Published in:
01-02-2014 | Original Article
Performance of EuroSCORE II compared to EuroSCORE I in predicting operative and mid-term mortality of patients from a single center after combined coronary artery bypass grafting and aortic valve replacement
Published in: General Thoracic and Cardiovascular Surgery | Issue 2/2014
Login to get accessAbstract
Objective
The performance comparison of the recently introduced European System
for Cardiac Operative Risk Evaluation II in predicting operative as well as mid-term mortality, with its previous version in patients after combined aortic valve replacement and coronary artery bypass grafting surgery.
Methods
This retrospective analysis included 216 patients operated on at one institution from 01/1999 to 12/2005. Accuracy and calibration of EuroSCORE I and II were assessed by plotting the areas under the receiver operator curves and comparing observed and predicted mortalities.
Results
EuroSCORE II showed, regarding early mortality, a slightly higher discriminatory accuracy with an area under the receiver operator curve of 0.77, while additive and logistic EuroSCORE I areas were 0.749, 0.75, respectively. The highest specificity and sensitivity level was approached for EuroSCORE II at a predicted mortality of 4.4 %. Receiver operator curves concerning mid-term mortality revealed areas for additive, logistic EuroSCORE and EuroSCORE II of 0.745, 0.739 and 0.718 with the highest accuracy levels at predicted mortalities of 6.5, 6.48 and 3.88 %, respectively. Mean predicted mortalities by logistic EuroSCORE and EuroSCORE II were 8.35 and 3.99 %, respectively, while overall observed operative mortality was 6.3 %. In “high-risk” patients (EuroSCORE > 13), EuroSCORE II underestimated early and mid-term outcomes.
Conclusions
Regarding operative mortality, EuroSCORE II showed in this study a slightly higher discriminatory accuracy than EuroSCORE I. There were no significant differences in the calibration of the two model versions in “low-” and “moderate-risk” patients regarding early as well as mid-term mortality. Analyses in larger patient populations will contribute to further model improvement.