with obesity. The strongest evidence linking diet to VTE comes from the Women’s Health Initiative, in which women who were randomly assigned to receive vitamin E supplementation (600 IU) were at 21% lower risk of VTE than women receiving placebo112. Among observational studies, some, but not all, have found a lower risk of VTE associated with the intake of fruit and vegetables, low or moderate quantities of alcohol, vitamin E, vitamin B6 and fibre, and a higher risk of VTE with greater intake of a ‘Western’ diet, red and processed meat, and trans fatty acids111,113,114,115,116,117,118. Given the inconsistent and neutral findings, public health messages are unlikely to promote consumption of a healthy diet for VTE prevention. However, dietary patterns that might be beneficial for reducing the risk of VTE are also advantageous for more common conditions, such as coronary heart disease and cancer, and are therefore encouraged in current dietary guidelines.


Tobacco smoking has been postulated to increase the risk of VTE, given the potential for smoking to increase coagulability and chronic inflammation119,120. However, data evaluating smoking and VTE incidence are less convincing. Smoking was associated with provoked but not with unprovoked VTE events in a large, individual participant data meta-analysis that included validated VTE events13. Specifically, fully adjusted hazard ratios for the association between current smoking and provoked or unprovoked VTE events were 1.36 (95% CI 1.22–1.52) and 1.08 (95% CI 0.90–1.29), respectively13, with similar findings in other studies8,121. The stronger association between smoking and provoked VTE might be explained by the well-known association between smoking and numerous health conditions (such as cancer, respiratory illnesses, myocardial infarction and stroke), which can lead to hospitalization and/or immobilization.

Anthropometric risk factors


Taller height is associated with greater risk of VTE8,9,122,123,124. In general, individuals in the highest height quantile are at 1.5–2.0 times greater risk of VTE than those in the lowest height quantile8,9,122,123,124. Possible explanations for this association include taller people having a greater venous surface area, a higher number of venous valves and greater hydrostatic pressure125,126. In one analysis, which included leg length measures, leg length was a VTE risk factor independent of total body height, but height was no longer associated with VTE risk after accounting for leg length122. This finding supports the hypothesis that leg length and height are associated with the risk of VTE through blood stasis or changes to the vascular wall rather than through hypercoagulability. A Mendelian randomization analysis that used data from the ARIC, CHS and the Mayo Clinic VTE studies found that single-nucleotide polymorphisms associated with greater height were associated with elevated VTE risk127, strongly suggesting a causal relationship. A Swedish study that used a co-sibling design, specifically full sibling pairs discordant for height, also suggested a causal association128.


High BMI has been consistently associated with elevated risk of VTE, in a dose–response manner6,7,8,9. For example, in the LITE study, the hazard ratio for VTE was 1.58 (95% CI 1.20–2.09) for BMI 25 to ≤30 kg/m2, 2.10 (95% CI 1.57–2.82) for BMI 30 to ≤35 kg/m2 and 3.09 (95% CI 2.26–4.23) for BMI ≥35 kg/m2 compared with normal weight (BMI ≤25 kg/m2)6. Weight gain has also been associated with excess VTE risk129,130. A 2020 analysis estimated that the population-attributable risk of VTE associated with obesity in the USA was 30%110. An important additional assumption of population-attributable risk calculations is that the association is causal. Physiologically, obesity might increase the risk of VTE through venous stasis, elevated concentrations of haemostatic and inflammatory biomarkers, and increased risk of other conditions known to predispose to VTE (such as cancer)110. Strong evidence of causality was derived from two Mendelian randomization studies that demonstrated that single-nucleotide polymorphisms associated with increased BMI were linked to a higher risk of VTE131,132. Large-scale, randomized, controlled trials have not yet tested whether prevention of weight gain or interventions to lose weight reduce the risk of VTE.

