TRI-Score-based mortality prediction after transcatheter tricuspid valve intervention: additive role of right ventricular strain

J. Schlegl (Bernau bei Berlin)¹, C. Edlinger (Bernau bei Berlin)¹, M. Bannehr (Bernau bei Berlin)¹, T. Kücken (Bernau bei Berlin)¹, M. Lichtenauer (Salzburg)², A. Krutz (Berlin)³, V. Paar (Salzburg)⁴, M. Neuß (Bernau bei Berlin)¹, A. Haase-Fielitz (Bernau bei Berlin)¹, C. Butter (Bernau bei Berlin)^1
1Immanuel Klinikum Bernau Herzzentrum Brandenburg / Kardiologie Bernau bei Berlin, Deutschland; ²Landeskrankenhaus Salzburg Universitätsklinik für Innere Medizin II, Kardiologie, internistische Intensivmedizin und Notaufnahme Salzburg, Österreich; ³Deutsches Herzzentrum der Charite (DHZC) Klinik für Kardiologie, Angiologie und Intensivmedizin, Campus Charité Mitte Berlin, Deutschland; ⁴Universitätsklinik der Salzburger Landeskliniken Klinik für Innere Med. II, Kardiologie u. intern. Intensivmedizin Salzburg, Österreich


Background
Patients undergoing transcatheter tricuspid valve intervention (TTVI) remain at relevant risk for death and recurrent heart failure events despite reduction of tricuspid regurgitation. The TRI-Score is a dedicated clinical risk score for severe tricuspid regurgitation and reflects advanced right-sided heart failure, clinical vulnerability and end-organ impairment. Right ventricular (RV) strain quantifies longitudinal RV myocardial deformation and may capture RV functional reserve beyond conventional, load-dependent echocardiographic parameters. We investigated the prognostic value of the pre-interventional TRI-Score and the additive role of pre-interventional RV strain after TTVI.
 
Methods
Patients undergoing TTVI with available pre-interventional TRI-Score and RV strain were included in a complete-case analysis. Pre-interventional TRI-Score was evaluated alone and in combination with RV strain for prediction of 12- and 24-month mortality. Discrimination was assessed using receiver operating characteristic analysis, and paired DeLong testing was performed to compare TRI-Score models with corresponding TRI-Score plus RV strain models. Long-term survival was evaluated by Kaplan-Meier analysis according to predefined TRI-Score risk categories. Cox regression was performed using survival time up to 24 months.


Results
Complete-case analysis included 56 patients for 12-month mortality and 54 patients for 24-month mortality, with 11 and 16 deaths, respectively. Baseline characteristics reflected a high-risk TTVI population and were summarized in Table 1. Patients in the TRI-Score high risk category showed significantly impaired 24-month survival compared with low/intermediate risk categories (log-rank p<0.001; Figure 1). In Cox regression, high TRI-Score risk category was strongly associated with 12-month mortality (HR 6.84, 95% CI 1.42–32.97; p=0.017) and 24-month mortality (HR 5.59, 95% CI 1.75–17.88; p=0.004; Figure 3). Discriminative performance for 12-month mortality was acceptable for TRI-Score (AUC 0.778) and was numerically higher after integration of RV strain (AUC 0.806), corresponding to an absolute AUC increase of 0.028 without statistical significance in paired DeLong testing (p=0.322; Figure 2). For 24-month mortality, TRI-Score yielded an AUC of 0.746, while the combined TRI-Score/RV strain model reached an AUC of 0.771 (absolute AUC increase 0.025; DeLong p=0.387). In Cox regression models combining TRI-Score risk category and RV strain, high TRI-Score risk category remained associated with both 12-month mortality (HR 6.13, 95% CI 1.19–31.63; p=0.030) and 24-month mortality (HR 4.90, 95% CI 1.46–16.51; p=0.010; Figure 3).


Conclusion
Pre-interventional TRI-Score risk category strongly predicted mortality after TTVI, while RV strain added complementary echocardiographic information and may further refine postprocedural risk stratification.