The aim of this retrospective study was to review the outcome of 216 patients (pts), investigated for severe mitral regurgitation (MR) between January 1980 and December 1987. Definition of 'severe' was mainly clinical: MR sufficiently advanced for the cardiologist to investigate whether surgery should be imminent. One hundred and sixty two pts (group 1) were operated on; 54 (group 2) did not undergo surgery. Baseline characteristics were similar in the two groups, except for aetiology (less dystrophic and more ischaemic MR in group 2), functional class (88.3% class III or IV in group 1 vs 48.1% in group 2; P = 0.001), pulmonary pressures (lower in group 2) and left ventricular ejection fraction (group 1: 0.66 +/- 0.13; group 2: 0.56 +/- 0.17; P = 0.001). Mean follow-up after hospital admission or surgery was comparable in the two groups (group 1: 3.9 +/- 2.5 years; group 2: 3.5 +/- 2.7 years). Three pts (group 2) were lost to follow-up. Sixty-three pts died; 35 of the 162 operated on; 28 of the 54 non-operated on. Actuarial survival rate at 8 years was 74.0 +/- 4.3% in group 1 vs 33.2 +/- 9.2% in group 2 (P = 0.001). These results confirm that the prognosis of severe MR is poor if it is not operated on, whereas the postoperative course of severe MR, when operated on in time, is good.