Efficacy and safety of anticoagulant treatment in elderly patients with venous thromboembolism beyond the first 3 to 6 months of therapy: A narrative review

Extended anticoagulation remains a crucial consideration in patients with venous thromboembolism (VTE), especially in older adults. In this population, the two main challenges are the increased risk of bleeding and the underrepresentation in clinical trials. Current guidelines recommend indefinite anticoagulation for unprovoked VTE, except in patients at high bleeding risk, where the benefits and risks must be carefully balanced. In contrast, VTE triggered by a transient risk factor typically does not require prolonged treatment once the factor has resolved. For patients with chronic risk factors, indefinite anticoagulation is often recommended, but decisions should be individualized based on patient preference, risk of recurrence, and bleeding risk. Emerging data highlight the potential role of D-dimer testing in guiding extended therapy. However, in older adults, D-dimer levels may be frequently elevated for reasons unrelated to VTE, limiting its specificity. Real-world studies show wide variability in practice, with many clinicians opting for anticoagulation beyond six months regardless of whether the VTE was provoked. Recent observational data indicate that extended therapy beyond two years may significantly lower recurrence rates, especially following pulmonary embolism, with acceptable bleeding risks. New tools, such as the VTE-PREDICT risk score, aim to support personalized decision-making by estimating both recurrence and bleeding risks over 5 years. Nevertheless, validated bleeding risk scores in older patients remain suboptimal, demonstrating limited predictive power. Alternative strategies, including low-dose DOACs and adjunctive measures such as proton pump inhibitors, may reduce bleeding risk. Balancing recurrence prevention against bleeding complications remains central to optimizing extended-phase VTE treatment, particularly in elderly or comorbid populations.