Mitral valvuloplasty has become a widely accepted and, in many instances, preferred option for the surgical treatment of mitral valve regurgitation. But the results are not as reproducible and predictable as those of valve replacement. Reoperation is more frequent than after valve replacement but is, for practical purposes, the only significant complication of the procedure. Several series have demonstrated the superiority of the long-term results of mitral valvuloplasty for most disease aetiologies. However, some doubts have been cast on the utilisation of valvuloplasty in rheumatic mitral valve regurgi tation. Nonetheless, my own experience in Johannesburg has demonstrated better global and event-free survival in patients who underwent valvuloplasty, even though these results in rheumatic patients are, in everybody's experience, inferior to those in cases of other types of pathology, especially in degenerative (myxomatous) disease. Of particular concern are eases of acute rheumatic carditis but, in my view, valvuloplasty may still be applied in selected cases. In the past decade several modifications of the technique initially developed by Carpentier in the 1970s have contributed to the improvement of the results. Among these are the utilisation of leaflet (partial) and chordal substitutes and new annuloplasty rings. However, the procedure remains dependent on a striking learning curve and, as a consequence, the results vary widely between different surgical teams. Nonetheless, it is fair to say that the improvements obtained thus far have changed the indication to an earlier referral for surgery whenever the surgeon can safely predict reasonably good feasibility of the repair. In my experience, this is possible in the vast majority (> 95%) of cases of myxomatous disease and in over 85% of cases of pure rheumatic regurgitation. Hence, it is increasingly important to maintain the attitude that the needless excision of a mitral valve is a defeat to the surgeon.