Strong risk factors
•Recent surgery
•Recent trauma or fracture
•Immobilization/bedbound >1 week
Minor risk factors
•Estrogen therapy
•Pregnancy or within 3 months postpartum
•Prolonged travel
Fig. 23.1
Algorithm for treatment of VTE
The Role of Anticoagulation
The goals of anticoagulation are twofold: first, to stop the acute thrombotic period to block extension of the existing thrombus and improve symptoms and, second, to prevent the formation of new clots outside the initial thrombotic phase. Active treatment, which addresses the first goal, describes the time period from initiation of anticoagulation to inactivation of the acute thrombus , during which time pharmacologic anticoagulation protects from further clot deposition while endogenous systems stabilize and dissolve the existing thrombus [18]. This is a slow process, with about 50% of patients demonstrating persistent venous impairment at 6 months post-diagnosis and treatment. Surgically provoked clots resolve faster, while patients with cancer-associated thrombosis and those with larger clot burdens have slower clot dissolution [15, 19]. Following completion of active anticoagulant treatment, secondary prevention in the form of extended anticoagulation may follow if the risk of recurrent thrombosis after discontinuation of anticoagulation is deemed sufficiently high.
In patients unable to safely receive systemic anticoagulation, including those with absolute contraindications such as active hemorrhage (involving the central nervous system, gastrointestinal tract, or retroperitoneum), massive hemoptysis, severe thrombocytopenia, head trauma, or a history of life-threatening bleeding while on anticoagulation, inferior vena cava (IVC) filters are an alternative therapeutic consideration [20]. IVC filters may prevent PE and decrease short-term mortality but are associated with increased risks of recurrent DVT and long-term complications including filter fracture and migration and IVC thrombosis [21–23]. If placed, retrievable filters should be removed after anticoagulation can be safely initiated [23].
The choice of anticoagulant agent is guided by numerous clinical parameters including age, renal function, coexisting medical conditions (e.g., cancer, pregnancy), and patient preference. Unfractionated heparin (UFH) , low-molecular weight heparin (LMWH) , vitamin K antagonists (VKA) , and direct oral anticoagulants (DOAC) are all acceptable treatment options; descriptions and dosing of these anticoagulants are summarized in Tables 23.2 and 23.3. In the acute setting, hemodynamically stable patients, who are reliable, in secure social situations, and without severe symptoms, renal impairment, or high bleeding risk, may be treated safely on an outpatient basis [24]. Those with massive DVT (as defined by limb ischemia, thrombosis of the iliofemoral veins or IVC, or swelling of the entire limb), concomitant symptomatic PE, a high bleeding risk, or other select comorbidities limiting safe administration of anticoagulant therapy in the outpatient setting are suitable candidates for hospitalization [25].
Table 23.2
Common anticoagulation agents for outpatient VTE treatment
Anticoagulant | Mechanism of action | Half-life (time to peak)a | FDA approved indications | Specific reversal agent |
---|---|---|---|---|
Dabigatran (Pradaxa®) | Direct thrombin inhibitor | 14–17 h (2–4 h) | •Stroke prevention in non-valvular atrial fibrillation •Acute and extended treatment of DVT/PEb | Idarucizumab (Praxbind®) |
Rivaroxaban (Xarelto®) | Factor Xa inhibitor | 7–11 h (2–4 h) | •Stroke prevention in non-valvular atrial fibrillation •Acute and extended treatment of DVT/PE •DVT prophylaxis post hip/knee surgery | Andexanet alfa c |
Apixaban (Eliquis®) | Factor Xa inhibitor | 8–15 h (0.5–2 h) | •Stroke prevention in non-valvular atrial fibrillation •Acute and extended treatment of DVT/PE •DVT prophylaxis post hip/knee surgery | Andexanet alfa b |
Edoxaban (Savaysa®) | Factor Xa inhibitor | 10–12 h (1–2 h) | •Stroke prevention in non-valvular atrial fibrillation •Treatment of DVT/PE acute | |
