To examine late outcome of mitral valve repair in patients with preoperative atrial fibrillation, we reviewed the cases of 323 consecutive patients who underwent mitral valvuloplasty for mitral regurgitation from 1980 to 1991; average age of 215 men and 108 women was 64 years (range 14 to 88 years), and 224 patients (70%) were in New York Heart Association class III or IV before operation. The main indications for operation were severe mitral regurgitation (76%), coronary artery disease with associated mitral regurgitation (15%), and aortic valve disease (6%). At the time of mitral valve repair, coronary artery bypass grafting was done in 35% of patients, aortic valve replacement was done in 7%, and multiple other procedures were done in 10%. For all patients, the 30-day mortality rate was 2.5% (70% confidence limits 1.6% to 3.4%) and survivorships at 3 and 5 years were 81% and 76%, respectively. Before operation, 216 patients were in sinus rhythm and 97 patients had atrial fibrillation; in the latter group, 11 had recent onset of atrial fibrillation within 3 months preceding mitral valve repair. Comparing patients with preoperative atrial fibrillation to those with sinus rhythm, we found no significant difference in operative mortality (3% versus 1.9%) or 5-year survivorship (74.3% +/- 6.3% versus 76.9% +/- 4.0%). At late follow-up, atrial fibrillation was present in 5% of patients with preoperative sinus rhythm, 80% of patients with preoperative chronic atrial fibrillation, and 0% of patients with preoperative recent onset atrial fibrillation (p < 0.01). The left atrial size by echocardiography was larger in patients with preoperative atrial fibrillation compared with that in those with sinus rhythm (59 +/- 1.4 mm versus 50.9 +/- 0.7 mm; p < 0.05). There was, however, only a weak correlation between preoperative left atrial size and late atrial fibrillation. Further, age, gender, and associated coronary artery disease did not correlate with presence of atrial fibrillation at late follow-up. Prevalence of late thromboembolic events was similar in patients with preoperative sinus rhythm compared with that in those with atrial fibrillation. These data suggest that mitral valve repair should be done before or soon after the onset of atrial fibrillation to maximize the chance of postoperative sinus rhythm and avoid long-term anticoagulation with warfarin. However, the early and late results of mitral valve repair in patients with chronic atrial fibrillation are good, and concomitant operation for supraventricular arrhythmia must have negligible morbidity and no adverse effect on operative mortality.