Performance of contemporary surgical risk scores for transcatheter aortic valve implantation: A meta-analysis

Background: Transcatheter aortic valve implantation (TAVI) is considered for severe aortic valve disease at high and now intermediate risk for surgical aortic valve replacement. Risk stratification plays a critical role decision-making for intervention and modality. We compared the prognostic utility of surgical risk scores for TAVI in this meta-analysis. Methods: MEDLINE, Embase, Cochrane and Web of Science databases from 1 January 1980 to 31 December 2015 were searched. Studies were systematically reviewed for inclusion, and data extracted for pooled analyses. Results: Amongst 1688 articles searched, 47 full-text articles were screened and 24 studies (12,346 TAVI cases) included for analyses. Pooled c-statistics (95% confidence interval) for operative mortality were EuroSCORE 0.62 (0.57-0.67), EuroSCORE II 0.62 (0.59-0.66), STS Score 0.62 (0.59-0.65). Pooled calibration odds ratios (95% CI) were EuroSCORE 0.31 (0.25-0.38), EuroSCORE II 1.26 (1.06-1.51), STS 0.95 (0.72-1.27). C-statistics (95% CI) for 1-year mortality were EuroSCORE 0.62 (0.57-0.67), EuroSCORE II 0.66 (0.61-0.71) and STS Score 0.58 (0.53-0.64). Conclusion: Surgical risk scores at most modestly discriminated operative and 1-year mortality. The EuroSCORE grossly over-estimated operative mortalitywhile the EuroSCORE II and STS Scores fitted better to TAVI outcomes with their own limitations. There is a need for the development and validation of TAVI-specific risk models. (C) 2017 Elsevier B.V. All rights reserved.