Progress in surgical technique and decreased early and late postoperative risk should lead to a modification in the indications for aortic valve replacement, before irreversible myocardial changes have definitively compromised the result of surgery. One hundred and seventy aortic valve replacements (100 pure or predominant aortic stenoses, 70 cases of aortic insufficiency) were carried out by the same surgeon over a period of 4 years with a minimum follow-up of one year and average of 25.4 months, using à Bjork prosthesis or a homograft. Early postoperative mortality was 5.3%, not differing from that associated with other types of valve replacement (mitral, polyvalvular). It is related more directly to surgical technique than to preoperative prognostic factors. Late mortality was 8.1%. Almost one third of these late deaths were related to the surgical technique or to the model of aortic prosthesis used. Stage IV cardiac failure plays a pejorative role in this late mortality, whilst no prognostic role could be demonstrated with respect to angina, meancardiac surface, Sokolow index, mean pulmonary artery pressure or diastolic pressure in the left ventricule. One year after surgery there was found to be a significantly important decrease in the Sokolow index and a modest decrease in mean radiological cardiac surface area. Beyond one year, no further improvement was seen. The majority of the patients surviving surgery had a good functional result since only 1% of the aortic stenosis patients and 7% of the aortic insufficiency group remained in stage III or IV cardiac failure. 78% of the patients who were working before operation were able to resume their professional activity after an average period of 6.2 months. However only 40% of the patients with stage IV failure who underwent surgery could return to work.