PL03: Thromboembolic events following cardioversion for acute atrial fibrillation and flutter: a systematic review and meta-analysis

Abstract

Introduction: Several recent observational studies have presented concerning data regarding the safety of cardioversion (CV) for acute atrial fibrillation and flutter (AAFF). We conducted this systematic review to determine whether it is safe to cardiovert AAFF patients without prescribing oral anticoagulation (OAC) post-CV for those who are CHADS-65 negative. Methods: We conducted a librarian assisted search of MEDLINE, Embase, and Cochrane from inception through November 23, 2019. We included observational studies and randomized trials reporting thromboembolic (TE) events (i.e. stroke, transient ischemic attack, or systemic thromboembolism) within 30 days following CV in patients with AAFF, where onset of symptoms was <48 hours. Two reviewers independently screened studies and extracted data. The main outcome was risk of TE events within 30 days post-CV, stratified by OAC use. Risk of bias was assessed with the Quality in Prognostic Studies (QUIPS) tool. The primary analysis was based on prospective studies and the secondary analysis was based on retrospective studies. We performed meta-analyses for TE events where 2 or more studies were available, by applying the DerSimonian-Laird random-effects model. We implemented analyses stratified by study design using Open MetaAnalyst and generated the forest plots. Results: Our search yielded 969 titles; 74 were selected for full-text review and 20 studies were included in the review. The primary meta-analysis of 6 prospective studies, including two randomized trials, found a TE event rate of 0.15% (2 TE events/1,314 CVs). Within this prospective group, lack of OAC use was associated with a decreased risk of TE events (RR = 2.15 where RR >1 indicates increased risk of TE events with OAC compared to no OAC; 95% CI 0.50 to 9.31; I2 = 0%). Five of the 6 prospective studies had a low or moderate risk of bias in all QUIPS domains. Secondary meta-analysis of 6 retrospective studies revealed a TE event rate of 0.53% (56 TE events/10,521 CVs). This subgroup showed a trend favouring OAC use with decreased risk of TE events (RR = 0.34 where RR <1 suggests decreased risk of TE events with OAC; 95% CI 0.17 to 0.72; I2 = 0%). Conclusion: In the primary analysis of prospective studies, we found a low TE event rate following CV of AAFF, irrespective of OAC use. This contradicts previous analyses of retrospective studies. Our study supports the longstanding practice of not necessarily prescribing OAC post-CV in the ED for AAFF patients who are CHADS-65 negative.