Risk scores for prediction of 30‐day mortality after transcatheter aortic valve implantation: Results from a two‐center study in Norway

OBJECTIVES: Transcatheter aortic valve implantation (TAVI)‐specific risk scores have been developed based on large registry studies. Our aim was to evaluate how both surgical and novel TAVI risk scores performed in predicting all cause 30‐day mortality. In addition, we wanted to explore the validity of our own previously developed model in a separate and more recent cohort. METHODS: The derivation cohort included patients not eligible for open surgery treated with TAVI at the University Hospital of North Norway (UNN) and Oslo University Hospital (OUS) from February 2010 through June 2013. From this cohort, a logistic prediction model (UNN/OUS) for all cause 30‐day mortality was developed. The validation cohort consisted of patients not included in the derivation cohort and treated with TAVI at UNN between June 2010 and April 2017. EuroSCORE, Logistic EuroSCORE, EurosSCORE 2, STS score, German AV score, OBSERVANT score, IRRMA score, and FRANCE‐2 score were calculated for both cohorts. The discriminative accuracy of each score, including our model, was evaluated by receiver operating characteristic (ROC) analysis and compared using DeLong test where P< .05 was considered statistically significant. RESULTS: The derivation cohort consisted of 218 and the validation cohort of 241 patients. Our model showed statistically significant better accuracy than all other scores in the derivation cohort. In the validation cohort, the FRANCE‐2 had a significantly higher predictive accuracy compared to all scores except the IRRMA and STS score. Our model showed similar results. CONCLUSION: Existing risk scores have shown limited accuracy in predicting early mortality after TAVI. Our results indicate that TAVI‐specific risk scores might be useful when evaluating patients for TAVI.