Reliability of transthoracic echocardiography in the assessment of aortic valve morphology: pathological correlation in 178 patients.
OBJECTIVE: To examine patient characteristics affecting the utility of transthoracic echocardiography in determining aortic valve morphology, particularly for the diagnosis of congenital bicuspid aortic valve (BAV). DESIGN: A retrospective comparison of preoperative echocardiographic determination of aortic valve morphology with pathological findings of the explanted valves. SETTING: A tertiary referral centre. PATIENTS: Consecutive patients who had aortic valve replacement between July 1994 and April 1996, and had preoperative echocardiograms. RESULTS: Of 313 patients, 181 (58%) had preoperative echocardiography. Three of the valves were excluded because they were too fragmented for pathological determination of valvular morphology. In the remaining 178 patients, aortic valvular morphology was determined by echocardiography in 104 (58%). Multivariate analysis showed that echocardiography was successful less often in women (odds ratio 0.44, P = 0.03) and in patients with densely calcified valves (odds ratio 0.69, P = 0.02), whereas age had no effect (odd ratio 0.99, P = 0.42). In those with adequate echocardiographic images, echocardiography had both a high sensitivity (0.92) and a high specificity (0.96) for the diagnosis of BAV. CONCLUSIONS: Echocardiography is a useful tool for the diagnosis of BAV, although suboptimal images may pose a problem in many patients, particularly women and patients with heavily calcified valves. When adequate images are obtained, transthoracic echocardiography can reliably identify aortic valvular morphology in most patients.