Hypercoagulability in Venous and Arterial Thrombosis

Venous Thrombosis

The current interest in a molecular approach to hypercoagulability is exemplified by patients who present with venous thrombosis that has no apparent cause. These patients are examined for a hereditary defect of an anticoagulant pathway, and assays are done for antithrombin III, protein C, protein S, and resistance to activated protein C [2-5]. The percentage of such patients who have a demonstrable genetic defect is now considerably higher than it was even a few years ago; in the past 2 years alone, a dramatic increase has been seen in the number of reports of patients presenting with venous thromboemboli who have been found to possess the inherited defect of resistance to activated protein C.

Factor V Leiden Mutation

In a case–control study of 301 unselected patients who presented with venous thrombosis, 21% were reported to have resistance to activated protein C associated with a mutation at nucleotide 1691 (Arg506) of the factor V gene, in which the arginyl residue is replaced with glutamine: the factor V Leiden mutation. Other laboratories have reported higher percentages (≤ 60%), but these studies were done on highly selected patient samples. In a recent Dutch study of 471 consecutive patients with a first presentation of objectively confirmed deep venous thrombosis, 85 patients (18%) were heterozygous for resistance to activated protein C and 7 (1.5%) were homozygous [6]. The researchers estimated that the relative risk for venous thrombosis was increased 7-fold for persons who were heterozygous for the factor V Leiden mutation and 80-fold for those who were homozygous. They also suggested that most homozygous persons will have at least one thrombotic episode during their lifetime. The prevalence of the factor V Leiden mutation was reported to be high (38%) among 16 symptomatic patients with protein S deficiency [7], suggesting that the combination of the factor V Leiden mutation and protein S deficiency may be associated with a higher risk for thrombosis. The same group [8] had previously reported an increased risk for thrombosis in patients who had both factor V Leiden mutation and protein C deficiency compared with siblings who had only a single defect.

In a French study [9] of 321 patients with the factor V Leiden mutation, 303 patients were heterozygous for this mutation and 18 were homozygous. Of the latter group, 15 patients had venous thromboembolism at a median age (±SD) of 29 ± 11 years. Most of these homozygous patients had at least one precipitating cause of venous thromboembolism, and the most common was the use of oral contraceptives. The authors comment that use of estroprogestogen contraceptives carries a significant risk for venous thromboembolism in homozygous women. In users of contraceptives containing desogestrel, the relative risk was 6.0 (95% CI, 1.9 to 19.0) among carriers of the factor V Leiden mutation, superimposed on the 8-fold increased risk for venous thrombosis in carriers of this mutation [10]. Consequently, the risk of carriers who use this third-generation oral contraceptive compared with noncarriers who do not use it is increased almost 50-fold [10]. A striking increase in thrombotic risk is therefore seen when the factor V Leiden mutation exists in the presence of other prothrombotic stimuli, such as the use of oral contraceptives. Hellgren and colleagues [11] studied 34 pregnant women who had previous thromboembolic complications in connection with pregnancy and 28 women who developed thrombosis while taking oral contraceptives. They found resistance to activated protein C in almost 60% of the former group and about 30% of the latter. Their results led them to ask whether it would be reasonable to do general screening for resistance to activated protein C during early pregnancy or before prescribing oral contraceptives. Because of the numbers of persons involved, however, such widespread screening could have a considerable effect on health care budgets.

The prevalence of the factor V Leiden mutation is much lower in normal patients than in patients who present with overt thrombosis. In 618 European persons, the allele frequency was reported to be 4.4% (CI, 2.6% to 6.3%), and only 4% of the carriers were homozygous [12]. However, this mean figure disguised a wide range of values in different ethnic groups, varying from a prevalence of 0% in a small group of Italians to a carrier rate (double the allele frequency) of 14% in a Greek population. It is interesting that this study did not find the allele in native African, Asian, or Australasian blood donors.

In an important large epidemiologic study by Ridker and coworkers [13], almost 15 000 apparently healthy men in the United States provided baseline blood samples and were subsequently followed for a mean of 8.6 years. In those who developed myocardial infarction, strokes, or deep venous thrombosis, the presence of the factor V Leiden mutation was associated with an increased risk for venous thrombosis. In contrast, the presence of the mutation was not associated with an increased risk for myocardial infarction or stroke. The carrier rate for the factor V Leiden mutation was 6.0% in men who remained free of vascular disease and 6.1% in those who developed myocardial infarction.

