Chapter 130: Hereditary Thrombophilia

To our knowledge, the term thrombophilia was first used by Nygaard and Brown in 1937, when they described sudden occlusion of large arteries, sometimes with coexistent venous thrombosis.¹ In 1956, Jordan and Nandorff extensively reviewed their own and previously published cases on the familial tendency in thromboembolic disease.² The term thrombophilia was then used to describe patients with prominent manifestations of venous thromboembolism (VTE; venous thrombosis in any site or pulmonary embolism) such as recurrent spontaneous VTE, VTE at young age, a strong family history of VTE, or thrombosis in an unusual site, such as the splanchnic veins or cerebral sinuses. Currently, the term thrombophilia is generally used for laboratory abnormalities, usually in the coagulation system, which increase the risk of VTE. Thrombophilia can be either acquired or hereditary. An example of acquired thrombophilia is the antiphospholipid syndrome, which is characterized by a tendency toward venous or arterial thrombosis or pregnancy complications, in combination with persistent lupus anticoagulant or antibodies to cardiolipin or β₂-glycoprotein-1 (Chap. 131). Furthermore, there are many acquired and transient conditions that lead to a prothrombotic state, including cancer, surgery, strict immobilization, pregnancy and the postpartum period, and the use of estrogen-containing medication, such as oral contraceptives and hormone replacement therapy.

Patients with hereditary thrombophilia have an increased risk of developing VTE and just like in patients without thrombophilia, approximately half of patients will develop their first episode in relation to an acquired prothrombotic risk factor. Moreover, despite young age being a criterion for thrombophilia and the mean age at time of a first thrombosis being approximately 10 years lower than in the general population, the majority of patients with thrombophilia will have the first episode later in life.³ The theoretical concept is that patients with thrombophilia have an intrinsic prothrombotic state, which in itself is insufficient to cause thrombosis, but may lead to an event when superimposed upon (clinical) risk factors, including increasing age.³

The role of hereditary thrombophilia in arterial cardiovascular disease has been extensively studied.^4,5 Most of those studies did not demonstrate significant associations between hereditary thrombophilia and arterial disease with the exception of patients with events before the age of 55 years.⁵ Moreover, the relative risk increase was very modest (odds ratios [OR] of 1.1 to 1.8) in studies that did find significant associations, indicating that hereditary thrombophilia is not a major risk factor for arterial cardiovascular disease.⁴

Like the acquired antiphospholipid syndrome, most hereditary thrombophilias are also modestly associated with pregnancy related disorders such as (recurrent) miscarriage, stillbirth, intrauterine growth retardation, preeclampsia and the hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome of pregnancy, although for later pregnancy complications this association is controversial.^6,7,8

In the past decades, hereditary thrombophilia has evolved from a very rare genetic disorder to a prevalent trait. This evolution is a direct consequence of increasing insight into the blood coagulation system, as well as advanced genetic research tools that allowed the search for abnormalities in candidate coagulation proteins and their encoding genes. At present, some form of thrombophilia can be identified in about half of the patients presenting with VTE. Likely inspired by the high yield of thrombophilia testing, testing has increased tremendously for various indications,⁹ but whether the results of such tests help in the clinical management of patients has still not been settled.^10,11 This chapter provides an overview of the most important hereditary thrombophilias and of the history of thrombophilia research. It reviews the risks associated with the most commonly tested thrombophilias and provides guidance on the indications and potential implications of the results of thrombophilia testing in various patient groups.

HISTORY OF THROMBOPHILIA RESEARCH

Research into thrombophilia began with the investigation of candidate coagulation proteins and their genes in highly thrombophilic families and linking abnormalities with the clinical phenotype within these families. As a next step, findings were confirmed in case-control studies, which yielded risk increases compared to controls, often derived from the general population. For clinicians and patients however, absolute risk estimates were needed to guide decisions regarding prevention or treatment. These were sought again in family studies of consecutive probands with a specific thrombophilic defect. The major progress in genetic and bioinformatics techniques now allows investigation in populations of patients with VTE, as well as in thrombophilic families.^12,13,14