May–Thurner syndrome

Patients with May–Thurner syndrome have an elevated risk of developing an extensive left iliofemoral DVT (localized blood clot) owing to an anatomical variant in which the right common iliac artery compresses the left common iliac vein against the lumbar spine133. Determining the population burden of May–Thurner syndrome is challenging because of the specialized imaging techniques required for its diagnosis134. However, the extremely limited data available suggest that this syndrome might account for 2–5% of DVT cases in the USA and Europe134. Given the complexity of diagnosing May–Thurner syndrome, the relative and absolute risks of VTE in individuals with this anatomical variant are unknown.

Genetic risk factors

Heritability and family history

Studies have estimated that the heritability of VTE is approximately 23%60 to 30%135. These heritability estimates are lower than those that were based on family studies (40–60%)136,137. In a multi-generation study of Swedish registry data, the familial odds ratio for VTE was 2.49 (95% CI 2.40–2.58) in siblings, 2.65 (95% CI 2.50–2.80) in children, 2.09 (95% CI 2.03–2.15) in parents, 1.52 (95% CI 1.26–1.85) in maternal half-siblings, 2.34 (95% CI 2.00–2.73) in paternal half-siblings, 1.69 (95% CI 1.57–1.82) in nieces or nephews, 1.47 (95% CI 1.33–1.64) in cousins and 1.14 (95% CI 1.09–1.18) in spouses of individuals diagnosed with VTE137.

Genetic risk factors

Numerous genetic risk factors have been identified that predispose individuals to VTE, as has been extensively reviewed previously136,138,139. A few of the more notable established genetic risk factors are highlighted below. Importantly, genetic discovery work continues60,135.

The factor V Leiden variant is the cause of the most common monogenetic inherited thrombophilia among populations of European origin, but is rare in non-European populations140. In a pooled analysis, the risk of VTE was increased fourfold in individuals who were heterozygous for the factor V Leiden variant and increased 11-fold in those who were homozygous for the factor V Leiden variant compared with non-carriers141. The lifetime risk of VTE in those with the factor V Leiden variant is ~17%2. Likewise, the prothrombin F2 gene variant G20210A is often cited as a risk factor for VTE and is associated with a twofold to threefold relative increase in the risk of VTE142.

Sickle cell disease and trait are genetic risk factors for VTE2,143,144,145. In the USA, 0.2% of Black American individuals have sickle cell disease and 8% of Black American individuals have sickle cell trait146. Whereas sickle cell disease is characterized by substantial morbidity, individuals with sickle cell trait are often unaware of their status147. Both conditions are also associated with procoagulant abnormalities148,149. Several studies have demonstrated that sickle cell trait is a moderately strong risk factor for VTE2,143,144,145, with 18% of individuals with sickle cell trait developing VTE in their lifetime2.

More common genetic variants associated with VTE tend to be associated with a lower magnitude of risk than rare mutations, but might have a substantial effect on the population-level risk of VTE. Some examples include non-O blood group and variants in the F2, F5, F11, FGG or ZFPM2 genes150.

Genetic risk scores

Genetic risk scores have been created to predict the risk of VTE. In an anlysis of data from the LITE study, an a priori weighted score was created based on five established VTE-associated single-nucleotide polymorphisms in the ABO, F2, F5, F11 and FGG genes151. When this score was used, the risk of VTE in white individuals increased 1.41-fold (95% CI 1.27–1.56) per allele increment. Another study calculated a genetic risk score based on 37 VTE susceptibility gene variants, which showed a robust dose–response association with the risk of VTE60. Compared with individuals in the 25th to 75th percentiles of the genetic risk score, the odds ratios for VTE were 0.51 (95% CI 0.43–0.61) in individuals at or below the 5th percentile and 3.19 (95% CI 2.89–3.52) in individuals at or above the 95th percentile60. Genetic risk scores are not currently used for VTE risk assessment, but given the familial studies suggesting heritability of VTE risk, there is potential for the use of genetic risk scores to help refine an individual’s VTE risk. A huge barrier to implementing genetic risk scores in practice are the limitations in understanding genetic VTE risk in populations of non-European origin as well as how to use them in racially mixed populations.

