Echocardiographic Predictors of Aortic Stenosis Outcomes

Study Questions:

Are echocardiographic parameters of aortic stenosis (AS) severity and left ventricular (LV) systolic function useful predictors of mortality?


Clinical and echocardiographic data were retrospectively analyzed for a cohort of 1,065 patients with at least mild AS (maximum velocity [Vmax] >2.0 m/s) referred to the echocardiography laboratory at a single institution between 1999 and 2007; after having excluded patients with at least moderate aortic regurgitation, mitral stenosis or regurgitation, or prior aortic valve intervention. Follow-up was available for an average of 5.7 years. Study endpoints were aortic valve replacement (AVR; n = 584), composite of AVR or death (n = 932), all-cause mortality (n = 550), and cardiovascular mortality (n = 398).


The most powerful echocardiographic predictors of valve-related events were parameters of AS severity, such as peak aortic jet velocity (Vmax), mean gradient, and aortic valve area (AVA) (all p < 0.001). The main predictors of mortality were LV ejection fraction (LVEF) and stroke volume index (p < 0.05). After multivariable adjustment, LVEF (p < 0.001) and stroke volume index (p = 0.02) remained the only echocardiographic predictors of mortality, even after adjustment for symptomatic status. AVA also was associated with mortality, whereas Vmax and mean gradient were not.


The most powerful echocardiographic predictors of mortality among patients with AS are low LVEF and low flow (low stroke volume index), whereas parameters of AS severity predict valve-related events, but not overall mortality. The authors concluded that low flow should be integrated in the risk stratification and in therapeutic decision making among patients with AS.


This is a somewhat complicated study to interpret. The authors sought to determine whether parameters of LV function versus parameters of AS severity best predict outcomes among patients with AS. However, the retrospective, observational study included patients with as little as mild AS, likely creating a mixed study population. It might not come as a surprise that, among patients with less than severe AS, the degree of AS (mild vs. moderate) was less important than the degree of LV systolic dysfunction; but that measures of AS severity best predicted referral for AVR. It would be of interest to know whether the same findings would be reproduced in a population restricted to patients with severe AS.

< Back to Listings