Incidence and clinical correlates of de-novo aortic regurgitation with a fully magnetically levitated left ventricular assist device: a MOMENTUM 3 trial portfolio analysis

Aims: We assessed the incidence, predictors and clinical correlates of de-novo aortic regurgitation (AR), which physiologically reduces left ventricular assist device (LVAD) effectiveness due to recirculation syndrome, in the MOMENTUM 3 trial portfolio of the fully magnetically levitated HeartMate 3 (HM3) pump using the randomized pivotal trial (PT) and post-trial continued access protocol (CAP). Methods and results: De-novo aortic regurgitation incidence at 2 years was analysed in the randomized PT and validated in the first 1000 implanted patients of the CAP. Patients with concomitant/prior aortic valve surgery or without baseline or post-implant echocardiograms were excluded from this analysis. AR severity was assessed qualitatively by site-adjudicated echocardiograms (significant AR was defined as moderate or severe grade on echocardiogram). Of 1028 patients enrolled in the PT, 918 were eligible for inclusion in this analysis (HM3, n = 465; HMII, n = 453). At 2 years of LVAD support, freedom from significant AR was greater in the HM3 (92%) than HMII (82%) (hazard ratio 0.45, 95% confidence interval 0.27–0.75, p < 0.01). Of 907 HM3 patients analysed from the first 1000 implanted CAP patients, the rate of freedom from significant AR was 90%, consistent with the PT (p = 0.3). In the combined HM3 group (n = 1372), multivariable Cox modelling identified increasing age and female sex as significant predictors. Survival free of urgent transplant or AR corrective procedure was similar between HM3 patients with and without significant de-novo AR. Conclusions: The development of moderate or severe grade de-novo AR is reduced with the fully magnetically levitated HM3 LVAD compared to the axial-flow HMII pump. The occurrence of significant de-novo AR with the HM3 pump is not associated with a worse outcome at 2 years of follow-up.