Acquired risk factors


Cancer is the quintessential risk factor for VTE152. After the cancer itself, thrombosis is the next leading cause of death in people with cancer, and complicates the care of people with cancer153. The mechanism of thrombosis in patients with cancer relates to the cancer itself as well as to cancer treatments and complications of cancer treatment, as has been reviewed previously153.

The annual incidence of VTE in patients with cancer ranges from 0.5% to 20% depending on the cancer type and cancer treatments offered153,154. Risk factors, such as cancer stage, certain cancer treatments, central catheters and surgical procedures, all interact to influence the risk of VTE153,154. Specific thrombosis risk scores exist for patients with cancer in general155 as well as for patients with specific cancer types undergoing selected treatments156,157. None of these risk scores incorporate the temporal nature of many risk factors that patients with cancer encounter, such as hospitalization and surgical procedures, which have been shown to be important for VTE risk158.

Prevention of thrombosis in patients with cancer is under active investigation, and in some individuals with cancer primary prophylaxis might be warranted. Professional medical societies frequently publish guidelines for the management and prevention of thrombosis in patients with cancer and, as we improve our understanding of the risk of VTE in these patients, the indications for VTE prevention will evolve159.

Autoimmune diseases

Antiphospholipid syndrome (APS), and the related entity catastrophic APS, is an autoimmune disorder that predisposes individuals to thrombosis and is the quintessential thrombotic autoimmune disease160. The diagnosis of this clinicopathological entity requires both the appropriate clinical setting, defined by unexplained pregnancy loss or thrombosis, as well as pathological confirmation of the presence of persistent antiphospholipid antibodies (as measured, for example, by anti-cardiolipin antibodies) and/or lupus anticoagulants (as measured, for example, by the dilute Russell’s viper venom time)161. In APS, autoantibodies disrupt phospholipids protecting the blood vessel wall, leading to thrombosis161. The incidence of APS in the general population is approximately 2 per 100,000 person-years162. However, the presence of antiphospholipid antibodies is common in the general population, ranging from 1–5% in healthy young populations to 50% in older individuals with multiple comorbidities163. In VTE-free individuals with certain autoimmune diseases such as systemic lupus erythematosus, antiphospholipid antibodies and/or lupus anticoagulants might increase the risk of thrombosis by up to threefold161,162. About 1% of people with APS develop catastrophic APS164. Catastrophic APS can develop in individuals with APS who have uncontrolled activation of the complement system, leading to catastrophic thrombosis, often despite the use of anticoagulation therapy160.

Numerous other autoimmune diseases have been linked to an increased risk of VTE165. For example, a Swedish nationwide study found that the risk of pulmonary embolism was elevated in the first year after hospitalization for 33 different autoimmune disorders166. Among the most studied are rheumatoid arthritis, systemic lupus erythematosus and inflammatory bowel disease165,166,167,168,169. Some studies have estimated that patients with systemic lupus erythematosus have up to a 26% risk of having a thrombosis event throughout their disease course165,169. These autoimmune diseases are thought to increase VTE risk by upregulating the body’s procoagulant factors while simultaneously decreasing the levels of physiological anticoagulants via systemic inflammation165. Other potential causes include increased VTE risk associated with some treatments for autoimmune diseases as well as a higher prevalence of antiphospholipid antibodies leading to increased incidence of APS in this patient population165,170.