In 1965, deficiency of the natural anticoagulant antithrombin became the first hereditary thrombophilia when Egeberg reported a Norwegian family with a remarkable tendency to VTE.¹⁵ Deficiencies of the other anticoagulant proteins, that is, protein C and protein S, were discovered as hereditary risk factors for VTE in the early 1980s.^16,17 By that time, genes could be cloned and numerous mutations in the genes encoding antithrombin, protein C, and protein S had been identified as underlying causes of low plasma levels of the anticoagulant proteins.^18,19,20 Another decade later, in 1993, Dahlbäck and colleagues described the phenomenon of activated protein C (APC) resistance, a poor anticoagulant response to APC, in a Swedish family with an increased tendency to develop VTE.²¹ The genetic basis for this APC resistance was discovered independently in several laboratories in 1995 and is caused by a single point mutation in the factor V gene which was termed factor V Leiden.^22,23,24,25 In 1996, genetic analysis of prothrombin revealed a G-to-A transition at position 20210 that was more common in patients with VTE and a strong family history of this disease than in healthy controls without VTE.²⁶ In the 1970s, it was found that individuals with non–O blood group have an increased risk of VTE.²⁷ Individuals with non–O blood group have higher levels of von Willebrand factor (VWF) and factor VIII than people with blood group O, which was the presumed mechanism of increased risk. In 1995, data from the Leiden Thrombophilia Study, a case-control study of patients with VTE and matched healthy controls, demonstrated that increased factor VIII (FVIII) activity, but not VWF activity, was independently associated with an increased risk of VTE.²⁸ Homocysteine is an intermediary amino acid formed by the conversion of methionine to cysteine. Homocystinuria or severe hyperhomocysteinemia is a rare autosomal recessive disorder characterized by severe elevations in plasma and urine homocysteine concentrations. This disease is characterized by developmental delay, osteoporosis, ocular abnormalities, and severe occlusive vascular disease. About half of the vascular complications are of venous origin.²⁹ Mild hyperhomocysteinemia was therefore studied as a risk factor for VTE in the 1990s and homocysteine levels exceeding the 95th percentile of the normal population were confirmed to be a risk factor for VTE.³⁰

Since then, numerous genetic variants that increase the risk of VTE to a more or lesser extent have been identified and are variably included in diagnostic panels of thrombophilia testing.³¹ Essentially, the majority of hereditary thrombophilias exert their effect either by upregulation of procoagulant clotting factors, or by downregulation of anticoagulant factors (Fig. 130–1). An overview of the common hereditary thrombophilias that increase the risk at least twofold, and their prevalence in patients with VTE and in the general population is presented in Table 130–1. For these more common thrombophilias a large number of clinical studies provided reliable estimates of the relative and absolute risk for VTE.

CLASSIFICATION, PATHOPHYSIOLOGY AND PREVALENCE OF COMMON HEREDITARY THROMBOPHILIA

Deficiencies of the Natural Anticoagulants Antithrombin, Protein C, and Protein S

Deficiencies of the natural anticoagulants antithrombin, protein C, and protein S, were among the first established hereditary thrombophilias. For antithrombin and protein C, two types of deficiencies are distinguished. In type I deficiency, levels of both antigen and activity are reduced and in type II, antigen levels are normal, but one or more functional defects in the molecule lead to a decreased activity. Type II antithrombin deficiencies are further subdivided according to the site of the defect in antithrombin. The defect is located in the thrombin binding domain (i.e., the reactive site) in type IIa deficiency, in the heparin binding domain in type IIb deficiency, and type IIc deficiency comprises a pleiotropic group of mutations.³² Interestingly, patients with type IIb deficiency seem to have a significantly lower risk of VTE than other types.³² Protein S circulates in two forms: free protein S (approximately 40 to 50 percent) which functions as a cofactor for APC, and protein S bound to complement component C4b-binding protein. In type I deficiency, total and free antigen levels and activity are all reduced, in type II deficiency, total and free antigen are normal, but activity is reduced, and in type III deficiency, activity and free antigen are reduced, while total antigen is low to normal. Type I and type III are probably phenotypical variants of the same disease as family members with the same DNA mutation can present with either type I or type III deficiency.³³ Whether this classification into various subtypes is truly clinically relevant for any of the deficiencies of the natural anticoagulants is largely unknown. Moreover, most laboratory panels now only test the activity of antithrombin, protein C, or protein S, and thereby do not distinguish between different types of deficiencies. Homozygous antithrombin deficiency is extremely rare and the only reported cases involve type IIb deficiencies.³⁴ Homozygous type I deficiency has never been described in humans and is believed to be incompatible with life. Complete antithrombin deficiency in knockout mice leads to embryonic death.³⁵ Homozygous protein C and protein S deficiencies are also very rare and these are associated with neonatal purpura fulminans and massive thrombosis.^36,37 In a similar fashion, warfarin-induced skin necrosis has been described in patients with heterozygous protein C or S deficiencies after initiation of vitamin K antagonists (VKAs).^38,39 The concept is that after VKA initiation vitamin K–dependent protein C and S levels drop sooner than levels of factors II, IX, and X, thereby temporarily causing a paradoxal procoagulant state.⁴⁰ This is, however, a rare clinical complication, possibly resulting from concomitant treatment with (low-molecular-weight) heparin in the acute phase of VTE. Deficiencies can be caused by a large number of mutations, that are recorded in occasionally updated databases.^18,19,20 Overall, deficiencies of the natural anticoagulants are rare (see Table 130–1). In cohorts of consecutive patients with VTE, the prevalence of a deficiency of one of the natural anticoagulants is below 10 percent.⁴¹ In the general population, the prevalence of the deficiencies combined is approximately 1 percent.^42,43,44