Emmerich and associates [14] also found no statistically significant difference in the prevalence of the mutation between 609 men who had myocardial infarction (5.1%) and 692 age-matched controls (4.6%). However, they cautioned that their results cannot be extrapolated to persons who are homozygous for the factor V Leiden trait. In men older than 60 years of age, the prevalence of the mutation exceeded 25% in those who developed primary venous thrombosis, giving them an adjusted risk seven times higher than that of men homozygous for the wild-type allele (CI, 2.6 to 19.1 times) [13].

In a study of 814 apparently healthy Germans, 56 heterozygous and 2 homozygous mutant alleles were found among 1628 alleles tested, for an allele frequency of 3.6% [15]. In a general population in Italy, 124 of 4703 persons studied were classified as heterozygous (2.7%) and only 4 were classified as homozygous [16]. This relatively low prevalence of the factor V Leiden mutation in Italy was recently confirmed by Mannucci and colleagues [17], who found an allele frequency of 1.3% (CI, 0.6% to 2.6%) in 344 persons studied by polymerase chain reaction [17].

In a study reported from New Haven, Connecticut [18], the factor V Leiden mutation was found in 3 of 214 unselected black persons in the United States (1.4%); this prevalence was not significantly different from the prevalence found in a local sample of white persons (1.6%). The investigators concluded that their data did not support a difference in the prevalence of the mutation between black and white populations, at least in the United States. However, a prevalence rate of 1.6% for the mutation in a white population in the United States seems unusually low compared with the figure of 6% found in the largest study in the United States [13]. The genetic heterogeneity of U.S. black persons, as compared with African black persons studied by Rees and coworkers [12], suggests that the factor V Leiden mutation will be found occasionally in the U.S. black population.

Other Inherited Defects of the Anticoagulant Pathways

Resistance to activated protein C due to heterozygosity for the factor V Leiden mutation is clearly the most prevalent predisposing cause of thrombosis that can be defined today in molecular terms. The mutation is found much more frequently than other congenital abnormalities of the anticoagulant pathways, such as protein C, protein S, or antithrombin III deficiency [18-20]. For example, of 277 unselected patients with objectively diagnosed deep venous thrombosis, 1% were deficient in antithrombin III, 3.2% were deficient in protein C, and 2.3% were deficient in protein S, for a total of 6.5% [21]. The prevalence of the factor V Leiden mutation in patients presenting with deep venous thrombosis is therefore approximately three times that of the patients with antithrombin III, protein C, and protein S deficiency combined.

In a later study of 474 consecutive patients who presented with objectively diagnosed episodes of deep venous thrombosis, Koster and associates [22] found that 4.6% of the patients were deficient in protein C on the basis of a single measurement. The proportion decreased to 3.1% when two measurements were taken. When stringent diagnostic criteria were applied, this study found that the relative risk for thrombosis in patients with protein C deficiency compared with normal controls was 6.5. The results for patients deficient in antithrombin III were similar, although less pronounced. This study did not find an increased risk for thrombosis in patients with low protein S levels.

In their controls, Koster and associates [22] found that the prevalence of protein C deficiency in the Netherlands was 0.2% on the basis of two measurements. This prevalence is close to the figure reported in more than 5000 healthy blood donors in the United States who had no evidence of thrombosis [23], and it contrasts strikingly with the 5% to 10% prevalence of the heterozygous factor V Leiden mutation seen in healthy populations of European descent [5]. Allaart and colleagues [24] found that about half of all first episodes of venous thromboembolism in patients with a heterozygous protein C deficiency occurred in the absence of a predisposing event, such as surgery or pregnancy. By the age of 45 years, half of the 77 heterozygotes in their study had developed a manifestation of venous thromboembolism. They concluded that a protein C mutation was associated with an increased risk for venous thrombosis and that anticoagulant prophylaxis should be considered on an individual basis for patients with such a mutation.

