2019 Volume 83 Issue 8 Pages 1755-1761
Background:The purpose of this study was to investigate if the transcatheter aortic valve replacement (TAVR) risk score can independently predict outcomes following TAVR, and to evaluate its predictive performance.
Methods and Results:This retrospective multicenter study comprised 682 patients with severe aortic stenosis who underwent TAVR. The primary endpoint was all-cause death following TAVR. The clinical model was established using variables identified as independent predictors in the multivariate analysis. Incremental values were assessed after adding atrial fibrillation, body mass index (BMI), and serum albumin to the TAVR risk score in receiver-operating characteristic analysis. The median TAVR risk score was 2.1% with a mean follow-up of 505 days. On Kaplan-Meier analysis, a TAVR risk score >4% had lowest survival rate, whereas TAVR risk score ≤2% had the highest survival rate at 3 years (log-rank P<0.001). The multivariate Cox regression analysis found the TAVR risk score was independently associated with all-cause death, and demonstrated moderate predictive performance for predicting all-cause death at 3 years. However, if each independent predictor is added to the model, it significantly increased discriminatory performance, particularly with BMI and serum albumin level.
Conclusions:We demonstrated that the TAVR risk score could independently predict mid-term death following TAVR, and had greater predictive performance for predicting all-cause death at 3 years with BMI and serum albumin level.
