Objective: To correlate atrial fibrillation with the incidence of new thromboembolic (TE) stroke in older patients with and without prior TE stroke.
Design: In a prospective study of 2101 older patients, electrocardiograms showed that atrial fibrillation was present in 283 patients (13%). At 42-month mean follow-up, atrial fibrillation was associated with the incidence of new TE stroke in patients with and without prior TE stroke.
Setting: A large long-term health care facility where 2101 older patients were studied.
Patients: The 2101 patients included 1451 women and 650 men, mean age 81 +/- 8 years (range 60 to 103).
Measurements and main results: Atrial fibrillation was present in 283 of 2101 patients (13%). The mean age was 84 +/- 7 years in patients with atrial fibrillation and 81 +/- 8 years in patients with sinus rhythm (P = .0001). The prevalence of atrial fibrillation was 5% in patients aged 60 to 70 years, 14% in patients aged 71 to 80 years, 13% in patients aged 81 to 90 years, and 22% in patients aged 91 to 103 years (P < .0001). Mean follow-up was 31 +/- 18 months in patients with atrial fibrillation and 44 +/- 27 months in patients with sinus rhythm (P = .0001). Previous TE stroke occurred in 123 of 283 patients (43%) with atrial fibrillation and in 431 of 1818 patients (24%) with sinus rhythm (P < .0001). New TE stroke occurred in 131 of 283 patients (46%) with atrial fibrillation and in 303 of 1818 patients (17%) with sinus rhythm (P < .0001). The log-rank test showed that patients with atrial fibrillation had a significantly higher probability of developing new TE stroke than those with sinus rhythm (P < .0001). The multivariate Cox regression model showed that independent risk factors for new TE stroke were male sex (relative risk = 1.3), prior TE stroke (relative risk = 3.1), and atrial fibrillation (relative risk = 3.3).
Conclusions: Atrial fibrillation, prior TE stroke, and male sex are independent risk factors for the development of new TE stroke in older patients.