Postintervention duration of anticoagulation in venous surgery

Initial anticoagulant therapy

In patients with DVT an initial course of low-molecular-weight heparin (LMWH) or fondaparinux is recommended by the ACCP guidelines, over the use of unfractionated heparin. ⁶ LMWH or fondaparinux should preferably been given once daily by subcutaneous injection. While oral anticoagulation with vitamin K antagonists (VKA) should be started early, the parenteral therapy should be continued for a minimum of five days and until the international normalized ratio (INR) is 2.0 or above for at least 24 hours (Grade 1B). Initial management in patients with an isolated distal DVT of the leg provoked by surgery or by a non-surgical transient risk factor is basically identical to the one for patients with proximal DVT. The long-term anticoagulant management is different for those with a provoked versus those with an unprovoked DVT.

Duration of anticoagulant therapy

The duration of anticoagulation for a first DVT depends mainly on the character and extend of the original event, and the associated risk for recurrence.^7,8 In patients with a provoked DVT or with isolated calf vein thrombosis the risk of recurrent DVT is low. ⁹ Consequently, the ACCP recommends treatment with anticoagulation for three months over treatment of a longer time-limited period (e.g. 6 or 12 months) (Grade 1B) or extended therapy (Grade 1B regardless of bleeding risk). In patients with an unprovoked proximal DVT of the leg the risk of recurrent DVT is high with incidences up to 20% within a period of two years.^4,8 Therefore, treatment with anticoagulation for at least three months is recommended over treatment of a shorter duration (Grade 1B).

Risk—benefit assessment for prolonged anticoagulation

After three months of treatment, the risk-benefit ratio of extended therapy should be assessed. One of the critical elements in the ACCP risk-benefit assessment is the patient's bleeding risk. Thus, in patients with a first, unprovoked proximal DVT of the leg and who have a low or moderate bleeding risk, extended anticoagulant therapy is recommended over three months of therapy (Grade 2B). ⁶ In contrast, if such patients have a high bleeding risk three months of anticoagulant therapy is recommended over extended therapy (Grade 1B). In patients with a first unprovoked isolated distal DVT of the leg, three months of anticoagulant therapy is suggested over extended therapy in those with a low or moderate bleeding risk (Grade 2B) and recommended in those with a high bleeding risk (Grade 1B). ⁶ The other important issue in risk-benefit assessment is the risk of recurrent thrombosis. Tailoring of treatment duration is based on a number of risk factors, including sex, age, residual thrombosis, persistent thrombophilia risk factors and biomarkers. The contribution of the individual risk factors or combinations thereof has not been fully established yet.

Persistent thrombophilia risk factors

While common genetic thrombophilia traits like factor V Leiden and the prothrombin 20210 variant increase the risk of a first DVT, they do hardly bear relevance for the risk of recurrent DVT (RR 1.4–1.5). ¹⁰ While other rare thrombophilia defects including deficiencies in antithrombin, protein C or S, are associated with an increased risk of a first DVT, occurring at a young age, as well as with recurrent thrombosis. ¹¹ Although elevated coagulation factors pose a risk for a first thrombosis, only an elevated factor VIII level is also associated with a risk of recurrent DVT. ¹²

Practically, it means that standard testing for thrombophilia is not generally warranted, but still could be considered in young subjects (<50 years) and mostly confined to women (child bearing potential). The presence of factor V Leiden or prothrombin 20210 variants should not affect the duration of anticoagulation after a first unprovoked DVT. In young individuals and particularly those with a positive family history the finding of a deficiency in antithrombin, protein C or S may be sufficient reason to continue anticoagulation indefinitely, provided this is also the patient's preference. Acquired risk factors include antiphospholipid antibodies (lupus anticoagulant [LAC] and anticardiolipin antibodies [ACA]). When found in a patient with thrombosis, laboratory testing for LAC and ACA should be repeated at three months and if confirmed a diagnosis of antiphospholipid syndrome is essentially confirmed. While there is substantial heterogeneity both in the nature and the biological effects of the antibodies and in the clinical presentation of the antiphospholipid syndrome (with a wide range is estimated recurrent rates of DVT, between 10% and 70%), the decision on duration of anticoagulation is difficult. ¹⁰ In general, when we find and confirm APS we will recommend continued anticoagulation unless the bleeding risk or other arguments prevent this. Routine testing for homocysteine levels has been abandoned since, except for the exceptional case of homozygous homocysteinuria, the finding of an elevated homocysteine level as such, usually does not have therapeutic consequences (unless vitamin deficiency is suspected). ¹³ In all patients, also the elderly, we routinely test for LAC, ACA and factor VIII levels as these are the only risk factors that may influence the decision to continue anticoagulation.^10,14

