Thrombophilia: Clinical and Laboratory Assessment and Management

Introduction

Venous thromboembolism (VTE) poses a number of challenges and opportunities to the practicing hematologist. The duration of anticoagulation is a frequent reason for consultation because it poses difficult questions of risk and benefit. Although the practical frustrations of anticoagulation to patients have been slightly mitigated by the arrival of alternatives to warfarin, weighing the risks of bleeding compared with clotting remains as challenging as ever. The quest for biomarkers to predict the risk of recurrence has thus far been inadequate despite the discovery of hereditary thrombophilias (factor V Leiden, prothrombin [PT] G20210A mutation, and deficiencies of antithrombin III [ATIII], protein C, and protein S). The use of hereditary thrombophilia testing to help better predict which patients will benefit from long-term anticoagulation held out great promise, but as detailed in this chapter, the practical applications are limited and nuanced. Unfortunately, results from thrombophilia testing have not resulted in more tailored management to reduce risk of recurrent thrombosis. Hereditary thrombophilias only marginally increase the risk of recurrent VTE. Conversely, there is no evidence that a negative laboratory thrombophilia panel is associated with a lower risk of recurrence. Thus the management of patients with unprovoked VTE is unchanged by the presence or absence of an identifiable hereditary thrombophilia in most cases. Furthermore, inappropriate thrombophilia testing or misinterpretation of test results can cause harm through overtreatment or false reassurance. Clinical risk factors typically outweigh genetic ones in management. No study has conclusively proven a benefit or harm to thrombophilia testing, and an attempt to settle the question prospectively with a multicenter randomized control trial was terminated due to low enrollment. However, in this chapter the available prospective and retrospective data will be reviewed as a means to advise the practicing hematologist on a number of frequently encountered scenarios.

This chapter will review the basic laboratory aspects of thrombophilia testing. In addition, the use of these tests in the clinical context of VTE in adults will be discussed. The role of thrombophilia testing in other clinical settings is discussed elsewhere ( Chapter 15, Chapter 17, Chapter 20, Chapter 22, Chapter 23, Chapter 33 ).

Assessment for thrombophilia should not be thought of as merely a panel of laboratory tests, but rather as a comprehensive review of the totality of clinical and family history that is undertaken to understand the factors that contributed to the patient's tendency to clot (thrombophilia in the truest sense) and what interventions might mitigate risks of VTE in the future.

Should All Patients With Unprovoked Venous Thromboembolism Receive Lifelong Anticoagulation?

The rationale for indefinite anticoagulation for most patients with unprovoked VTE is based on the observation that recurrent VTE is common after unprovoked VTE. Five to fifteen percent of patients recur in the first year if anticoagulation therapy is discontinued after 3 to 6 months. By 10 years, approximately 40% of patients with unprovoked VTE will have experienced a recurrence. The recommendation for indefinite anticoagulation after unprovoked VTE is predicated on the premise that a recurrence risk of greater than 5% in the first year would outweigh the bleeding risk associated with warfarin prophylaxis for most patients. However, indefinite anticoagulation for unprovoked VTE has never been proven to have a mortality benefit, and older estimates suggested that the risk of death from bleeding might be similar to risk of death from VTE. However, this risk calculus preceded the arrival of direct oral anticoagulants (DOACs) that, in both clinical trials as well as real-world use, cause major bleeding less frequently than warfarin does. Trials comparing DOACs with placebo for extended anticoagulation after a first unprovoked VTE have been performed and showed favorable reductions in recurrent VTE compared with their effect on major bleeding. Although none were continued long enough to detect an effect on mortality, the safety signals were favorable. As an example, with full-dose or prophylactic-dose apixaban, the number needed to treat to prevent recurrent VTE in the first year was 14, whereas the number needed to harm for major or clinically relevant nonmajor bleeding was 200. In fact, no statistical difference could be detected in major bleeding compared with placebo. However, the patients in the study were generally without significant risk factors for bleeding, and major bleeding was sufficiently rare that precise estimates of relative risk for major bleeding could not be made. Along the same lines, extended anticoagulation with prophylactic dosing of rivaroxaban (10 mg daily) appeared to be as effective as rivaroxaban 20 mg daily but had a major bleeding risk similar to a prophylaxis with 100 mg daily aspirin. In summary, prophylactic dosing of DOACs (apixaban 2.5 mg BID, rivaroxaban 10 mg daily) is effective at reducing recurrent VTE (after the initial treatment period with full strength anticoagulation) with favorable bleeding risk. This is an important contrast with warfarin, where reduced intensity treatment (INR goal 1.5 to 2) was less effective than conventional warfarin dosing while maintaining similar bleeding risk. Table 14.1 summarizes these options. Therefore for patients with an unprovoked VTE without an increased bleeding risk, indefinite anticoagulation is recommended for secondary prophylaxis. Accordingly, testing for hereditary thrombophilia is not recommended for these patients because the results will not change management.