The same group [25] also examined the survival of patients with heterozygous protein C type 1 deficiency in several families studied over many years and found no evidence that this group had an increased mortality rate compared with the general population. Earlier, they had come to similar conclusions in patients with hereditary antithrombin III deficiency and had concluded that although anticoagulant prophylaxis is likely to reduce morbidity by preventing thromboembolism in patients with antithrombin III and protein C deficiency, no evidence currently suggests that such treatment improves overall survival [26]. It is well to bear this in mind when advising patients who have an identified anticoagulant defect. Although these patients should clearly be given prophylaxis during periods of acquired hypercoagulability, such as pregnancy or surgery, whether they need life-long anticoagulant therapy is still unsettled.

High plasma homocysteine levels have recently been reported as a risk factor for deep venous thrombosis in a population study [27]; patients with these levels had a matched odds ratio of 2.5 (CI, 1.2 to 5.2) compared with controls. The risk was found to increase with age and was more pronounced in women than in men [27]. The researchers concluded that the risk for deep venous thrombosis in patients with hyperhomocystinemia and the factor V Leiden mutation did not potentiate each other. In contrast, in 45 members of seven unrelated consanguineous kindreds in which at least one member was homozygous for homocystinuria, venous or arterial thrombosis (or both) developed in 6 of 11 patients with homocystinuria; all 6 also had the factor V Leiden mutation [28]. Three women who were heterozygous for both homocystinuria and the factor V Leiden mutation had recurrent fetal loss and placental infarctions. The investigators concluded that screening for factor V Leiden mutation may be indicated in patients with homocystinuria and their family members before they have any procedures that carry a prothrombotic risk.

These recent findings in patients with venous thromboembolism raise the hope that more cases of thrombosis will eventually be susceptible to molecular analysis. It has been suggested that because of the marked phenotypic variability of the thrombophilias, multigene interactions are involved; prothrombotic mutations in two or more genes create an inherited predisposition to thrombosis [29]. In the presence of such prothrombotic stimuli as pregnancy or surgery, these inherited predispositions manifest as clinical thrombosis [1]. The search for genetic molecular defects continues apace, and further defects will undoubtedly be identified. However, many if not most patients with thrombosis have no demonstrable blood abnormality and are still a puzzling challenge.

Assessing the Risk

Both epidemiologic and genetic approaches are valid in assessing hypercoagulable patients, but the basis on which a course of action is recommended should be recognized. An assessment of known risk factors, such as age, immobility, and family history, should be combined with laboratory testing for specific inherited coagulation factor defects. When a particular genetic defect has been identified, adapting the knowledge of what happens in populations to an individual patient requires careful judgment. For example, a young woman with a factor V Leiden mutation who is taking oral contraceptives may be as much at risk for venous thromboembolism as an elderly woman having her hip replaced. In the former example, the risk is lifelong and the patient should be dissuaded from using oral contraceptives; in the latter, the risk is finite, and prophylaxis can be discontinued after the patient is mobile. Usually, a patient presents after an initial thrombotic episode has occurred and, if a genetic defect is identified, a decision has to be made about how extensively members of the family should be investigated. Most physicians treat individual persons and not populations, and they are rightly concerned that treatment should not add to their patients' burden. Unfortunately, there is currently no certain answer to the question of whether the potential hazard of recurrent thromboembolism outweighs the risk for bleeding from indefinite warfarin therapy.

Bauer [30] has drawn up useful guidelines for the management of patients with hereditary defects predisposing to thrombosis. He recommends that persons at high risk-such as patients who have had two or more spontaneous thromboses, one spontaneous life-threatening thrombosis, one thrombosis at an unusual site, or one thrombosis in the presence of more than one genetic defect-should be given anticoagulant therapy indefinitely. Persons at moderate risk are defined as those who have had one thrombosis in response to a prothrombotic stimulus and those who have remained asymptomatic. In the latter, vigorous prophylaxis should be given during high-risk situations [30]. A joint Norwegian-Swedish expert group recommended the use of oral anticoagulant therapy on an indefinite basis in patients with a demonstrated deficiency of antithrombin III, protein C, and protein S and in patients with recurrent venous thromboemboli [31]. Although these are all helpful recommendations, they are inevitably empirical and are not based on the results of prospective clinical trials. Many would argue that the case for indefinite prophylaxis has not been proven, especially in patients who have not had thromboemboli. In patients with a genetic defect who have no prothrombotic stimuli and have remained asymptomatic, the clinician should assess and discuss the possible risks individually. The approach that the authors adopt in patients with a personal or family history of venous thromboembolism is to screen for deficiencies of antithrombin III, protein C, and protein S; resistance to activated protein C; and homocystinuria. If the results of screening for protein C or protein S deficiency are negative, anticoagulant prophylaxis is used for patients having surgery (especially if the surgery is for cancer, hip replacement, or knee replacement and if the patient is older than 40 years of age). If the results of screening for these deficiencies are positive and if the patient is having surgery or is pregnant, anticoagulant prophylaxis is instituted and continued at least until the patient is fully mobile or has delivered the baby. If the results of screening for resistance to activated protein C are positive and the patient has had spontaneous multiple episodes or life-threatening episodes of venous thromboembolism, indefinite anticoagulant therapy is considered. If the results of screening for resistance to activated protein C are positive and the patient requests oral contraceptives, alternate methods of contraception are recommended.

