Abstract
OBJECTIVE: To determine in patients with first ever stroke whether atrial fibrillation influences clinical features, the need to perform computed tomography, and prognosis. DESIGN: Observational cohort study with maximum follow up of 6.5 years. SETTING: Primary care, based on 10 general practices in urban and rural Oxfordshire. SUBJECTS: Consecutive series of 675 patients with first ever stroke registered in the Oxfordshire community stroke project. MAIN OUTCOME MEASURES: Prevalence of atrial fibrillation by type of stroke; effect of atrial fibrillation on case fatality rate and risk of recurrent stroke, vascular death, and death from all causes. RESULTS: Prevalence of atrial fibrillation was 17% (95% confidence interval 14% to 20%) for all stroke types (115/675), 18% (15% to 21%) for cerebral infarction (97/545), 11% (4% to 11%) for primary intercerebral haemorrhage (7/66), and 0% (0 to 11%) for subarachnoid haemorrhage (0/33). For patients with cerebral infarction the 30 day case fatality rate was significantly higher with atrial fibrillation (23%) than with sinus rhythm (8%); the risk of early recurrent stroke (within 30 days) was 1% with atrial fibrillation and 4% with sinus rhythm. In patients who survived at least 30 days the average annual risk of recurrent stroke was 8.2% (5.9% to 10.9%) with sinus rhythm and 11% (6.0% to 17.3%) with atrial fibrillation. CONCLUSIONS: After a first stroke atrial fibrillation was not associated with a definite excess risk of recurrent stroke, either within 30 days or within the first few years. Survivors with and without atrial fibrillation had a clinically important absolute risk of further serious vascular events.