A simple preoperative score including the surgeon's experience to predict the probability of a successful mitral valve repair

Abstract

OBJECTIVES According to recent guidelines, mitral valve (MV) repair is preferable to replacement. However, replacement is sometimes inevitable. Aims of the study were to identify variables that predict the risk of an unsuccessful MV repair and to evaluate a score that could help in planning MV surgical procedures. METHODS Clinical data of all consecutive adult mitral valve surgical procedures, performed during a 50-month period, were extracted from our clinical registry, and combined with echocardiographic variables. The variables identified by a univariable analysis, together with factors known from the literature as indicating a possible risk of an unsuccessful MV repair, were compiled in a multivariable logistic regression analysis. The surgeon's experience was also taken into account. RESULTS Of 545 MV procedures, 162 (29.7%) were MV replacements. Seven variables were identified as independent predictors of MV replacement (odd ratio; 95% CI): Endocarditis (7.8; 3.7-16.5), absence of annular dilatation (3.6; 2.2-5.9), leaflet calcification (6.1; 3.0-12.7), annular calcification (3.7; 1.9-7.3), mitral stenosis (29.6; 9.3-93.8), mitral sclerosis (3.0; 1.7-5.3), surgeon's limited experience (3.9; 1.4-11.0). The ability of this model to discriminate between repair and replacement was calculated, and an area under the ROC curve of 0.87 was shown. A score was calculated for each patient and distributed into four risk categories: Low risk (0-6), moderate risk (7-10), high risk (11-16) and very high risk (>16) of MV replacement with, respectively, 10.2 vs 10.0%, 40.5% vs 38.9%, 66.7 vs 70.4% and 93.2 vs 93.2% observed vs predicted probability of MV replacement. CONCLUSIONS Preoperative assessment of seven variables can accurately predict the risk of an unsuccessful MV repair.