Factor V Leiden/Factor V G1691A

In 1993, Dahlbäck and colleagues first described APC resistance in a Swedish family with a high tendency of VTE.²¹ In the original paper, Dahlbäck proposed that APC resistance was best explained by an hereditary deficiency of a previously unrecognized cofactor to APC, after having ruled out several possible mechanisms, including deficiencies of protein S and protein C, or linkage with polymorphisms in the FVIII or VWF genes. He then showed that this "cofactor" was identical to coagulation factor V.²¹ Soon thereafter, several laboratories independently from each other reported the underlying genetic defect, a single G-to-A substitution in the gene of factor V at nucleotide position 1691, resulting in an amino acid change from Arginine (Arg) to Glutamine (Gln) at position 506, the first cleavage site of factor Va for APC^22,23,24,25 (Fig. 130–2). The mutation was named factor V Leiden after the city in the Netherlands in which the group with the first publication was located.²² The proteolytic inactivation of activated factor V (FVa) is approximately 10 times slower for Gln506-FVa compared with Arg506-FVa, which explains the partial, but not full, resistance to APC.⁴⁵ Factor V Leiden is the most common hereditary thrombophilia. In unselected consecutive patients with VTE, the prevalence is 20 to 25 percent.²² The prevalence in the general population varies considerably between different ethnic groups. Factor V Leiden is very rare among Asians and Africans but has a high prevalence (approximately 5 percent) among whites (see Table 130–1).⁴⁶ Within Europe, the prevalence is higher in the north than in the south.⁴⁶ This implies a founder effect that suggests that the mutation occurred after the separation of non-Africans from Africans and after the divergence of whites and Asians. Studies using linkage disequilibria between factor V Leiden and specific markers indicate that the mutation occurred around 21,000 years ago.⁴⁷ The high prevalence of factor V Leiden suggests an evolutionary benefit.⁴⁸ The presumed mechanism is reduced peripartum and menstrual blood loss in affected female carriers.^49,50 More recently, factor V Leiden was associated with increased sperm counts and a shorter conception time in affected male and female carriers, suggesting another evolutionary benefit.^51,52,53 The mechanism for this increased fertility is unknown.

Prothrombin G20210A

The prothrombin G20210A mutation was discovered in 1996 by genetic analysis of the prothrombin gene in families with a strong tendency of VTE and confirmed in a case-control study of patients with VTE.²⁶ The mutation is located in the 3′-untranslated region of the prothrombin gene and augments translation and stability of prothrombin messenger RNA.⁵⁴ This leads to an average 32 percent higher blood levels of the zymogen prothrombin that is structurally identical to the protein produced by patients with the wild-type gene.²⁶ In unselected patients with VTE, the prevalence of prothrombin G20210A is approximately 6 percent.⁵⁵ Like factor V Leiden, the prothrombin G20210A mutation is found largely in whites and genetic linkage disequilibrium studies date the mutation back to 24,000 years ago.^47,56 Within Europe, the prevalence of the mutation increases from 3 to 5 percent in Southern Europe and the Middle East to 1 to 2 percent in Northern Europe, which opposes the observed geographical gradient of factor V Leiden.⁵⁶

Increased Levels of Factor VIII

Most of coagulation FVIII circulates in complex with VWF. Therefore, determinants of VWF, including ABO-blood group and endothelial stimulation, indirectly also determine FVIII levels.⁵⁷ Furthermore, many other factors have been associated with high FVIII levels, including increasing age, high body mass index, diabetes mellitus, and hypertriglyceridemia.⁵⁷ High FVIII levels are part of acute phase reactions and sustained increases are observed during pregnancy, surgery, chronic inflammation, malignancy, liver disease, hyperthyroidism, renal disease. In most of these conditions, there is a concordant increase of FVIII and VWF levels.⁵⁷ Apart from ABO blood group, the genetic causes of high FVIII levels are largely unknown. Nevertheless, persistence over time and familial clustering of high FVIII levels in patients with VTE implies that genetic factors are prevalent. In a family study of consecutive patients with persistently elevated levels of FVIII of at least 150 percent and venous or arterial thrombosis, the prevalence of high FVIII in first degree family members was 40 percent, which is almost in the range to what one would expect from an autosomal dominant inheritance.⁵⁸ Despite the uncertainty about the mechanisms of high FVIII levels, the association between high FVIII levels and venous thrombosis is well established. The very clear dose-dependent relationship between FVIII levels and the risk of VTE suggests that elevated FVIII is causative for thrombosis.²⁸ Interestingly, increased FVIII levels in patients with VTE are usually not influenced by acute-phase reactions.⁵⁷ The prevalence of elevated FVIII levels is high: 25 percent of patients with a first episode of deep-vein thrombosis and 11 percent of healthy control subjects have FVIII levels of 150 percent or higher.²⁸ Hence, an increased level of FVIII is the most common familial, albeit not monogenetic, thrombophilia with an estimated population attributable risk of 15 percent.⁵⁷ Whether measurement of FVIII levels should be part of a hereditary thrombophilia panel is debatable as FVIII levels may be transiently increased by a multitude of external factors including the acute VTE episode itself (Table 130–2). This leads to a significant proportion of "false-positive" test results, a necessity for repeat testing, and potentially unnecessary concern among tested patients.

