Introduction: The requirement for right ventricular-assisted device (RVAD) implantation portends poor outcomes after left ventricular-assisted device (LVAD) implantation. We investigate RVF prediction models to help prediction of RV failure requiring RVAD implantation after centrifugal-flow LVAD HeartMate3 (HM3) placement. Methods: A total of 124 patients who underwent HM3 implantation at our institution from 2015-2023 were included in our analysis. We collected clinical variables used in EUROMACS RVF Risk Score (Circulation 2018), Utah Model (Am J Cardiology 2010), Michigan RVF Risk Score (JACC 2008), and CRITT (ATS 2013). Baseline scoring systems were compared with our cohort of patients with and without RVAD implantation after HM3 placement with ROC Area Under the Curve (AUC) statistical analysis. In addition, we investigated the pulmonary artery pulsatility index (PAPi) as a predictor for RV failure requiring RVAD. Results: There were 17 patients (13.7%) who underwent RVAD implantation after HM3 placement. Patients who required RVAD implant had worse survival (p = 0.037). The EUROMACS score differentiated those patients who required RVAD from those who did not (AUC: 0.66; 95% C.I: 0.51-0.81). In contrast, CRITT (AUC: 0.64; 95% C.I: 0.49-0.78), Michigan score (AUC: 0.37; 95% C.I: 0.25-0.51), and Utah (AUC: 0.53; 95% C.I: 0.40-0.68) scores did not differentiate those who required RVAD. PAPi was significantly lower in patients requiring RVAD (2.5±1.2 vs 3.9±3.0, p=0.001). Also, even as a single parameter, PAPi had an AUC of 0.63 (95% C.I: 0.50-0.76). Conclusion: RVAD implantation is associated with high early mortality. Existing RVF prediction models were not able to predict future RVAD requirements with high AUC, with the EUROMACS system having the only statistical significance. There is a lack of a strong predictive model for RV failure in HM3. Future efforts are needed to establish predictors for RV failure requiring RVAD implants in the HM3 population, especially utilizing novel hemodynamic markers such as PAPi.
Read full abstract