Background: Non-rheumatic atrial fibrillation (NRAF) is a very common arrhythmia but its role in the prognosis and cardiovascular mortality is controversial. In particular, cause and predictors of death are not completely known.
Methods: We analyzed the cause of death and the possible predictors of cardiovascular mortality in 664 outpatients (mean age 72 +/- 9 years old) enrolled in the "Trieste Area Study on Non-Rheumatic Atrial Fibrillation" (TASAF), a prospective community study, after a follow-up of 27 +/- 9 months. The mean duration of the arrhythmia at enrollment was 59 months (range 1-360 months). Only 42 patients (6.3%) were on anticoagulants by general practitioners and 205 (30.8%) were on antiplatelet drugs.
Results: Of these patients, 110 (16.5%) died: 28 (25.5%) due to a cerebral or peripheral thromboembolism, 10 (8.2%) of sudden death, 46 (42.7%) of expected cardiac death and 25 (22.7%) of non-cardiac causes. In one patient, the cause of death was uncertain. Sixty-nine patients underwent postmortem examination. In univariate analysis, left ventricular dysfunction (p = 0.03) and an enlarged left atrium (p = 0.03) proved to be directly related to increased cardiovascular mortality. Both in univariate and Cox proportional hazards model analysis, aging (odds ratio 1.09, IC 95% 1.05-1.12, p = 0.00001), history of heart failure (odds ratio 1.27, IC 95% 1.01-1.60, p = 0.036), cardiomegaly (odds ratio 1.35, IC 95% 1.01-1.81, p = 0.040), diabetes mellitus (odds ratio 1.35, IC 95% 0.99-1.84, p = 0.058) and previous myocardial infarction (odds ratio 1.56, IC 95% 1.20-2.03, p = 0.0007) were all independent risk factors for cardiovascular mortality. A history of cerebral or systemic embolism (23 versus 12%, p = 0.09) and, above all, one or more recurrences before enrollment (11 versus 2.3%, p = 0.04), were associated with embolic mortality.
Conclusions: Patients with NRAF have an increased risk of cardiovascular death. Aging, the presence of diabetes, cardiomegaly on chest x-ray, heart failure and a previous myocardial infarction were independent risk factors for cardiovascular mortality. A history of embolism at enrollment significantly conditioned the embolic mortality rate but above all, embolic events during follow-up determined a very high percentage of total deaths (25.5% of all causes). A proper anticoagulant therapy should strongly be advised to all patients with no contraindications.