When aortic valve replacement is performed in patients with a small aortic annulus, prosthesis-patient mismatch is of concern because it may affect postoperative clinical status. We conducted a retrospective study of outcomes in 65 patients with aortic stenosis requiring valve replacement. Fifty were given a 17-mm or 19-mm St. Jude Regent mechanical valve, and 15 were given a 19-mm Medtronic Mosaic bioprosthesis. Echocardiography was carried out preoperatively, at discharge, and at follow-up. There was 1 (2%) operative death in the Regent group and none in the Mosaic group. There was no valve-related event. Follow-up echocardiography in both groups revealed a significant increase in the mean effective orifice area index, a decrease in the mean left ventricular-aortic pressure gradient, and a decrease in the mean left ventricular mass index. Prosthesis-patient mismatch (effective orifice area index <0.85 cm(2) · m(-2)) existed in 13 (26%) patients in the Regent group and 11 (73%) in the Mosaic group at discharge. All patients improved to New York Heart Association functional class II or better. A small-sized prosthesis may provide satisfactory clinical and hemodynamic results in patients with a small aortic annulus.