There have been several attempts to identify a subset of patients with unprovoked VTE who have a lower risk of recurrent VTE and for whom indefinite anticoagulation might not be required. Because hereditary thrombophilia testing is unhelpful for this purpose, as discussed previously, other clinical and laboratory markers have been studied. Among the factors identified, d -dimer appears to have the best discriminatory power for identifying a lower risk group. This group is confined to women because men with unprovoked VTE have a higher risk of VTE recurrence after anticoagulation is stopped, irrespective of d -dimer results. Although a negative d -dimer may identify a group of women at lower risk of recurrence, it is unclear that this risk is sufficiently lower to justify stopping anticoagulation. To further refine the selection of lower risk individuals, d -dimer and sex have been combined in some risk prediction models with other factors that may affect VTE recurrence risk, including obesity. The most well-known scoring systems, including some or all of these factors, are the DASH, Vienna, and HERDOO2 scores. Although increasing age is clearly a risk factor for first VTE incidence, there are discordant data on whether it is a risk factor for recurrence. To this effect, the DASH score has younger age (<50) as a risk factor for recurrence, whereas HERDOO2 has older age (≥65). The HERDOO2 rule was prospectively validated in a cohort of 2785 patients, and the elements are H yperpigmentation, E dema, or R edness in either leg; d -dimer level greater than or equal to 250 µg/L during treatment with anticoagulants; O besity with body mass index greater than or equal to 30; and O lder age, 65 years or older. Women with 0 to 1 criteria were deemed appropriate to discontinue anticoagulation after the initial 5- to 12-month treatment. However, it should be noted that this group was still at increased risk of recurrence. The recurrence rate for low-risk women with non–hormone-related VTE was 3.1% per patient-year compared with 8.1% for men and high-risk women. Therefore if stopping anticoagulation based on a low HERDOO2 score, appropriate measures to prevent recurrence should be undertaken for all periods of increased risk, such as pregnancy, prolonged airplane travel, and surgery. In addition, patients with hereditary thrombophilia other than heterozygosity for factor V Leiden and PT G20210A mutation were excluded from these studies; hence if relying on these scoring systems, testing for ATIII deficiency and antiphospholipid antibodies may be appropriate. Another challenge with using d -dimer–based risk stratification is that the majority of the data come from patients on warfarin, and it is unclear whether these data can be extrapolated to patients on DOACs. At this time, no consensus panel has incorporated any of these d -dimer–based scoring systems into their recommendations. With the option of prophylactic dosing of a DOAC improving the safety of long-term anticoagulation, it remains to be seen whether these scoring systems will temper the recommendation that all patients with an unprovoked pulmonary embolism (PE) or proximal deep vein thrombosis (DVT) without an increased bleeding risk should receive indefinite anticoagulation.

One group of patients with unprovoked thrombosis for whom indefinite anticoagulation may not be required is isolated distal (calf vein) DVT. The rate of recurrence for isolated distal DVT appears to be approximately half the recurrence rate for proximal DVT and PE. In addition, it has been observed that patients who present with DVT as their initial VTE are more likely to recur with their second events as DVT, whereas patients who present with PE are more likely to recur with PE. Although symptomatic DVT can cause significant morbidity and long-term sequelae, it rarely poses a direct mortality risk.

