Mortality Risk Score In Aortic Surgery. Initial Experience With The Implementation Of A New Predictive Model

Introduction and objectives: Predictive stratification scores of hospital mortality, logistic EuroSCORE and EuroSCORE II are reasonable tools in cardiac surgery. However, its utility in aortic procedures is controversial. The aim of this study is to design a new specific mortality risk model adjusted to patients undergoing thoracic aortic surgery.Methods: We analyzed retrospectively 310 consecutive patients who underwent surgery of proximal aorta (root, ascending and/or arch) between 2006 and 2013. A predictive model was developed using multiple logistic regression, entering variables that were significant in univariate analysis. Discriminatory performance was calculated by area under ROC curve and was compared with those of the logistic EuroSCORE and EuroSCORE II.Results: The observed death in-hospital mortality was 8.1%. Hospital mortality related variables were NYHA class, creatinine clearance, critical preoperative status, and active endocarditis. Logistic EuroSCORE, EuroSCORE II and new model predicted mortality were 14.98, 8.12, and 8.06, respectively. Area under ROC curve was 0.665 (0.541-0.789) for logistic EuroSCORE, 0.718 (0.588-0.848) for EuroSCORE II and 0.775 (0.675-0.876) for the new developed model (Hosmer-Lemeshow P=.405).Conclusions: The new developed model predicts hospital mortality more accurate and with more discriminatory power than logistic EuroSCORE and EuroSCORE II in thoracic aortic surgery, also offering an easier method for the calculation based on only 4 preoperative parameters. (C) 2014 Sociedad Espanola de Cirugia Toracica-Cardiovascular. Published by Elsevier Espana, S.L.U. All rights reserved.