Warfarin (Coumadin®) | Vitamin K antagonist | 2–5 daysd (72–96 h)e | •Prophylaxis and acute and extended treatment of DVT/PE •Prophylaxis for thromboembolic complications associated with atrial fibrillation, cardiac valve replacement •Post myocardial infarction | Vitamin K, prothrombin complex concentrate, fresh frozen plasma |
Enoxaparin (Lovenox®) | Binds to antithrombin leading to FXa, FII inactivationf | 4.5–7 h (3–5 h) | •DVT prophylaxis in abdominal and orthopedic surgeries or in mobility limited hospitalized patients during acute illness •Treatment of DVT/PE •Unstable angina, non Q wave myocardial infarction •Treatment of ST-elevation myocardial infarction | Protamine sulfate (reverses about 60% of the anti-Xa activity) [131] |
Table 23.3
Anticoagulant dosing for VTE treatment
Anticoagulant | Dosing for acute treatment | Dosing for extended treatment |
---|---|---|
Dabigatran (Pradaxa®) | •Initial treatment with a parenteral anticoagulant (LMWH, UFH) for 5–10 days •CrCl >30 mL/min: 150 mg twice daily •CrCl <30 mL/min: contraindicated | •CrCl >30 mL/min: 150 mg twice daily •CrCl <30 mL/min: contraindicated |
Rivaroxaban (Xarelto®) | •CrCl >30 mL/min: 15 mg twice daily with food × 21 days, then 20 mg daily with food •CrCl <30 mL/min: contraindicated | •CrCl >30 mL/min: 20 mg daily with food •CrCl <30 mL/min: contraindicated |
Apixaban (Eliquis®) | •10 mg twice daily × 7 days, then 5 mg twice dailya | •2.5 mg twice dailya |
Edoxaban (Savaysa®) | •Initial treatment with a parenteral anticoagulant (LMWH, UFH) for 5–10 days •CrCl >50 mL/min: 60 mg daily •CrCl 15–50 mL/min or weight ≤60 kg: 30 mg daily | Not applicable |
Warfarin (Coumadin®) | •Overlap with LMWH or UFH for 4–5 days and until attainment of target INR •Adjust dose for goal INR 2–3 | •Adjust dose for goal INR 2–3 |
Enoxaparin (Lovenox®) | •1 mg/kg SC every 12 h or 1.5 mg/kg SC once dailyb •CrCl <30 mL/min: 1 mg/kg SC once daily | •1 mg/kg SC every 12 h or 1.5 mg/kg SC once dailyb •CrCl <30 mL/min: 1 mg/kg SC once daily |
Anticoagulation for Proximal DVT Due to a Strong Provoking Risk Factor
VTE caused by a reversible risk factor has a lower risk of recurrence than unprovoked VTE. The actual risk of clot recurrence varies according to the type of provoking factor. Patients with VTE after major surgery or trauma carry a very low risk of recurrence, with less than 1% of such patients experiencing a recurrent thrombotic event in the year following completion of anticoagulation and 3% having recurrent thrombosis at 5 years [26–28]. Based on these numbers, patients with a proximal DVT due to a “strong” provoking risk factor such as surgery, trauma, or profound immobilization are anticoagulated for a defined period of 3 months (Fig. 23.1) [29, 30]. The duration of anticoagulation in such patients is generally independent of any additional factors, including body mass index, patient comorbidities, or massive or life-threatening VTE. An exception is DVT patients with persistence of an otherwise reversible major thrombotic risk factor, in whom extended anticoagulation may be considered as long as the risk factor remains [31]. Patients with DVT arising in the context of spinal cord injury are usually given therapeutic anticoagulation for at least 3–6 months, with consideration of prophylactic-dose anticoagulation afterwards [32].
Anticoagulation for Proximal DVT Due to a Minor Provoking Risk Factor
Patients with VTE associated with a minor, nonsurgical thrombotic risk factor such as exogenous estrogen exposure or prolonged travel have a higher risk of recurrent thrombosis (6–8% clot recurrence at 1 year, 15% at 5 years) than those with VTE due to a strong thrombotic risk factor [26–28]. Because the recurrence rates for DVT and PE due to minor thrombotic risk factors are still lower than those observed for unprovoked thrombosis, such patients are usually anticoagulated for 3 months [30]. In many cases, however, additional factors may weigh into this decision, including age-appropriate and symptom-directed cancer screening, D-dimer testing, gender, thrombophilia testing, and thrombotic risk recurrence scores, like patients with unprovoked VTE (Fig. 23.1 and discussed further below).
