Antithrombotic treatment at onset of stroke with atrial fibrillation, functional outcome, and fatality: a systematic review and meta-analysis

Background: In atrial fibrillation-associated stroke, conflicting data exist regarding association between therapeutic vitamin K-antagonist anticoagulation (International Normalized Ratio 2-3) and early death and functional outcome, and few data exist relating to late outcome in ischemic and haemorrhagic atrial fibrillation-stroke.

Aim: We performed a systematic review and meta-analysis of oral anticoagulation at stroke onset, death and functional outcome.

Methods: We performed a systematic review, searching multiple sources. Studies were included if outcomes in atrial fibrillation-associated stroke were reported stratified by pre-stroke antithrombotic status, with documented International Normalized Ratio at onset. Outcomes were survival and good functional outcome (modified Rankin score 0-2) at discharge/30 days, and at one-year.

Results: Of eight studies (3552 patients) in ischemic stroke, International Normalized Ratio ≥ 2 compared with other treatments (International Normalized Ratio < 2, antiplatelet, or no antithrombotic) was associated with good outcome [pooled odds ratio 1·9 (95% confidence interval) 1·5-2·5, P < 0·001] and improved survival at 30 days discharge (pooled odds ratio for death 0·4, confidence interval 0·2-0·5, P < 0·001). The net benefit remained after inclusion of haemorrhagic stroke (odds ratio for good outcome 1·89, confidence interval 1·45-2·46, P < 0·001). At one-year, improved functional outcome for International Normalized Ratio ≥ 2 (pooled odds ratio 1·7, confidence interval 1·0-2·7, P = 0·04) and survival (odds ratio for death 0·5, confidence interval 0·4-0·8, P = 0·001) were also observed.

Conclusions: Therapeutic International Normalized Ratio at stroke onset was associated with early and late improved survival and functional recovery suggesting sustained benefit for warfarin anticoagulation for stroke outcome in atrial fibrillation patients. Long-term outcome data following stroke in patients taking new oral anticoagulants is required.