Cessation of oral anticoagulation in relation to mortality and the risk of thrombotic events in patients with atrial fibrillation

Bleeding risk (often perceived, rather than actual) is a common reason for cessation of oral anticoagulation with Vitamin K antagonists (VKA). We investigate clinical outcomes in a consecutive population of VKA naive atrial fibrillation (AF) patients, who initiated VKA therapy in 'our clinic. We included consecutive VKA-naive patients with non valvular AF, initiated on VKA therapy in our anticoagulation outpatient, clinic in 2009. During follow-up, adverse events [thrombotic/vascular events (stroke, acute coronary syndrome, acute heart failure and cardiac death), major bleeding and death], and VKA cessation were recorded. At the end of the follow-up, we determined time within therapeutic range (TTR), using a linear approximation (Rosendaal method). We studied 529 patients (49% male, median age 76), median follow, up 835 days (IQR 719-954). During this period 114 patients stopped VKA treatment. 63 patients suffered a thrombotic/cardiovascular event (5.17%/year, 27 thrombotidischaemic strokes), 51 major bleeding (4.19%/year) and 48 died (3.94 %year). Median TTR was 54% (34-57). On multivariate analysis (adjusted by CHA(2)DS(2)-VASc score), VKA cessation was associated with death [Hazard Ratio (HR) 3.43; p<0.001], stroke [4.21; p=0.001] and thrombotic/cardiovascular events [2.72; p<0.001]. Independent risk factors for major bleeding were age 11.08; p<0.001], previous stroke [1.85; p=0.049], and TTR [0.97; p=0.001], but not VKA cessation. In conclusion, in AF patients AF, VKA cessation is independently associated with mortality stroke and cardiovascular events. Specifically, VKA cessation independently increased the risk of stroke, even after adjusting for CHA2DS2-VASc score. TTR was an independent risk factor for major bleeding following initiation of VKA therapy.