Long-term outcomes in patients with non-valvular atrial fibrillation and left atrial appendage closure in a large UK tertiary centre: 10-year experience

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. Introduction Left atrial appendage occlusion (LAAO) is a well-established stroke prevention strategy in patients with non-valvular atrial fibrillation (AF) and high risk of bleeding or contra-indication to oral anticoagulation (OAC). Despite encouraging randomised control trial and international registry safety and efficacy data, long-term outcome data remains sparce. Purpose This study sought to evaluate the long-term outcomes in ‘real-world’ AF patients undergoing left atrial appendage occlusion in a large UK tertiary centre. Methods This retrospective study included all patients that had a LAAO device implanted in our institution from January 2010 to December 2020. Medical notes, electronic patient records, procedural and imaging reports were reviewed. Annual bleeding risk was extrapolated from the Swedish National Cohort study according to CHA2DS2-VASc and HASBLED score. Results During the study period a total of 225 patients underwent LAAO device implant. Seventy-two percent were male, age 74 ± 8 years, BMI 27 ± 6 kg/m2, CHA2DS2-VASc score 4.4 ± 1.2, HASBLED score 3.2 ± 0.8 and at high risk of stroke (98 ischaemic strokes and 129 haemorrhagic strokes) and bleeding (151 life-threatening bleeding episodes). Three different LAAO devices were used: 136 Watchman, 54 Watchman FLX and 35 Amplatzer Cardiac plugs. Three patients (1.3%) had fatal complications related to the procedure. At discharge, 10% were taking single antiplatelet (ATP), 79% dual-antiplatelet (DAPT), 1.4% OAC, 3.6% ATP and OAC, 3.1% DAPT and OAC, 1.3% were not taking any anti-thrombotic. Nine (4%) patients had device-related thrombus on follow-up transoesophageal echocardiography with no significant difference between devices (5.0%, 2.8% and 6.7% p = 0.8, respectively) and anticoagulation strategy (p = 0.7). Over a total follow-up of 889 patient-years (mean follow-up 3.9 ± 3.7 years), 24 (10.4%) patients died, 55 patients (6.2/100 patient-years) suffered an adverse event, 15 ischaemic strokes (1.7/100 patient-years) and 20 non-procedural major bleeding episodes (2.3/100 patient-years) occurred. Compared to estimated annual stroke and bleeding risk adjusted for CHA2DS2-VASc and HASBLED score, our cohort had a 79% and 65% relative risk reduction in ischaemic stroke and major bleeding, respectively. Conclusion In this cohort of "real-world" high-risk patients, major bleeding and thromboembolic rate remained low on long-term follow-up. Abstract Figure 1