171: Risk scores versus pragmatic clinical assessment to predict operative risk in aortic valve replacement for aortic stenosis

BackgroundPreoperative risk assessment of cardiac surgery is based on international validated scores. However their additional value above simple clinical assessment (CA) remains controversial. The aim of this study was to compare CA by cardiologists with the 5 most commonly used scores (additive and logistic EuroSCORE, EuroSCORE II, STS-score, Ambler-score) to predict perioperative mortality in patients undergoing aortic valve replacement for aortic stenosis.MethodsFrom October 2009 to November 2011, 314 consecutive patients (73±9,7 years; 29% octogenarians) were included. A surgical coronary revascularization was associated to aortic valve replacement in 22%. According to the expected mortality by CA, patients were split in 4 groups: “low” mortality risk [0–3.9%], “intermediate” [4–6.9%], “high” [7–9.9%] and “very high” ≥10%. The 5 scores were calculated for all the patients.ResultsObserved total operative mortality was 5,7%. The distribution of predicted mortality in the 4 groups was highly different according to the method. The positive predictive value (PPV) of each method was calculated for the 21% most at risk patients (corresponding to the 64 patients ranked in “high” and “very high” mortality risk groups by CA) resulting in PPV=17.2% for EuroSCORE II, 14.1% for CA and STS-score, 10.9% for additive EuroSCORE and logistic EuroSCORE and 10% for Ambler score. Predictive values of “low” and “intermediate” mortality risks were not significantly different depending on the methods (PPV between 2.8 and 4.4%).Conclusionpragmatic CA remains useful to predict operative risk in patients with surgical aortic valve replacement and to balance the different international scores