Patients with malignancies

A distinct management is recommended in patients with a DVT of the leg in the presence of active cancer. If the risk of bleeding is deemed ‘not high’, the ACCP recommends extended anticoagulant therapy over three months of therapy (Grade 1B), and if there is a high bleeding risk, extended anticoagulant therapy is suggested (Grade 2B). With regard to the type of anticoagulation, in patients with DVT and cancer (treatment) LMWH is the first-choice-suggested medication over VKA, because of the associated lower risk of recurrence (relative risk reduction compared with VKA approximately 50%). The risk for bleeding does not differ. ¹⁵

Duration of anticoagulant therapy for a recurrent event of DVT

High weight is given to bleeding risk for the choice of treatment in patients with a second episode of unprovoked DVT. The ACCP recommends extended anticoagulant therapy over three months of therapy in those who have a low bleeding risk (Grade 1B), and suggests extended anticoagulant therapy in those with a moderate bleeding risk (Grade 2B). In patients with a second unprovoked DVT, who have a high bleeding risk, three months of anticoagulant therapy is suggested over extended therapy (Grade 2B). These recommendations raise questions about the definition of ‘extended’ therapy. In practice, this oftentimes means lifelong anticoagulant therapy. However, upon ageing co-morbidity and other problems occur that may result in an increased risk of bleeding and an associated shift in the benefit-risk ratio. Our policy is therefore to follow patients at periodic intervals (e.g. annually) and tailor long-term anticoagulant treatment on a more individual basis. This follow-up is also ACCP recommended and includes risk assessment based on common risk factors for thrombosis and bleeding to which biomarkers may be added, including d-dimer and thrombophilia factors like factor VIII and LAC (see next sections).

Hence, although technically ‘extended’ is appropriate, a more individually based assessment may be needed to determine whether anticoagulation should be prolonged for a certain period or be prescribed lifelong. As also considered by the ACCP, patient preferences are increasingly important in assessing the duration of anticoagulation. In practice, many patients have rather clear opinions on whether they intent to risk a recurrent DVT (or pulmonary embolism) or not, i.e. whether they prefer to continue the burden of anticoagulation or not. One element that may be hard to estimate though is the bleeding risk. In patients who never experienced bleeding complications a true risk-benefit assessment may be confined to estimating the burden of recurrent thrombosis.

Patient characteristics

Patients who undergo surgical or radiological interventions in the acute phase of DVT usually are patients with extensive leg complaints or are selected based on their increased risk for PTS.^16,17,18 Currently, this means that mainly patients with iliofemoral DVT will be eligible for intervention. ¹⁹ The iliofemoral localization of the DVT is also associated with an increased prevalence of anatomical lesions. ²⁰

Surgical interventions in the sub acute phase are usually initiated to relieve post-thrombotic sequels. ²¹

Initial anticoagulant therapy

In patients that undergo venous thrombolytic removal of the clot in the acute phase of thrombosis, irrespective of the technique and whether or not stenting is applied, immediate anticoagulation following this procedure is pertinent to reduce the risk of recurrent thrombosis and thrombus propagation. According to recent ACCP guidelines, the anticoagulant management in patients who have had any method of thrombus removal performed, the same intensity and duration of anticoagulant therapy as in comparable patients who do not undergo thrombosis removal is recommended (Grade 1B).

