Predictors of thromboembolism in atrial fibrillation: II. Echocardiographic features of patients at risk. The Stroke Prevention in Atrial Fibrillation Investigators.

OBJECTIVE

To identify echocardiographic predictors of arterial thromboembolism in patients with nonrheumatic atrial fibrillation and to determine whether these add to clinical variables for risk stratification.

DESIGN

Cohort study of patients assigned to placebo in a randomized clinical trial.

SETTING

Five hundred sixty-eight inpatients and outpatients with nonrheumatic atrial fibrillation assigned to placebo therapy at 15 U.S. medical centers from 1987 to 1989 in the Stroke Prevention in Atrial Fibrillation study. Patients were followed for a mean of 1.3 years.

MEASUREMENTS: M-mode and two-dimensional (2-D) echocardiograms performed at study entry and interpreted by local cardiologists. The predictive value of 14 echocardiographic variables for later ischemic stroke or systemic embolism was assessed by multivariate analysis.

MAIN

RESULTS

Left ventricular dysfunction from 2-D echocardiograms (P = 0.003) and the size of the left atrium from M-mode echocardiograms (P = 0.02) were the strongest independent predictors of later thromboembolism. Multivariate analysis of these two independent echocardiographic predictors with the three independent clinical predictors of thromboembolism (history of hypertension, recent congestive heart failure, previous thromboembolism) identified 26% of the cohort with a low risk for thromboembolism (1.0% per year; 95% Cl, 0.2% to 4.0%). Compared with risk stratification using clinical variables alone, echocardiographic results altered thromboembolic risk stratification in 18% of the entire cohort and in 38% of those without clinical risk factors.

CONCLUSIONS

Both left ventricular and left atrial variables are significant predictors of thromboembolism in patients with nonvalvular atrial fibrillation. Our results challenge traditional views of the pathogenesis of ischemic stroke in patients with atrial fibrillation and suggest that standard echocardiography contributes to risk stratification, differentiating the one third of patients without clinical risk factors who are at increased risk for stroke from the remainder who may not need antithrombotic prophylaxis.