From 1967 to 1988, 339 patients with mitral stenosis underwent surgical commissurotomy, 103 with a closed and 236 with an open technique. The 1 month and 1, 5, 10 and 20 year overall survival rate was 99.7%, 99%, 95%, 87% and 59%, respectively, and the technique (open versus closed) was not a risk factor. Technique was also not a risk factor for a second mitral commissurotomy, subsequent mitral valve replacement, thromboembolism or poor functional status. Risk factors were older age at commissurotomy, black race, higher pulmonary vascular resistance, mitral leaflet calcification, left ventricular enlargement and postcommissurotomy mitral incompetence. The closed technique was a risk factor for mitral incompetence immediately after commissurotomy, but important incompetence developed in only 2 of the 103 patients undergoing closed commissurotomy. Mitral valve replacement was not required within 10 years in 78% of patients and within 20 years in 47%. Despite some increased prevalence of postcommissurotomy mitral incompetence and particularly in view of the considerable long-term likelihood of mitral valve replacement, initial therapy for most patients with mitral stenosis should probably be surgical closed (or percutaneous balloon) mitral commissurotomy. The equations developed in the present study can be used to predict and compare outcome probability after percutaneous balloon commissurotomy with that after surgical commissurotomy and to compare these with outcome probability after mitral valve replacement.