Mild Hyperhomocysteinemia

Homocysteine is an intermediate in the metabolism of the amino acids methionine and cysteine and participates in several metabolic pathways. Some of the enzymes involved in homocysteine metabolism are dependent on vitamin B₆, folic acid, and vitamin B₁₂, and deficiencies lead to hyperhomocysteinemia. A polymorphism in the enzyme methylenetetrahydrofolate reductase (MTHFR), c.C677T, leads to an alanine to valine substitution at position 222 resulting in a variant enzyme with reduced activity and increased thermolability. Homozygosity for this polymorphism leads to 24 percent increased homocysteine levels and is the most common genetic cause of mild hyperhomocysteinemia.⁵⁹ The prevalence is 10 to 20 percent in whites and 10 percent in Orientals, but it is rare in Africans.⁶⁰ Many other conditions are associated with hyperhomocysteinemia, including renal failure, hypothyroidism, smoking, excessive coffee consumption, inflammatory bowel disease, psoriasis, and rheumatoid arthritis.⁶¹ Severe hyperhomocysteinemia (plasma levels exceeding 100 μmol/L), also named homocystinuria, is a rare autosomal recessive disorder clearly associated with vascular occlusive disease.²⁹ Because of this association, mildly increased levels (exceeding the 95th percentile of the normal population) were studied as a risk factor for venous and arterial thrombosis. In the Leiden Thrombophilia Study mild hyperhomocysteinemia was associated with a 2.5-fold increased risk of first VTE.³⁰ Interestingly, the association of homocysteine levels with VTE was stronger in men than in women and increased with age.³⁰ The mechanism by which hyperhomocysteinemia would lead to thrombosis is unknown. A 2005 meta-analysis that included more than 50 studies with more than 8000 cases of VTE confirmed the association of VTE with hyperhomocysteinemia and also demonstrated a 20 percent higher VTE risk in MTHFR 677TT carriers compared with 677CC carriers.⁶² The association between homocysteine levels is again under scrutiny in newer studies. The single largest case-control study of more than 4000 patients with a first VTE found no association between MTHFR 677TT and VTE.⁶³ Furthermore, a family study of probands with mild hyperhomocysteinemia and venous or arterial thrombosis found that the association between homocysteine and VTE disappeared after adjustment for FVIII levels.⁶⁴ Finally, homocysteine lowering by B vitamins did not reduce the incidence of recurrent events both in patients with VTE as well as in patients with arterial cardiovascular disease.^65,66 As a result of these studies, homocysteine testing (either its levels of polymorphisms) as part of a thrombophilia panel has been largely abandoned.

PITFALLS IN LABORATORY TESTING FOR HEREDITARY THROMBOPHILIA

Testing for hereditary thrombophilia is performed in many patients, or family members of patients, with various thromboembolic diseases or pregnancy complications.⁹ However, it is important to realize that many acquired, often transient, conditions may affect the test results. Most known is the use of VKAs, which reduces the levels of anticoagulant factors protein C and protein S, which can thereby mimic severe deficiencies. Another example is pregnancy that reduces free protein S levels and increases APC resistance and FVIII levels.⁶⁷ However, other factors can also affect test results and should be considered in the interpretation of results or, better yet, in the timing of testing. Table 130–2 presents an overview of acquired conditions that can yield false-positive thrombophilia tests. For deficiencies of the natural anticoagulants as well as for elevated FVIII levels, repeated testing should be performed to exclude spuriously abnormal tests. Genetic testing for deficiencies of the natural anticoagulants is not performed because of the large number of known mutations. The hereditary nature of deficiencies must be established by confirming the abnormality in a first-degree family member.

