Abstract
BACKGROUND: Operative risk scoring algorithms identify patients with severe AS for transcatheter valve implantation in whom the anticipated operative mortality for conventional surgery would be considered prohibitive. We compared the three risk scores EuroScore 1 (LES), society of thoracic surgeons' (STS) score and ACEF (age-creatinine-ejection fraction score) to the readjusted EuroScore 2 recently presented. METHODS: We reviewed all consecutive patients receiving either isolated conventional aortic valve replacement (cAVR) or transapical aortic valve implantation (TA-TAVI) in a two-year period (n = 206). 30-days mortality was considered as primary endpoint. RESULTS: TA-TAVI was performed in 76 patients, isolated cAVR in 130 patients. Overall mortality was 4.4% (TA-TAVI: 7.9%; cAVR: 2.3%). EuroScore 2 showed a good estimation for the entire population as well as within the subgroups: 4,02 ± 5,36% (TA-TAVI: 6.16 ± 7.14%, cAVR: 2.77 ± 3.42%). Predicted mortalities as assessed by LES were largely overestimated (TA-TAVI: 27.4 ± 20.9% cAVR: 10.6 ± 10.6%, sensitivity: 0.89, specificity: 0.71). STS predicted mortality was 6.3 ± 4.4% for TA-TAVI patients as to 3.2 ± 3.1% for cAVR patients (sens.: 0.22, spec.: 0.96) and ACEF predicted a mortality of 1.16 ± 0.36% for cAVR and 1.58 ± 0.59% for TA-TAVI patients (sens.: 0.78, spec.: 0.89). CONCLUSION: The newly refined EuroScore 2 showed a good correlation within the studied population. For the individual patient, new cut-offs will have to be defined to triage patients for TAVI procedure. A drawback for complex score systems such as EuroScore and STS is the lack of recalibration to smaller populations as encountered in even large single centers.