Neil A. Zakai, Departments of Medicine and Pathology, College of Medicine, University of Vermont, Colchester Research Facility, 208 South Park Drive, Colchester, VT 05446, USA. Tel.: +1 802 656 8968; fax: +1 802 656 8965. E-mail: Neil.email@example.com
Summary. The incidence of venous thrombosis (VTE) varies by race, with African-Americans having over 5-fold greater incidence than Asian-ancestry populations, and an intermediate risk for European and Hispanic populations. Known racial differences in genetic polymorphisms associated with thrombosis do not account for this gradient of risk, nor do known racial variations in environmental risk factors. Data on the incidence of and risk factors for VTE outside of Europe and North America and in non-European ancestry populations are sparse. Common genetic polymorphisms in European-Ancestry populations, such as factor V Leiden and prothrombin G20210A, and environmental risk factors, such as obesity, may account for some of the increased risk in European populations, and high factor VIII, high von Willebrand factor and low protein C levels and increased prevalence of obesity may explain some of the increased risk in African-Americans. The low rates in Asian populations may be partially explained by low clinical suspicion in a perceived low-risk population and lack of access to healthcare in other populations. As risk factors for thrombosis, such as surgery and treatment for cancer, are applicable to more people, as obesity increases in prevalence in the developing world, and as surveillance systems for VTE improve, VTE may increase in previously low-risk populations. While differences in VTE by race due to genetic predisposition will probably always be present, understanding the reasons for racial differences in VTE will help providers develop strategies to minimize VTE in all populations.
Venous thromboembolism (VTE) consisting of deep venous thrombosis (DVT) and pulmonary embolism (PE) is a common, potentially preventable, and often missed cardiovascular disease. The incidence varies by race, with individuals of African and European ancestry having the highest risk and individuals of other ethnicities thought to have a lower risk [1,2]. VTE results from a complex interplay between genetic and environmental risk factors, which are poorly studied in non-European populations and outside of North America and Europe [3,4]. While racial differences in the incidence of VTE have been appreciated for many years and both genetic and environmental VTE risk factors are known to vary by race, the impact these differences have on the incidence of VTE is unknown [3,5].
Genetics is thought to account for up to 60% of the risk of VTE, with the two known clinically evaluated single nucleotide polymorphisms associated with VTE risk (factor V Leiden and prothrombin gene polymorphism) found predominantly in European ancestry populations [3,6–9]. Abnormalities in coagulation such as deficiencies of protein C, protein S and antithrombin, and elevated factor (F) VIII and elevated FXI are at least partially heritable (heritability between 30% and 60%) and widely found in diverse populations but their role in the differential incidence of VTE by race is unknown [3,7,10–17]. Environmental risk factors for VTE such as hospitalization, surgery, cancer and obesity vary by race and geography, again with unknown consequences for the incidence of VTE [18–21].
In this manuscript, we will review (i) the epidemiology of VTE in diverse racial groups and (ii) the genetic and environmental risk factors for VTE and how they may contribute to the differences in VTE by race. To accomplish these goals, we performed a systematic review of the literature using OVID to discern all epidemiologic manuscripts with subject headings of ‘venous thrombosis’, ‘deep vein thrombosis’ or ‘pulmonary embolism’. By understanding why there are differences and who is at most risk for VTE, we can begin to develop better and more targeted preventive measures to help reduce VTE in everyone.
Epidemiology of race and VTE
Population-based estimates are rare outside of North America, Europe and several centers in Asia, with few studies having sufficiently diverse populations to make direct comparisons. The epidemiology of VTE is challenging: older observational and administrative database studies may underestimate VTE incidence due to imperfect or invasive studies on diagnosis of VTE; estimates from hospitalized VTE miss DVTs treated as an outpatient; and observational cohort studies and randomized controlled trials enroll specific populations, which may skew results [22,23]. Table S1 presents a summary of the incidence rates for VTE overall, and for DVT and PE separately, across several different studies.
