Lawrence L.K. Leung M.D.1
1Professor of Medicine and Chief, Division of Hematology, Department of Medicine, Stanford University School of Medicine
The author has no commercial relationships with manufacturers of products or providers of services discussed in this chapter.
Thrombosis is more than excessive blood clotting; it also involves vascular inflammation. The classic triad of Virchow identifies three major elements in the pathophysiology of thrombosis: endothelial injury, a decrease in blood flow, and an imbalance between procoagulant and anticoagulant factors.
Endothelial cells can be activated or injured by a variety of stimuli, including mechanical trauma, endotoxins and cytokines, proteases, inflammatory mediators, immune complex deposition, oxygen radicals, and hypoxia. Each of these stimuli affects multiple facets of endothelial cell function, ultimately changing the cell from its natural antithrombotic state to a prothrombotic one.
The vascular endothelium, in its unique location in the vessel wall, is capable of sensing and responding to the different mechanical forces in the blood circulation. The shear stress caused by the friction from blood flow seems to be particularly important in modulating endothelial functions. In areas of linear flow, the blood moves in ordered laminar patterns in a regular pulsatile fashion. Such a steady, laminar blood flow apparently promotes an antithrombotic endothelial phenotype. In areas of disrupted flow, such as at vascular bifurcations or stenoses, the endothelium may be exposed to significant changes in shear gradients, and the cells may become activated and prothrombotic.
Imbalance between procoagulant and anticoagulant factors can be hereditary or acquired [see Table 1]. Some clotting factors, such as factor VIII and fibrinogen, are acute-phase reactants: their plasma levels increase significantly with acute inflammation, possibly conferring a transient prothrombotic state. Some hereditary deficiencies of anticoagulant proteins, such as factor V Leiden and antithrombin (AT), are associated with recurrent thrombosis; these are among the best understood clinical hypercoagulable states.
Table 1 Inherited and Acquired Hypercoagulable States
Although the hypercoagulable state is systemic, thrombosis occurs locally (e.g., in the lower extremities). The clinical outcome likely reflects a complex interaction between the systemic prothrombotic predisposition and local hemostatic control mechanisms specific to the vascular bed. Specific and distinctive gene transcript expression patterns in different vascular beds have now been demonstrated.1
Assessment of Patients with Thrombotic Disorders
PRIMARY CLINICAL ISSUES IN THROMBOTIC DISORDERS
Important questions in the assessment of thrombosis include the following: (1) How likely is it that the thrombosis is caused by an underlying hypercoagulable state? (2) How extensive a workup is indicated? (3) When should the workup be done? Answers to these questions come from a consideration of the patient's age at the time of the first thrombosis; presence or absence of a provoking factor; family history and past medical history of response to situations associated with high risk of thrombosis; and the site, type, and severity of the thrombosis.
Age of Onset of First Thrombosis
In a retrospective study involving 150 families with an inherited predisposition to recurrent thrombosis (thrombophilia), the mean age at the time of the first thrombosis was 35 to 40 years. However, the first episode of thrombosis can occur as early as the second decade of life if the patient has more than one hereditary risk factor.2
Presence or Absence of a Provoking Factor
Common triggers of thrombosis are surgery, trauma, pregnancy, malignancy, prolonged immobilization, and infection. Malignancy or infection can be clinically overt or subclinical. Those circumstances can provoke thrombosis even in persons with a normal coagulation system, and they often uncover a thrombophilia that had been clinically silent.2 However, sometimes no provoking factors can be identified. Such a spontaneous, idiopathic thrombosis, especially when it occurs in a young person, strongly suggests an underlying hereditary hypercoagulable state.
If thrombosis develops in a patient who has had previous pregnancies or surgeries (especially orthopedic procedures) without any thrombotic complications, an acquired hypercoagulable state should be considered. Likely conditions in such cases are antiphospholipid antibody syndrome or Trousseau syndrome (see below).
Objectively documented venous thromboembolism before 50 years of age in a first-degree family member strongly suggests a hereditary thrombotic disorder. However, a negative family history does not exclude a hereditary condition. Clinical thrombosis is frequently the culmination of multiple thrombogenic risk factors, only one of which may be hereditary and irreversible. In patients with symptomatic thrombosis and well-documented hereditary hypercoagulable states, it is not uncommon to find other family members with the same deficiency but no clinical thrombosis.
A patient who experiences recurrent thrombosis likely has a hypercoagulable state (hereditary or acquired). However, if a patient who initially presented with deep vein thrombosis (DVT) in a lower extremity returns with symptoms involving the same leg, the problem may be postphlebitic syndrome rather than recurrent thrombosis. Acute exacerbation of the postphlebitic syndrome, with its increased leg edema and pain, can be difficult to distinguish from recurrent acute DVT. As an anticipatory measure, it is sometimes useful to obtain a repeat compression ultrasound study of the lower extremity after resolution of an acute episode of DVT; the repeat scan can provide a baseline for future comparison.
Site of Thrombosis
Most commonly, thromboses involve the deep veins of the lower extremities. Thrombosis at an atypical site, such as the hepatic, mesenteric, or cerebral veins (or skin necrosis after warfarin administration), increases the likelihood of an underlying hypercoagulable state. Spontaneous axillary vein thrombosis may also indicate the presence of an underlying hypercoagulable state, but this association is controversial.2,3,4
Recurrent thrombosis at arterial sites has a differential diagnosis—and therefore a workup—that is quite different from that for recurrent venous thrombosis [see Table 2]. Most of the common hereditary hypercoagulable states (e.g., AT deficiency or factor V Leiden) are associated with venous thromboses, such as DVT in the lower extremities. They are seldom associated with arterial thromboses, such as transient ischemic attack, stroke, digital ischemia, and myocardial infarction. A few hypercoagulable states, such as the antiphospholipid syndrome and hyperhomocysteinemia, are associated with both types of thrombosis.
Table 2 Screening Tests for Patients with Suspected Hypercoagulable State
The Hypercoagulable Workup
Extent of the workup
On the basis of the above clinical considerations, one may estimate the likelihood of an underlying thrombophilia in a given patient with thrombosis [see Table 3]. Because studies of cost-effectiveness and outcomes are not available, it is difficult to list strict practice guidelines regarding the extent of the hypercoagulable workup. In general, however, if the likelihood of an underlying hypercoagulable state is high, an extensive workup is warranted.
Table 3 Clinical Features That Suggest Thrombophilia
A limited workup is appropriate for mild to moderate DVT of the lower extremities with an obvious provoking factor. For example, in a young woman who experiences DVT in the superficial femoral vein while on an oral contraceptive, evaluation of factor V Leiden, prothrombin mutation 20210A, AT, protein C, protein S, homocysteine, anticardiolipin antibodies, and lupus anticoagulant may be sufficient. On the other hand, an acquired hypercoagulable state should be considered in an elderly patient with a spontaneous DVT and no history of previous thrombosis. Diagnostic possibilities in such cases would include antiphospholipid antibody syndrome, acquired AT deficiency (if the patient has evidence of nephrotic syndrome), or Trousseau syndrome.
When the clinical history strongly suggests thrombophilia—as in a patient with recurrent thrombosis or thrombosis at atypical sites—one may argue that a workup for an underlying hypercoagulable state is unnecessary because the result will not alter the management of the case. However, the identification of any underlying risk factors will improve the understanding of the disease for both the patient and the treating physician; and it will guide the counseling of the patient, especially regarding the need for screening of related family members. In the case of hyperhomocysteinemia and elevated lipoprotein(a) levels, identification of risk factors will permit the use of specific therapies (see below).