Hyperlipidaemia, hypertension and diabetes

The role of traditional risk factors for cardiovascular disease, such as hypertension, hyperlipidaemia and diabetes, in the risk of VTE has long been debated. In crude or minimally adjusted analyses, associations are generally present13. However, after demographic factors and BMI are accounted for, hypertension, hyperlipidaemia and diabetes are not associated with the risk of VTE13. Arguably, the best data evaluating these associations come from a large, individual-level meta-analysis that included nine prospective cohorts including 244,865 participants with a total of 4,910 VTE events occurring during a mean follow-up of 4.7–19.7 years13. Hazard ratios adjusted for age, sex and BMI for overall VTE were 0.98 (95% CI 0.89–1.07) for hypertension, 0.97 (95% CI 0.88–1.08) for hyperlipidaemia and 1.01 (95% CI 0.89–1.15) for diabetes13. These neutral findings are also supported by a 2021 study showing that a polygenic risk score for coronary artery disease is not related to venous thrombosis171. In summary, convincing evidence indicates that hypertension, hyperlipidaemia and diabetes are not VTE risk factors independently of age, sex and obesity.

Other clinical conditions

Numerous other clinical conditions and risk factors have also been associated with an increased risk of VTE. Although some of the increase in VTE risk is likely to be due to confounding comorbidities, immobility and adverse environmental conditions (such as air travel)172, evidence suggests the involvement of specific pathological mechanisms underlying conditions such as kidney disease173,174, heart failure175,176 and atrial fibrillation177,178.

Previous VTE

Although treatment and secondary prevention of VTE is extraordinarily effective, previous VTE is one of the strongest risk factors for a VTE event179. Even superficial thrombophlebitis has been shown to be a risk factor for recurrent VTE180.

Primary treatment failure occurs within 3 months of the initial VTE diagnosis20. In early studies on VTE treatments, radiographic progression of VTE based on various criteria is common with treatment (up to 15%), although usually asymptomatic181. In clinical trials comparing direct oral anticoagulant treatment with other anticoagulant strategies, the rate of recurrent symptomatic VTE was ≤3.5% with either treatment strategy182,183,184,185,186,187. Establishing risk factors for primary treatment failure is challenging owing to the differing definitions of treatment failure. In a secondary analysis of clinical trial data, risk factors for the failure of primary treatment with parental heparin bridged to an oral vitamin K antagonist were cancer, chronic cardiovascular disease, chronic respiratory disease, concurrent diseases (renal, hepatic, gastrointestinal, neurological, haematological or multisystem diseases) and younger age188. After 3–6 months of primary treatment, individuals should be risk-stratified for secondary prevention of VTE with anticoagulation therapy or observation. In the absence of long-term anticoagulation treatment, approximately 30% of patients have VTE recurrence within the next 10 years8,189,190. Predictors of VTE recurrence within 180 days include active cancer and inadequate initial anticoagulation treatment191.



Mortality from VTE is higher after pulmonary embolism events than after DVT events3,192. In a study using data from ~100,000 patients from 26 countries who were included in the RIETE registry, 30-day mortality from VTE was 2.6% for distal DVT, 3.3% for proximal DVT and 5.2% for pulmonary embolism193. The 30-day mortality was higher for DVT (5.1%)194 and pulmonary embolism (9.1%)195 among US Medicare beneficiaries aged ≥65 years, which might be attributable to the older age of this patient population. The same researchers found that the 1-year mortality for DVT was 19.6%194 and the 6-month mortality for pulmonary embolism was also 19.6%195. The 1-year mortality for pulmonary embolism was not reported.

After VTE, 1-year survival varies greatly by underlying VTE cause. In an analysis using administrative data for first-time VTE in Quebec, Canada, the 1-year survival rate for VTE was 77% overall, but when stratified by VTE-provoking status, it was 47% for cancer-associated VTE, 84% for provoked VTE and 93% for unprovoked VTE196. Long-term mortality remains elevated after an unprovoked VTE event, as shown by the MEGA study197.

Additional research is needed to understand the burden of VTE associated with specific health conditions as well as the need for and optimal approaches to VTE prophylaxis. For example, as noted in the section ‘Hormonal states’, VTE accounts for ~9% of pregnancy-associated deaths in the USA70. Additional efforts are need to identify and prevent VTE in the peripartum period.

