Background An increasing number of continuous flow pumps are currently under clinical studies, however very little data exist on the hemodynamic and cellular responses of the failing heart to continuous flow support. The purpose of this investigation was to characterize the response of the failing myocardium to continuous flow support. Methods We compared echocardiographic and cellular markers of failing myocardium at the time of left ventricular assist device (LVAD) implantation and explantation in 20 consecutive patients (12 pulsatile flow [Novacor] and 8 continuous flow [DeBakey-Noon]). Results The use of mechanical support with both continuous- or pulsatile-type LVADs resulted in a reduction of left ventricular end-diastolic dimension (LVEDD), end-diastolic volume (EDV), end-systolic volume (ESV) and left atrial volume (LAV), as well as a decrease in mitral E/A ratio, tricuspid regurgitation velocity (TRV) and pulmonary valve acceleration time (PVAT). Comparative analyses for patients treated with a continuous- vs pulsatile-type LVAD support showed a greater degree of unloading with the latter type, as shown by the effect on LVEDD (−13.7% vs −33.7%, p = 0.0.004), EDV (−23.5% vs −41.2%, p = 0.015), ESV (−25.6% vs −57.6%, p = 0.001) and LAV (−25.2% vs −40.4%, p = 0.071). The hemodynamic effects of continuous vs pulsatile LVAD support were similar, as shown by their effect on mitral E/A ratio (−23.9% vs −39.9%, p = NS), TRV (−26.4% vs −23.8%, p = NS) and PVAT (28.5% vs 38.5%, p = NS). Only pulsatile support demonstrated a statistically significant percent change in mass (−6.3% vs −20.6%, p = 0.038). Continuous and pulsatile forms of mechanical support demonstrated equivalent reductions in myocardial tumor necrosis factor-α (TNF-α), total collagen and mycocyte size. Conclusions Our findings show that, although there are differences between these 2 devices in magnitude of unloading, both forms of support effectively normalize cellular markers of the failing phenotype.