Timing of the workup
The clinician needs to know not only what tests to order but when to order them. In acute thrombosis, many inhibitors of the clotting cascade (e.g., AT and protein C) are consumed. Immediately after the episode, their plasma levels may be decreased, even in patients who do not have a hereditary deficiency. Heparin therapy can reduce antithrombin levels up to 20%, whereas warfarin treatment reduces the levels of protein C and protein S. Usually it is best to postpone measurement of these inhibitors until the acute thrombotic episode is completely resolved, preferably 4 weeks after termination of oral anticoagulation therapy. Tests for specific genotypes (e.g., factor V Leiden) can be performed at any time, however.
Frequency and Relative Risk of Venous Thromboembolism
The frequency of various hypercoagulable states in unselected patients who present with venous thrombosis ranges from 1% to 25% [seeTable 4]. It should be recognized that these thrombophilias do not confer equivalent thrombotic risk. Factor V Leiden and prothrombin mutation 20210A, the two most prevalent risk factors, confer only a modest increase in relative risk of thrombosis, approximately threefold to sevenfold above normal. Moderate hyperhomocysteinemia, another common risk factor, also carries a modest increase in risk. Heterozygous deficiencies of AT, protein C, and protein S are generally considered more significant risk factors than factor V Leiden. AT deficiency and the antiphospholipid antibody syndrome are probably the greatest risk factors. The recurrence rate in patients with antiphospholipid antibody syndrome is as high as 50% to 70% in some studies.
Table 4 Frequency and Relative Risk of Venous Thrombosis in Selected Hypercoagulable States*
Patients with symptomatic thrombosis frequently have more than one risk factor, which may have a synergistic effect in increasing the thrombosis risk. For example, women with factor V Leiden who use oral contraceptives have a risk of venous thromboembolism that is 35-fold higher than that in the general population.
Hereditary Hypercoagulable States
Epidemiology and Etiology
The frequency of symptomatic inherited AT deficiency in the general population has been estimated to be approximately 1 per 2,000 people.5 The deficiency is transmitted in an autosomal dominant pattern. Homozygous AT deficiency has not been reported, presumably because the condition is incompatible with normal fetal development.
There are two types of inherited AT deficiency. Type I is quantitative, as measured by antigenic and functional assays. A large number of molecular mutations have been characterized in type I AT deficiency, including partial gene deletions and single-nucleotide substitutions that cause nonsense or missense mutations leading to premature stop signals in the protein-translation process.
Type II deficiency is qualitative; plasma levels of AT antigen are normal. The underlying defect is generally a single nucleotide change that causes missense mutations, giving rise to a dysfunctional protein. Many of these proteins have decreased affinity for heparin binding.
In rare cases, AT deficiency is acquired. This condition may occur after administration of intravenous heparin for more than 3 days or after asparaginase therapy. It may also develop in patients with disseminated intravascular coagulation (DIC), severe liver disease, or the nephrotic syndrome.
AT inactivates factor Xa and thrombin by forming a stable stoichiometric complex with each of them. AT is present in sufficient amounts in plasma to inactivate all the thrombin formed in a given plasma volume, but it does so slowly unless it is activated by endothelial cell surface heparan sulfate or by administered heparin [see 5:XIII Hemorrhagic Disorders]. Patients with hereditary AT deficiency have evidence of continuous factor X activation and thrombin generation (as supported by elevated plasma levels of prothrombin fragment F1.2) even when they are clinically asymptomatic.
Patients with AT deficiency show an increased incidence of venous thrombosis, usually triggered by a prothrombotic stimulus such as surgery, infection, immobilization, or trauma. This association suggests that the superimposition of a prothrombotic stimulus on an underlying subclinical hypercoagulable state leads to clinical thrombosis.
Typical clinical presentations are DVT of the legs, pulmonary embolism, and occasionally mesenteric vein thrombosis. There is no convincing evidence to suggest that AT deficiency increases the risk of arterial thrombosis.6
Affected patients usually have a family history of recurrent thromboses, generally beginning in youth and often associated with surgery or trauma. Pregnancy and the use of oral contraceptives also increase the risk of thromboses in AT-deficient patients. The tendency to thrombosis increases with advancing age: by age 50, only 10% of AT-deficient patients are free of symptoms.
The AT level should be determined by a functional assay rather than an antigenic assay, so that both type I and type II deficiency can be evaluated. Patients with AT deficiency have a surprisingly modest reduction in the protein: values measured by both bioassay and immunoassay range from 25% to 60% of normal in type I disease.
Study of a large AT-deficient kindred indicates that long-term anticoagulant prophylaxis is not warranted in asymptomatic carriers of this deficiency.7 Asymptomatic carriers should receive prophylactic anticoagulation only in situations known to increase the risk of thrombogenesis, such as abdominal surgery.7 Once such patients have experienced a thrombotic event, however, they probably require lifelong warfarin therapy. Warfarin is the mainstay of long-term therapy for patients with AT deficiency and recurrent thromboembolism.
Acute episodes of thrombosis must be treated with heparin. Because AT deficiency may render heparin relatively ineffective, the physician should be alert to heparin resistance. In patients receiving unfractionated heparin, resistance is manifested by minimal prolongation of the partial thromboplastin time (PTT) despite the administration of therapeutic doses. If low-molecular-weight heparin (LMWH) is used, as is commonly the case, the level of anti-factor Xa (anti-FXa) should be checked to ensure that a therapeutic anticoagulant effect is achieved [see 5:XII Hemostasis and Its Regulation].
If heparin resistance occurs despite increased doses of heparin, heparin plus purified AT concentrates or fresh frozen plasma should be given. AT has a half-life of about 60 hours. These preparations can be used to carry an AT-deficient patient through surgery or delivery and should bring the AT level up to nearly 100%, depending on the patient's baseline AT level. The AT level should be checked and the infusion repeated at 24-hour intervals to maintain a normal AT level for 5 to 7 days after delivery or surgery.
Pregnancy in an AT-deficient patient is difficult to manage. Because warfarin may cause fetal malformations and neonatal hemorrhage, patients should be treated with full-dose unfractionated heparin or LMWH; those receiving LMWH should be switched to unfractionated heparin 1 to 2 weeks before delivery so that rapid reversal of anticoagulation, if necessary, can be more easily attained. If a therapeutic effect cannot be achieved (as measured by the PTT with unfractionated heparin or the anti-FXa level with LMWH), an AT infusion can be given.8 Generally, this is not necessary. Anticoagulation should be promptly reinstituted after delivery.
PROTEIN C AND PROTEIN S DEFICIENCY
Pathophysiology and Clinical Presentation
Deficiency or defect in protein C or protein S results in a loss of ability to inactivate excess factor VIIIa and factor Va, the two major cofactors that regulate amplification of the clotting cascade. Protein C levels are low in patients with DIC and liver disease, probably because the activation of hemostasis consumes this factor [see 5:XIII Hemorrhagic Disorders].
Homozygous protein C deficiency causes lethal thrombosis in infants. Heterozygous protein C deficiency probably occurs with a prevalence of 1 per 200 to 300 in the general population. Clinical expression of heterozygous protein C deficiency varies: many persons with heterozygous deficiency, as well as persons with low-normal protein C levels from other causes, do not experience thrombosis,9 whereas other patients with heterozygous deficiency exhibit a definite tendency toward venous thrombosis even though their protein C levels are 40% to 50% of normal. This phenotypic variability suggests multiple gene interactions and supports the hypothesis that clinical thrombosis in such patients may result from a combination of protein C deficiency and one or more other prothrombotic mutations.10 Cerebral venous thrombosis presumably accounts for cases of cerebral hemorrhagic infarction that occur in young adults with protein C deficiency.
