Predicting the risk of late futile outcome after transcatheter aortic valve implantation

BACKGROUND: Transcatheter aortic valve implantation (TAVI) for aortic stenosis (AS) risk assessment is still developing and mostly concerned with mortality in the periprocedural period. We therefore sought to develop and then validate a score to predict 1-year adverse outcome. METHODS: Patients that underwent TAVI for severe AS in the Israeli registry. Patients with unsuccessful/suboptimal implantation were excluded. The cohort was split to derivation/validation cohorts by a ratio of 70:30. The outcome was defined as 1-year composite of mortality, stroke, and no improvement in NYHA class (vs. baseline). Logistic regression was used to fit the prediction model. RESULTS: Out of 2,440 patients meeting inclusion criteria, 276 were excluded, leaving 2,160 patients for both cohorts. At 1 year, 299 (14%) patients experienced the adverse ("futile") outcome. The derived prediction model included mean aortic valve (AV) gradient, previous pacemaker, previous oncological disease, need for diuretics, baseline NYHA class, hemoglobin and creatinine levels, and nonfemoral access site. The model's area under the curve (AUC) was 0.69 in the derivation and 0.70 in the validation cohort. Performance of other scores in the validation cohort were lower (0.60 for STS, 0.55 for Euroscore2, 0.56 for TVT score, and 0.53 for TAVI2-score, p = .03). Based on three risk tiers, patients had a low risk (20/306, 7% futility), a medium risk (50/304, 17%), and high risk (18/37, 49%) for futility. CONCLUSIONS: The TAVI futility risk model can be used to provide further insight regarding prediction measures and/or patients' outcomes outside of the periprocedural period (NCT02023060).