Bridge or Burden? Heart Transplant Outcomes Following Left Ventricular Assist Device Explant in the Contemporary Allocation Era.

Background

Whether left ventricular assist device (LVAD) explant during heart transplantation confers additional risk remains unsettled in the contemporary allocation era.

Methods

Heart transplant recipients at a single center (January 2020-December 2024) were retrospectively reviewed. Patients undergoing multi-organ transplants, with congenital heart disease, ≥2 prior sternotomies, non-HeartMate 3 LVADs or pre-transplant ECMO were excluded. Recipients were grouped as primary sternotomy, non-LVAD redo sternotomy, or LVAD explant. Outcomes included transfusions, vasoplegia, and 30-day, 1-year, and 2-year mortality. IPTW-weighted logistic regression and Cox models assessed vasoplegia and mortality.

Results

Among 432 recipients (primary sternotomy 237; non-LVAD redo sternotomy 90; LVAD explant 105), transfusion volumes increased stepwise intraoperatively and 48 hours postoperatively (both p<0.001). Vasoplegia differed across cohorts (12.1% vs 24.3% vs 22.9%; p=0.014); after IPTW adjustment, LVAD explant remained associated with higher odds versus primary sternotomy (OR 2.2 [1.1-4.4]; p=0.03). Unadjusted survival declined across groups at 1 year (98.3% vs 93.3% vs 87.6%, p<0.001) and 2 years (95.0% vs 89.1% vs 85.2%, p=0.003). Versus primary sternotomy, LVAD explant had higher adjusted hazards of death at 1 year (HR 5.8 [1.8-19], p=0.003) and 2 years (HR 2.9 [1.3-6.7], p=0.01). In an exploratory lower-risk LVAD explant subgroup (n=37), observed mortality differences were attenuated.

Conclusions

LVAD explant during heart transplant was associated with greater transfusions, vasoplegia, and early-to-intermediate mortality compared with primary sternotomy. These findings suggest LVAD explant identifies a higher-risk transplant phenotype shaped by surgical burden, device-associated physiologic effects, and allocation-era transplant timing, rather than by intrinsic device-related risk alone.