DVT Associated with Exogenous Estrogen
Combined oral contraceptive pills (OCP) , hormone replacement therapy (HRT) , and selective estrogen receptor modulators (SERM) are associated with a two- to fourfold increased risk of thrombosis [33–37]. Such hormone exposure modulates levels of fibrinogen, antithrombin, proteins C and S, and numerous coagulation factors (including factors II, VII, and VIII), leading to an increased propensity toward thrombosis [34, 38–40]. For combined OCPs, thrombotic risk varies according to the specific types and doses of estrogen and progesterone components, with “total estrogenicity” (defined by the ratio of estrogen and progesterone) being of greater thrombotic significance than absolute dosage amounts [33, 34, 40]. Certain progesterone-only options are safer alternatives from a thrombotic standpoint, as studies report no significant increase in thrombosis for progesterone-only oral pills and intrauterine devices, although injectable depot progesterone may pose some thrombotic risk [41, 42]. For HRT, hormonal formulation may also affect thrombotic risk although debate remains as to which modes of delivery are safest, as some studies indicate increased risk with oral estrogen [35, 43] while others suggest a higher risk with the transdermal form [44].
Older literature suggested that women with OCP- or HRT-associated thrombosis had similar rates of recurrent VTE as those whose thrombosis was unprovoked [45–47]. More recent data, however, report a low risk of clot recurrence, similar to DVT and PE due to strong thrombotic risk factors [48]. Based on this, current guidelines recommend 3 months of anticoagulation in patients with estrogen-associated VTE [30, 49], although many clinicians, including those at our institution , opt to pursue age-appropriate and symptom-directed cancer screening in such patients as well [50].
DVT Associated with Pregnancy
The incidence of VTE in pregnancy is estimated at about 1–2 cases per 1000 women, about 5–10 times higher than nonpregnant women, with the highest risk during the 6–12 weeks following delivery, at which time the risk rises to 15 to 35-fold [51–56]. In women with acute DVT or PE during pregnancy, anticoagulation with UFH or LMWH, neither of which crosses the placenta, is recommended from the time VTE is diagnosed until 6 weeks postpartum, for a minimum total duration of 3 months [54, 57, 58].
Women without active DVT or PE but with a history of prior estrogen-associated VTE should receive antepartum and postpartum prophylactic anticoagulation to mitigate risk of another event, due to a 6–9% chance of a recurrent VTE during pregnancy [52, 57]; postpartum thromboprophylaxis typically continues to 6 weeks after delivery. Postpartum prophylactic anticoagulation may also be indicated in pregnant women with other thrombotic risk factors; antepartum thromboprophylaxis is sometimes given as well, although recent data have called into question its utility [59].
DVT Associated with Prolonged Travel
The risk of DVT due to prolonged travel is very small [60]. For unclear reasons, such DVT, when it occurs, tends to involve the distal rather than proximal leg [14, 61]. As with estrogen-associated DVT, patients with travel-related DVT are typically anticoagulated for 3 months [30], with many clinicians also recommending cancer screening.
For patients without active thrombosis but with thrombotic risk factors (e.g., prior VTE, obesity, recent surgery, OCP or HRT use, pregnancy, cancer, heritable thrombophilia), graduated compression stockings may be beneficial in reducing the risk of DVT [62]. Prophylactic anticoagulation is also increasingly being given to such patients, although data for this is less strong [63].
May–Thurner Syndrome
May–Thurner syndrome (compression of the left common iliac vein by the overlying right iliac artery) is an anatomic abnormality commonly affecting young women in their third through fifth decades. Risk factors include pregnancy, exogenous estrogen exposure, obesity, or heritable thrombophilia. Because the DVT arises from vascular compression, treatment typically involves a combination of catheter-directed thrombolysis, stent placement, and anticoagulation; the latter of which may last a year or longer, depending on stent patency and the presence or absence of PTS [64].
Anticoagulation for Unprovoked DVT
Unprovoked VTE occurs in the absence of identifiable thrombotic risk factors (Table 23.1). The risk of clot recurrence is 2- to 2.5-fold higher for unprovoked than for provoked VTE, with 10% of patients experiencing a recurrent VTE after 1 year and at least 30% having recurrent thrombosis at 5 years [27, 65]. Consensus guidelines therefore recommend a minimum of 3 months of anticoagulation in patients with unprovoked DVT or PE, with consideration of extended or indefinite anticoagulation in those with an acceptable bleeding risk (Fig. 23.1) [30]. Several additional clinical and laboratory factors should be considered when weighing the risks and benefits of extended anticoagulant therapy.