The incidence of VTE among European-origin populations is better studied than in other populations. Among North Americans of European origin, administrative databases estimate an incidence of VTE between 103 and 149 per 100 000 [24–26]. In the European Union, using a modeling approach, the occurrence (both incident and recurrent) of VTE is estimated to be 245 per 100 000, with rates of DVT and PE of 148 per 100 000 and 95 per 100 000, respectively . The incidence rate in the Malmo Thrombophilia Study (MATS) was found to be lower at 66 per 100 000, though the Swedish Multigenerational Register and the HUNT2 study in Norway support a higher incidence of VTE [9,28,29]. Observational cohort data from the Longitudinal Investigation of Thromboembolism (LITE) reveal an annual incidence rate of 196 per 100 000 among relatively healthy US adults > 44 years old [30,31].
There are few if any population-based estimates of VTE incidence in African-origin populations outside of North America. Among North Americans of African descent older than 18, administrative databases estimate the overall incidence of VTE at between 141 and 155 per 100 000 while LITE, in an older population, estimated the incidence of VTE to be 318 per 100 000 [24,25,31,32]. When compared directly with European-ancestry individuals within the same study and accounting for age and gender, African-Americans have a 30–60% higher incidence of VTE than European-Americans [23–25,31] and approximately 2- to 3-fold increased incidence of VTE as compared with Asian-Americans [24,25], Native Americans  and Hispanic-Americans [24,25]. In some case series and administrative-database studies, African-Americans are found to have a higher overall and/or relative incidence of PE; however, observational cohort studies have not confirmed this finding [24,25,31,33,34]. There are no population-based estimates for the burden of VTE among African ancestry populations in Africa, Latin America or Europe.
In most studies, individuals of Asian descent appear to have a lower rate of VTE than other racial groups, though all studies reporting population-level data use administrative databases or hospital discharge data. In Singapore, the incidence of VTE in 2006 was 51 per 100 000 among Chinese, 56 per 100 000 among Indians, and 70 per 100 000 among Malay (57 per 100 000 among the total population) . In Hong Kong, the rates of hospitalized DVT and PE were reported at 17.1 and 3.9 per 100 000 people in 2000 . The overall rate of VTE among North Americans of Asian descent > 18 years old was 29 per 100 000, approximately a quarter that seen in North Americans of European descent in the same study and between the rates seen in Singapore and Hong Kong [24,25,35]. There are no population-based estimates of VTE in the Middle East or South Asia; however, the incidence of hospitalized VTE at one South Asian hospital was comparable to that seen in other countries and the rate of VTE among Indian ancestry individuals in Singapore was 56 per 100 000 [35,37].
Incidence of VTE among members of other ethnic-racial groups is sparsely represented in the literature. In the California Discharge Dataset, the incidence of VTE in Hispanic-Americans was 61 per 100 000, significantly lower than that for North Americans of both European and African descent [24,25]. Using administrative data from the Indian Health Service, the incidence of VTE among Native Americans/Alaskan Natives varies between 33.1 and 71 per 100 000, also significantly lower than for North Americans of both European and African descent, though comparing across different administrative datasets [32,38].
In summary, VTE appears to be most common in individuals of African descent in North America, with the incidence among Europeans in North America and Europe nearly as high, and a much lower incidence among people of Hispanic descent in the US, and Asian populations in both the US and Asia and Native Americans. Unfortunately, the incidences of VTE in Latin America, Africa, the Middle East, South Asia and large portions of East Asia are not known, but VTE is probably not rare in these populations.
Risk factors for VTE and race
VTE represents a complex interrelation of genetic and environmental risk factors. Racial differences in VTE incidence may be explained by (i) differential risk factor prevalence and (ii) differential impact of risk factors. Further, healthcare facilities equipped to diagnose and treat VTE or perform complex surgeries and treat cancers associated with VTE are not available to a large portion of the population of the developing world or to underserved populations in the developed world . Thus the impact of VTE and the risk factors for VTE are poorly studied for much of the world’s population. Figure 1 proposes how existing and novel VTE risk factors may mediate the observed racial differences in VTE.
