Comparative prognostic performance of risk scores for 12-month mortality and rehospitalization after transcatheter tricuspid valve intervention.

Background: Interventional treatment of tricuspid regurgitation (TR) is a rapidly evolving field, but selecting patients most likely to benefit remains challenging. Currently, no dedicated or widely accepted risk score exists for transcatheter tricuspid valve interventions (TTVIs). In clinical practice, established scores such as Model for End-stage Liver Disease excluding INR (MELD-XI), European System for Cardiac Operative Risk Evaluation II (EuroSCORE II), GLIDE (Gap, Location, Image quality, Density, and En-face tricuspid regurgitation morphology), Tricuspid Regurgitation Impact Score (TRI-SCORE), and TRIVALVE are applied to estimate the risk of mortality and rehospitalization. This study analyzed the prognostic performance of these risk scores for predicting 12-month outcomes following TTVI. Methods: In this prospective single-center cohort study, 60 consecutive patients undergoing either edge-to-edge repair (n=47) or heterotopic minimally invasive bicaval valve implantation (n=13) were enrolled. Five established risk scores [MELD-XI, EuroSCORE II, GLIDE (Gap, Location, Image quality, Density, and En-face tricuspid regurgitation morphology), TRI-SCORE, and TRIVALVE] were calculated prior to intervention according to their original definitions. The primary outcome was a composite of all-cause mortality or rehospitalization for heart failure at 12 months. Discriminatory ability was assessed using receiver operating characteristic (ROC) curves and the corresponding area under the curve (AUC). Using the score with the best performance, patients were additionally stratified into risk categories based on the optimal cut-off value derived from ROC analysis. Survival analyses were performed using the Kaplan-Meier method and compared using the log-rank test. Results: No procedure-related deaths occurred. In-hospital mortality was 2.8%. Thirteen patients (21.7%) were rehospitalized for decompensated heart failure, and 11 patients (18.3%) died during the 12-month follow-up period. Among all evaluated risk scores, the TRI-SCORE demonstrated the strongest prognostic performance for the primary endpoint (AUC 0.76, 95% confidence interval: 0.61-0.91; P=0.004) and significantly discriminated survival across predefined risk categories (log-rank P=0.002). All other scores showed lower predictive ability for 12-month mortality or rehospitalization (AUC ≤0.65). Conclusions: The TRI-SCORE demonstrated the best discriminatory capacity among the evaluated models for predicting 12-month mortality and rehospitalization after TTVI.