Anticoagulation-associated bleeding in patients randomised to screening for atrial fibrillation or usual care

Abstract

Abstract Background Atrial fibrillation (AF) prevalence is rising, strongly aided by population ageing and increased awareness. Detection and treatment of subclinical AF may help mitigate stroke risk, but could increase iatrogenic complications. Purpose To evaluate whether detection of subclinical AF using implantable loop recorder (ILR) and subsequent initiation of oral anticoagulant (OAC) enhances the risk of major bleeding in elderly individuals with stroke risk factors. Methods This was a post-hoc analysis of a trial which randomised 6004 participants without known AF but with stroke risk factors in a 1:3 ratio to either ILR-screening for AF and OAC initiation upon AF episodes lasting ≥6 minutes or usual care (Control group). We examined the risk of major bleeding using OAC initiation as a time-varying exposure (figure 1) in Cox regression models, and similarly for the use of antiplatelets. Further, we estimated the incidence rates of OAC/antiplatelet initiation and discontinuation along with the distribution of major bleeding events according to drug exposure and subgroups. Results A total of 1019 participants (17%) initiated OAC; 578 (12.8%) in Control group vs. 441 (29.4%) in ILR group, with a mean age of 78.8 (±4.7) years in Control vs. 77.0 (±4.8) in ILR group, p<0.0001. Major bleeding events totalled 221 (3.7%) with 47 of these occurring after OAC initiation (4.6% of all OAC-users); 26 (4.5%) vs. 21 (4.8%) in Control and ILR group, respectively. Without the use of OAC or antiplatelets the overall rate of major bleeding was 0.44 (0.33-0.57) per 100-perosn years compared to 1.54 (1.13-2.05) after OAC initiation. For the individual person, the hazard ratio for major bleeding after OAC initiation was 2.08 (1.50-2.90) p<0.0001 compared to before initiation (Control: 2.81 (1.82-4.34) p<0.0001, ILR: 1.32 (0.78-2.23) p=0.31, p=0.07 for interaction)(table 1), similarly was usage of antiplatelet associated with a hazard ratio of 1.3 (0.96-1.75) p=0.09. For OAC users ≥75 years, the rate of major bleeding was 1.96 (1.30-2.83)(Control: 2.20 (1.23-3.63), ILR: 1.73 (0.92-2.96)), compared to 1.17 (0.70-1.83) among users <75 years (Control: 1.64 (0.82-2.93), ILR: 0.84 (0.36-1.66)). After OAC initiation, the discontinuation rate in the Control and ILR group was 9.60 (7.9-11.6) and 6.15 (4.95-7.55), respectively, while the rate of antiplatelet discontinuation was 7.44 (6.92-7.99) and 12.24 (10.9-13.4) in the Control and ILR group. Conclusion This post-hoc study of a randomised study found a two-fold risk increase of major bleeding after OAC initiation on individual basis compared with before, but this increase did not seem augmented by AF screening and treatment. A weaker signal was seen for antiplatelets. The OAC-associated bleeding risk was markedly higher in the elderly population. Finally, discontinuations of OAC and antiplatelets were common throughout follow-up.