Propensity Analysis and Adjusted Outcomes
Given the significant differences in baseline characteristics between the RAT and ST groups, seven baseline variables were used to create a logistic regression model for selection to a RAT or ST approach.Figure 4 shows the results of the logistic regression model predicating selection to a RAT or ST approach. Those more likely to undergo a sternotomy were older, of non-white race, more likely to have had a previous cardiac operation, and to be having a concomitant procedure, while those undergoing a small right anterior thoracotomy were more likely to have posterior leaflet pathology, degenerative MV disease, and to have diabetes. The area under the receiver operating curve (c-statistic) for the model was 0.84 and Homer and Lemeshow Goodness-of-fit test was not statistically significant (p=0.25), demonstrating good model discrimination and limited collinearity and interactions (Supplemental Figure 1 ).
Adjusted multivariable Cox-Proportional hazard outcomes are shown inFigure 5 . The presence of any anterior or posterior repair was not predictive of mortality, however, a small right anterior thoracotomy approach was associated with decreased mortality (Hazard Ratio (HR) 0.32, 95% CI 0.13-0.82, p=0.018). Neither the need for MV reoperation or progression of mitral regurgitation were affected by the type of repair or operative approach.