Comparative analysis of cardiac function before LVAD implantation in patients with and without early, acute right heart failure

\(\bf Background:\) Early acute right heart failure (eaRHF) during left ventricular assist device (LVAD) implantation significantly impacts patient survival and complicates perioperative management. Although numerous clinical, echocardiographic, and hemodynamic risk factors have been identified, accurately predicting eaRHF remains challenging. Cardiac magnetic resonance (CMR) provides a precise, non-invasive evaluation of cardiac structure and function and may enhance risk stratification eaRHF. This study aims to assess the predictive value of preoperative CMR-derived parameters, comparing their utility to established echocardiographic and right heart catheterization (RHC) markers for identifying eaRHF. \(\bf Methods:\) This retrospective analysis was conducted on 55 patients who received CMR before LVAD implantation at our center between 2018 and 2024. Of these 55 patients, 40 had image quality sufficient for offline analysis. Patients receiving a temporary right ventricular assist device (tRVAD) intraoperatively were defined as having eaRHF. Receiver Operating Characteristic (ROC) analysis was used to evaluate the predictive capability of CMR, echocardiographic, and RHC parameters. \(\bf Results:\) Ten patients (25%) developed eaRHF. Preoperative bilirubin levels were significantly higher in the eaRHF group (1.6 mg/dl vs. 1.1 mg/dl, \(\it p\) = 0.010). Echocardiographic Tricuspid Annular Plane Systolic Excursion (TAPSE) tended to be lower in eaRHF patients (12 mm vs. 18 mm, \(\it p\) = 0.080). RHC parameters, specifically right ventricular stroke work index (RV-SWI; \(\it p\) < 0.001), cardiac output (CO; \(\it p\) = 0.003), and cardiac index (CI; \(\it p\) = 0.004), were significantly lower in eaRHF patients. CMR showed significantly higher RV end-diastolic volumes (RV-EDV, 288.4 ml vs. 216.7 ml, \(\it p\) = 0.046) and indexed RV-EDV (RV-EDVi, 135.4 ml/m\(^{2}\) vs. 104.7 ml/m\(^{2}\), \(\it p\) = 0.033) in the eaRHF group. ROC analysis identified CO (AUC = 0.90, sensitivity = 100%, specificity = 72%, \(\it p\) < 0.001), CI (AUC = 0.88, sensitivity = 83%, specificity = 83%, \(\it p\) < 0.001), and RV-SWI (AUC = 0.86, sensitivity = 83%, specificity = 86%, \(\it p\) < 0.001) as strong predictors. Moderate predictive values were observed for RV-EDVi (AUC = 0.73, \(\it p\) = 0.040) and RV global radial strain (RV-GRS; AUC = 0.70, \(\it p\) = 0.044). \(\bf Conclusion:\) Hemodynamic parameters from RHC demonstrated the strongest predictive capability for eaRHF. However, selected CMR-derived parameters, especially indexed RV-EDV and RV GRS, offer moderate predictive value and may serve as adjunctive tools in preoperative risk stratification for LVAD candidates.