Accuracy of two-dimensional echocardiography in determining aortic valve structure in patients >50 years of age having aortic valve replacement for aortic stenosis.

We sought to measure the accuracy of 2-dimensional transthoracic echocardiography in determining aortic valve structure in patients with aortic stenosis (AS) undergoing aortic valve replacement (AVR). Few studies have compared aortic valve structure determined by echocardiogram to that determined by examination of the operatively excised stenotic aortic valve. Two-dimensional echocardiograms were reviewed and interpreted by an expert echocardiographer in blinded fashion in 100 patients >50 years of age (mean 70) who had undergone AVR for isolated AS ± aortic regurgitation and the aortic valve structure (unicuspid, bicuspid, tricuspid) was compared to that from examination of the operatively excised stenotic valve. After excluding 14 cases in which echocardiograms were uninterpretable because of heavy calcium and/or poor image quality, congenitally malformed valves were present in 44 patients (51%) and tricuspid valves in 42 of the 86 patients (49%). Ten of the 14 patients (71%) with uninterpretable echocardiograms had congenitally malformed valves. Valve structure by echocardiogram was concordant with morphologic interpretation in 57 of 86 patients (66% accuracy, kappa = 0.33). Accuracy trended toward improvement as degree of AS decreased. In patients with valve areas similar to those enrolled in the recent transcatheter aortic valve implantation trial (PARTNER; 0.7 ± 0.2 cm(2)), aortic valve structure was accurately determined by echocardiography in 21 of 35 patients (60%). In conclusion, aortic valve structure was interpretable by transthoracic echocardiogram in 86 of 100 patients and accurate in 57 of these 86 patients (66%).