Discontinuation of oral anticoagulation and risk of worse clinical outcomes in atrial fibrillation patients: a report from the Murcia AF Project III

Background Oral anticoagulation (OAC) is central for the appropriate management of atrial fibrillation (AF) and significantly reduces the risk of stroke. However, OAC is often discontinued in the elderly, especially when high bleeding risk, history of bleeding or after side effects of OAC. Purpose To investigate the rate of OAC withdrawal and its impact on the risk of severe clinical outcomes in a "real world" cohort of AF patients. Methods Prospective study including AF outpatients starting vitamin K antagonists (VKAs) or direct-acting OACs (DOACs) from January, 2016 to November, 2021. During 2-years follow-up, we recorded all changes and discontinuations of OAC, as well as clinical outcomes including ischemic stroke/transient ischemic attack [TIA], cardiovascular outcomes (composite of myocardial infarction, ischemic stroke/TIA, systemic embolism, and venous thromboembolism), major bleeds, and all-cause deaths. The risk for every clinical outcome of interest associated to discontinuation of OAC was assessed by competing risk regression, using the Fine and Gray model, considering the first clinical event as a competing risk (i.e. ischemic stroke/TIA or death, cardiovascular outcomes or death, and major bleed or death). Results We included 3259 patients (52.8% women; median age 77 [70-83] years), 1050 on VKAs and 2209 on DOACs. During follow-up, 156 (4.8%) patients stopped OAC (21 [0.6%] started antiplatelets alone, 30 [0.9%] switched to heparin and 105 [3.3%] were prescribed no antithrombotic therapy). There was a higher rate of OAC cessation among VKA users when compared to DOAC users (7.0% vs. 3.7%, p<0.001). The most common reasons for stopping OAC were the bleeding events during the follow-up (35.9%), starting chemotherapy during follow-up (7.7%), and frailty or palliative care (7.7%). In proportion, patients who withdrawn OAC had more subsequent ischemic stroke/TIA (7.7% vs. 0.9%, p<0.001), composite cardiovascular outcomes (8.3% vs. 1.5%, p<0.001), major bleeding events (8.3% vs. 1.0%, p<0.001) and death (32.7% vs. 12.4%, p<0.001). Competing risk regression analyses adjusted for several risk factors demonstrated that discontinuation of OAC was significantly associated to ischemic stroke/TIA (adjusted sHR 9.73, 95% CI 4.65-20.37) (Figure). Similar results were found for composite cardiovascular outcomes (adjusted sHR 6.10, 95% CI 3.15-11.80), and even major bleeding events (adjusted sHR 7.96, 95% CI 4.09-15.51). The risk of death was explored by adjusted Cox regression analysis showing that OAC cessation also increased mortality (adjusted HR 2.15, 95% CI 1.58-2.90). Conclusions In this real-world cohort of AF patients under OAC therapy, discontinuation of either VKAs or DOACs was significantly associated with ischemic stroke/TIA, composite cardiovascular outcomes, major bleeding and death. Education about OAC should emphasize the need for and importance of maintaining OAC. Caution is advised when considering OAC discontinuation.Survival curves for ischemic stroke/TIA