Decision Making in Venous Thromboembolism

Determining the duration of anticoagulation is more complex in patients with idiopathic/unprovoked VTE. Kearon and colleagues found that in patients with first idiopathic VTE, patients who were anticoagulated for 24 months versus 3 months had lower risk of recurrent VTE (1.3% per patient-year with 24 months versus 27.4% per patient-year with 3 months) [69].Similar studies and meta-analyses have demonstrated decreased recurrence rates in patients anticoagulated for a prolonged period of time. However, one study of prolonged anticoagulation revealed that at 3 years there was no difference in recurrence rate in patients with PE who were anticoagulated for 6 months versus 1 year [70].The likelihood of recurrent DVT in patients with first episode of idiopathic proximal DVT treated with either 3 months or 12 months of warfarin was similar after treatment was discontinued [71].Prolonged periods of anticoagulation do not directly influence risk of recurrence but instead may only delay occurrence of a second event [72].For that reason, the decision is essentially whether to anticoagulate for 3 months or to continue therapy indefinitely [73]. Current guidelines recommend continuing anticoagulation for 3 months in those at high risk of bleeding, and continuing for an extended duration in those at low or moderate bleeding risk [51]. Patients' values and perferences should be entertained and decisions made on a patient-by-patient basis.

For patients at high risk of recurrent VTE, we generally recommend indefinite anticoagulation unless the patient has a significantly elevated bleeding risk or strongly prefers to discontinue anticoagulation and compliance concerns are evident. High-risk patients are those who have suffered from multiple episodes of recurrent VTE, those who have clotted while being anticoagulated, and those with acquired risk factors, such as antiphospholipid antibodies and malignancy. Other high-risk groups are those with high-risk thrombophilias such as deficiency of protein S, protein C, or antithrombin, homozygous factor V Leiden or prothrombin gene mutations, and compound heterozygous factor V Leiden/prothrombin gene mutation in the setting of an unprovoked event. Further discussion of models for risk assessment of recurrence is provided below.

Assessment of Bleeding Risk

The bleeding risk associated with the use of anticoagulation must be weighed against the risk of clotting events when determining duration of anticoagulation, especially in those patients for whom indefinite anticoagulation is a consideration. Risk of bleeding while on anticoagulation is approximately 1–3% per 100 patient-years [74],but concomitant medical conditions such as renal failure, diabetes-related cerebrovascular disease, malignancy, advanced age, and use of antiplatelet agents all increase the risk of bleeding. Bleeding risk is highest when patients first initiate anticoagulation and is approximately 10 times the risk in the first month of therapy than after the first year of therapy [75].

Risk assessment models such as the RIETE score may be helpful when indefinite anticoagulation is a possibility [76].The RIETE score encompasses 6 risk factors (age > 75 years, recent bleeding, cancer, creatinine level > 1.2 mg/dL, anemia, or PE at baseline) to categorize patients into low risk (0 points, 0.3% risk of bleeding), intermediate risk (1–4 points, 2.6% risk of bleeding) and high risk (> 4 points, 6.2% risk of bleeding) within 3 months of anticoagulant therapy. The ACCP has developed a more extensive list of 17 potential risk factors for bleeding to categorize patients into low risk (no risk factors, 0.8%/year risk of bleeding), intermediate risk (1 risk factor, 1.6%/year risk of bleeding) and high risk (2 or more risk factors, >6.5%/year risk of bleeding) categories [77].The RIETE score is simpler to use but was not developed for assessing risk of bleeding during indefinite therapy, while the ACCP risk categorization predicts a yearly risk and is therefore applicable for long-term risk assessment but is more cumbersome to use. In practice, we generally use a clinical gestalt of a patient’s clinical risk factors (particularly age, renal or hepatic dysfunction, and frequent falls) to assess if they may be at high risk of bleeding and if the risk of indefinite anticoagulation may thus outweigh the potential benefit.