VENOUS THROMBOEMBOLISM

The relative risk of a first episode of VTE in individuals with a form of common hereditary thrombophilia ranges from 2 to 11 (Table 130–3). These figures were derived from family and population-based cohort or case-control studies.¹¹ Individuals homozygous for the prothrombin G20210A or factor V Leiden mutations are at higher risk for VTE than heterozygotes. Likewise, patients with combined thrombophilic disorders have a higher risk of VTE than those with a single defect.¹¹ One can estimate the absolute risk of VTE by multiplying the relative risk with the absolute incidence in cohorts from the general population, in which the risk of first VTE is approximately 0.2 to 0.3 per 100 person-years.⁶⁸ Because this may however lead to imprecise estimates, it is preferable to use absolute risk estimates derived from cohort studies. Retrospective cohort studies may produce less-reliable incidence estimates because retrospective studies carry the risk of (unconscious) selection of patients and data and the clinical diagnosis of VTE may not have been confirmed by objective tests. Prospective cohort studies of asymptomatic carriers of hereditary thrombophilic defects are probably better suited to estimate the true incidence of thrombosis in thrombophilic patients. It is important to note that cohort studies have been mainly performed in relatives of (consecutive) patients with a particular thrombophilic defect. Absolute risk estimates from family studies are higher than from population based studies. Even in the absence of any hereditary thrombophilia, the risk of VTE is still twofold increased in first degree family members of patients with VTE.⁶⁹ This suggests cosegregation of other, unmeasured or unknown thrombophilias. Table 130–4 presents absolute risk estimates for a first-episode VTE for asymptomatic carriers with a family history of VTE. These risk estimates can be used to counsel both affected and unaffected family members about their risk of VTE. These studies also provide absolute risk estimates for VTE associated with exogenous additional risk factors such as surgery, trauma or immobilization, and pregnancy or use of hormonal contraception (Table 130–4). These incidences are derived from retrospective family studies, as prospective studies are limited by shorter followup duration and reduced power.

Thrombophilia and the Risk of Recurrent Venous Thromboembolism

Regardless of thrombophilia, the absolute risk of a recurrent episode is much higher than the risk of a first episode of VTE. The most important determinant of recurrence is the presence of transient clinical risk factors during the time of the first episode.⁷⁰ After an unprovoked first episode of VTE, the risk of recurrence is approximately 10 percent in the first year after cessation of anticoagulation and approximately 5 percent per year thereafter.⁷⁰ The risk is lower after VTE that was associated with a transient risk factor, with an incidence in the first 2 years of 0.7 percent per year for surgery-provoked VTE and 4.2 percent per year for VTE provoked by estrogen use, pregnancy, temporary immobilization, or trauma.⁷⁰ Other determinants for recurrence are male sex, proximal (vs. distal) deep venous thrombosis (DVT), and elevated D-dimer levels after stopping anticoagulation.⁷⁰

Although individuals with hereditary thrombophilia have a higher risk of developing a first episode of VTE, the risk increases for recurrent events in patients with prior VTE are much lower. Numerous case-control studies in VTE patients with a specific thrombophilia with nonthrombophilic VTE patients as controls yield relative risks of 1.5 to 2.0 for most hereditary thrombophilias.^{71,72,73,74,75,76} Patients develop VTE when the combination of individual susceptibility (including hereditary thrombophilia) and acquired, sometimes transient, risk factors are sufficient to cause VTE. This individual susceptibility is constant throughout life. Therefore, patients with a previous episode of VTE have proven to have sufficient individual susceptibility, irrespective of hereditary thrombophilia being part of that susceptibility. This explains why the risk for recurrence is at best only marginally increased for carriers of hereditary thrombophilia.⁷⁷

ARTERIAL THROMBOEMBOLIC DISEASE

The increased risk of VTE in patients with hereditary thrombophilia has led to many studies investigating the association of thrombophilia with arterial thromboembolic disease. Factor V Leiden and prothrombin G20210A are the most extensively studied as risk factors for arterial disease. The largest meta-analysis of case-control studies of patients with myocardial infarction found an OR of 1.17 (95% confidence interval [95% CI] 1.08 to 1.28) for factor V Leiden (60 studies with 42,390 patients) and 1.31 (95% CI 1.12 to 1.52) for prothrombin G20210A (40 studies with 26,087 patients).⁴ The association between these mutations and myocardial infarction is stronger when analyses are limited to patients with myocardial infarction below the age of 55 years with ORs of 1.34 (95% CI 0.94 to 1.91) for factor V Leiden and 1.86 (95% CI 1.00 to 3.51) for prothrombin G20210A.⁵ Meta-analyses of studies in patients with ischemic stroke have shown similar modest risk increases in patients with factor V Leiden or prothrombin G20210A.⁷⁸

Deficiencies of the natural anticoagulants are less prevalent than factor V Leiden and prothrombin G20210A and as a result, the association with arterial thromboembolic disease has not been extensively studied. Although various case reports of patients with antithrombin, protein C, or protein S deficiency have been published, most case-control studies have failed to demonstrate a significant association with myocardial infarction and ischemic stroke.⁷⁹ A retrospective cohort study of 552 first-degree family members of patients with venous or arterial thrombosis and a deficiency of either antithrombin, protein C or protein S found no increased risk of arterial cardiovascular disease in affected family members older than age 55 years.⁸⁰ However, in persons younger than age 55 years, protein C and protein S were associated with a five- to ninefold increased risk, whereas antithrombin deficiency did not confer an increased risk.⁸⁰

