A study of the human aortic valve orifice by transesophageal echocardiography.

The transverse short-axis plane of the aortic valve was imaged by transesophageal echocardiography at a relatively high frame rate in 25 anesthetized patients undergoing heart surgery. The effective, time-averaged aortic valve area (a-AVA) was compared with areas obtained with triangular and circular valve orifice models (t-AVA and c-AVA, respectively). The aortic valve orifice was circular during 33.6% +/- 17.5% of systole. The relations between the triangular or circular aortic valve areas and a-AVA were as follows: t-AVA = 1.04 x a-AVA - 0.14 (r = 0.90; standard error of the estimate = 0.24 cm2) and c-AVA = 1.37 x a-AVA + 0.00 (r = 0.90; SEE = 0.30 cm2). Bias analysis showed no significant difference between a-AVA and t-AVA (bias = -0.04 +/- 0.23 cm2; difference not significant) but a significant overestimation of the average valve area by c-AVA (bias = +0.88 +/- 0.30 cm2; p < 0.001). Thus the aortic valve orifice was not circular for the entire duration of systole and valve area calculations based on a triangular model approximated a-AVA more closely than did those based on a circular model. These findings suggest that, for echocardiographic measurements that incorporate the aortic valve orifice area (e.g., stroke volume determinations), the use of a triangular valve area model, rather than a circular model, may produce more accurate results in anesthetized patients with heart disease.