Abstract P357: A Simple Mortality Risk Score for Calcific Aortic Stenosis

Background: Calcific aortic stenosis (CAS) of the elderly is associated with high mortality. It would be useful to have a simple and accurate tool (risk score) to identify risk. Methods: We retrospectively reviewed all echocardiograms at our hospital from October 2006 to March 2008 and identified 208 veterans aged 65 years or older with CAS, aortic valve peak gradient >30 mmHg, but no prior aortic valve surgery. Aortic valve peak gradient, age, clinical co-morbidities, tobacco and alcohol use, creatinine (Cr) and hemoglobin (Hb) data were analyzed. Variables with significant odds ratios for 1-year mortality were entered into a regression model. Results: All-cause 1-year mortality was 16.3%. Significant odds ratios for mortality were found for Hb =< 10.0 g/dL (OR=13.3, p<0.0001); alcohol abuse (OR=3.76; p=0.008) and heart failure (HF) history (OR=2.88, p=0.027). The c-statistic for this model was good (c=0.848). Aortic valve peak gradient did not add to the model, but all subjects were required to have a gradient >30 mmHg to be included. A risk score was developed by assigning 2 points to Hb of 10 g/dL or lower, 1 point to alcohol abuse, and 1 point to HF. The figure plots survival vs. risk scores and predicts very low (2.9%), low (14.4%), intermediate (41.2%), high (52.4%), and very high (100%) 1-year mortality for scores of 0, 1, 2, 3 and 4, respectively. This represents a >30-fold range of mortality risk. Conclusion: For CAS patients 65 years or older with aortic valve peak gradient >30 mmHg and no prior aortic valve surgery, our simple risk score identifies a wide range in 1-year mortality risk and allows prediction of mortality risk with a high degree of predictive accuracy (c-statistic=0.85).