Deficiency of protein S also leads to venous thrombosis, including mesenteric vein thrombosis. Pregnancy and the use of oral contraceptives lower the protein S level, which may account for some cases of thromboembolism that occur under such circumstances.11 Acquired protein S deficiency also occurs in patients with the nephrotic syndrome, who lose protein S in urine.12 Case reports have associated protein S deficiency with warfarin-induced skin necrosis.13
Functional and antigenic assays for protein C and protein S are now available in most coagulation laboratories. Functional assays are preferable for diagnosis. It is important to measure free protein S because some patients who have low free protein S levels have normal or borderline total protein S levels. Coagulation assays for protein C and protein S can give falsely low values in patients with factor V Leiden.14
Warfarin is the treatment of choice for preventing thrombosis, even though it lowers protein C levels still further. Because the half-life of protein C is only 6 to 7 hours, much shorter than that of prothrombin and factor X, a period of enhanced hypercoagulability follows initiation of warfarin therapy in patients with protein C deficiency. Heparin should be given along with warfarin during the initiation of anticoagulation; it can be withdrawn afterward. Warfarin-induced skin necrosis is a rare complication of anticoagulation therapy.
FACTOR V LEIDEN
Epidemiology and Etiology
Factor V Leiden is a mutated form of factor V (first identified by researchers in Leiden, The Netherlands) that, once activated, is relatively resistant to the anticoagulant effects of activated protein C (APC). The defect is transmitted as an autosomal dominant trait. Approximately 5% of the general white population is heterozygous for factor V Leiden; the defect is almost absent in other ethnic groups.15 Factor V Leiden is now considered to be the most common hereditary hypercoagulable state. Its prevalence in patients with thrombophilia is as high as 20% to 50%.16 In a large cohort study of unselected patients with a first episode of symptomatic DVT, factor V Leiden was found in 16% of patients.17 In women who have thrombosis while taking oral contraceptives, the frequency of factor V Leiden is about 23%.18 The relative risk of DVT in a factor V Leiden homozygote (estimated incidence, 0.5% to 1% a year) is approximately 80-fold higher than in a normal person.19 The risk of thrombosis in persons who are heterozygous for factor V Leiden is estimated to be fourfold to eightfold higher than that in normal persons; the relative risk increases to more than 30-fold when factor V Leiden is combined with oral contraceptive use. The absolute risk of thrombosis, however, is low.20 Association of factor V Leiden with deficiencies of protein C, protein S, or AT has been reported in some families.21,22,23 Overall, although factor V Leiden is highly prevalent, it is a relatively weak risk factor for thrombosis.
Approximately 5% of cases associated with inherited resistance to APC are attributable to other mutations and defects.24,25,26 Conditions such as factor VIII elevation, pregnancy, oral contraceptive use, and lupus anticoagulant may result in APC resistance.27 APC resistance that is not caused by factor V Leiden may be a risk factor for stroke28,29 and venous thrombosis.30,31 The overall risk of venous thrombosis from APC resistance is similar to or less than that posed by factor V Leiden.32
Resistance to the anticoagulant effects of APC is caused by a specific mutation in factor V (factor V Leiden or factor V R506Q) that results from a single-nucleotide substitution that leads to the replacement of arginine with glutamine at position 506.32 Arginine 506 is located at one of the two major APC cleavage sites of activated factor V. Activated factor V Leiden expresses normal procoagulant activity, but its degradation by APC is approximately 10 times slower than that of normal activated factor V (factor Va). This slowing leads to increased thrombin generation.33 In addition, recent evidence suggests that factor V (but not factor Va), together with protein S, serves as a cofactor of APC in the inhibition of the factor VIIIa/factor IXa complex and that factor V Leiden has a poor APC cofactor function [see Figure 1].
Figure 1. Degradation of Factor V Leiden
Degradation of thrombin-activated factor V Leiden by activated protein C (APC) is significantly slower than that of normal activated factor V (factor Va), which leads to enhanced thrombin generation (left). Recent evidence suggests that normal factor V, together with protein S, serves as a cofactor of APC in the inhibition of factor VIIIa (right). This APC cofactor function of factor V requires the cleavage of factor V by APC at arginine 506; therefore, factor V Leiden has a poor cofactor function.
Clinical manifestations of factor V Leiden are similar to those of deficiencies of AT, protein C, and protein S—mainly, venous thrombosis. However, the first thrombotic manifestation in factor V Leiden often occurs later than in the other hereditary thrombophilic states. Approximately 25% of apparently healthy men older than 60 years who experience a first episode of venous thrombosis have factor V Leiden.34 There are conflicting data on whether factor V Leiden is associated with an increased risk of recurrent deep vein thrombosis. Several studies reported a slightly enhanced recurrence risk (twofold to fourfold), but more recent studies have shown that the risk of recurrence is similar to that in persons without the mutation.17,35,36
Factor V Leiden can be identified rapidly and precisely with simple DNA-based tests. These tests allow the diagnosis to be made in patients receiving anticoagulation therapy with warfarin and in those who have coexisting antiphospholipid antibodies. Because factor V Leiden is not the sole cause of APC resistance, it may be worthwhile to pursue the diagnosis with an APC-resistance test in selected cases.
Management of factor V Leiden is similar to that of AT, protein C, and protein S deficiencies. Patients with a first episode of venous thrombosis should receive anticoagulation therapy for 6 months. Thereafter, they should be given prophylactic anticoagulation therapy in situations known to provoke thrombosis. Long-term anticoagulation should be considered in patients with recurrent thrombosis.37
Young women known to be factor V Leiden carriers should avoid the use of oral contraceptives, which increases the relative risk of thrombosis (although the risk remains low in terms of absolute incidence). The optimal treatment of carriers during pregnancy has not been established. The rate of venous thromboembolism is low, about 2% without thrombosis prophylaxis.20 My practice is not to use thrombosis prophylaxis routinely during pregnancy, but I will consider postpartum prophylaxis for 6 weeks, especially when the family history of thrombosis is strong. Routine screening of family members of patients with factor V Leiden is not cost-effective.38
PROTHROMBIN GENE MUTATION 20210A
A G-to-A mutation at nucleotide position 20210 in the 3′ untranslated region of the prothrombin gene is associated with an increased incidence of venous thrombosis. The prevalence of the mutation in healthy persons is about 2.3%. Like factor V Leiden, this mutation is very rare in Asians and Africans. Unlike factor V Leiden, it is more common in southern Europeans than in northern Europeans.39 The relative risk of thrombosis in persons with this mutation is 2.8, which is similar to the relative risk in those with factor V Leiden.40 The mutation can be found in up to 18% of patients with thrombosis and family histories of thrombosis. The most common presentation is DVT of the lower extremities. Prospective studies have not shown an increased risk of recurrent DVT in patients with this mutation.41 However, carriers who are heterozygous for both factor V Leiden and the prothrombin mutation have a higher risk of recurrent thrombosis.35 The combination of oral-contraceptive use and the prothrombin gene mutation is associated with an increased incidence of cerebral vein thrombosis in young women.42
Homocysteine is a highly reactive amino acid that is normally found in blood at levels of 5 to 15 mmol/L. Normally, homocysteine is derived from methionine by a transmethylation process and is remethylated to methionine or converted to cysteine [see Figure 2]. Metabolism of homocysteine requires betaine, cobalamin (vitamin B12), folate, and pyridoxine (vitamin B6).
Figure 2. Intracellular Metabolism of Homocysteine
Homocysteine's intracellular metabolism occurs through remethylation to methionine or transsulfuration to cysteine.142 Elevation in plasma homocysteine levels can result from hereditary deficiency in cystathionine β-synthase (CBS); a thermolabile mutant of methylene-tetrahydrofolate reductase (MTHFR); or low dietary levels of cobalamin (vitamin B12), folate, or pyridoxine (vitamin B6), which are essential cofactors in the metabolic process. (BHMT—betaine-homocysteine methyltransferase)
Homocysteine can promote oxidation of low-density lipoprotein (LDL) cholesterol and presumably is toxic to vascular endothelium.43,44 It may also inhibit thrombomodulin expression and protein C activation and suppress endothelial heparan sulfate expression; both of these effects lead to hypercoagulability.45,46 Homocysteine also enhances the binding of lipoprotein(a) (an atherogenic lipoprotein) to fibrin, which may provide a link between hyperhomocysteinemia, thrombosis, and premature atherosclerosis [see Lipoprotein(a), below].47 The vascular damage caused by high homocysteine levels leads to arterial and venous thrombosis and, perhaps, accelerated atherosclerosis.