Weighing Bleeding Risk

For patients without an increased bleeding risk, an unprovoked VTE is an indication for indefinite anticoagulation. However, for patients with increased bleeding risk, a more nuanced assessment of risk to clot and risk to bleed is appropriate. Antiphospholipid antibodies and ATIII deficiency are the two main laboratory findings that would portend a clinically significant increased risk of recurrence should anticoagulation be stopped. Several scoring systems have been proposed for weighing the risk of bleeding while on therapeutic anticoagulation. All were validated in the warfarin era, contain similar risk factors, and none has proven clearly superior. The HAS-BLED score ( Table 14.2 ) is one such system for estimating annual risk of major bleeding on warfarin treatment for atrial fibrillation, whereas the RIETE score was validated in patients with VTE, most of whom were taking a vitamin K antagonist. The annual risk of major bleeding can be used, along with the estimated annual risk for recurrent VTE, to weigh the risks and benefits of extended anticoagulation. However, it is important to note that DOAC treatment has a lower risk of major bleeding than warfarin treatment, and therefore even patients at increased risk of warfarin-associated bleeding may derive a net benefit from a DOAC. Low-dose apixaban (2.5 mg BID) or low-dose rivaroxaban (10 mg daily) could be considered an option in these scenarios because these agents reduced the risk of recurrent VTE compared with placebo or low-dose aspirin, respectively, without an observed increased risk of major bleeding, albeit in cohorts with low propensity to bleed. Further complicating risk:benefit analysis is that some risk factors for bleeding such as active malignancy are also concomitant risk factors for thrombosis. In many real-world situations, there are inadequate data to predict whether continuing anticoagulation after a bleeding event is more likely to be beneficial or harmful. In these situations, it is helpful to engage in shared decision-making. One potentially useful question to ask patients is: “What keeps you up at night? Are you more worried about having another clot or about having another bleed?”

Pursuit of Thrombophilia Testing May Divert Attention From More Important Clinical Risk Factors

Many large cohorts have confirmed the heritable nature of thrombophilia, but this is only partially accounted for by the known mutations. This may be due to the fact that, like many traits, thrombophilia is influenced by a large number of polymorphisms, each having only a small independent effect. Other factors, such as height, race, blood type, and obesity, run in families and are associated with risk of VTE. Thus the currently available testing does not account for all the risk of any one inherited thrombophilic condition. Unprovoked VTE at a young age (<45 years) is associated with an increased risk of VTE in family members, whereas factor V Leiden and PT gene mutations are rather weak predictors. Having a first-degree relative with an unprovoked VTE doubles one's risk of developing VTE compared with the relative having a provoked VTE. This risk is nearly 3 times higher if the family member experienced VTE before age 50 compared with later in life. The impact of these factors on VTE risk was independent of PT gene mutation or factor V Leiden mutation status in the family member. The risk was further increased when multiple family members were affected with VTE.

What Are the Pitfalls of Genetic Testing?

When considering genetic testing for hereditary thrombophilia, physicians should follow best practices regarding genetic testing. Genetic counseling with informed consent should precede genetic testing. Informed consent discussion should include an understanding of the risks and benefits of both courses of action, as well as what risk modification tools are available. Patients may voice concerns about the use of genetic information. However, in the United States the Genetic Information Nondiscrimination Act (GINA) of 2008 protects against the use of genetic information by group health insurance plans and employers. Knowledge of GINA is not widespread and even after the bill was passed, concerns about health insurance discrimination continued to be a frequent reason for patients declining genetic testing. However, patients should be aware that GINA does not include protections regarding life or disability insurance.

When counseling asymptomatic patients regarding the impact of hereditary thrombophilia, risks should be presented as absolute risks rather than relative risks. Studies show that patients are more likely to overestimate their own risk if presented with relative risks. As an example, if an individual without prior VTE is told that their relative risk of VTE has quadrupled due to factor V Leiden heterozygosity, they may perceive a high risk because they may be unaware or have difficulty conceptualizing the relatively low absolute overall risk of VTE in the general population. They will perceive this differently than hearing that they have an 11% absolute lifetime risk of VTE. Physicians can reduce framing bias by offering outcomes in both positive and negative forms : an 11% lifetime risk of VTE should also be framed as an 89% chance they will never be affected by VTE. Other key elements of risk communication include using a consistent denominator (e.g., express all risks as percentages or as out of 1000 persons). Visual formats such as the “One Thousand People Palette” can be used to represent outcome estimates for testing decisions. Hematologists should also be aware of the varied psychological effects that genetic testing results may have on a family, including anxiety and guilt.

When Is It Helpful to Test Unaffected Family Members for Genetic Mutations?

