The echo score revisited: Impact of incorporating commissural morphology and leaflet displacement to the prediction of outcome for patients undergoing percutaneous mitral valvuloplasty.

Background: Current echocardiographic scoring systems for percutaneous mitral valvuloplasty (PMV) have limitations. This study examined new, more quantitative methods for assessing valvular involvement and the combination of parameters that best predicts immediate and long-term outcome after PMV.

Methods And Results: Two cohorts (derivation n=204 and validation n=121) of patients with symptomatic mitral stenosis undergoing PMV were studied. Mitral valve morphology was assessed by using both the conventional Wilkins qualitative parameters and novel quantitative parameters, including the ratio between the commissural areas and the maximal excursion of the leaflets from the annulus in diastole. Independent predictors of outcome were assigned a points value proportional to their regression coefficients: mitral valve area ≤1 cm(2) (2), maximum leaflets displacement ≤12 mm (3), commissural area ratio ≥1.25 (3), and subvalvular involvement (3). Three risk groups were defined: low (score of 0-3), intermediate (score of 5), and high (score of 6-11) with observed suboptimal PMV results of 16.9%, 56.3%, and 73.8%, respectively. The use of the same scoring system in the validation cohort yielded suboptimal PMV results of 11.8%, 72.7%, and 87.5% in the low-, intermediate-, and high-risk groups, respectively. The model improved risk classification in comparison with the Wilkins score (net reclassification improvement 45.2%; P<0.0001). Long-term outcome was predicted by age and postprocedural variables, including mitral regurgitation, mean gradient, and pulmonary pressure.

Conclusions: A scoring system incorporating new quantitative echocardiographic parameters more accurately predicts outcome following PMV than existing models. Long-term post-PMV event-free survival was predicted by age, degree of mitral regurgitation, and postprocedural hemodynamic data.