Epidemiology and Etiology
Hyperhomocysteinemia can be divided into three classes: severe (homocysteine plasma concentration > 100 mmol/L), moderate (25 to 100 mmol/L), or mild (16 to 24 mmol/L). Severe hyperhomocysteinemia is usually caused by a homozygous deficiency of the enzyme cystathionine β-synthase. Affected persons have severe mental retardation, ectopic lens, skeletal abnormalities, and severe early-onset arterial and venous thrombotic disease.48
Mild or moderate hyperhomocysteinemia results from either hereditary or acquired defects in the homocysteine metabolic pathway. Heterozygous deficiency in cystathionine β-synthase is quite common in the general population, with a frequency of 0.3% to 1.4%.48 A defect in the remethylation pathway is commonly caused by a thermolabile mutant of the methylene-tetrahydrofolate reductase (MTHFR) enzyme whose activity is approximately 50% of normal; the homozygous state has a prevalence of 5% in the general population.49 However, the homozygous form of the MTHFR thermolabile enzyme isoform is not clinically relevant in patients whose diet includes adequate folate.
Common causes of acquired hyperhomocysteinemia are deficiencies of dietary cobalamin, folate, or pyridoxine. A prospective study found that mild hyperhomocysteinemia is quite common in the elderly, despite normal serum vitamin concentrations.50 Acquired hyperhomocysteinemia is also common in patients with end-stage renal disease.
Mild to moderate hyperhomocysteinemia is associated with cerebrovascular disease, coronary artery disease, and peripheral vascular disease in persons younger than 55 years and with carotid artery stenosis in the elderly.51,52 It is found in 10% of patients with a first episode of DVT.53 In a prospective study, a graded relationship was found between elevated plasma homocysteine levels and mortality in patients with coronary artery disease.54
Severe hyperhomocysteinemia should be suspected in patients with the characteristic phenotype (see above). Mild to moderate hyperhomocysteinemia should be suspected in cases of arterial and venous thrombotic disease—including cerebrovascular disease, peripheral arterial disease, and DVT—especially in young persons.
Plasma homocysteine exists in free and protein-bound forms and is generally measured and reported as total plasma homocysteine (normal range, 5 to 15 mmol/L). Diagnosis of hyperhomocysteinemia is usually made by measuring plasma homocysteine levels after an overnight fast. Because as many as 40% of patients with hyperhomocysteinemia may have a normal fasting level, a methionine-loading test should be considered when indicated.55 However, methionine is not generally available in most pharmacies. Plasma folate and vitamin B12 levels should also be measured to exclude hyperhomocysteinemia caused by folate or B12 deficiencies.
Daily use of oral pyridoxine (250 mg) and folic acid (5 mg) brings elevated homocysteine levels down to normal in most cases.56 Patients who have vitamin B12 deficiency should be given B12 supplements. Repeat measurement of plasma homocysteine levels (generally done 1 month after starting supplementation) may be prudent to ensure that the treatment with pyridoxine and folate is working. Betaine (3 g p.o., b.i.d.) is sometimes effective in patients with hyperhomocysteinemia that is resistant to pyridoxine and folate. It is currently unknown whether correction of hyperhomocysteinemia by these measures leads to clinical benefit.
Lipoprotein(a) [Lp(a)] is an independent risk factor for coronary artery thrombosis.57 A prospective case-control study associated elevated plasma Lp(a) levels with an approximately threefold increase in risk of coronary artery disease in men.58 The association between high Lp(a) and ischemic stroke in young adults is controversial.59,60 Distributions of Lp(a) are skewed in the general population—especially among whites, in whom the median is 3.7 mg/dl but the mean is 6.9 mg/dl.61 The 95th percentile for plasma Lp(a) is estimated to be in the 25 to 30 mg/dl range.
The Lp(a) class of lipoproteins is formed by the assembly of LDL particles and apoprotein(a), a protein that has some structural similarities to plasminogen (specifically, in the kringle domains) and competes with plasminogen for the endothelial cell binding site, thereby displacing plasminogen and downregulating plasmin generation at the endothelial cell surface.62 High plasma concentrations of Lp(a) may therefore suppress the endothelial fibrinolytic response. Lp(a) is found in the intima of human atherosclerotic vessels, and transgenic mice expressing human Lp(a) develop extensive atherosclerosis.63 Measurement of Lp(a) levels can be done in commercial laboratories and should be considered in young patients with arterial thrombosis. Elevated LDL cholesterol levels appear to elicit or exacerbate the risk factors associated with high Lp(a), and therefore, diet, exercise, and standard pharmacologic approaches should be used in patients with high LDL cholesterol levels.64 In small studies, niacin at high doses (2 to 4 g p.o. daily) and tamoxifen (20 mg daily) have lowered elevated Lp(a) levels by 30% to 40%.65,66 High doses of niacin are frequently associated with facial flushing and headaches. These unpleasant side effects can be ameliorated by starting niacin at a low dose (e.g., 300 mg daily) and then increasing the dose incrementally over time or through the use of extended-release niacin. Liver function should be checked periodically.
Approximately 300 abnormal fibrinogens (dysfibrinogens) have been reported, and about 85 structural defects have been identified in dysfibrinogenemia. These are most commonly characterized by functional defects of fibrinopeptide A release and fibrin polymerization and less commonly by defective plasminogen binding and activation. About half of the fibrinogen mutations are not associated with any clinical symptoms. Mild bleeding or recurrent thrombosis occurs in about equal numbers in the remaining mutations.67 In rare cases, patients experience both bleeding and thrombosis. Acquired dysfibrinogenemia may complicate hepatocellular carcinoma or chronic liver disease. Evaluation in a general laboratory usually shows a discrepancy between antigenic and functional levels of fibrinogen, because most patients with dysfibrinogens have suboptimal clotting function, with prolonged thrombin time (TT) and reptilase time (RT). The abnormal fibrinogens form fibrin clots that are resistant to clot lysis. Precise identification of the structural defect requires substantial effort in a research laboratory. Management of recurrent thrombosis caused by dysfibrinogenemia is the same as that in other patients with thrombophilia.
DYSPLASMINOGEN AND ABNORMAL FIBRINOLYSIS
In rare cases, abnormal plasminogens (dysplasminogens), which are defective in their activation to plasmin, are associated with thrombosis. Patients with such a disorder have a low plasma plasminogen level on functional assays.68 Increased levels of plasminogen activator inhibitor-1 (PAI-1) and decreased plasma fibrinolytic activity have been reported in patients with preeclampsia.69 Acquired impairment of fibrinolytic activity may be associated with postoperative thrombosis.70 However, more studies are required to establish the role of abnormal fibrinolysis in recurrent clinical thrombosis.71 Antigenic assays for tissue plasminogen activator (t-PA) and PAI-1 are available in some commercial laboratories, but specific functional assays are available only in research laboratories.
ELEVATED FIBRINOGEN, FACTOR VII, AND FACTOR VIII LEVELS
A high plasma fibrinogen level is an independent risk factor for coronary artery disease.72 An elevated factor VII level has also been associated with the development of heart disease.73 A factor VIII level above the 90th percentile of normal is associated with an approximately fivefold increased risk of a first episode of DVT74,75; it also increases the risk of recurrence.76 Additional studies are required to establish the clinical utility of measuring these parameters in patients with thrombosis.
