Genetics of venous thrombosis
Frits R. Rosendaal, Department of Clinical Epidemiology, C7-P, Leiden University Medical Center, PO Box 9600, 2300 RC Leiden, the Netherlands.
Tel.: +31 71 526 4037; fax: +31 71 5266994.
Summary. Venous thrombosis (deep vein thrombosis, pulmonary embolism) is a common and serious disorder, with genetic and acquired risk factors. The genetic risk factors can be subdivided in to those that are strong, moderate and weak. Strong risk factors are deficiencies of antithrombin, protein C and protein S. Moderately strong are factor V Leiden, prothrombin 20210A, non-O blood group and fibrinogen 10034T. There are many weak genetic risk factors, including fibrinogen, factor XIII and factor XI variants. Even for moderately strong risk factors (relative risks 2–5), the majority of carriers will never develop thrombosis.
Venous thrombosis usually occurs as deep vein thrombosis of the leg or pulmonary embolism, with an age-dependent incidence of one to three individuals per 1000 per year . The case–fatality rate is over 5%, mainly caused by pulmonary embolism . Both sexes are equally afflicted by a first venous thrombosis, but the risk of recurrent thrombosis is higher in men than in women [1,2].
Risk factors for venous thrombosis
The occurrence of venous thrombosis requires the presence of several risk factors, genetic and acquired (Table 1). The effect of these varies depending on which other risk factors are present. This renders an evaluation of individual risk factors, as well as a classification as ‘idiopathic’ and ‘provoked’ thrombosis somewhat arbitary.
Table 1. Risk factors for venous thrombosis
|Immobilization||Antithrombin deficiency||High levels of factor VIII|
|Plaster cast||Protein C deficiency||High levels of factor IX|
|Trauma||Protein S deficiency||High levels of factor XI|
|Major surgery||Factor V Leiden (FVL)||High levels of fibrinogen|
|Orthopedic surgery||Prothrombin 20210A||High levels of TAFI|
|Malignancy||Dysfibrinogenemia||Low levels of TFPI|
|Oral contraceptives||Factor XIII 34val||APC-resistance in the absence of FVL|
|Hormonal replacement therapy||Fibrinogen (G) 10034T|| |
|Antiphospholipid syndrome||Non-O blood group||Hyperhomocysteinemia|
|Myeloproliferative disorders|| ||High levels of PCI (PAI-3)|
|Polycythemia vera|| || |
|Central venous catheters|| || |
|Age|| || |
|Obesity|| || |
Genetic causes of thrombosis
In 1965, Egeberg  identified the first defect leading to thrombophilia, when he described a family with hereditary antithrombin deficiency. In 1969, it was reported that ABO-blood group was related to the risk of venous thrombosis . Deficiencies of protein C and protein S were identified as causes of heritable thrombophilia in the early 1980s [5,6]. In the 1990s, factor V Leiden and prothrombin 20210A were discovered [7,8]. While the deficiencies of the natural anticoagulants may be viewed as strong risk factors (risks 5- to tenfold increased), and FV Leiden and prothrombin 20210A as moderate risk factors (risk increase 2- to 5-fold), more recently reported variants are very common, but weak risk factors for thrombosis, increasing risk no more than 1.5-fold [9,10]. None of the genetic risk factors for first venous thrombosis described here have much effect on the chance of recurrence.
Strong genetic risk factors
Deficiencies of natural coagulation inhibitors
Deficiencies of antithrombin, protein C, and its cofactor protein S are found in less than 1% of the population (antithrombin deficiency in only 1 per 5000) [11–13]. There are hundreds of mutations responsible for defective genes (http://www.hgmd.cf.ac.uk/ac).
Because the deficiencies are so rare, most reports on risk come from family studies, from which it has been observed that these deficiencies can lead to a highly penetrant phenotype with over tenfold increased risks for heterozygous carriers [14–16]. In studies among unselected consecutive patients with thrombosis, i.e. not stemming from the rare families with penetrant thrombophilia, the risks conferred by these deficiencies appear lower in the unselected patients than in the families with penetrant thrombophilia [17,18], because such thrombophilic families harbor additional defects that may as yet be unknown [19–21].
Less well-studied proteins that are part of anticoagulant pathways are tissue factor pathway inhibitor (TFPI), thrombomodulin and the endothelial protein C receptor (EPCR). While genetic deficiencies are not known, low levels of TFPI have been associated with thrombotic risk, with a doubling of risk for those in the lowest 10% of levels . Thrombomodulin and EPCR are membrane-bound receptors and their levels cannot be readily measured. Studies in which shedded soluble thrombomodulin (TM) and EPCR were measured suggested a relationship of levels with venous thrombosis [23,24]. A frequent polymorphism in EPCR appears related to levels and risk [24,25]. The presence of auto-antibodies directed against EPCR leads to a doubling of the risk of thrombosis .
