Optimal Duration of Anticoagulation for Unprovoked Venous Thromboembolism: A Threshold Decision Analysis Model

Background: The optimal duration of long-term anticoagulation (AC) following an unprovoked venous thrombosis event (VTE) after the initial 3-6 months is unknown. Despite the providers careful consideration of the bleeding risk, probability recurrence off AC, and the patient's values and preferences (VP) about avoiding a recurrent VTE and bleed, the final decision remain subjective. When presented with the same patient, multiple providers may make different recommendations about the duration and intensity of AC. Using threshold decision analysis to model the effects of extended AC treatment we created and tested a model that quantifies the recommendations of the American Society of Hematology (ASH) thrombophilia for long-term AC over 10 years (10 Pr(VTE)). Methods: We developed a threshold decision model incorporating VTE recurrence risk, AC efficacy, bleeding risks, and patient values and preferences (V&P). Key parameters included: - VTE risk: 10% in year 1, then 5%/year - AC efficacy: Relative Risk Reduction (RRR) = 0.85 [0.77 to 0.9] - Bleeding risk: Low (0.5%/year) and High (1.5%/year) - Major bleeding on AC: RR 2.17 - Relative value (RV): patient's weighting of VTE vs. bleeding risk Results: The model calculates a threshold probability (T) at which AC benefits equal bleeding harms. AC is recommended when the probability of VTE without treatment (Pr(VTE)) exceeds the risk of bleeding. In most scenarios, the 10 Pr(VTE) consistently exceeded T, favoring long-term AC. This held true across various RV assumptions (RV=0.5, 1, 2) for both low and high bleeding risk scenarios. The only exception was in the high bleeding risk, worst-case scenario, where patients valuing avoiding bleeding twice as much as VTE (RV=2) might opt against long-term AC. Conclusions: Based on current evidence of AC efficacy, bleeding risks, and estimated VTE recurrence rates, long-term AC appears to be the optimal management strategy for most patients with unprovoked VTE. Only in cases of high bleeding risk combined with strong patient preference for avoiding bleeding (RV>2) might discontinuation of AC be considered. This model provides a framework for personalized decision-making, considering individual patient risks and preferences in determining the optimal duration of AC for unprovoked VTE.