339 patients underwent aortic valve replacement. Data collection was prospective and the logistic and additive EuroSCORE algorithms were calculated according to published guidelines.
The observed-over-expected mortality ratio was 0.096. In the ROC curve analysis, the asymptotic significance was greater than 0.05. On multivariate analysis, only critical preoperative state remained significant independent predictor of in-hospital mortality (Odds Ratio 1.6, CI 1.2–2.1).
These outcomes suggest that EuroSCORE models may fail in predicting hospital mortality in subsets of cardiac surgery patients and dedicated risk models for isolate aortic valvular surgery may be useful to provide more precise estimates of hospital mortality.