A comparison between surgical risk scores for predicting outcome in patients undergoing transcatheter aortic valve implantation`

BACKGROUND: Surgical risk scoring systems are of modest value in predicting outcome following transcatheter aortic valve implantation (TAVI). There is no consensus as to which scoring system is the most useful. We examined the efficacy of all four currently available global scorings systems (Additive EuroScore [aES], Logistic EuroScore [LES], EuroScore II [ES II], Society of Thoracic Surgeons [STS] Score) in predicting outcome, both in the short (30 day) and longer term, in patients undergoing TAVI. METHODS: Additive EuroScore (aES), Logistic EuroScore (LES), EuroScore II (ES II), and Society of Thoracic Surgeons (STS) score were assessed in 312 consecutive TAVI patients. Mortality tracking was obtained from the Office of National Statistics as of May 2014. RESULTS: Mean age was 81.2 +/- 7.0 years; 53.2% were male. Mean aES, LES, ESI and STS were 9.7 +/- 1.9, 17.4 +/- 9.4, 6.1 +/- 4.3 and 4.6 +/- 2.8, respectively. Thirty-day mortality was 4.8%; long-term mortality (maximum 5.8 years, mean 2.2 +/- 1.5 years) was 25.3%. Using both univariate and multivariate modelling all four scoring systems failed to predict outcomes in the short term. In the longer term, univariable analysis demonstrated that higher aES and LES scores were associated with higher mortality (aES P= 0.0048, LES P= 0.054, respectively). After multivariable analysis, only higher aES remained significant. CONCLUSIONS: Surgical risk scores are poor at predicting outcomes in patients undergoing TAVI, particularly in the short term.