Long-term risk of major bleeding after discontinuing anticoagulation for unprovoked venous thromboembolism: a systematic review and meta-analysis

Background  The long-term risk of major bleeding after discontinuing anticoagulant therapy for a first unprovoked venous thromboembolism (VTE) is uncertain. Objectives  To determine the incidence of major bleeding up to 5 years after discontinuing anticoagulation for a first unprovoked VTE. Methods  We searched MEDLINE, EMBASE, and Cochrane CENTRAL (from inception to January 2021) to identify relevant randomized controlled trials (RCTs) and prospective cohort studies reporting major bleeding after discontinuing anticoagulation in patients with a first unprovoked or weakly provoked VTE who had completed (IMAGE_)3 months of initial treatment. Unpublished data on major bleeding events and person-years were obtained from authors of included studies to calculate study-level incidence rates. Random-effects meta-analysis was used to pool results across studies. Results  Of 1,123 records identified by the search, 20 studies (17 RCTs) and 8,740 patients were included in the analysis. During 13,011 person-years of follow-up after discontinuing anticoagulation, the pooled incidence of major bleeding (n = 41) and fatal bleeding (n = 7) per 100 person-years was 0.35 (95% confidence interval [CI]: 0.20-0.54) and 0.09 (95% CI: 0.05-0.15). The 5-year cumulative incidence of major bleeding was of 1.0% (95% CI: 0.4-2.4%). The case-fatality rate of major bleeding after discontinuing anticoagulation was 19.9% (95% CI: 10.6-31.1%). Conclusion  The risk of major bleeding once anticoagulants are discontinued in patients with a first unprovoked VTE is not zero. Estimates from this study can help clinicians counsel patients about the incremental risk of major bleeding with extended anticoagulation to guide decision making about treatment duration for unprovoked VTE.