Despite technological advances in ventricular assist devices (VADs) to treat end-stage heart failure, hemocompatibility remains a constant concern, with supraphysiological shear stresses an unavoidable reality with clinical use. Given that impeller rotational speed is related to the instantaneous shear within the pump housing, it is plausible that modulation of pump speed may regulate peak mechanical shear stresses and thus ameliorate blood damage. The present study investigated the hemocompatibility of the HeartWare HVAD in three configurations typical of clinical applications: standard systemic support left VAD (LVAD), pediatric support LVAD, and pulmonary support right VAD (RVAD) conditions. Two ex vivo mock circulation blood loops were constructed using explanted HVADs, in which pump speed and external loop resistance were manipulated to reflect the flow rates and differential pressures reported in configurations for standard adult LVAD (at 3150 rev⸱min-1 ), pediatric LVAD (at 2400 rev⸱min-1 ), and adult RVAD (at 1900 rev⸱min-1 ). Using bovine blood, the mock circulation blood loops were tested at 37°C over a period of 6 hours (consistent with ASTM F1841-97) and compared with a static control. Hemocompatibility assessments were conducted for each test condition, examining hematology, hemolysis (absolute and normalized index), osmotic fragility, and blood viscosity. Regardless of configuration, continuous exposure of blood to the VAD over the 6-hour period significantly altered hematological and rheological blood parameters, and induced increased hemolysis when compared with a static control sample. Comparison of the three operational VAD configurations identified that the adult LVAD condition - associated with the highest pump speed, flow rate, and differential pressure across the pump - resulted in increased normalized hemolysis index (NIH; 0.07) when compared with the lower pump speed 'off-label' counterparts (NIH of 0.04 in pediatric LVAD and 0.01 in adult RVAD configurations). After normalizing blood residence times between configurations, pump speed was identified as the primary determinant of accumulated blood damage; plausibly, blood damage could be limited by restricting pump speed to the minimum required to support matched cardiac output, but not beyond.