OBJECTIVES--(a) To assess the echocardiographic incidence of restenosis after successful balloon dilatation of the mitral valve at a mid-term follow up of one year among a population of predominantly United Kingdom patients. (b) To identify any factors, assessed before or during dilatation, which may predict the development of restenosis. DESIGN--Successful dilatation of the mitral valve was defined as an increase in mitral valve area of > 25% and a final valve area of at least 1.5 cm2. Echocardiographic restenosis was defined at follow up as a loss of 50% of initial gain and a valve area of less than 1.5 cm2. Mitral valve area was assessed by transthoracic echocardiography before, during, 48 hours after, and one year after successful balloon dilatation of the mitral valve. Echo score before dilatation (an assessment of valvar and subvalvar calcification, thickening, and mobility), age, rhythm, echocardiographic mitral valve area before and after dilatation, left atrial pressure before and after dilatation, and end diastolic mitral valve gradient before and after dilatation were compared in those patients with and without echocardiographic restenosis at one year. SETTING--A regional cardiothoracic centre in the United Kingdom that performs 20-30 balloon dilatations of mitral valves each year. PATIENTS--39 patients, with symptomatic dominant mitral stenosis, who had undergone successful balloon dilatation of the mitral valve, and in whom echocardiographic assessment of mitral valve area was available at one year. 92% of patients were citizens of the United Kingdom. INTERVENTIONS--Balloon dilatation of the mitral valve by the Inoue technique. MAIN OUTCOME MEASURES--Mitral valve area and patient symptom class (New York Heart Association) one year after successful dilatation of the mitral valve. RESULTS--The incidence of echocardiographic restenosis was eight of 39 patients (21%). Of the eight patients with restenosis four underwent mitral valve replacement, two had repeat dilatation of the mitral valve, and two remained on medical treatment. With univariant analysis, factors associated with restenosis were increased age, higher echo score before dilatation, and a lower mitral valve area immediately after the operation. The only independent risk factor for restenosis, shown by multivariant analysis, was a high echo score before dilatation. There was no significant fall in mitral valve area at one year in those patients without restenosis. Most (28/31) of these patients had echocardiographic evidence of splitting of at least one commissure after dilatation compared with only two of eight patients who developed restenosis. Of 10 patients with an echo score before dilatation > or = 10 only two had an initially successful operation and no restenosis at one year. CONCLUSIONS--The echocardiographic incidence of restenosis after dilatation of the mitral valve by the Inoue technique in patients of the United Kingdom is 21%. The principal factor associated with restenosis is a high echo score before dilatation. Increases in mitral valve area are maintained in those patients without restenosis and it is likely that the mechanism of initial increase in valve area is different in the two groups, being commissural splitting in those patients who do not get restenosis and valve stretching in those that do. In patients with an echo score > or = 10 dilatation of the mitral valve should be considered only as a palliative procedure.