Even when not clinically recognized, VTE might be an important marker of an increased risk of death. Among patients with acute medical illness, asymptomatic DVT events diagnosed with compression ultrasonography were associated with a threefold increased risk of short-term all-cause death compared with those with no evidence of DVT198.



Bleeding is a complication of every anticoagulant therapy. Historically, the incidence of bleeding after diagnosis of VTE has been shown to be >10%, and early bleeding is associated with a greater risk of VTE recurrence and death199,200. Data from phase III, randomized, controlled trials suggest that the use of direct oral anticoagulants, instead of warfarin, for the primary treatment of VTE could reduce the risk of bleeding201,202, but patient characteristics might influence the magnitude of the benefit203,204. The risk of bleeding drives many of the recommendations for the use of anticoagulation therapy for the secondary prevention of VTE20,21.

Post-thrombotic syndrome

Post-thrombotic syndrome refers to symptoms and signs of chronic venous insufficiency that develop after DVT205. Post-thrombotic syndrome is a common complication of DVT that develops in 20–50% of patients after proximal DVT and is severe in 5–10% of patients180. Moreover, approximately 4% of patients with DVT develop venous stasis ulcers206.

Post-thrombotic syndrome has specific risk factors that can be identified at the time of or after DVT: recurrent ipsilateral DVT, obesity, chronic kidney disease, more extensive DVT, poor quality of initial anticoagulation therapy, ongoing symptoms or signs of DVT 1 month after diagnosis, and elevated plasma D-dimer levels at 1 month180,207. Additionally, in a meta-analysis of patients with DVT who underwent ultrasonography at least 6 weeks after their DVT event, the ultrasonography parameters of residual vein thrombosis and venous reflux at the popliteal level were both predictive of post-thrombotic syndrome27,208. Methods for the primary prevention of post-thrombotic syndrome are uncertain209.

Pulmonary hypertension

Survivors of VTE are at an elevated risk of pulmonary hypertension, which is also referred to as chronic thromboembolic pulmonary hypertension210,211. Pulmonary hypertension can occur when pulmonary emboli or thrombi do not resolve and instead obstruct major pulmonary arteries, leading to increased pulmonary artery pressures, right-sided heart remodelling and pulmonary hypertension210.

The true incidence of pulmonary hypertension after VTE is unknown211,212, although some of the strongest studies suggest an incidence of 1–4% in the 2 years after the pulmonary embolism event212,213,214. Few studies have evaluated risk factors for pulmonary hypertension in patients with VTE. However, in the UK Clinical Practice Research Datalink analysis, chronic obstructive pulmonary disease, atrial fibrillation, heart failure and splenectomy were associated with a greater risk of pulmonary hypertension212. In a USA-based study using MarketScan administrative claims data, a higher Charlson Comorbidity Index score was associated with greater risk of pulmonary hypertension, and the individual comorbidities associated with the strongest elevation in pulmonary hypertension risk were concomitant heart failure, chronic pulmonary disease and alcohol abuse214.

Post-VTE quality of life and physical function

Quality of life and physical function after VTE are also crucial considerations215. VTE is often viewed as an acute condition, but evidence indicates that survivors of VTE are more likely to be granted permanent work-related disability216, experience a decline in physical function in the years subsequent to their index event217, and be frail and have poor physical function long after their initial VTE event compared with similarly aged individuals who did not have VTE218. Individuals who have VTE might be a group with a high risk of functional decline who should receive screening and potentially interventions to preserve physical function.