Bleeding Risk
Extended anticoagulation with VKA effectively reduces risk of thrombosis recurrence by about 90%, with similar results seen in extended treatment using LMWH and TSOACs (Dabigatran, Apixaban, Rivaroxaban) [29, 66–69], but at a cost of a two to threefold increased risk of bleeding [28]. Patients on extended anticoagulation have an annual major bleeding risk of about 1–3%, which rises to 4–5% in older individuals [66, 70, 71]. Factors associated with increased bleeding include advanced age (greater than 65 years old), cancer, previous bleeding, thrombocytopenia, renal or liver failure, concomitant use of antiplatelet agents or nonsteroidal anti-inflammatory drugs, recent surgery, or frequent falls [30].
Bleeding risk assessment tools such as the HAS-BLED score have been developed for patients on chronic VKA for atrial fibrillation. While these scores have not been validated in VTE, some data suggests that they might be an accurate predictor of early bleeding risk in such patients [72].
Case fatality rates, defined as the percentages of fatal events among patients with a particular disease, offer an additional tool to weigh the potential risks and benefits of long-term anticoagulation [73, 74]. In VTE patients on anticoagulation, case fatality rates due to recurrent VTE and bleeding are similar during the first 3 months of anticoagulant therapy, following which the case fatality rate for recurrent VTE drops significantly [73, 74]. The case fatality rates for recurrent DVT are half that of recurrent PE and one-third that of major hemorrhage, so in order to benefit from long-term anticoagulation, the estimated rate of recurrent DVT in an individual patient must be three times that of major hemorrhage [73, 75].
In patients with unprovoked VTE who have a high long-term risk of bleeding, where the risk of serious hemorrhage outweighs the projected benefit of ongoing anticoagulation, anticoagulant therapy is usually discontinued after 3 months [29, 75]. For patients with low to moderate bleeding risk, discussions regarding extended vs. short-term anticoagulation are more complicated and highly individualized. This is an area of active investigation, and in hopes of individualizing recommendations, many studies have focused on identifying factors that separate patients with the highest risk of recurrent thrombosis , who have the most to gain from extended treatment, from those with lower recurrence risk, in whom extended anticoagulation may reasonably be avoided.
Estimating Recurrent VTE Risk
D-dimer: D-dimer is a by-product of fibrinolysis and is used for diagnostic purposes as a noninvasive marker to exclude VTE [76, 77]. In the multicenter prospective PROLONG trial, D-dimer levels were measured before and 1 month after stopping anticoagulation in 608 patients with unprovoked VTE (the vast majority of whom had proximal limb DVT without PE) who completed at least 3 months of VKA therapy [78]; those with D-dimers within normal range remained off treatment, whereas those with elevated values after cessation of anticoagulation were randomized to either resume anticoagulation or remain off it for the next 18 months. The highest recurrence rate (15%) occurred in patients with elevated D-dimer levels who remained off anticoagulation, compared to those with abnormal D-dimers who resumed anticoagulation (2.9%) and those with normal D-dimers (6.2%). On extended follow-up, patients with negative D-dimers continued to have lower risks of VTE recurrence compared to those with positive D-dimers (estimated annual risk, 3.5% vs. 8.9%, respectively) [79, 80]. Surveillance monitoring of D-dimers in patients who have stopped anticoagulation may also have utility [81].
Gender : Men with unprovoked VTE have a 1.5–2.5 times higher risk of recurrent VTE than women [29, 30, 45, 82, 83]. The increase in recurrent thrombotic risk attributed to gender is independent of D-dimer status [84, 85]. Men with unprovoked VTE who have a negative D-dimer after stopping anticoagulation have a higher risk of clot recurrence than women (9.7% vs. 5.4% per patient year, respectively), indicating that a negative D-dimer is not as reassuring in men as it may be for women [85].