Several case-control studies have found associations with elevated FVIII levels and myocardial infarction.⁵⁷ Furthermore, prospective cohort studies of healthy individuals have demonstrated marginally increased risks of myocardial infarction and stroke in patients with elevated FVIII levels (ORs between 1.0 and 1.4).⁵⁷ In a prospective family study of asymptomatic first-degree family members of patients with elevated FVIII and either VTE or premature arterial thrombosis, the risk of arterial thromboembolism was increased 4.5-fold compared with family members with normal FVIII levels.⁸¹ However, elevated FVIII levels are associated with several well-known risk factors for arterial cardiovascular disease, including obesity, high glucose levels, increasing age, chronic inflammatory diseases, and renal disease.⁵⁷ It is possible that studies evaluating the association between elevated FVIII and arterial thromboembolic disease have not been able to sufficiently adjust for known and unknown confounders. Moreover, acute myocardial infarction and ischemic stroke may cause acute-phase reactions that transiently increase FVIII levels, hampering interpretation of case-control or retrospective cohort studies. However, patients with hemophilia A, a genetic cause of decreased FVIII levels, have an approximate 80 percent lower risk of death from ischemic heart disease, indicating a potential causal relation between FVIII levels and arterial thrombosis.⁸²

Mild hyperhomocysteinemia and MTHFR 677TT have been extensively studied in relation to arterial thromboembolic disease. A meta-analysis of studies that included more than 5000 patients with ischemic heart disease and more than 1000 patients with ischemic stroke demonstrated a significant correlation between homocysteine level and the risk of arterial thrombosis.⁸³ The risk increase in was higher in retrospective studies than in prospective studies in which homocysteine levels are measured before the thrombotic episodes. This could, in part, be explained by the observed association between hyperhomocysteinemia and other well-known risk factors for arterial cardiovascular disease, including smoking, chronic inflammatory disorders, and renal failure.⁶¹ Like with elevated FVIII levels, it is uncertain whether studies that investigate the relation between homocysteine and arterial cardiovascular disease have been able to sufficiently adjust for confounding variables. The association between MTHFR 677TT and ischemic heart disease has shown mixed results with no association in studies in North American patients and a modest 16 percent risk increase in studies in European patients.⁸⁴ It was initially hypothesized that this difference is attributable to a lower dietary folate intake in Europe. However, this hypothesis conflicts with the results of trials in patients with vascular disease in whom homocysteine lowering with folic acid and B vitamins did not reduce the risk of recurrent episodes.⁶⁶

PREGNANCY COMPLICATIONS

Although many studies have observed a relationship between hereditary thrombophilia and pregnancy complications, including recurrent miscarriage, late pregnancy loss, preeclampsia, intrauterine growth restriction, and placental abruption, this should be regarded as controversial. Most associations are modest in strength and vary with type of thrombophilia and type of pregnancy complication.^8,85,86 Furthermore, the most recent and larger prospective cohort studies found lower ORs for hereditary thrombophilia than older and smaller case-control studies, which may point to a bias in the observed associations.^8,87,88 The mechanisms of how thrombophilia would lead to pregnancy complications remain largely unknown. It is unlikely that mere hypercoagulability with thrombosis of placental vasculature is the pathophysiologic substrate for an association with thrombophilia. Animal and in vitro studies have implicated a role for both procoagulant and inflammatory pathways in pregnancy failure and interesting effects of acetylsalicylic acid (ASA) and heparin.⁸⁹ For instance, in a murine high-risk pregnancy model, heparin rescued factor V Leiden–associated placental failure, but this was independent of anticoagulation.⁹⁰

GENERAL CONSIDERATIONS OF THROMBOPHILIA TESTING

Several arguments against testing for thrombophilia should be considered.¹⁰ First, an obvious disadvantage of testing for thrombophilia is the high cost. Although two studies concluded that testing for thrombophilia in some scenarios could be cost-effective, the underlying assumptions from inconsistent observational studies seriously hamper their interpretation.^91,92 Second, although the psychological impact and consequences of knowing that one is a carrier of a (genetic) thrombophilic defect are considered limited, a qualitative study described several negative effects of both psychological and social origins.^93,94 Difficulties in obtaining life or disability insurance are frequently encountered by individuals who are known carriers of thrombophilia, regardless of whether they are symptomatic or asymptomatic.⁹³ Third, the most compelling argument against testing is the potential false reassurance that may arise from a negative thrombophilia test for individuals who come from families with a thrombotic tendency. For example, Table 130–4 indicates that in these families, women without thrombophilia have a markedly increased risk of oral contraceptive-related VTE compared to pill users from the general population (0.7 percent in women with a natural anticoagulant deficiency versus 0.04 percent per year of use), reflecting a selection of families with a strong thrombotic tendency in which yet unknown thrombophilias have co-segregated.

The following paragraphs discuss the potential scenarios for thrombophilia testing in more detail.