In other venous surgical conditions, like venous reconstruction for relieving PTS by revascularization and/or stenting, which is becoming more popular these days in addition to clot removal procedures, the anticoagulant management is not evidence based. In such situations, in parallel to the situation without interventions, individual patient characteristics and risk factors will have to determine the optimal anticoagulant management.

Increased levels of d-dimer

The assessment of the d-dimer concentration in blood is a sensitive marker for activated clotting and is commonly used as a diagnostic tool for the exclusion of VTE in conjunction with a clinical decision rule. ²² In patients who are being anticoagulated for DVT, the d-dimer level will in general be suppressed. An elevated level of d-dimer was associated with an increased risk of recurrent DVT in patients during anticoagulant treatment, as well as in patients who had already stopped anticoagulation for one month.^23,24 A systematic review on 2000 patients showed an annual risk of recurrent DVT of 3.5% when a d-dimer after cessation of anticoagulation was below the cut-off level, while the risk was nearly 9% in those above this level. ²⁵ High levels of d-dimer are also associated with an increased risk for PTS.^26–28 Consequently, PTS has been associated with a twofold increased risk for recurrent DVT. ²⁹ Palareti et al. investigated the effect of resuming anticoagulation in patients with an elevated d-dimer level. He could show that the risk of recurrent DVT was significantly reduced in those subjects with 2 events/100 patient-years in those on anticoagulation versus 11 events/100 patient-years in those with an elevated d-dimer level, but not resuming anticoagulation. Since d-dimer is a reflection of activated clotting, as well as an indication of an activated fibrinolytic system, the use of d-dimer as a biomarker to assess the risk for recurrent DVT in patients after surgical procedures and stenting may be advocated. ³⁰

Residual venous thrombosis

Residual venous thrombosis (RVT) indicates residual thrombosis or vessel wall abnormalities (re-modelling) following DVT. The presence of residual thrombosis was associated with an increased risk of recurrent thrombosis in the DACUS trial, with its follow-up recently reported as the extended DACUS trial; the relative risk for recurrent thrombosis was 7.4 (95% confidence interval [CI] 4.9–9.5).^31,32

In the study by Le Gal et al., ³³ however, no significant association between RVT and recurrent VTE was noted with a HR of 1.4 (95% CI 0.9–2.1). In a systematic review and meta-analysis on this topic Carrier et al. ³⁴ noted only a significant association between RVT and (any) recurrent DVT (provoked and unprovoked together, while the association was not significant for unprovoked DVT only). Tan et al. ³⁵ showed in another systematic review that the association of RVT with recurrent VTE was dependent on additional factors like the presence of malignancy and on the criteria used. In summary, RVT does not provide unequivocal weight in determining the duration of anticoagulation. Since there is limited evidence that prolonged anticoagulation may reduce the volume of RVT, it is our policy to prolong anticoagulation for another six months, in particular in patients with unprovoked DVT, since they have a higher risk of recurrent DVT as compared with the provoked DVT patients. For the relation between RVO and PTS the same holds true. There is no unequivocal evidence linking RVT to an increased risk for PTS.

In summary, a quantitative d-dimer assay may be helpful in detecting an increased risk of recurrent DVT. Additional factors like timing of blood sampling, patient age and assay cut-off point do not appear to influence the value of d-dimer interpretation. ³⁶ While an elevated d-dimer level may indicate an increased risk, in patients free of symptoms of recurrence the decision to restart anticoagulation can be difficult. Patient and doctor preferences of acceptability of risks are important. In those who feel at risk of recurrent DVT, an elevated d-dimer level may be sufficient to resume anticoagulation indefinitely, while in those with a value below cut-off anticoagulation it can probably be rather safely withheld. In contrast, in patients that are reluctant to take anticoagulants (or in case of an expected high bleeding risk) a policy of repeated d-dimer testing might be considered with a low threshold of objective testing in case of symptoms of VTE. The latter strategy is however not evidence based.