Moderate genetic risk factors
Factor V Leiden
Factor V Leiden (rs6025) is a common unique gain-of-function mutation, with a prevalence of carriers among Caucasians of approximately 5% [7,27]. Among patients with venous thrombosis it is found in 20%, and in approximately 50% of patients with familial thrombophilia . The variant leads to resistance to activated protein C (APC-resistance) . The mutation is located in the part of the gene encoding for one of the cleavage sites in factor V, where APC inactivates factor Va . The risk of thrombosis is 5-fold increased in heterozygotes, and fiftyfold in homozygotes [27,29].
Because factor V Leiden is common, it is responsible for a large proportion of all venous thrombotic events (20%–25%). Still, it has limited relevance for the individual: in thrombophilic families approximately 50% of the carriers will have developed thrombosis by age 65 , but in relatives with factor V Leiden who are related to unselected probands with factor V Leiden, life-time risk of thrombosis in only approximately 25% [31,32]. The risk is probably even lower in carriers without symptomatic relatives.
A mutation in the 3′-untranslated part of the prothrombin gene (prothrombin 20210A, rs1799963) leads to increased prothrombin levels, which are associated with an increased risk of venous thrombosis . This gain-of-function mutation is also frequent (several percent carriership) and also almost exclusively found in Caucasians. Carriers have a 2- to 3-fold increased risk of venous thrombosis, and the variant is found in approximately 6% of patients with venous thrombosis . As factor V Leiden and prothrombin 20210A both are common, compound heterozygotes are not exceedingly rare, and have a twentyfold increased risk of thrombosis compared to individuals with neither mutation .
Individuals with non-O ABO blood groups have a 2- to 4-fold higher risk of thrombosis than those with blood group O . O blood group is associated with reduced levels of von Willebrand factor – because of increased clearance – and factor VIII, which is related to thrombotic risk . The risk is increased for all non-OO genotypes (A1A1, A1A2, A1O1/A1O2, BB/BO1/BO2, A1B/A2B), except for A2O1/A2O2/A2A2 . The high prevalence of non-O blood groups and the magnitude of the risk makes blood group one of the more important genetic risk factors for venous thrombosis.
Fibrinogen gamma 10034T
A C to T variant at position 10034 in the fibrinogen gamma chain (rs2066865) reduces the fraction of gamma-fibrinogen (an alternative splice product) in plasma, which is related to the risk of venous thrombosis . Approximately 6% of individuals carry the variant, which increases thrombotic risk approximately 2-fold.
Weak genetic risk factors
Several genetic factors are known that have a weak effect on thrombosis, i.e. with relative risks between 1.0 and 1.5. Some of these were found by genome-wide searches, while other ones, such as the variants in methylenetetrahydrofolate reductase (MTHFR) and factor XIII were reported before such large genetic studies were performed. There are many of these variants, of which we will discuss only a few.
The C>T variant at position 677 (rs1801133) in MTHFR renders the enzyme thermolabile, resulting in slightly elevated homocysteine levels. The variant is common (10% of the general population are homozygous carriers), but the effect on homocysteine levels is small, which may explain the conflicting results about the effect of carriership on the risk of thrombosis. A meta-analysis suggested a weak effect (relative risk 1.20), but a recent large single study (the MEGA study) showed no association at all [37,38].
Homozygous carriers of the leu-allele of this FXIII variant (rs5985), who form approximately 10% of the population, have a 30% reduced risk of venous thrombosis, which effect is strongest (50% risk reduction) in those with high fibrinogen levels [39,40]. Several single nucleotide polymorphisms (SNPs) in all three fibrinogen genes have been associated with the risk of thrombosis, albeit with small risk increases (relative risks 1.2–1.3) . Variants in the promoter region of protein C, that are associated with mildly reduced levels of protein C (position 2404 and 2418), are associated with a 1.3-fold increased risk of venous thrombosis [9,41,42].
In a recent study, nearly 20 000 SNPs distributed over 11 000 genes were tested in a serial approach in three independent datasets (from the Leiden Thrombophilia Study and the MEGA study, divided in two datasets), totaling over 3000 patients with a first deep vein thrombosis and over 5000 controls . Several SNPs were associated with venous thrombosis in all three datasets, located in CYP4V2 (rs 13146272), SERPINC1 (antithrombin, rs 2227589), GP6 (Glycoprotein VI, rs1613662) and F9 (Factor IX, rs 6048, with a variant protein that is known as Factor IX Malmö). The variant in CYP4V2 was located close to the prekallikrein and factor XI genes, and was associated with factor XI levels. The risk-enhancing alleles for these variants were all common, with 18%–97% of the population carrying a variant allele, and were associated with mild increases in thrombotic risk, with relative risks varying from 1.2 to 1.5.
At the moment several ‘genome-wide association studies’ (GWAS) are underway, in which up to a million SNPs are tested in association studies. It is likely that these will add to the list of common and weak genetic risk factors. Individually, these risk factors have no clinical utility at all. It is possible, however, that comprehensive knowledge of all genetic risk factors in an individual may lead to relevant risks. It should be noted, however, that many acquired risk factors confer much higher risks than the genetic variants associated with thrombosis. As these are often transient, they offer the best opportunity to reduce the burden of thrombosis by improved and individualized anticoagulant prophylaxis.
Disclosure of Conflict of Interests
The authors state that they have no conflicts of interest.