BACKGROUND Surgical valve repair for mitral regurgitation has significant advantages over valve replacement, but the durability of the technique varies according to the cause of mitral valve disease. In this study, we examined the long-term performance of this procedure in a young rheumatic population and also attempted to identify factors predicting a poor outcome. RESULTS Between January 1981 and 1989, 308 patients underwent primary mitral valve repair for rheumatic mitral regurgitation at our institution. Forty-nine patients who failed to report after surgery and another 5 with discordant data were excluded from the analysis. Mitral regurgitation was pure in 182 patients (72%) and associated with mild commissural fusion in 72 patients (28%). Patient ages ranged from 6 to 52 years (mean, 18 +/- 9 years). A total of 243 patients (96%) were in New York Heart Association class III or IV before surgery, and 66 (26%) had atrial fibrillation. Mean follow-up period was 60 +/- 35 months (range, 1 to 132 months). Rheumatic activity was present clinically in 30% and macroscopically during surgery in 32%. Surgical techniques included insertion of a Carpentier ring (99%), chordal shortening (88%), leaflet resection (14%), chordal transposition (7%), and commissurotomy (28%). Operative mortality was 2.6%, late mortality was 15%, and the reoperation rate was 27%. At 5 years, 96.8% of the patients were free from thromboembolism, 97.7% were free from endocarditis, 74.9% were free from reoperation, 66% were free from valve failure, and 66.2% were free from major events. Multivariate analysis identified active rheumatic carditis as a significant predictor of reoperation, valve failure, and future events, while sinus rhythm and shorter bypass time at initial surgery were the only predictors of long-term survival. Patients with pure mitral regurgitation, sinus rhythm, and no active carditis at initial operation had the best overall 5-year results. Among the 148 survivors without reoperation, 142 (96%) were in New York Heart Association class I and II, and 107 (72%) were in sinus rhythm. Doppler echocardiographic studies showed absence of mitral regurgitation in 34 patients (23%), severe regurgitation in 23 (16%), and severe mitral stenosis in 6 (4%). CONCLUSIONS Mitral valve repair in this young rheumatic population is associated with a high long-term morbidity. Presence of active rheumatic carditis has a significantly adverse effect on the success of mitral valve repair.