0029 : Is the Euroscore II reliable to estimate operative mortality of valvular surgery?

Purpose Concerns have been raised regarding the reliability of the logistic Euroscore (Euroscore I). The Euroscore II has been recently described to achieve a more accurate prediction of operative mortality. However its performance has not been specifically studied in valvular surgery. Methods The predictive performance of the Euroscore I and II was compared in 2931 consecutive patients who underwent valvular surgery in our institution between 2006 and 2010. Discrimination was assessed using the cindex and calibration by comparing predicted and observed mortality. Results Mean age was 64±16 years, mean Euroscore I 8.8±9.9% and mean Euroscore II 5.9±7.5%. Half of the patients underwent more than a single valve procedure. Mortality at 30 days was 5.5% (160 patients). C-index was 0.77 (95% CI 0.74-0.81) for the Euroscore I and 0.81 (95% CI 0.77-0.84) for the Euroscore II. The overall comparison between predicted and observed mortality showed a significant difference for the Euroscore I (p Conclusion The Euroscore II achieves a good predictive performance in the estimation of 30-day mortality after valvular surgery. Calibration is markedly improved as compared with the Euroscore I. These results support the use of the Euroscore II for valvular surgery.