Objective We hypothesised that left atrial emptying fraction (LAEF) would predict long-term cardiovascular outcome in the general population and better so than left atrial (LA) volumes.

Methods A community-based sample (n=740) in sinus rhythm prospectively underwent clinical evaluation, echocardiography and blood analyses including N-terminal pro B-type natriuretic peptide (NTproBNP). The LA maximum (LAVmax) and minimum volumes (LAVmin) were indexed to body surface area (LAVImax and LAVImin, respectively). LAEF was calculated as LAVmax-LAVmin divided by LAVmax. The participants were followed for a median of 4.9 years regarding incident cardiovascular events (cardiovascular death or hospitalisation because of myocardial infarction, heart failure or stroke). Cox regression models were used to evaluate the associations of LA volumes and LAEF with the outcome.

Results In a multivariable beta regression model, including clinical and echocardiographic baseline characteristics, higher plasma levels of NTproBNP, higher E/e′ and left ventricular systolic dysfunction remained as independent determinants of a lower LAEF. After adjustment for baseline characteristics, including NTproBNP levels, LAEF (HR for 1 SD decrease 1.33, 95% CI 1.04 to 1.70, p=0.022), but not LAVImax (HR for 1 SD increase 0.88, 95% CI 0.70 to 1.10, p=0.25) or LAVImin (HR for 1 SD increase 1.02, 95% CI 0.83 to 1.27, p=0.83) remained independently associated with outcome.

Conclusions In this community-based cohort, LAEF was a powerful predictor of incident cardiovascular events and its predictive ability was stronger than for LA volumes. Our findings suggest that LA dysfunction may represent a more advanced state of LA remodelling than LA enlargement.