TESTING FOR THROMBOPHILIA TO MODIFY THE RISK OF A FIRST VENOUS THROMBOEMBOLISM

Having a family history of VTE is a poor predictor of the presence of thrombophilia.^69,95 Still, a potential advantage of testing patients with VTE for thrombophilia may be the identification of asymptomatic family members in order to take preventive measures if tested positive, and to withhold such measures if relatives have tested negative. An important requisite is that a test result indeed dichotomizes carriers and noncarriers in terms of their risk for a first episode of VTE.

Based on the absolute risks for a first episode of VTE (see Table 130–4), it is clear that the 1 to 3 percent annual major bleeding risk associated with continuous oral anticoagulant treatment outweighs the risk of VTE.^96,97 Table 130–4 also shows that during high-risk situations such as surgery, immobilization, trauma, pregnancy, and the postpartum period, and during the use of oral contraceptives, the absolute risk is generally low, with the exception of women with a natural anticoagulant deficiency who use oral contraceptives or are pregnant.

Estimates of the effect of avoidance of oral contraceptives on the number of prevented episodes of VTE by means of thrombophilia testing can be calculated for women who have a positive first-degree relative with VTE in whom the thrombophilic defect is known.⁹⁸ To avoid one VTE event, 28 women with antithrombin, protein C or protein S deficiency, and a positive family history for VTE would need to refrain from oral contraceptives, and to identify these women, 56 female relatives would need to be tested.⁹⁸ For factor V Leiden or the prothrombin 20210A mutation, approximately 333 women would need to avoid oral contraceptives and 666 female relatives would need to be tested. Although the number of tested women for the natural deficiencies seems quite acceptable, the major argument against this scenario is that a normal level of antithrombin, protein C, or protein S in women from these families does not exclude a strongly increased risk of VTE during oral contraceptive use, as compared to the general population (see Table 130–4). The same, but to a lesser extent, is true for women from thrombophilic families who do not carry either the factor V Leiden or prothrombin mutation, but here also the number needed to screen is unacceptably high.

Table 130–5 indicates the estimated number needed to test to initiate prophylactic measurements around pregnancy, again applicable to women from thrombophilic families. Only for women with antithrombin, protein C, or protein S deficiency, or those who are homozygous for factor V Leiden (Table 130–5), the risks of 4 and 16 percent respectively during pregnancy and the postpartum period may outweigh the nuisance of daily subcutaneous low-molecular-weight heparin (LMWH) injections with frequently occurring skin reactions, and the very small risk for severe complications of anticoagulant therapy during pregnancy.^{99,100,101,102} Furthermore, the numbers in Table 130–5 underestimate the number of women that need to use prophylaxis (and be tested prior to this decision) in order to avoid pregnancy-related VTE, because a 100 percent efficacy of prophylaxis is assumed in these calculations. Whether the absolute risks of pregnancy-related VTE justifies prophylaxis for 8 months during pregnancy, or the shorter postpartum period of 6 weeks is a matter of the physicians' and patients' preference.¹⁰² The risk of pregnancy-related VTE in women from these families who do not have the hereditary thrombophilic defect is approximately 0.5 percent, compared to 0.2 percent in the general population.¹⁰³ Hence, withholding prophylaxis to women from thrombophilic families who do not have the defect is supported by evidence from well-designed studies of individuals in the same clinical context.

THROMBOPHILIA TESTING IN PATIENTS WITH VENOUS THROMBOEMBOLISM

Thrombophilia testing is most often considered in patients with VTE, particularly if they are young, have recurrent episodes, have thrombosis at unusual sites, or have a positive family history for the disease. However, although such a strategy may lead to an increased yield of testing, the main question is whether a positive test result should alter management. As previously discussed, thrombophilia is a very poor predictor of recurrent VTE, and whether the very modest risk increase warrants prolongation of the duration of anticoagulation, particularly after provoked VTE, is a matter of debate.^70,104 Furthermore, given the rarity of homozygous or double heterozygous thrombophilias in unselected patients with VTE, the efficiency of testing is obviously very low.^10,105 A randomized controlled trial of testing for thrombophilia in patients with a first episode of VTE would provide the ultimate evidence to decide whether this is justified, but no such trials have been successfully performed.¹⁰⁶ To investigate whether testing for thrombophilia reduced the risk of recurrent VTE, 197 patients who had had a recurrent event were compared to 324 patients who did not have a recurrence.¹⁰⁷ The OR for recurrence was 1.2 (95% CI 0.9 to 1.8) for tested versus nontested patients, indicating that testing, with real life clinical decisions based on the outcome of testing, did not reduce the risk of recurrent VTE in patients after a first episode.