Genetic risk factors
VTE is often thought of as a genetic disease, with evidence of heritability approaching 60% in European-ancestry populations [8,39]. Genetic factors thought to relate to VTE include FV Leiden, prothrombin G20210A, deficiencies in protein C, protein S and antithrombin, and elevations of some procoagulant factors (von Willebrand factor, factors VII, VIII, IX and XI) [10,17].
Factor V Leiden and prothrombin G20210A are common but modest risk factors for VTE in populations of European origin . The frequency of these mutations is up to 10% for FV Leiden and 3% for prothrombin G20210A in European populations, with a lesser prevalence in populations with European admixture (e.g. Latin Americans and African-Americans), but they are virtually non-existent in populations without European admixture [6,7,12,33]. In a registry study in the US, 14.7% of European-ancestry and 1.5% of African-ancestry participants with VTE had FV Leiden, and 3.6% vs. 0.3% of European vs. African-ancestry participants with VTE had prothrombin G20210A. The population attributable risks of VTE for factor V Leiden and prothrombin G20210A are modest at best in European populations and do not explain all the increased risk in European-ancestry populations vs. Asian and Hispanic populations and certainly do not explain the increased risk in African-Americans compared with other US populations [10,40]. Non-O blood group is also heritable and a risk factor for VTE, independent of FVIII and von Willebrand factor levels, but does not explain increased VTE in African-Americans as European-Americans have a higher prevalence of non-O blood groups than African-Americans .
While the prevalence of single nucleotide polymorphisms can be known in various populations, abnormalities of coagulation factors are difficult to compare as most laboratories set their own reference ranges based on local population distributions. The epidemiology of anticoagulant protein deficiency (protein C, protein S and antithrombin) across racial and ethnic groups is complex. Protein C levels appear lower in African-Americans compared with European-Americans but the role lower protein C levels have in VTE in African-Americans is unknown [31,34,42]. A high prevalence of protein C and protein S deficiency have been reported in Japanese, Taiwanese and Thai populations with VTE and are reported in Brazilian populations with VTE [7,12,43–46]. In a healthy population, levels of procoagulant proteins FVIII and von Willebrand factor are higher in African-Americans as compared with European-Americans, but are similar among Americans of Chinese, Hispanic and European descent, and elevated FVIII is not rare among some Asian populations [31,46,47]. Intriguingly, high FVIII and von Willebrand levels may mediate some of the increased risk of VTE seen in African-Americans compared with European-Americans . Factor VII, IX and XI levels may be a VTE risk factor but the impact these proteins have on the incidence of VTE across races is not known .
Elevated levels of D-dimer, a byproduct of fibrinolysis of cross-linked fibrinogen, are thought to represent activation of the coagulation system. Baseline levels are a strong predictor of future VTE and other thrombotic diseases . D-dimer levels are influenced by both genetic and environmental factors and may represent a balance between multiple genetic and environmental risk factors for thrombotic risk [17,48,49]. African-Americans have the highest D-dimer levels and Chinese-Americans the lowest D-dimer levels, with Hispanic-Americans and White-Americans having intermediate levels [47,48]. Another global coagulation test, the activated partial thromboplastin time, is also related to VTE risk, but racial differences have not been studied in depth . These data suggest that the coagulation system taken as a whole is related to VTE risk and may partially reflect the interactions of multiple lesser environmental and genetic VTE risk factors.