Acquired Hypercoagulable States
ANTIPHOSPHOLIPID ANTIBODY SYNDROME
The antiphospholipid antibody syndrome is caused by autoantibodies to proteins associated with negatively charged phospholipids. The terms antiphospholipid and anticardiolipin are used synonymously. Antiphospholipid antibodies also include lupus anticoagulant, which is an inhibitor that was first identified in patients with systemic lupus erythematosus. Many patients who have this inhibitor do not have lupus, and it is sometimes called lupuslike anticoagulant.
Antiphospholipid antibody syndrome occurs secondary to systemic lupus erythematosus and, less commonly, to rheumatoid arthritis, temporal arteritis, and other connective tissue disorders. It is also associated with HIV-1 and hepatitis C infections, lymphoproliferative diseases, and certain drugs (e.g., phenothiazine and procainamide). When no risk factor can be identified, the syndrome is regarded as primary. In a large cohort study of 1,000 patients with antiphospholipid antibody syndrome, 53% of patients were classified as having primary antiphospholipid antibody syndrome, and 47% had secondary antiphospholipid antibody syndrome.77
The two most common protein targets for the antiphospholipid antibodies appear to be β2-glycoprotein I (β2-GPI) and prothrombin. β2-GPI is a plasma protein that binds anionic phospholipids with high affinity. It has weak anticoagulant function in vitro. β2-GPI can induce cardiolipin from its usual bilaminar form to a hexagonal form that is highly immunogenic.78 The anticardiolipin antibody enzyme-linked immunosorbent assay (ELISA) usually detects antibodies directed against the cardiolipin/β2-GPI complex. Lupus anticoagulant antibodies have been purified that specifically react with prothrombin but not thrombin. These purified antibodies can enhance the binding of prothrombin to the cultured endothelial cell surface.79 Other protein-phospholipid targets may also be involved. Antiphosphatidylethanolamine antibodies are found in many patients with antiphospholipid antibody syndrome, and some of these antibodies inhibit activated protein C function.80 Antibodies to heparin and heparan sulfate, which inhibit the heparin-dependent neutralization of thrombin by AT, have been found.81 On the basis of this heterogeneity of antiphospholipid antibodies, it seems likely that multiple mechanisms are involved in the pathogenesis of thrombosis in this syndrome.
Thrombotic events occur in approximately 30% of patients with antiphospholipid antibodies (overall incidence, 2.5 events per 100 patient-years).82 In the cohort study cited above, 37% of patients presented with venous thrombosis, 27% with arterial thrombosis, 15% with both venous and arterial thrombosis, and 12% with fetal loss only77 [see 15:IV Systemic Lupus Erythematosus].
The diagnosis of antiphospholipid antibody syndrome should be considered in any patient who presents with an idiopathic arterial or venous thrombosis or in a woman with a history of recurrent miscarriages. The diagnosis is confirmed by the presence of anticardiolipin antibodies on ELISA (see above) or lupus anticoagulant on clotting assays.
The general criteria for the diagnosis of lupus anticoagulant are (1) prolongation of at least one phospholipid-dependent clotting assay; (2) proof, by mixing studies, that the prolongation is caused by an inhibitor and not a clotting factor deficiency; and (3) confirmation that the inhibitor is phospholipid dependent [see Table 5]. The clotting tests commonly used are activated PTT (aPTT) and dilute Russell viper venom time (RVVT). The reagents in aPTT are variably sensitive to the lupus anticoagulant and are influenced by concentrations of some plasma clotting factors (e.g., factor VIII). Therefore, an aPTT reagent that is sensitive to the lupus anticoagulant should be used in the screening test. Dilute RVVT is much more sensitive than aPTT but is a manual test and not as well standardized. Other tests, such as kaolin clotting time and the tissue thromboplastin inhibition test, are useful when available. The presence of an inhibitor necessitates a mixing study to demonstrate lack of correction with normal plasma. Correction of the prolongation by addition of phospholipid in the form of platelet lysates or as hexagonal-phase phospholipid will confirm the diagnosis. Clotting factor assays can be carried out in equivocal cases. A lupus anticoagulant will cause functional deficiency of several phospholipid-dependent clotting factors, not just one particular factor.
Table 5 Proposed Clinical and Laboratory Criteria for the Antiphospholipid Antibody Syndrome93
Anticardiolipin antibodies are reported as IgG (in IgG phospholipid [GPL] units) and IgM (in IgM phospholipid [MPL] units). The prevalence of elevated anticardiolipin IgG and IgM antibodies in normal populations is approximately 5%; with repeated testing, the prevalence is less than 2%.83 High titers of anticardiolipin IgG antibodies (> 33 GPL) are associated with an approximately fivefold increase in overall thrombotic risk.82,84 The importance of low titers of IgG antibodies (< 20 GPL), isolated IgM antibodies, and IgA antibodies has not been established.85,86 Both functional and antigenic assays should be ordered in the evaluation of a patient, because these two assays do not completely overlap. In one study of antiphospholipid antibody syndrome, 88% of patients had anticardiolipin antibodies (IgG, IgM, or both) and 54% of patients had lupus anticoagulant. Lupus anticoagulant was typically found in association with anticardiolipin antibodies, but it occurred in isolation in about 12% of patients.77 Certain infections and drug exposures may lead to a transient appearance of antiphospholipid antibodies, which disappear after the resolution of infection or discontinuance of the drug [see Table 6]. Therefore, laboratory tests should be repeated at least once (6 weeks after the first tests) to confirm the diagnosis. Conversely, approximately 20% of patients with low titers of anticardiolipin IgG antibodies will have higher titers upon repeat testing. Retesting is also warranted in patients with new or recurrent thrombosis.85
Table 6 Classification of Antiphospholipid Antibodies147
The current therapeutic recommendations for antiphospholipid antibody syndrome are mostly based on observational studies that support an association between antiphospholipid antibodies and thrombosis, particularly recurrent thrombosis.82,83,84,85,86 In the acute treatment of DVT in patients with antiphospholipid antibody syndrome, monitoring the effect of unfractionated heparin can be problematic because lupus anticoagulant prolongs the aPTT. The use of LMWH circumvents this problem because LMWH does not require dose titration and monitoring. The patient should be treated with LMWH and warfarin in the usual fashion, with an overlap of at least 5 days before discontinuing LMWH.
Retrospective analysis shows that patients with the antiphospholipid antibody syndrome and a history of thrombosis have a high rate of recurrent thrombosis (in the range of 50% to 70%) if they are not given prolonged warfarin therapy.87,88 The site of the first thrombotic event (i.e., arterial or venous) tends to predict the site of the recurrent event.88 High-intensity warfarin therapy, to an international normalized ratio (INR) of 3.0 to 3.5, had been advocated for these patients. This recommendation was based primarily on a retrospective analysis,87 however, and two prospective clinical trials have now demonstrated that most of these patients can be adequately treated with conventional levels of anticoagulation (i.e., an INR of 2.0 to 3.0).89 On the other hand, there are clearly some patients with antiphospholipid antibody syndrome who experience recurrent thrombosis with conventional anticoagulation and hence require a higher level of treatment.
The optimal duration of oral anticoagulation therapy for antiphospholipid antibody syndrome has not been fully established. Recurrent venous thrombosis or ischemic stroke usually justifies long-term warfarin. In a patient with a first episode of DVT who is found to have antiphospholipid antibodies, warfarin therapy is indicated for at least 6 months and perhaps for life.90 The severity of the specific thrombotic episode, the coexistence of any reversible thrombotic risk factors, and the risks of long-term oral anticoagulation therapy should also be considered. It should be noted that the lupus anticoagulant may occasionally increase the prothrombin time (PT) and, in turn, the INR, thus posing a problem for the monitoring of warfarin therapy.91 When PT and INR increase, use of a lupus anticoagulant-insensitive thromboplastin reagent is helpful.