Surveillance ultrasonography : Although commonly performed, a role for surveillance ultrasonography in assessing clot recurrence risk in patients with DVT has not been established. In most prospective studies and meta-analyses, residual vein occlusion following an initial period of anticoagulation either was not associated with increased VTE recurrence or demonstrated only a minor association [86–89], while a positive D-dimer appeared to be a stronger predictor [88, 89]
Thrombophilia testing : Five major thrombophilias have been described: factor V Leiden (FVL), prothrombin gene mutation, and deficiencies of antithrombin, protein C, and protein S [90]. Minor thrombophilias include elevations in other coagulation factors, such as factor VIII, von Willebrand factor, or plasminogen activator inhibitor [91–93]. The most common among these is FVL, present in about 5% of Caucasians and associated with DVT more than PE (the so-called “FVL paradox ” [94]).While heritable thrombophilias increase the overall lifetime risk of VTE, they exert at most only a minor effect on recurrent VTE in patients with an unprovoked DVT or PE [26, 47, 95], although such effects may be amplified when present in combination [96]. Consensus guidelines advise against thrombophilia testing in patients with provoked DVT or PE but are uncertain as to its role in those with unprovoked VTE [90, 97, 98], as such testing has not been shown to change outcomes [99].
Other factors: Several other factors may be associated with an increased risk of recurrent thrombosis, including older age, PTS, and obesity [29, 47, 93, 100–103]. None of these on an individual level play strongly into decisions about duration of anticoagulation.
Risk assessment models : Several multivariable risk assessment models have been developed to aid in estimation of recurrent thrombosis risk in patients with unprovoked DVT and PE [71, 104–106]. The goal of all of these models is to aid clinicians in identifying patients with projected recurrence risks low enough to justify stopping anticoagulation. Male gender and abnormal D-dimers are the only variables identified in all risk assessment models as adverse predictors of VTE recurrence risk. In the absence of sufficient external validation studies in diverse patient populations (the risk models described below were derived from predominantly Caucasian populations) and prospective studies assessing the clinical impact of management decisions using these scores, it remains unclear how to best utilize these tools [24, 106].
The Vienna prediction model was derived from a study of 929 Austrian patients with unprovoked thrombosis who were treated with at least 3 months of anticoagulation [104]. Patients with cancer, estrogen-associated VTE, or inherited thrombophilias were excluded. Three variables—male gender, location of clot (proximal DVT vs. PE), and elevated D-dimer measured 3 weeks after discontinuation of anticoagulation—were significantly associated with an increased rate of recurrent VTE and compiled to estimate an individual patient’s risk of recurrence at 12 and 60 months (Table 23.4). An updated version was published in 2014 enabling ongoing risk assessments based on serial D-dimer measurements up to 15 months post-anticoagulation [107]. The original Vienna prediction model was externally validated in a large cohort of 904 patients [65], although the updated model failed to predict recurrent VTE rates in a multicenter study of older adults [108].
The DASH score was derived from a study of 1818 patients with unprovoked VTE who completed at least 3 months of VKA therapy. Four factors predicted clot recurrence: an abnormal D-dimer (measured 3–5 weeks after stopping treatment), age less than 50 years, male sex, and VTE not associated with hormonal therapy (Table 23.4). The DASH score has not been externally validated.
The men continue and HERDOO2 rule was derived from a multicenter prospective study of 646 patients with unprovoked VTE, which found that upon discontinuation of anticoagulation, men faced a 13.7% annual risk of recurrent VTE with no identifiable low-risk group compared to a 5.5% annual risk for women [105]. Clinical predictors were identified to stratify women into low- and high-risk groups, including signs of venous stasis (hyperpigmentation, edema, or redness of either leg), D-dimer ≥250 μg/L while on warfarin, obesity (body mass index ≥30 kg/m2), and age 65 years or older (Table 23.4). Preliminary results from the REVERSE II trial presented at the European Society of Cardiology meeting in 2016 validate the HERDOO2 rule [109].
The DAMOVES score was derived from a prospective study of 398 Spanish patients with unprovoked VTE 124. Among preselected variables, the authors found that abnormal D-dimer (while on anticoagulation), advanced age, inherited thrombophilic mutation (FVL, prothrombin G20210A mutation), obesity, the presence of varicose veins, elevated factor VIII (eight), and male sex were associated with increased VTE risk (Table 23.4). A limitation of this model is its reliance on a single factor VIII level, which may fluctuate in inflammatory states.
Table 23.4
Proposed risk prediction tools