Thromboembolic and haemorrhagic events in patients with atrial fibrillation: a prospective cohort study in UK primary and secondary care.

BACKGROUND: Strong evidence on the long-term safety and efficacy of different types of anticoagulants would help clinicians to prevent thromboembolic events among patients with atrial fibrillation (AF) while minimising the risk of haemorrhages. AIM: To estimate the risk of thromboembolic and haemorrhagic events for patients with AF on antiplatelets or anticoagulants. DESIGN AND SETTING: This was a cohort study using routinely collected UK primary and secondary care clinical data from patients with AF, aged ≥18 years, and with an indication to receive anticoagulation before April 2012. METHOD: The risk of ischaemic stroke or transient ischaemic attack (TIA), coronary heart disease (CHD), peripheral artery disease (PAD), or gastrointestinal (GI) haemorrhage, between 1 April 2012 and 1 April 2017, was estimated using multivariate Cox regression models for patients on antiplatelets only, a combination of antiplatelets and vitamin K antagonists (VKAs), or novel oral anticoagulants (NOACs), and compared with those on VKAs only. RESULTS: Compared with VKAs, antiplatelets were associated with a higher risk of stroke or TIA, hazard ratio (HR) 1.51, 95% confidence interval (CI) = 1.09 to 2.09, and GI haemorrhage, HR 1.79, 95% CI = 1.01 to 3.18. The risk of thromboembolic and haemorrhagic events was similar for those on a combination of antiplatelets and VKAs, or those on VKAs only. The risk was also similar for those on NOACs or VKAs, except for CHD, where it was higher for patients on NOACs, HR 2.07, 95% CI = 1.35 to 3.19. CONCLUSION: Anticoagulants are associated with lower risk of thromboembolic and haemorrhagic events among patients with AF than antiplatelets. More research is required on the risk associated with VKAs or NOACs.