In asymptomatic patients with anticardiolipin antibodies or lupus anticoagulant but no history of thrombosis, anticoagulation is not required.
Pregnancy Loss in Antiphospholipid Antibody Syndrome
Several prospective studies confirm an association between recurrent miscarriages and antiphospholipid antibodies. The antibodies presumably cause pregnancy loss by promoting placental thrombosis.92 Antiphospholipid antibodies should be measured in patients with a history of unexplained second- or third-trimester loss, fetal demise, early-onset severe preeclampsia, and intrauterine growth retardation.93In contrast, antiphospholipid antibodies are not associated with sporadic early pregnancy loss,94 which is frequently the result of genetic abnormalities in the fetus. The relationship of antiphospholipid antibodies with infertility is uncertain at present.
The management of pregnant women with antiphospholipid antibody syndrome is difficult because of the syndrome's association with thrombosis and the increased risk of bleeding with antithrombotic therapy. In a prospective, randomized, placebo-controlled trial, a combination of prednisone and aspirin was demonstrated to be ineffective in promoting live birth; in fact, it increased the risk of prematurity.95 On the other hand, two prospective trials have demonstrated that heparin and low-dose aspirin (81 mg a day) provide a significantly better pregnancy outcome than low-dose aspirin alone, with viable infants being delivered in 70% to 80% of cases.96,97Furthermore, low-dose heparin (given initially as 5,000 units subcutaneously twice daily and adjusted to maintain the aPTT within the upper limits of the normal range) seems to be as effective as higher-dose heparin combined with low-dose aspirin.98 Treatment should begin as soon as pregnancy is confirmed. LMWH is preferable to unfractionated heparin for long-term use because LMWH can be given once or twice daily and may reduce the risk of osteopenia and heparin-induced thrombocytopenia (see below). Enoxaparin (40 mg once daily) and aspirin (100 mg daily) has been given from week 12 of gestation until 6 weeks postpartum with good results.99
HEPARIN-INDUCED THROMBOCYTOPENIA AND THROMBOSIS
Heparin-induced thrombocytopenia (HIT) is a relatively common antibody-mediated drug reaction, occurring in about 1% of patients receiving porcine heparin and 5% of patients receiving bovine heparin.100 The incidence of HIT is much lower in patients treated with LMWH. In a subset of patients, HIT progresses to a potentially fatal disorder characterized by venous and arterial thrombosis. Interestingly, both the frequency of HIT antibody formation and the clinical manifestations of HIT vary considerably in different patient populations. The incidence of HIT antibody formation is much higher after cardiac surgery than after orthopedic surgery (50% versus 15%); however, the incidence of clinically significant postoperative HIT appears to be lower in cardiac surgical patients than in orthopedic patients, in whom the incidence is 5%.101,102
The pathogenesis of HIT is attributable to the presence of an IgG antibody that recognizes a complex of heparin and platelet factor 4 (PF4) [see Figure 3].103 PF4 is a cationic protein found in platelet α-granules; when released from the granules, PF4 binds to the negatively charged heparin molecule with high affinity. The IgG antibody binds to the PF4-heparin complex on platelet membranes, forming a ternary complex that in turn binds to the platelet membrane FcγRII receptor. This binding activates the platelets, leading to further release of PF4 and formation of PF4-heparin complex. The immune complex-coated platelets are cleared rapidly by the reticuloendothelial system, giving rise to thrombocytopenia. The thrombotic complications in HIT are caused by activation of platelets by the immune complex, which leads to the formation of platelet microparticles and enhanced thrombin generation.104 PF4 also binds to heparinlike sulfated glycosaminoglycans (e.g., heparan sulfate) on the endothelial cell surface. In vitro evidence indicates that the antibody in HIT is able to bind to endothelial cells. The cells may then become activated, giving rise to thrombosis.105 Given that only a subset of patients who form HIT antibodies experience clinical HIT,102 the induction of HIT antibodies and the development of thrombocytopenia and subsequent thrombosis should be regarded as a continuum. Concomitant thrombotic risk factors probably play a major role in determining the clinical progression and manifestations of HIT.
Figure 3. Heparin-induced Thrombocytopenia
In a proposed explanation for heparin-induced thrombocytopenia, IgG antibodies recognize platelet factor 4 (PF4)-heparin complexes. The resulting PF4-heparin-IgG immune complexes bind to Fc receptors on circulating platelets. Fc-mediated platelet activation releases PF4 from α-granules in platelets, establishing a cycle of platelet activation and formation of prothrombotic platelet microparticles. Removal of immune complex-coated platelets by the reticuloendothelial system results in thrombocytopenia. PF4 also binds to heparan sulfate on the surface of endothelial cells, leading to immune-mediated injury, thrombosis, and disseminated intravascular coagulation.145
HIT typically develops 5 to 10 days after the initiation of heparin therapy. However, in patients who received heparin within the previous 100 days and are being retreated, the onset can be rapid—within hours after starting heparin.106 Conversely, onset of HIT may not occur until as long as 19 days after heparin therapy is stopped.107 This delayed-onset HIT appears to be associated with a higher titer of IgG antibodies against the PF4-heparin complex.
HIT is generally defined as a platelet count below 150 × 109/L or a drop in the platelet count by more than 50% from the postoperative peak at 5 to 14 days after heparin is started. The mean platelet count in HIT is about 60 × 109/L. Severe thrombocytopenia, with platelet counts below 20 × 109/L, occurs in fewer than 10% of patients; in 10% to 15% of patients, despite the 50% drop from peak levels, the platelet count nadir is above 150 × 109/L.102
The risk of HIT-associated thrombosis was once thought to be quite small; however, it is now recognized that thrombosis occurs in about one third to one half of patients with HIT, with venous thrombosis occurring more frequently than arterial thrombosis.104 Thrombosis can occur at any platelet count, even at a very low one.
DVT, with or without pulmonary embolism, is the most common event leading to the diagnosis of HIT. The disorder may be further complicated by limb gangrene, especially in the setting of warfarin treatment without concomitant alternative anticoagulant coverage (see below). Cerebral vein thrombosis and adrenal hemorrhagic necrosis are uncommon but well-documented complications of HIT, and early diagnosis and urgent therapy can be lifesaving. Arterial thrombosis may present as limb ischemia, stroke, myocardial infarction, or, less commonly, mesenteric thrombosis and renal arterial thrombosis. Some patients have laboratory findings that support a diagnosis of DIC. Heparin-induced skin lesions may occur at heparin injection sites and range from painful erythematous papules to extensive dermal necrosis.108
Unlike antibodies induced by quinine, quinidine, or sulfonamides, which can persist for years, heparin-induced antibodies appear to be quite transient. They fall to undetectable levels at a median of 50 to 85 days, depending on the assay performed.106
The diagnosis of HIT is supported by the finding of heparin-induced platelet aggregation in the presence of the patient's serum. However, the sensitivity of this test can be as low as 50%.109 Sometimes, the patient's serum can cause spontaneous aggregation of donor platelets in the absence of heparin, most likely caused by the presence of immune complexes unrelated to HIT, which makes proper interpretation of the test result impossible. Heparin-induced platelet serotonin release using washed platelets has high sensitivity and specificity for HIT but is available only in a few specialized laboratories. ELISAs to detect antibodies that are reactive to the PF4-heparin complex are commercially available and have become the most commonly used test for HIT. These ELISAs have a higher sensitivity than the platelet aggregation assay and can be more easily performed in a general clinical diagnostic laboratory. However, false positive results (i.e., positive tests in the absence of HIT or thrombocytopenia) occur in 10% to 15% of medical patients and in more than 20% of patients receiving heparin for peripheral vascular surgery. A seroconversion rate as high as 50% has been reported in patients undergoing cardiopulmonary bypass surgery, limiting its usefulness in that situation.110 Because HIT can be complicated by serious thrombotic problems, however, diagnosis of HIT should be based primarily on appropriate clinical findings, and management should be started while laboratory confirmation is awaited.
