Analysis of Pre and Perioperative Factors Associated with Mortality
To assess the effect of type of aortic repair on overall mortality, a multivariable Cox proportional hazard model of pre and intraoperative factors was built (Figure 2 and Appendix Table 3) . Adjustments were made for age, gender, presence of antegrade cerebral perfusion, prior myocardial infarction, preoperative renal failure, peripheral artery disease, chronic lung disease, hyperlipidemia, hypertension, and concomitant coronary artery bypass grafting, the presence of a total arch repair was associated with a 2.53 time increase in mortality (HR 2.53, 95% CI 1.38 – 4.62, p = 0.003) as compared to a hemiarch repair. The other strong predictor of mortality in this analysis was increasing age, with a 1.76 time increase per 10 years of age (HR 1.76, 95% CI 1.37 – 2.28, p<0.001). Testing for an interaction between age and type of aortic repair was not statistically significant. While the other factors were not significant, they were left in the final model due to their level of significance on univariable analysis and clinical plausibility as a confounder. However, removal of each non-significant covariate did not alter the overall assessment of the relationship of either type of aortic arch repair or age in relationship to mortality. Predicted survival curves demonstrating the strong relationship with mortality between type of aortic repair (total arch or hemiarch) and age are shown in Figure 3 .