Environmental risk factors
Surgery is a known risk factor for VTE . The incidence of different surgeries differs by race and geography, as does the use of VTE prophylaxis [2,18,27]. In the US, VTE following surgery is highest in African Americans, with decreasing frequencies in European-Americans and Hispanic-Americans, and the lowest frequency in Asian-Americans . Looking in-depth at VTE following hip and knee replacement surgeries, administrative data in the US suggests that 2.4% of patients undergoing total hip arthroplasty and 1.7% of patients undergoing total knee arthroplasty suffer VTE within 91 days of surgery . In contrast, in Singapore, the rate of symptomatic DVT and PE following total knee arthoplasty without VTE prophylaxis is 1.13% and 0.17% [19,52]. Similarly, a multicenter study of symptomatic VTE in 11 Asian nations found a symptomatic VTE rate of 1.2% (95% CI, 0.7, 1.8) within 1 month of hospital discharge among 2420 patients undergoing total knee replacement, total hip replacement or hip fracture surgery . In marked contrast, in a multinational Asian study, 41% of 408 patients who were not given VTE prophylaxis experienced a venographic-confirmed DVT, a rate approaching that seen in North American and European populations [18,54,55]. Rates of screen-detected VTE also approach 40–60% in Indian patients undergoing major orthopedic procedures . VTE following surgery in Asian populations is not rare, and as wealth, access to medical care and awareness of VTE increases in these societies, the rates of documented VTE provoked by surgery are likely to increase .
Obesity as defined by a body mass index > 30 kg m−2 is associated with at least a 2-fold increased risk of VTE [30,57]. The prevalence of obesity differs by race and geography, with higher prevalence in Europe and North America, but with increasing rates in Asian populations [58–60]. Recent data from the US suggest that some of the increased incidence of VTE in African-Americans compared with European-Americans is mediated by the increased prevalence of obesity in African-Americans, but this does not explain the lower rate of VTE in the Hispanic-American population, who have high rates of obesity [24,25,31]. The role of gene-environment interactions is highlighted by the potential interaction between obesity, FV Leiden and VTE in European populations, with obese carriers of FV Leiden having a higher incidence of VTE than predicted by either obesity or FV Leiden alone . Similar gene-environment interactions may exist for other populations but have not been studied.
Other risk factors
Other risk factors, such as cancer, hospitalization, pregnancy, use of oral contraceptives and hormone replacement therapy, and antiphospholipid antibody syndrome, may influence race and region-specific VTE rates. More in-depth study is needed to clarify the role these risk factors play in racial differences in VTE incidence.
Novel risk factors
In African populations, sickle cell trait and an increased prevalence of human immune deficiency virus are candidate risk factors for the increased incidence of VTE [62,63]. Recent genome-wide association studies have found additional single nucleotide polymorphisms potentially related to VTE risk, but most have been in predominately European-ancestry populations and the frequency of these polymorphisms and the association with VTE in non-European populations is not known [64–66]. Recent advances in genetic epidemiology and a focus on the relationship between genetics and disease in non-European populations may improve our knowledge regarding the risk factors for VTE.
The complexity of VTE is mirrored by the complexity of racial differences in VTE. While it is safe to conclude that there are racial differences in VTE incidence, with African-Americans having the highest incidence and Asian populations having the lowest incidence, some of this difference may come from lack of surveillance for VTE, lack of clinical suspicion in ‘low-risk’ populations, and lack of access to medical care. We can conclude that genetic predispositions to thrombosis, such as FV Leiden and prothrombin G20210A in European populations, high FVIII, high von Willebrand factor and low protein C in African populations, may play a role in racial differences in VTE. Further, environmental risk factors, such as obesity, access to complex surgeries and treatments for cancer, and prevalence of novel VTE risk factors, such as human immunodeficiency virus and sickle cell trait, are likely to emerge as important mediators of the racial difference in VTE. While differences in VTE by race due to genetic predisposition will probably always be present, understanding the reasons for racial differences in VTE will help providers develop strategies to minimize VTE in all populations and minimize these differences.
Both authors were involved in all stages of manuscript preparation.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.