Association of preceding antithrombotic therapy in atrial fibrillation patients with ischaemic stroke, intracranial haemorrhage, or gastrointestinal bleed and mortality

AIMS: To analyse 90-day mortality in atrial fibrillation (AF) patients after a stroke or a severe bleed and assess associations with the type of antithrombotic treatment at the event. METHODS AND RESULTS: From the Stockholm Healthcare database, we selected 6017 patients with a known history of AF who were diagnosed with ischaemic stroke, 3006 with intracranial haemorrhage, and 4291 with a severe gastrointestinal bleed (GIB). The 90-day mortality rates were 25.1% after ischaemic stroke, 31.6% after intracranial haemorrhage, and 16.2% after severe GIB. We used Cox regression and propensity score-matched analyses to test the association between antithrombotic treatment at the event and 90-day mortality. After intracranial haemorrhage, there was a significantly higher mortality rate in warfarin compared to non-vitamin K oral anticoagulant (NOAC)-treated patients [adjusted hazard ratio (aHR) 1.36, 95% confidence interval (CI) 1.04–1.78]. After an ischaemic stroke and a severe GIB, patients receiving antiplatelets or no antithrombotic treatment had significantly higher mortality rates compared to patients on NOACs, but there was no difference comparing warfarin to NOACs (aHR 0.84, CI 0.63–1.12 after ischaemic stroke, aHR 0.91, CI 0.66–1.25 after severe GIB). Propensity score-matched analysis yielded similar results. CONCLUSION: Mortality rates were high in AF patients suffering from an ischaemic stroke, an intracranial haemorrhage, or a severe GIB. NOAC treatment was associated with a lower 90-day mortality after intracranial haemorrhage than warfarin.