Acquired Hypercoagulable States

Other examples of persons with hypercoagulable states include patients with the antiphospholipid syndrome who can develop both arterial and venous thrombi. However, the relation between the presence of antiphospholipid antibodies and subsequent thrombosis are complex [32, 33], and the heterogeneity of the clinical syndromes of patients who develop lupus anticoagulants make it difficult to generate a general recommendation about management. Derksen and associates [34] have recommended that patients with antiphospholipid antibodies who develop even one episode of venous thrombosis should be given long-term anticoagulant therapy. Patients who have the presence of a persistent lupus anticoagulant or a high-titer antiphospholipid antibody are candidates for anticoagulant prophylaxis during periods of acquired hypercoagulability.

Summary

The importance of a genetic component in the pathogenesis of venous thrombosis has been increasingly recognized in recent years, and about one in five patients of European descent who present with venous thrombosis have a specific defect in their anticoagulant pathways. Patients with a personal or family history of venous thromboembolism should now be screened for these defects before surgery or childbirth and should be given anticoagulant prophylaxis when they are exposed to prothrombotic stimuli.

Arterial Thrombosis

In patients with arterial thrombosis, tests for defects in the anticoagulant pathways are of little help. Although arterial thrombi have been described, most patients with positive results on tests for antithrombin III, protein C, or protein S deficiency and resistance to activated protein C develop venous thromboembolism. Conversely, patients with arterial thrombosis tend to have normal results on such tests because they have a different disease, namely, atherosclerosis. They may develop thrombosis as a complication of diffuse or localized atheroma; however, such patients do not primarily have disorders of their anticoagulant pathways, and screening for these defects produces a low yield. Patients with arterial thrombosis are tested for hypercholesterolemia and other lipoprotein abnormalities, and evidence is sought for accelerated atherosclerosis resulting from such conditions as hypertension, hyperlipidemia, diabetes, or homocystinuria. Genetic abnormalities in platelet function were also recently reported to play a part in arterial thrombosis. A linkage between one allele of a polymorphism in a platelet glycoprotein IIIa polypeptide and the risk for early-onset coronary artery disease has been described [35]. The odds ratio for coronary thrombosis was sixfold higher in patients younger than 60 years of age who have this polymorphism compared with a control group.

Hypercoagulability and Coronary Heart Disease

Increasing evidence indicates that hypercoagulability contributes to the pathogenesis of coronary heart disease. Several prospective studies [36, 37] have consistently shown that a direct, independent, and statistically significant association exists between fibrinogen levels and the subsequent incidence of heart disease. Evidence is accumulating to show that this association is about as strong as the relation between cholesterol levels and heart disease [38, 39]. In a recent prospective study of more than 3000 patients having coronary angiography who were followed for 2 years [40], an increased incidence of myocardial infarction or sudden death was associated with higher baseline concentrations of fibrinogen, von Willebrand factor antigen, and tissue plasminogen activator antigen. The researchers concluded that an increased fibrinogen concentration is a strong and independent predictor of cardiovascular risk in apparently healthy persons as well as in persons with manifest coronary artery disease. Low fibrinogen concentrations were associated with a low risk for new coronary events, even in patients with high serum cholesterol levels. The authors make the point that the association of higher concentrations of both fibrinogen and C-reactive protein with increased coronary risk suggests that the fibrinogen concentration is elevated as a consequence of inflammatory reactions occurring in progressive atherosclerosis. The elevation of tissue plasminogen activator antigen and von Willebrand factor antigen, both released by endothelial cells, also points to vessel-wall perturbation in such patients. However, whether elevated fibrinogen is a consequence of active vascular disease or whether such elevations mediate the development of coronary artery disease remains unclear.