THROMBOPHILIA TESTING IN PATIENTS WITH ARTERIAL CARDIOVASCULAR DISEASE

The association between hereditary thrombophilia and arterial cardiovascular disease is questionable, or at least much weaker than for VTE. The association is stronger in patients with events before the age of 55 years. As a result, thrombophilia test panels are often ordered in with arterial cardiovascular disease. In a survey among Dutch physicians that ordered thrombophilia tests in 2000 consecutive patients in 2003 and 2004, arterial cardiovascular disease, mainly ischemic stroke, was the indication for testing in 23 percent of patients.⁹ Interestingly, only 54 percent of those patients were younger than 50 years of age. Testing for hereditary thrombophilia in patients with (premature) arterial cardiovascular disease could only be justified if the test results would mandate different secondary prevention. However, more vigorous secondary prevention such as long-term dual antiplatelet therapy or oral anticoagulation instead of ASA monotherapy is not beneficial for most patients with arterial cardiovascular events, mainly because of the increased risk of bleeding. Whether such a strategy is beneficial for patients with hereditary thrombophilia has never been tested, but is very unlikely given the very modest risk increases associated with hereditary thrombophilia. Therefore, testing in this setting is not justified.

THROMBOPHILIA TESTING IN WOMEN WITH PREGNANCY COMPLICATIONS

Thrombophilia testing in women with pregnancy complications would be indicated if a test result would alter management. However, to date, testing for hereditary thrombophilia in this setting cannot be justified^10,102 for the following reasons. For women at moderate to high risk of preeclampsia, ASA provides a modest benefit in reducing the risk of preeclampsia, but this is regardless of the presence of hereditary thrombophilia.^102,108 Whether anticoagulant treatment with heparin or LMWH improves the chance of a successful pregnancy outcome in women with pregnancy complications is presently unknown as results from randomized clinical trials are extremely inconsistent.^{109,110,111,112,113} It is also uncertain whether presence of hereditary thrombophilia is a prerequisite for an assumed beneficial effect, if any. Only three randomized controlled trials have been exclusively dedicated to women with hereditary thrombophilia and recurrent miscarriage, a single fetal loss, or late pregnancy complications. The first trial found promising results in women with heterozygous factor V Leiden mutation, prothrombin G20210A mutation, or protein S deficiency and a single previous pregnancy loss after 10 weeks gestation.¹¹⁴ Women who were allocated to enoxaparin had a much higher chance of a livebirth than those allocated to ASA (86 percent and 29 percent, respectively; OR 15.5; 95% CI 7 to 34), but several methodologic issues were raised, and the results of this single study have not been confirmed by other trials.^102,115 Second, in the FRUIT trial, women with hereditary thrombophilia and a history of preeclampsia or intrauterine growth restriction requiring delivery before 34 weeks of gestation, were randomized between dalteparin with ASA and ASA alone.¹¹⁶ The primary outcome (recurrence of a hypertensive disorder, e.g., preeclampsia, HELLP, or eclampsia) did not differ between the two groups, but none of the women in the LMWH/ASA group developed recurrent hypertensive disorders prior to 34 weeks gestational age, whereas six women in the ASA-only group delivered before 34 weeks because of recurrent hypertensive disorders (risk difference 8.7 percent; 95% CI 1.9 to 15.5 percent). Finally, the TIPPS study included thrombophilic women at high risk for pregnancy complications or at increased risk of VTE and randomized them between dalteparin and no dalteparin.¹¹³ The primary composite outcome (severe or early onset preeclampsia, small-for-gestational-age infant, pregnancy loss, or VTE) did not differ between the groups. Hence, to date, there is no evidence from sufficiently powered and adequately designed clinical trials that justify use of heparin to improve pregnancy outcome in women with hereditary thrombophilia, and heparin should only be given in the context of a clinical trial.¹¹⁷

The knowledge about the hereditary contribution to the etiology of VTE has increased tremendously over the past decades. Still, testing for thrombophilia serves only a limited purpose and should not be performed on a routine basis. Thrombophilia testing in asymptomatic relatives may be useful in families with antithrombin, protein C, or protein S deficiency, or for siblings of patients who are homozygous for factor V Leiden, and is limited to women who intend to become pregnant or who would like to use oral contraceptives. Careful counseling with knowledge of absolute risks helps patients make an informed decision in which their own preferences can be taken into account, and in which the clinician should be cautious to not provide false reassurance in case of a negative test result. Observational studies show that patients who have had VTE and have thrombophilia are at most at a slightly increased risk for reoccurrence. Other determinants, including circumstances during the first VTE, elevated D-dimer levels, and male sex, are better predictors of recurrent VTE.⁷⁰ Furthermore, no beneficial effect on the risk of recurrent VTE was observed in patients who had been tested for inherited thrombophilia. In the absence of trials that compared routine and prolonged anticoagulant treatment in patients testing positive for thrombophilia, testing for such defects to prolong anticoagulant therapy cannot be justified. Finally, there is at present no reason to test patients with arterial cardiovascular disease, or women with recurrent miscarriage or late pregnancy complications for hereditary thrombophilia, in the absence of evidence-based guidelines for changes in management.