Management of HIT consists of stopping heparin immediately and starting alternative anticoagulation therapy. It is important to discontinue all types of heparin: there are anecdotal reports of HIT caused by trace amounts of heparin used in heparin flushes of intravascular lines. Even if the patient has mild thrombocytopenia alone without any evidence of thrombosis, it is advisable to discontinue heparin and treat the underlying hypercoagulable state with an alternative anticoagulant. This aggressive approach is supported by a retrospective cohort study in which thrombosis developed within 30 days of heparin cessation in approximately half of patients who initially had no clinical symptoms from their HIT.100
HIT may develop in patients who were receiving heparin for preexisting thrombosis and are in the process of being switched over to warfarin. If the patient has been on warfarin for 4 or 5 days and the INR has reached an adequate therapeutic range, the clinician may rely on warfarin alone and monitor the patient carefully. However, if the patient has been on warfarin for less than 4 days or has evidence of a thrombotic complication from the HIT, an alternative anticoagulant should be used in addition to warfarin.
Alternative anticoagulant agents include lepirudin, argatroban, and fondaparinux. Although LMWH is much less immunogenic than unfractionated heparin in causing HIT,104 it cannot be used as a safe substitute when a patient develops HIT caused by unfractionated heparin. LMWH and unfractionated heparin have extensive cross-reactivity (> 90%) in terms of antibody recognition. LMWH is not an appropriate choice in patients with HIT.
Hirudin, a 65-amino-acid protein originally extracted from the salivary gland of the medicinal leech (Hirudo medicinalis), is a potent direct thrombin inhibitor. Hirudin binds directly to thrombin's active site, independently of AT. It is not neutralized by PF4. Hirudin's anticoagulant function is monitored by aPTT.
Lepirudin is a recombinant form of hirudin that is approved for treatment of HIT. In two prospective clinical trials, use of lepirudin reduced serious thrombotic complications to about 20% (compared with a rate of about 40% in historical control subjects).111,112 Lepirudin was given by intravenous bolus (0.4 mg/kg), followed by continuous infusion at 0.15 mg/kg/hr for 2 to 10 days as indicated. The dose was adjusted to maintain a target aPTT of 1.5 to 3 times normal. Patients experienced an increase of minor bleeding (from puncture sites, epistaxis, and hematuria) but no intracranial bleeding. Of note, 40% of patients developed antihirudin antibodies. These are not neutralizing antibodies and may actually enhance the drug's potency, perhaps by delaying its clearance—another reason to monitor aPTT levels. Judging from cardiology intervention trials, bleeding risk would be substantially increased with concomitant use of thrombolytics; therefore, it is not advisable to use the agents in combination. There is no effective antidote for lepirudin. It has a half-life of approximately 1.3 hours and is cleared by the kidneys. Thus, in patients with renal insufficiency, the dose of lepirudin needs to be adjusted carefully on the basis of creatinine clearance values.
Argatroban, a synthetic direct thrombin inhibitor, is approved for prophylaxis or treatment of thrombosis in patients with HIT. It is given by continuous intravenous infusion of 2 mg/kg/min to maintain an aPTT of 1.5 to 3 times baseline (not to exceed 100 seconds or 10 mg/kg/min). In one study, argatroban reduced the serious thrombotic complications of HIT by about 50%, as compared with historical controls, with a major bleeding rate of about 7%.113 Its half-life is only 40 to 50 minutes. In contrast to lepirudin, argatroban is cleared by the liver and therefore can be used more easily in patients with renal insufficiency. Like lepirudin, argatroban does not have a specific antidote, and bleeding complications need to be watched for carefully.
A third alternative anticoagulant is fondaparinux, a synthetic pentasaccharide that activates AT, leading to thrombin inhibition. Fondaparinux has been approved for prophylaxis against DVT in orthopedic surgery. Because of its small size and reduced negative charge, fondaparinux does not form a complex with PF4 and therefore does not react with the antibody directed against the heparin-PF4 complex in HIT. Case reports describe the successful use of fondaparinux in HIT at a fixed dose of 7.5 mg administered subcutaneously once daily.
Patients who have HIT without associated thrombotic complications should be treated with one of the alternative anticoagulants until the platelet count has returned to normal, which generally takes 5 to 7 days. It is my practice to discontinue the alternative anticoagulant at that time. However, some experienced clinicians choose to continue empirical anticoagulation for up to 1 month, on the rationale that HIT represents an intensive hypercoagulable state. In patients who require prolonged anticoagulation—whether because they have other indications for anticoagulation or because they have had thrombotic complications from HIT—warfarin is used for long-term treatment. Warfarin should be started only after the patient has received adequate anticoagulation therapy with one of the alternative anticoagulants. The two agents should be used concurrently for at least 5 days before the alternative anticoagulant is discontinued. Both lepirudin and argatroban may increase the PT and INR, thus interfering with warfarin dose adjustments. The PT should be rechecked 6 hours after the discontinuance of lepirudin or argatroban to ensure that an INR of 2 to 3 has been achieved. For patients who have had HIT-associated thrombosis, warfarin therapy for 3 months should be adequate.
Some patients with serologically confirmed HIT may, at some point in the future, require surgery that involves cardiopulmonary bypass. The use of recombinant hirudin or argatroban in that situation has been described anecdotally.112,114,115 Given the transient nature of the heparin-induced antibodies, subsequent reuse of heparin is theoretically reasonable, and indeed, a limited number of patients have been given heparin again after the disappearance of heparin-induced antibodies without any significant clinical sequelae.106,116 Nevertheless, the use of heparin in this situation should be restricted to patients with a compelling indication for it, such as cardiac or vascular surgery, and only after the absence of detectable heparin-dependent antibodies has been confirmed by a sensitive assay, such as a PF4-heparin ELISA. Also, because reexposure to heparin may elicit a recurrence of heparin-dependent antibodies, heparin should be used only during the procedure itself, and an alternative anticoagulant should be started postoperatively for prophylaxis against recurrence of HIT. The anti-heparin-PF4 antibody should be checked postoperatively.
Some patients with cancer—especially those with occult solid tumors of the pancreas, ovary, liver, brain, colon, lung, or breast—may experience spontaneous venous thrombosis of the upper and lower extremities, or Trousseau syndrome. These patients also have an increased propensity toward recurrent arterial thrombosis and thromboembolism.117 In a prospective study of patients who presented with idiopathic symptomatic DVT, cancer was diagnosed in approximately 8% of the patients during a 2-year follow-up (odds ratio, 2.3). In patients with recurrent thrombosis, the incidence of cancer was even higher (17%; odds ratio, 4.3).118
The underlying cause of Trousseau syndrome is a chronic, compensated form of DIC [see 5:XIII Hemorrhagic Disorders]. The activated procoagulants generated in DIC enhance thrombosis. Immunochemical staining of sections from tumors commonly associated with Trousseau syndrome often reveals tissue factor on the tumor surface.119,120 A cancer procoagulant has been purified from some adenocarcinomas and leukemic cells; this procoagulant, identified as a cysteine protease, can activate the clotting cascade by directly activating factor X.121,122Interaction of the tumor cells with monocytes, platelets, and endothelial cells may also generate inflammatory cytokines and induce endothelial and monocytic procoagulant activities, further exacerbating the thrombosis. Tumor cells may secrete soluble mucins—complex polysaccharides that can activate leukocytes, leading to platelet-leukocyte microthrombi and thrombin generation.123 Heparin blocks the tumor mucin activation of leukocytes, which may partially explain heparin's efficacy over warfarin in the treatment of this condition.