The Northwick Park Heart Study in London [37] was the first large-scale population study to report a significant correlation between elevated levels of clotting factor VII and the risk for subsequent ischemic heart disease. Other studies have also reported a relation between elevated plasma factor VII levels and arterial thrombosis [41]. Ruddock and Meade [42] suggested that factor VII activity may be more strongly related to fatal events of ischemic heart disease than to nonfatal events. They followed almost 1500 white men 40 to 64 years of age for a mean of 16 years. In these patients, increased factor VII activity was strongly related to fatal events of ischemic heart disease but not to nonfatal events. The investigators suggested that high levels of factor VII activity may influence outcome at the time of plaque rupture and tissue factor release by enhancing thrombin production and consequent fibrin deposition and platelet aggregability. Significant associations between factor VII and indices of thrombin generation, such as fibrinopeptide A and prothrombin activation peptides, have also been reported [43]. Miller and colleagues [44] have recently reported further evidence of increased activation of the hemostatic system in men at high risk for fatal coronary heart disease. In almost 3000 men aged 50 to 61 years, six markers of hemostatic status (factor VII coagulant activity, factor VII antigen, activated factor VII, factor IX activation peptide, prothrombin fragment 1+2, and fibrinopeptide A) were all positively and significantly associated with risk, providing further evidence for a hypercoagulable state in men at high risk for fatal coronary heart disease.

The assay of factor VII has complicated the interpretation of some of these findings, and it is still not known whether increased factor VII activity is due to increased amounts of the inactive zymogen, the presence of the activated enzyme, or both [45, 46]. New assay techniques based on soluble mutant tissue factor now permit the measurement of activated factor VII without interference from the normal large excess of zymogen factor VII in plasma [47]. It is hoped that ongoing prospective studies will clarify the position and define the relative roles of elevated concentrations of factor VII and activated factor VII as potential risk factors in coronary artery disease [48]. The results of large-scale, ongoing, prospective trials with low-intensity oral anticoagulant therapy aimed at modifying some of these risk factors are awaited with much interest [49].

Role of Fibrinolysis

Disorders of the fibrinolytic system are also a possible causal mechanism in the development of the hypercoagulable state in relation to both arterial and venous thrombosis. Impaired fibrinolysis may be due to 1) a defective synthesis or release, or both, of tissue plasminogen activator from the vessel wall; 2) a deficiency or functional defect in the plasminogen molecule; or 3) increased levels of inhibitors of tissue plasminogen activator or of the fibrinolytic enzyme plasmin. Although defective fibrinolysis, from whatever cause, can be considered to contribute to the prothrombotic state, deficiencies that lead to an impaired fibrinolytic response remain less well characterized than the anticoagulant pathways. Prins and Hirsh [50] concluded that a causal relation between impaired fibrinolysis and symptomatic venous thrombosis had not been established. However, they found good evidence for an association between impaired fibrinolytic activity measured either before or after surgery and increased risk for postoperative thrombosis. In arterial disease, the main mechanism for the fibrinolytic impairment seen frequently in patients with coronary artery disease is increased plasminogen activator inhibitor-1 activity in plasma. However, as Wiman [51] pointed out, before plasminogen activator inhibitor-1 can be regarded as a primary epidemiologic risk factor and not an acute phase reactant, its relation to myocardial infarction must be shown in prospective studies in healthy populations. He concluded that although the fibrinolytic enzyme system must be considered in forthcoming prospective studies, a reduced fibrinolytic potential cannot yet be considered an established risk factor in the conventional epidemiologic sense.

Summary

The case for measuring fibrinogen levels (as an additional indicator of the likelihood of subsequent coronary events) when assessing patients at risk for acute myocardial infarction is now a good one. Pending further clarification of assay methodology, the predictive value of factor VII determinations remains uncertain in general clinical practice, and the same is true for assays of the fibrinolytic system. No evidence indicates that the measurement of endogenous anticoagulant pathways has a useful role to play in assessing patients at risk for arterial thrombosis.