Euroscore II and STS Risk Model Scores in Aortic Stenosis: Can We Rely on Them?

Introduction The European System for Cardiac Operative Risk Evaluation II (EuroSCORE) and Society of Thoracic Surgeons (STS) risk models provide a method of predicating mortality of patients undergoing cardiac surgery. However, their validity in transcatheter aortic valve implantation (TAVI) remains controversial with some studies supporting its use as a good predictor of mortality whilst others find no association. We sought to investigate the validity of both EuroSCORE II and STS score as predictors of mortality in a real-life cohort of patients undergoing a TAVI.Methods Between 2010–2014, 115 (79 ± 8 years old; 56 male) consecutive patients with severe AS scheduled for TAVI had EuroSCORE and STS score calculated prior to intervention. The patients were followed up for a median 187 days (IQR 93,1520).Results During follow up, 27 patients died. Neither EuroSCORE nor STS were associated with prognosis in this cohort. EuroSCORE was not significantly associated with mortality, hazard ratio 1.33 per log unit (p = 0.28, 95% CI 0.90–2.20). This was similar to STS score, hazard ratio 1.08 per log unit (p = 0.78 95% CI 0.63 – 1.87). However, both confidence intervals are relatively wide indicating that more patients are required to substantiate this finding.Conclusions In this small cohort of patients, it would appear that neither EuroSCORE II nor STS are associated with overall survival. This cohort included many patients with poor mobility, previous CABG with LIMA graft and significant comorbidities not included in EuroSCORE/STS calculation. The scores might therefore have underestimated the true risk. More studies and more patients are required to further assess their validity. Although such scores have been accurately validated for open-heart surgery, we recommend that they should be interpreted with caution when attempting to predict risk in patients undergoing a TAVI.Abstract 43 Figure 1 Area under the curves (AUC) for both STS and Euroscore II. Both follow the diagonal line suggesting that they are not good predictors of outcome%U http://heart.bmj.com/content/heartjnl/102/Suppl_6/A31.full.pd