The impact of atrial fibrillation type on the risks of thromboembolic recurrence, mortality and major haemorrhage in patients with previous stroke: A systematic review and meta-analysis of observational studies.

INTRODUCTION: There is conflicting evidence on the impact of atrial fibrillation (AF) type, i.e. non-paroxysmal AF or paroxysmal AF, on thromboembolic recurrence. The consensus of risk equivalence is greatly based on historical evidence, focussing on initial stroke risks. We conducted a systematic review and meta-analysis to describe the impact of AF type on the risk of thromboembolic recurrence, mortality and major haemorrhage in patients with previous stroke.
METHODS: We systematically searched four multidisciplinary databases from inception to December 2018. We selected observational studies investigating clinical outcomes in patients with ischaemic stroke and AF, stratified by AF type. We assessed all included studies for risk of bias using the 'Risk of Bias In Non-randomised Studies - of Exposures' tool. The Comprehensive Meta-Analysis Software was used to calculate odds ratios from crude event rates.
RESULTS: After reviewing 14,127 citations, we selected 108 studies for full-text screening. We extracted data from a total of 26 studies, reporting outcomes on 23,054 patients. Overall, risk of bias was moderate. The annual incidence rates of thromboembolism in patients with non-paroxysmal AF and paroxysmal AF were 7.1% (95% confidence interval: 4.2-11.7) and 5.2% (95% confidence interval: 3.2-8.2), respectively. The odds ratio for thromboembolism in patients with non-paroxysmal AF versus paroxysmal AF was 1.47 (95% confidence interval: 1.08-1.99, p = 0.013). The annual mortality rates in patients with non-paroxysmal AF and paroxysmal AF were 20.0% (95% confidence interval: 13.2-28.0) and 10.1% (95% confidence interval: 5.4-17.3), respectively, and odds ratio was 1.90 (95% confidence interval: 1.43-2.52, p < 0.001). There was no difference in rates of major haemorrhage, odds ratio  = 1.01 (95% confidence interval: 0.61-1.69, p = 0.966).
CONCLUSION: In patients with prior stroke, non-paroxysmal AF is associated with significantly higher risk of thromboembolic recurrence and mortality than paroxysmal AF. Although current guidelines make no distinction between non-paroxysmal AF and paroxysmal AF for secondary stroke prevention, future guidance and risk stratification tools may need to consider this differential risk (PROSPERO ID: CRD42019118531). © European Stroke Organisation 2020.