Background. In response to the increasing use of percutaneous balloon mitral commissurotomy, the National Heart, Lung, and Blood Institute established the Balloon Valvuloplasty Registry in November 1987. Methods and Results. Between November 1, 1987, and October 31, 1989, 738 patients aged 18 or older underwent percutaneous balloon mitral commissurotomy at the clinical sites. Data were prospectively entered into the registry at the time consent was obtained. Serious complications occurred in 87 procedures, or 12%. Death in the laboratory occurred in eight patients, or 1%. Within 30 days there were 24 cumulative deaths, 18 cardiac and six noncardiac. Univariate analysis revealed that older age, a history of cardiac arrest, cerebrovascular disease, dementia, renal insufficiency, cachexia, class IV congestive heart failure, use of an intra-aortic balloon pump, use of sympathomimetic amines, and a high echo score (greater-than-or-equal-to 13) were associated with early death (p<0.01). Additional univariate predictors included a precommissurotomy mitral valve area of <0.7 cm2. Left atrial pressure >12 mm Hg and a mitral valve area of <1.5 cm2 after the procedure were also associated with higher 30-day mortality (p<0.05). Multivariate analysis identified higher echo score and smaller valve area before the procedure as the strongest predictors of early death (p<0.001). Centers that performed more than 25 procedures also had lower complication rates. Conclusions. Although percutaneous balloon mitral commissurotomy appears to be effective at relieving the hemodynamic effects of rheumatic mitral stenosis, it does have risks. In properly selected patients, however, it appears to have low morbidity and 30-day mortality. Individual center experience with the procedure also appears to have great impact on complications.