Risks induced by venous stenting

The risks associated with the insertion of venous stents are not analogous to the risks involved in the case of arterial stenting. First and foremost, the indication for arterial stenting is usually a stenosis or narrowing of the artery originating from atherosclerosis. The atherosclerotic lesion induces a hypercoagulable state caused by shear stress and endothelial dysfunction. ³⁷ This hypercoagulability is not a transient state, but continues to exist even after stenting. Therefore, anticoagulant treatment usually based on platelet aggregation inhibitors such as aspirin or Clopidogrel is necessary for longer periods of time. Venous stenting is performed in the acute phase of thrombosis when periprocedural a stenosis, mostly form an anatomic origin, is observed. The venous vessel wall may be temporally perturbed by the acute thrombosis, but will gradually return to its unperturbed state after the initial phase of the DVT. ³⁸ Therefore, based on the presence of a venous stent alone, prolonged anticoagulant treatment after the initial phase is not warranted. Whenever venous stenting is performed to relieve PTS, the necessity for anticoagulant treatment should depend on the individual risk profile for thrombosis for that specific patient.

New oral anticoagulants

New oral anticoagulants (NOACs) have been tested in large phase 3 trials in patients with symptomatic VTE. Rivaroxaban and dabigatran were shown to be non-inferior to warfarin with a comparable bleeding risk. In the Einstein study on rivaroxaban, no LMWH was given during the initial treatment period, while this was maintained in the dabigatran trial. Rivaroxaban has been registered for treatment of DVT in the USA and other countries. The advantages of the NOACs are in particular the lack of need for laboratory testing and dose-adjustment, the increased safety profile in terms of a lower rate of intracranial bleeding and, in case of rivaroxaban, the avoidance of parenteral LMWH dosing in the initial phase of DVT. The potential downsides to consider are the risk of lack of compliance, potentially more relevant for the unmonitored NOACs and the risk of problems in prescribing these drugs by physicians not aware of the restrictions that apply Importantly, impaired renal function should be seriously considered, not just at onset of prescribing an NOAC, but probably also at intervals afterwards, especially in the elderly. Moreover, the management of bleeding is still hindered by the absence of specific antidotes for NOACs.

The decisions of continued anticoagulation are similarly difficult and should be individually based. Extension studies have shown reduced rates of recurrent thromboembolism in patients on NOACs at more or less acceptable rates of bleeding. For the extension studies there is no direct comparison between NOACs and VKA available. In general, for the decision to prolong anticoagulation the same considerations as for VKA apply, i.e. individual assessment of the risks of recurrent thrombosis versus the bleeding risk.

Aspirin

Another way of reducing the risk of recurrent DVT is to switch from anticoagulation to fixed low-dose aspirin. There is an interesting indication for this type of medication, i.e. the increased rate of arterial thrombotic events, particularly myocardial infarction (MI) after cessation of anticoagulation in patients with unprovoked DVT. Several studies have indeed demonstrated an increased burden of atherosclerosis in patients with unprovoked DVT. Along this line cardiovascular risk management should be taken into consideration in all patients with unprovoked DVT, particularly in those at an age at risk of MI, roughly >50 years of age. This risk assessment could also include secondary prevention for DVT and primary prevention for arterial thrombosis with aspirin. The primary preventive effect of aspirin has been studied in the past. In the absence of DVT, aspirin would usually not be recommended to otherwise healthy persons. However, the presence of DVT provides the opportunity to combine venous and arterial risk reduction and may be worth considering. While older studies provided some evidence of risk reduction, a meta-analysis of two recent trials confirms this effect, showing a ± 30% reduced risk of recurrent VTE on low-dose aspirin (100 mg).^39,40 In these studies the rate of major bleeding complications was low. Hence, aspirin can be an important means of secondary prevention for VTE in those patients with unprovoked DVT in whom prolonged anticoagulation is not feasible.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not for-profit sectors.