Because thrombosis is a relatively infrequent event, most individuals with hereditary thrombophilia will never suffer VTE. Thus lifelong prophylactic anticoagulation is almost never recommended to unaffected individuals, regardless of their genotype. Because testing should only be undertaken if it will result in a change in management, a strong case can be made that unaffected family members of patients with hereditary thrombophilia should not be screened. However, there may be periods of transiently amplified risk that theoretically might warrant prophylactic measures beyond those recommended to the average individual, and this is the argument cited for screening some individuals. Because thrombotic events are infrequent, benefits to prophylactic measures are difficult to capture and most recommendations are based on theoretical risk frameworks. Next we will review the evidence regarding the four most commonly discussed (but nevertheless controversial) situations in which this might be warranted: pregnancy, estrogen use, major surgery, and extended airplane flight ( Table 14.3 ).

Estrogen Use

Supplemental estrogen is associated with a dose-dependent increased risk of thrombosis, thought to be related to altered levels of clotting factors. There also appears to be a synergistic effect with hereditary thrombophilias, particularly factor V Leiden. For example, the relative risk for VTE for factor V Leiden carriers taking combined oral contraceptives is 10.2 to 15.6, which is more than the sum of the separate risks. However, the absolute incidence of VTE in young women is low, less than 0.1% per patient year, so the event rates are still low. One study estimated that to prevent one death from VTE, 92,000 carriers of factor V Leiden would need to be identified and stopped from using oral contraceptives at a cost exceeding $300 million. Therefore screening all individuals for hereditary thrombophilia before starting oral contraceptives is not recommended. However, an unaffected individual with factor V Leiden identified through an affected family member has certain notable differences from an individual identified through population-based screening, including a higher risk of thrombosis (due to the family history ) and likely a higher motivation to take preventive measures such as avoiding estrogen-containing contraceptives. On the other hand, these measures can be recommended regardless of the outcome of thrombophilia testing, because a family history of unprovoked VTE is associated with increased risk of estrogen-related VTE even with negative thrombophilia testing. Given that there are now multiple effective alternatives with lower VTE risk, it is reasonable for family members of patients with unprovoked VTE to be counseled to avoid estrogen-containing contraceptives.

The dose of estrogen used for postmenopausal hormone replacement therapy in the Women's Health Study was 0.625 mg conjugated equine estrogen and was associated with a twofold increase in VTE risk. This is approximately equivalent to 10 µg of ethinyl estradiol, which is half the amount in most “low-estrogen” contraceptives currently prescribed in the United States. Accordingly, combined oral contraceptives are associated with a threefold to fivefold increased risk, depending on the dose of estrogen (20 to 35 µg ethinyl estradiol) and possibly also the type of progestin. Early contraceptives in the 1960s and 1970s had estrogen doses of 50 µg or more and were associated with even higher VTE risks, but these are now infrequently prescribed in the United States.

Progestin-only oral contraceptives (“mini-pills”) are associated with minimal risk for VTE but are less effective as contraceptives in typical use. However, depot progesterone implants do appear to increase risk twofold to threefold through an unclear mechanism. In contrast, intrauterine devices (IUDs) that contain the progestin levonorgestrel and nonhormonal (copper) IUDs do not increase risk for VTE and are the most effective nonpermanent methods of contraception.

The type of progestin appears to affect thrombotic risk. Second-generation progestins such as levonorgestrel seem to have the lowest risk. Third-generation progestins such as norgestimate, gestodene, and desogestrel were developed to have less androgenic side effects than levonorgestrel but appear to have a higher VTE rate, although this has not been shown in all studies.

The availability of alternatives to combined oral contraceptives has made it easier to recommend avoiding their use in family members of patients with unprovoked VTE. However, patients may be prescribed combined oral contraceptives for noncontraceptive indications such as endometriosis. In this case, risk estimation should take into account concurrent risk factors for VTE, including increasing age, obesity, and family history. It is important to note that pregnancy itself is associated with a higher risk than combined oral contraception; therefore not taking or discontinuing combined oral contraception cannot be recommended if the alternative is an ineffective method of birth control (see Chapter 31 ).

That age is a risk factor highly associated with VTE is well demonstrated by Fig. 14.1 .

Testing Considerations for Particular Disorders