The hemodynamic classification system for hemodynamic-related adverse events in left ventricular assist device patients

Abstract Background Durable left ventricular assist device (LVAD) physiology predisposes to right heart failure (RHF) and aortic insufficiency (AI), recently labelled hemodynamic-related adverse events (HDREs). However, the pathophysiology, incidence, and associated morbidity of HDREs remains largely unsettled due to heterogeneity in data reporting. Purpose We sought to validate the recently proposed LVAD Hemodynamic Classification System (HCS) for HDREs. Methods RHF and AI HDREs were retrospectively adjudicated, staged (mild to severe) and phenotyped (isolated or combined RHF/AI-HDREs) based on the HCS assessed at 6-month post-LVAD implant among 183 consecutive patients undergoing LVAD implant at a quaternary heart failure center, and either alive at 6-month on support (n=175) or deceased due to late HDREs 1 to 6 months post-LVAD implant (n=8). The primary outcome was a composite of first HF hospitalization or cardiovascular death between 6 and 24 months after LVAD-implant. Results At 6-month assessment, RHF-HDRE was adjudicated in all patients, while AI-HDRE could not be adjudicated in 22 (12.0%) patients. Overall, 53 (29.0%) patients developed an RHF-HDRE, and 26 (16.0%) an AI-HDRE (phenotypes: 39 isolated RHF, 15 isolated AI, 11 combined RHF-AI, 3 not-adjudicated). During pre-LVAD assessment, age and pulmonary artery pulsatility index were independently associated with RHF-HDRE; and pre-LVAD AI and systolic pulmonary artery pressure with AI-HDRE. Patients developing the combined RHF-AI phenotype required the most intensive bundle of critical care pre-LVAD implantation, had more frequent pre-LVAD RV dysfunction, the highest rate of early post-LVAD RHF, and the most prominent features of suboptimal left ventricular unloading, RV dysfunction, peripheral congestion, renal damage and impaired functional capacity at 6-month assessment. RHF-HDRE (adj-HR 2.69, 95%CI 1.14-6.34, p=0.024), AI-HDRE (adj-HR 4.21, 95%CI 1.70-10.40, p=0.002) and HDRE phenotype (isolated RHF or AI HDRE vs. no HDRE: adj-HR 2.80, 95%CI 1.03-7.62, p=0.044; combined RHF and AI HDRE vs. no HDRE: adj-HR 4.53, 95%CI 1.33-15.35, p=0.015) were associated with the primary outcome. A graded increase in primary outcome was observed with rising RHF and AI stages. Conclusions The LVAD Hemodynamic Classification System for HDREs identifies pathophysiological and clinical clusters of patients with significant prognostic implications. It may help the development of targeted strategies to mitigate HDREs and improve the success of long-term LVAD support.