The possibility of achieving effective mechanical ventricular assistance without the need for thoracotomy provides great clinical advantages. Two in vitro systems were used to assess left ventricular unloading by means of a small-diameter cannula inserted retrograde into the left ventricle by cannulation of the femoral artery. This cannula is connected to the inlet of a centrifugal blood pump (CP) that delivers the blood into the contralateral femoral artery. Steady-flow test circulation was used to pump fluid in a closed loop from a reservoir through the test cannula back into the reservoir. Pressure drops over cannulae with inner diameters of 4, 5, 6, 7, and 8 mm at flows of 2, 2.5, 3 L/min, against a pressure of 60, 80, 100, and 120 mmHg were calculated. A stationary pressure drop of 120 mmHg was measured at a flow of 3 L/min through a 100 cm cannula with an inner diameter of 6 mm. The second system was a pulsatile mock circulation composed of an atrial and an arterial reservoir linked by a pneumatic prosthetic ventricle. This system was coupled with a 100 cm cannula, 6.1 mm inner diameter, which was passed across the outflow valve of the pulsatile prosthetic ventricle and connected to a CP. Fluid was withdrawn from the ventricle and pumped back into the arterial reservoir. Pulsatile pressure drop over the cannula was measured at different CP flows for increasing systolic ventricular pressure; heart unloading was quantified as a function of CP flow under baseline and failing conditions of the prosthetic left ventricle model. At a constant CP flow the pressure drop over the cannula increased with the pulsatility inside the ventricle. The work of the prosthetic ventricle was reduced by more than 50% when the CP pump was set to 3 L/min; at the same flow setting, when the situation of a failing left ventricle was simulated, the CP was able to take over all the work of the prosthetic ventricle, establishing a stationary flow and a 25% higher mean aortic pressure. This approach to left ventricular assistance may have significant clinical relevance.