Percutaneous balloon mitral commissurotomy was attempted in Tunisia, where rheumatic fever is still endemic, in 463 consecutive patients with severe rheumatic mitral valve stenosis. Their mean age ±SD was 33 ± 12 years (range 8 to 68), 324 patients (70%) were women, and 327 (71%) were in sinus rhythm. Valvotomy was technically successful in 454 patients (98%). The mean mitral valve gradient decreased from 20 ± 7 to 6 ± 4 mm Hg, mean left atrial pressure decreased from 27 ± 8 to 15 ± 6 mm Hg, cardiac index increased from 3.0 ± 0.7 to 3.6 ± 0.8 L/min/m², and Gorlin mitral valve area, from 0.97 ± 0.19 to 2.2 ± 0.4 cm² (all p < 0.001). Two-dimensional echocardiographic mitral valve area increased from 1.03 ±0.18 to 2.15 ± 0.36 cm² (p < 0.00001). A final valve area of ≥1.5 cm² was achieved in 98% of patients. Multivariate analysis identified a pre-mitral valve area <0.8 cm² and an echocardiographic score (echo score) ≥12 as the strongest predictors of residual stenosis (final mitral valve area <1.5 cm²). Major procedural complications included mortality (0.4%), tamponade (0.7%), thromboembolism (2.0%), severe mitral regurgitation (4.6%), significant (pulmonary to systemic flow ratio ≥1.5) interatrial shunt (4.8%). Four hundred thirty patients were followed up between 6 and 82 months (mean 37 ± 22): 95% were in functional class I to II without reintervention, and 7 patients died (1.6%); restenosis (echocardiographic mitral valve area <1.5 cm²) occurred in 10.4% of patients. The 3-year Kaplan-Meier freedom from restenosis was 92%, and from reintervention 93%. Because fluoroscopic calcium and postprocedure mitral valve area <-1.8 cm² were the independent predictors of restenosis, patients with calcified valves should be selected for this procedure on a case-to-case basis.