Predictive value of the TRI-SCORE for in-hospital mortality after redo isolated tricuspid valve surgery

Objectives The TRI-SCORE reliably predicts in-hospital mortality after isolated tricuspid valve surgery (ITVS) on native valve but has not been tested in the setting of redo interventions. We aimed to evaluate the predictive value of the TRI-SCORE for in-hospital mortality in patients with redo ITVS and to compare its accuracy with conventional surgical risk scores. Methods Using a mandatory administrative database, we identified all consecutive adult patients who underwent a redo ITVS at 12 French tertiary centres between 2007 and 2017. Baseline characteristics and outcomes were collected from chart review and surgical scores were calculated. Results We identified 70 patients who underwent a redo ITVS (54±15 years, 63% female). Prior intervention was a tricuspid valve repair in 51% and a replacement in 49%, and was combined with another surgery in 41%. A tricuspid valve replacement was performed in all patients for the redo surgery. Overall, in-hospital mortality and major postoperative complication rates were 10% and 34%, respectively. The TRI-SCORE was the only surgical risk score associated with in-hospital mortality (p=0.005). The area under the receiver operating characteristic curve for the TRI-SCORE was 0.83, much higher than for the logistic EuroSCORE (0.58) or EuroSCORE II (0.61). The TRI-SCORE was also associated with major postoperative complication rates and survival free of readmissions for heart failure. Conclusion Redo ITVS was rarely performed and was associated with an overall high in-hospital mortality and morbidity, but hiding important individual disparities. The TRI-SCORE accurately predicted in-hospital mortality after redo ITVS and may guide clinical decision-making process ( www.tri-score.com ).