Predicting Significant Right Ventricular Failure Post-LVAD Implantation Using CMR Compared to Echocardiography and Right Heart Catheterisation

Abstract

Purpose Right ventricular failure (RVF) complicates 10-40% of LVAD implants. Significant RVF is associated with increased mortality and reduced chance of being bridged to transplant. It is therefore crucial to identify patients at high risk of RVF using pre-operative investigations. We investigated markers of RV function on pre-LVAD echocardiography (echo), right heart catheterization (RHC) and cardiovascular magnetic resonance imaging (CMR) in predicting significant post-LVAD RVF. Methods We performed a single centre retrospective analysis of 220 patients who underwent LVAD implantation between 2011-18. 33 patients had CMR within 15 months pre-LVAD implantation. CMR was used to derive RV volumes, TAPSE, RV ejection fraction, strain and scar burden. RV function parameters were derived from echo and RHC just prior the LVAD implantation. INTERMACS criteria was used to categorize RVF based on duration of pharmaceutical support (post-implant IV inotropes/vasodilators/inhaled nitrous oxide) - mild ( 14 days). The latter also included the need for RVAD or death due to RVF. Moderate and severe RVF were considered as significant RVF. Results The median age was 49 (IQR 39-56). 31 were males. 18 patients had no/mild RVF; 14 had significant RVF (8 moderate RVF, 6 severe RVF); 1 patient was not categorised due to incomplete data. 29 survived to discharge. Dilated cardiomyopathy was the commonest aetiology of heart failure (17/33) followed by ischaemic heart disease (11/33). Echo derived pulmonary hypertension (42 vs 36mmHg) and proportion of those with tricuspid regurgitation (83.3% vs 92.8%) were not different in the two groups. PVR on RHC was not different between the groups (2.82 vs 3.68 WU). CMR derived volumes, TAPSE (12.2 vs 12.1mm), Global longitudinal strain (-9.1% vs -7.2%) and radial strain (-12.4% vs -10.0%) and the presence of scar were not different between the two groups. CMR-RVEF however was lower in the group significant RVF (23.5 ± 8.2%) when compared to the group with no/mild RVF (31.0 ± 10.7%) (p=0.03). Conclusion When compared to indices from RHC and echo, pre-operative RVEF on CMR is the only marker that differentiated the group who developed significant post-LVAD RVF from the group that did not. A larger study with CMR closer to the time of LVAD implant is needed to validate this.