Venous thrombosis in Trousseau syndrome usually manifests as migratory superficial thrombophlebitis or DVT of the lower extremities. The recurrent arterial thrombosis in these patients arises from a nonbacterial thrombotic endocarditis in which sterile fibrin is deposited in the mitral valve. The fibrin clot may embolize to cause digital ischemia, transient ischemic attacks, and stroke.
Coagulation studies in Trousseau syndrome show evidence of chronic, low-grade DIC (i.e., slightly low or even high fibrinogen and platelet levels and high levels of D-dimer). The PT and PTT are generally not prolonged. Overt DIC in such patients is uncommon.
How aggressively should one pursue the diagnosis of an underlying cancer in patients with idiopathic DVT? Research has not yet demonstrated the benefit and cost-effectiveness of an extensive screening approach.124 In a large cohort study, cancer diagnoses after a primary thrombotic event were highest during the first 6 months of follow-up and declined rapidly to normal levels after the first year. Moreover, 40% of the patients who were diagnosed with cancer in the first year had distant metastases at the time of the diagnosis. It is unclear whether an earlier diagnosis after the thrombotic event would have changed the outcome in these patients. The researchers concluded that an aggressive search for a hidden underlying cancer in such patients is not warranted.125
At present, it is prudent to perform a careful history, physical examination, chest x-ray, routine blood counts, and chemistries. Some experts have also recommended multiple tests for fecal occult blood, prostate-specific antigen tests in men, and mammography and pelvic ultrasonography in women.126,127 Careful follow-up examination and tests should be done as indicated by the initial evaluation.128
The key to management of Trousseau syndrome is diagnosis and treatment of the underlying tumor. Unfortunately, tumors often present explosively in patients with Trousseau syndrome and may not respond to the usual therapies. In a prospective study, cancer patients had an approximately threefold increase in the rate of recurrent thrombosis and twofold increase in the rate of major bleeding during warfarin treatment of DVT. These complications occurred mostly during the first few months of anticoagulation and did not reflect underanticoagulation or over-anticoagulation but correlated with the extent and severity of the underlying cancer.129 The likely explanation for the increased thrombosis recurrence in these patients is relative warfarin resistance, whereas the increased bleeding may be related to bleeding at the primary tumor site. A prospective trial found that the risk of recurrent venous thromboembolism was 50% lower in cancer patients who received long-term treatment with the LMWH dalteparin than in those who received oral anticoagulation; there was no significant difference in the risk of major bleeding.130 If a patient is receiving chemotherapy for the underlying cancer, an exacerbation of the DIC associated with tumor lysis should be anticipated. An increase in the dose of heparin may be required.
THROMBOTIC REACTIONS TO ESTROGENS
Oral contraceptives increase the risk of thromboembolic disease approximately fourfold.131 Epidemiologic studies indicate that contraceptives containing a third-generation progestin (e.g., desogestrel) carry a twofold greater risk of thrombosis than those with a second-generation drug (levonorgestrel).132 For that reason, the preferred choice for first-time users of oral contraceptives is a compound containing a second-generation drug (e.g., Alesse, Levlite, Levora, Nordette, Triphasil, Trivora). The risk of venous thromboembolism disappears when the drugs are discontinued.
In postmenopausal women, estrogen replacement increases the risk of venous thromboembolism about threefold. The absolute risk is low, however—it is estimated to be approximately 3.2 per 10,000 patient-years. Therefore, estrogen replacement is not contraindicated in patients who require hormonal treatment to control severe postmenopausal symptoms. However, patients with a previous history of DVT or pulmonary embolism are at increased risk for recurrence and therefore should avoid hormone replacement therapy if possible.133
The association between estrogen use and thromboembolic disease remains unexplained.134 Estrogen treatment is known to produce changes in the plasma levels of many proteins involved in coagulation and fibrinolysis, including decreases in protein S and AT and increases in plasminogen. The changes are generally quite modest, however, and are not thought to account for the increased risk of thrombosis. On the other hand, in women who are heterozygous for factor V Leiden, oral contraceptive use increases the thrombotic risk synergistically to about 50 times normal.18 There is also a moderate synergistic increase in thrombotic risk (15-fold) with the combination of factor V Leiden and hormone replacement therapy.135
Management of Venous Thromboembolism
The acute management of an initial episode of DVT or pulmonary embolism in patients with proven or presumed underlying risk factors for thrombosis is the same as that in other patients: heparin and then warfarin [see 1:XVIII Venous Thromboembolism].
The optimal intensity and duration of warfarin treatment in DVT have been the subject of many large clinical trials over the past decade. As regards the intensity of oral anticoagulation, an INR of 2 to 3 has proved optimal, with a low rate of thrombosis recurrence and a rate of major bleeding of about 3% a year.136,137 In comparison, treatment to an INR of 1.5 to 1.9 is less effective in reducing recurrent thrombosis (although it is better than placebo) and provides no significant reduction in bleeding risk.136,137
Warfarin therapy usually should be continued for 3 to 6 months. In a prospective study of oral anticoagulation therapy in patients with a first episode of venous thromboembolism, 6 weeks of therapy was adequate for patients with temporary, reversible risk factors for thrombosis (e.g., surgery, trauma, temporary immobilization, or use of oral contraceptives). On the other hand, 6 months of oral anticoagulation therapy was clearly superior for patients with idiopathic venous thromboembolism (who are presumed to have intrinsic risk factors).138 The recurrence rate was quite high, however—approximately 12% at 2 years. On the basis of this evidence, at least 6 months of oral anticoagulation therapy is indicated in a patient with a first episode of idiopathic DVT or pulmonary embolism.
Further management should depend on results of the hypercoagulable workup (see above). A risk assessment for other predisposing factors will also be relevant.
Unfortunately, even when warfarin treatment is continued for 12 months after a first episode of idiopathic DVT, this does not reduce the risk of recurrent thrombosis once warfarin treatment is discontinued.139 Generally, there is a rapid rebound phase of recurrent DVT of about 10% during the first 6 to 12 months after warfarin therapy. This suggests that there is a subset of patients (10% to 20%) who have a stronger tendency toward thrombosis and thus experience recurrences fairly soon after discontinuance of oral anticoagulation.
Persistent elevation of D-dimer levels may help identify patients who are more likely to have recurrent thrombosis. In a prospective study, D-dimer levels that remained elevated 1 month after the discontinuance of warfarin were associated with a higher recurrence risk (approximately threefold to eightfold), whereas normal D-dimer levels had a high negative predictive value for recurrence.140 Thus, patients with elevated D-dimer levels merit more vigilant monitoring and consideration of long-term anticoagulation.
Because the optimal duration of long-term oral anticoagulation therapy for patients with thromboembolism remains undefined, this question is best addressed individually, on the basis of an estimation of the risk of recurrence [see Table 7].141,142 Before undertaking long-term anticoagulation therapy in a high-risk patient, the clinician must take the patient's risk of bleeding into account.
Table 7 General Guidelines for Management of Patients with Venous Thromboembolism
Of note, none of the published studies shows a significant difference in mortality between patients who receive long-term therapy and those who receive short-term therapy. There is also no evidence to suggest that prophylactic anticoagulation therapy improves overall survival. In historical studies, families with deficiencies of AT and protein C show no higher mortality than the general population displays.143,144Current clinical trials are studying the use of full-dose oral anticoagulation therapy for an extended period of time followed by an indefinite period of low-dose anticoagulation therapy. The results of these studies will help define the optimal long-term treatment for patients who are at high risk for recurrence of thromboembolism.
Figures 1 and 2 Seward Hung.
Figure 3 Dr. Rajeev Doshi.
Editors: Dale, David C.; Federman, Daniel D.