Operative risk of mitral valve replacement: discriminant analysis of 1329 procedures

The influence of 34 variables on the operative mortality rate for isolated mitral valve replacement (MVR) was assessed by univariate and multivariate logistic regression analysis. The physiologic lesions were classified as stenosis (20%, operative mortality rate 8 +/- 1%), regurgitation (44%, operative mortality rate 13 +/- 2%), and mixed (34%, operative mortality rate 8 +/- 1%). Functional class (NYHA), previous myocardial infarction, and hepatic dysfunction were powerful independent clinical determinants of operative mortality (p less than .001), along with age at operation and emergency operation (p = .001, p = .04). Concomitant coronary artery bypass grafting or tricuspid annuloplasty, angina, ischemic etiology, and physiologic lesion were not significant independent determinants of operative risk. Interestingly, year of operation, prosthetic valve dysfunction, and previous cardiac surgery had no important effect on operative mortality. Early operative risk for MVR was related to preoperative cardiac and hepatic function. Prior myocardial infarction substantially increased the risk even if the mitral valve disease was not ischemic in origin. Increased operative mortality rate in the subgroup with mitral regurgitation was related to advanced left ventricular failure and myocardial infarction rather than the etiology of the mitral regurgitation. These clinical factors coupled with more refined measurements of left ventricular systolic pump function (independent of loading conditions) should permit more intelligent decision making regarding the optimal timing of MVR, at least in terms of early operative risk.