SESSION TYPE: Cardiac Surgery Posters
PRESENTED ON: Wednesday, October 24, 2012 at 01:30 PM - 02:30 PM
PURPOSE: The most widely used risk scores to predict early mortality in cardiac surgery are The European System for Cardiac Operative Risk Evaluation (EuroSCORE) and The Society of Thoracic Surgeons Predicted Risk of Mortality (STS-PROM). The aim of this study was to compare the performance of logistic EuroSCORE (LES), EuroSCORE II (ES II), and STS-PROM in predicting early mortality after aortic valve replacement (AVR) for aortic stenosis (AS).
METHODS: Between 2002 and 2010, 258 consecutive patients underwent AVR, with or without CABG, for AS at Juntendo University Hospital. Observed versus predicted (O/E) mortality ratios were examined. Hosmer-Lemeshow goodness-of-fit test and area under the receiver operating characteristics curve (AUC) were examined to assess the performance of these models.
RESULTS: Observed early mortality was 4.2% (n=11). Predicted mortality rates for LES, ES II, and STS-PROM were 6.5%, 3.5%, and 4.7%, respectively, and thus the O/E ratios were 0.64, 1.20, and 0.89, respectively. Pearson correlation coefficient revealed a good linear relationship between STS-PROM and ES II (r = 0.76, p < 0.001). Hosmer-Lemeshow goodness-of-fit test indicated good accuracy for the prediction of mortality for these models (p= 0.76 for LES, 0.58 for ES II, 0.81 for STS-PROM). AUC curve was 0.79 (95% CI; 0.66 to 0.94) for LES, 0.69 (95%CI; 0.49 to 0.89) for ES II, and 0.79 (95% CI; 0.63 to 0.96), implying that the discrimination ability of ES II was worse than the other two models.
CONCLUSIONS: Overestimation of early mortality by LES (O/E ratio; 0.64) and underestimation by ES II (O/E ratio; 1.20) was observed, but STS-PROM (O/E ratio; 0.89) could more accurately predict early mortality. Discrimination ability of ES II assessed by AUC was worse than the others.
CLINICAL IMPLICATIONS: These results have implications for risk judgment in AVR for AS.
DISCLOSURE: The following authors have nothing to disclose: Kenji Kuwaki, Atsushi Amano, Hirotaka Inaba, Taira Yamamoto, Shizuyuki Dohi, Takeshi Matsumura, Terumasa Morita, Ryo Tsuruta, Atsumi Oishi, Yuichiro Sato, Kishio Kuroda
No Product/Research Disclosure InformationJuntendo University, Tokyo, Japan