VTE is divided epidemiologically and clinically into provoked and unprovoked events20. As described in the Introduction, the terminology is confusing and used inconsistently. Provoked events are those associated with conditions such as cancer, surgery, hospitalization and/or immobility, whereas unprovoked events are those associated with no identifiable trigger but include pregnancy and hormone-associated VTE. Approximately half of all VTE events are associated with a defined trigger and are, therefore, potentially preventable through interventions of finite duration27. Unprovoked VTE can theoretically be prevented by healthy lifestyle choices110,219, and even by off-target effects of cardiovascular disease therapies such as statins58. However, this strategy might lack a holistic and patient-centred approach to VTE prevention. One alternative approach is to view VTE risk as a temporally layered set of risk factors. Some risk factors are constant (such as genetics), whereas others increase over time (such as age), with subacute (obesity) and acute (hospitalization) events layering on the risk until a predefined individual thrombosis threshold is crossed (Fig. 1). In this section, we discuss prevention in terms of primordial, primary and secondary prevention.

Primordial prevention

Primordial prevention refers to preventing disease in otherwise healthy people by preventing the development of risk factors for the disease220,221. Not all risk factors are amenable to primordial prevention (such as age, sex and genetics). However, interventions that improve health have the potential to reduce the risk of VTE, although they might not have been studied specifically for VTE prevention220,221. Ideal levels of physical activity and dietary interventions (see relevant sections) can result in less obesity and fewer hospitalizations, which could reduce the risk of VTE222. Preventing individuals from starting to smoke could reduce the risk of VTE by reducing the incidence of and hospitalizations for respiratory diseases, cardiovascular diseases and certain cancers (see section ‘Smoking’)13,221. Equitable lifestyle interventions (by age, sex, race and socioeconomic status) to improve diet and physical activity and prevent smoking in the general population improve health without having to add VTE risk reduction into the equation223. Even for static risk factors, such as genetic factors, primordial prevention matters. Healthy lifestyle choices can attenuate the increased risk of VTE associated with adverse genetic variants224 (Fig. 3). Other opportunities for primordial prevention include a focus on gynaecological and obstetric health in younger women by selecting lower risk methods of contraception, such as cyclic oral contraceptives and intrauterine devices (see section ‘Hormonal states’).

Fig. 3: Cumulative incidence of VTE according to genetic risk score and the AHA Life’s Simple 7.
figure 3

Cumulative incidence of venous thromboembolism (VTE) per 1,000 person-years (95% CI) according to VTE genetic risk score category. The genetic risk score categories are classified according to American Heart Association (AHA) Life’s Simple 7 categories219. The cumulative incidence of VTE is greater among individuals with high VTE genetic risk scores. In each genetic risk score group, lifestyle is also associated with the risk of VTE. Specifically, regardless of genetic predisposition, individuals with optimal Life’s Simple 7 scores have the lowest risk of VTE, those with average Life’s Simple 7 scores have an intermediate risk and those with inadequate Life’s Simple 7 scores have the greatest risk of VTE. The graph is a re-analysis of data from ref.224.

Primary prevention

Primary prevention of VTE reduces VTE by specific interventions in those individuals at risk of VTE. The most common intervention is pharmacological prophylaxis with low doses of anticoagulation drugs. The two most discussed opportunities for primary prevention of VTE relate to hospitalization and cancer. Given the large number of VTE events attributable to hospitalization and the finite risk period, hospital-associated VTE risk has become a key target of VTE prevention measures to reduce health-care-associated complications225,226,227,228. Whereas guidelines exist for the prevention of VTE during the hospital stay, there is limited clarity on how long prophylaxis should be continued after discharge, except for certain orthopaedic procedures for which guidelines are well established60. A similar consensus on primary prevention of VTE is evolving for patients with cancer, although the best strategy to prevent VTE in these patients is not known because the risk of VTE varies over time159,229. Primary prevention with anticoagulation therapy increases the risk of bleeding, and the risk–benefit balance is not established in many populations at risk of VTE228.

Another intriguing pharmacological intervention for the primary prevention of VTE is the use of statins. The JUPITER trial62, which recruited individuals with elevated plasma C-reactive protein levels and normal blood lipid levels, demonstrated that an off-target benefit for rosuvastatin therapy was a reduction in VTE risk. Whether statins can be used in individuals with increased risk of VTE to reduce VTE events has not been tested.

