Validation of TRI-SCORE for Outcome Prediction After Isolated Tricuspid Valve Surgery in Asian Patients

Abstract

Background: TRI-SCORE was recently developed in Europe as a risk model for predicting in-hospital death after isolated tricuspid valve surgery. We aimed to validate TRI-SCORE in an Asian population and investigate its value for predicting long-term outcomes.

Methods and results: The TRI-SCORE was calculated for 202 patients (65±11 years, 61% women, 81% functional tricuspid regurgitation) who underwent isolated tricuspid valve surgery for severe tricuspid regurgitation at 2 Korean centers and was based on 8 parameters: age, New York Heart Association class, right-sided heart failure signs, furosemide daily dose, glomerular filtration rate, bilirubin, left ventricular ejection fraction, and moderate/severe right ventricular dysfunction. The primary outcome was all-cause death during follow-up; the secondary outcome was in-hospital death. During a median follow-up duration of 50 (interquartile range, 21-82) months after isolated tricuspid valve surgery, 23 (11.4%) patients experienced the primary outcome, and 7 (3.5%) patients experienced the secondary outcome. Observed all-cause death and in-hospital death increased by up to 50% in those with higher scores. Patients with the primary outcome had a higher TRI-SCORE (4.5±2.4 versus 2.9±2.1; P=0.001) than those without. The TRI-SCORE showed a significant association with the primary outcome (concordance index, 0.77, cutoff value, 4) and in-hospital death (area under the curve, 0.84; cutoff value, 3). Using the Kaplan-Meier analysis, patients with a high TRI-SCORE exhibited a poor outcome for all-cause death at follow-up (log-rank P<0.001) and in-hospital death (log-rank P=0.004).

Conclusions: TRI-SCORE was validated in an Asian population and helped predict long-term outcomes after isolated tricuspid valve surgery.

Keywords: outcomes; risk; surgery; tricuspid regurgitation.

Figure 1. Risk score distribution.

(A) Box plot of the TRI‐SCORE distribution according to the TR mechanism. (B) The number of patients for each group of TR mechanism. TR indicates tricuspid regurgitation.

Figure 2. ROC analyses for outcomes.

ROC analyses for (A) all‐cause death and (B) in‐hospital death. The TRI‐SCORE revealed a significant predictive value for both outcomes. AUC indicates area under the curve; and ROC, receiver operating characteristic.

Figure 3. Kaplan–Meier analysis.

Kaplan–Meier analysis of freedom from (A) all‐cause death and (B) in‐hospital death according to the 3 TRI‐SCORE grades. The subgroup with a higher TRI‐SCORE showed significantly worse outcomes.