Outcomes with oral anticoagulation resumption or discontinuation after an anticoagulation-related event in patients with atrial fibrillation

Abstract

Abstract Background There are uncertainties on the resumption of anticoagulation therapy in patients with non-valvular atrial fibrillation who experience a thrombotic or haemorrhagic event while on oral anticoagulation. Purpose We assessed the cumulative incidence of recurrent stroke, major bleeding and all-cause mortality associated with restarting antithrombotic treatment, after stroke or major bleeding during anticoagulation therapy for atrial fibrillation. Methods We included anticoagulated patients with atrial fibrillation that were discharged (2013-2020) from hospital for stroke, intracranial haemorrhage (ICH), and major bleeding. To adjust for competing risk of death and reduce confounding, we started the follow up after 120-days blanking period, counting events in patients resuming oral anticoagulation versus those that did not. Risks of all-cause mortality, thromboembolism, intracranial haemorrhage, and major bleeding were estimated with multivariable Cox proportional hazard models and propensity score to adjust for differences in baseline characteristics. Results 1029 patients (mean age 77 years) were included in the final cohort: 23% were ischemic stroke, 18% were ICH, and 59% major bleeding; of these, 77% resumed anticoagulation. Among patients with stroke as the index event (92% resumed anticoagulation) the cumulative incidence of events was significantly lower in patients resuming therapy (HR 0.2, 95%CI 0.1-0.4, p<0.01). In patients with ICH (64% resumed therapy) the cumulative incidence of events was lower in patients resuming anticoagulation (HR 0.4, 95%CI 0.2-0.7, p<0.01). Same was true for patients with major bleeding, with patients resuming therapy (76%), experiencing fewer events (HR 0.5, 95%CI 0.2-0.7, p<0.01). In the multivariable analysis considering age, sex and propensity score as covariates, resumption of anticoagulation significantly reduced the cumulative event rate (HR 0.5, 95%CI 0.4-0.6, p<0.01). Conclusion In patients with atrial fibrillation who experienced an anticoagulation-related event, resuming oral anticoagulation was associated with better outcomes for all-cause mortality and subsequent events as compared with patients who did not resume treatment.Stacked Area Charts and Cox regression