Regurgitant fraction by 3DE: outcome-based grading of secondary tricuspid regurgitation across atrial and ventricular phenotypes

Background The use of regurgitant fraction (RegFr) to grade the severity of tricuspid regurgitation (TR) by echocardiography relies on cardiac magnetic resonance criteria using cut-off values derived from patients with mitral regurgitation. Purpose To identify and validate outcome-based partition values of the RegFr determined with three-dimensional echocardiography (3DE) to grade the severity of secondary TR (STR), and to determine whether RegFr-based grading across different STR phenotypes (atrial vs ventricular) is associated with distinct prognostic profiles. Methods RegFr, defined as the ratio between regurgitant volume measured using the corrected proximal isovelocity surface area method, and the right ventricular stroke volume measured by 3DE, was determined in 586 patients (derivation cohort, 74±13 years; 52% women) with at least mild STR. An optimised multiway partitioning algorithm was applied to derive outcome-based RegF partition, defining mild, moderate, and severe STR. These thresholds were tested in an independent cohort of 200 patients (validation cohort; 73±12 years; 60% women). The algorithm was reapplied to patients with ventricular- (n = 355) and atrial-STR (n = 231) to test if phenotype-specific thresholds are needed. Results In the derivation cohort, the RegFr partition values for identifying patients with mild, moderate, and severe STR were <31%, 31%-54%, and >54%. These partition values stratified 3-year survival (69 ± 4%, 55 ± 4%, and 31 ± 5%; log-rank < 0.0001) and retained prognostic significance in the validation cohort (P < 0.001). In phenotype-specific analyses, optimal cut-points shifted to 31%/47% in ventricular STR (log-rank χ² = 34.7; permutation P < 0.001) and to 50%/56% in atrial STR (log-rank χ² = 37.1; permutation P < 0.001). Conclusion In patients with STR, RegFr is associated with outcomes. However, atrial and ventricular STR require specific threshold values to grade severity.