Secondary prevention

Secondary prevention refers to the prevention of recurrent VTE events in individuals with a history of VTE. Traditionally, the strategy for secondary prevention of VTE is anticoagulation therapy20. This approach reflects VTE being considered in clinical medicine as an event rather than a lifelong disease. Of note, in clinical practice, one of the strongest risk factors for VTE is a history of VTE. Even individuals with provoked VTE events often have multiple VTE recurrences throughout their life.

Although imperfect guidelines exist for who should be offered secondary VTE prevention with anticoagulation therapy, even, and perhaps especially, individuals who are not receiving anticoagulation therapy for secondary prevention are at risk of VTE recurrence with further provoking events such as pregnancy, hormone therapy, bone fracture, hospitalization and surgery20. In women, VTE provoked by exogenous hormone therapy or pregnancy warrants secondary prevention with anticoagulation therapy with subsequent pregnancy or hormone use. For hospitalized people, previous VTE is one of the most consistent risk factors for hospital-acquired VTE. The bottom line is that although the decision to offer anticoagulation therapy for secondary prevention of VTE is made once in the 3–6 months after the initial event, decisions on how best to prevent recurrent VTE during periods of VTE risk should be made throughout the life course. Although there are intriguing findings for non-antithrombotic agents for the secondary prevention of VTE, none are currently recommended for use in clinical practice61.


Improving awareness of VTE is another important avenue to VTE prevention230. Awareness of the disease is poor in the general population, and symptoms are often attributed to other disorders, thereby leading to a delay in diagnosis231. Unfortunately, even among individuals with a high risk of potentially provoking conditions, awareness of VTE is low. In a survey of patients with cancer, <20% of patients could name the symptoms of VTE and only 3% knew that cancer treatments are associated with a greater risk of VTE232. Continued education among medical professionals could also increase VTE awareness and implementation of guideline-based care230.

Reducing disparities in VTE burden

As noted above, VTE incidence and outcomes vary according to race and ethnicity. As with so many other health conditions, factors underlying these disparities in VTE incidence and outcomes include structural racism, differences in socioeconomic status and disparities in access to health care. Lower individual105,233 and neighbourhood4,233 socioeconomic status have been linked to greater VTE risk, even in regions with a universal health-care model (such as Denmark, Netherlands and Sweden). Although men in the general population have a higher risk of VTE recurrence than women, this risk differs by race, with Black and Latina women having a higher recurrence risk of VTE than white women in the USA234. For primary prevention of VTE, a US hospital-based project showed that care standardization through implementation of a mandatory computerized clinical decision support tool eliminated disparities in the use of VTE anticoagulation prophylaxis235. Also in the USA, Black individuals with VTE were fivefold more likely to experience complications of care after hospitalization for a VTE, including readmission, bleeding and death, than white individuals236,237. Reducing disparities in VTE incidence and outcomes will require individual, systems-based and societal commitments to equity.


VTE is an underappreciated source of morbidity and mortality worldwide. In the developed Western world, an estimated 8% of people will develop VTE in their lifetime2. However, global estimates of VTE incidence and burden are limited owing to the lack of VTE surveillance systems. VTE risk factors are multifactorial (Fig. 1), including acute triggers, subacute triggers and basal and acquired risk factors. Although the breadth of VTE risk factors make it difficult to determine who will develop VTE, it also yields many pathways amenable to targeting for VTE prevention. As with most conditions, the greatest potential to reduce the societal burden of VTE would be through primordial prevention to prevent the development of risk factors for VTE (such as obesity and common acute triggers). For individuals at risk of VTE or who have previous VTE, primary and secondary prevention are essential to reducing the burden of VTE. Differences related to race, ethnicity and socioeconomic status exist in VTE incidence and outcomes, and approaches to VTE prevention that improve health equity should be prioritized.