Echocardiography may allow preoperative selection of appropriately sized aortic homograft valves. This study compared measurement of aortic annulus diameter from the preoperative transthoracic two-dimensional echocardiogram with prosthetic valve annulus diameter in 178 consecutive patients under-going primary aortic valve replacement. Aortic annulus diameter was measured independently from the two-dimensional parasternal long axis view by two experienced observers blinded to the size of valve implanted. Mean annulus diameter from three end-diastolic frames was recorded. Of the 178 patients, 124 (70%) of Observer 1 measurements and 102 (57%) of Observer 2 measurements were within 2mm of valve annulus diameter (p < 0.0001 for both observers). Six percent of Observer 1 measurements and 12% of Observer 2 measurements were more than 4mm away from valve annulus diameter. The echocardiographic prediction of valve size corresponded to actual prosthetic valve annulus diameter in 69 patients (39%) for Observer 1 and 60 patients (34%) for Observer 2. The percentages of echocardiographic measurements within two millimeter of valve annulus diameter were significantly different for the two observers (p < 0.05). The percentages of echocardiographic measurements within two millimeter of valve annulus diameter in 39 patients without significant aortic valve calcification were not significantly different from the remaining 139 patients with calcified valves. The interobserver coefficient of variation was 6.2% and the intraobserver coefficient of variation was 4.3%. Although preoperative measurement of aortic annulus diameter from transthoracic echocardiography approximates prosthetic valve annulus diameter in the majority of patients undergoing primary aortic valve replacement, a significant minority of estimates may be inaccurate, even in the absence of heavy calcification.(ABSTRACT TRUNCATED AT 250 WORDS)