Liver dysfunction due to a low cardiac output state after cardiac surgery is associated with a poor prognosis, but whether one inotrope is superior to another in improving hepatic perfusion remains uncertain. This study compared the systemic and hepatic haemodynamic effects of levosimendan to dobutamine in patients with a low cardiac output state (cardiac index < 2.2 1/min/m2) after on-pump cardiac surgery. A total of 25 patients were randomised to receive either an intravenous bolus of levosimendan (12 μg/kg) over 15 minutes, followed by an infusion of 0.2 μg/kg/min for 24 hours, or an infusion of dobutamine 7.5 μg/kg/min for 24 hours and completed the study. The systemic and hepatic haemodynamics at 24 and 48 hours were all better after levosimendan than dobutamine (dobutamine group: cardiac index (1/min/m2)=2.51 [standard deviation ±0.29], 2.40±0.23; portal vein flow (ml/min): 614.0± 124.7, 585.9±144.8; pulsatility index: 2.02±0,28, 2.98±0.27 versus the levosimendan group: cardiac index: 3.02± 0.27, 2.98± 0.30; portal vein flow: 723.0± 143.5, 702.9±117.8; pulsatility index: 1.71 ±0.26,1.73 ±0.27). The improvement in portal vein blood flow at 48 hours was significantly better after levosimendan than dobutamine (41% vs. 11% increment from baseline, P<0.05). In addition, there was a significant reduction in hepatic artery resistance after levosimendan but not dobutamine (resistance index reduction 6.5% vs. 0%, P<0.05). In summary, levosimendan can be considered as a selective liver vasodilator and can improve hepatic blood flow through both the hepatic artery and portal venous system, whereas dobutamine can only improve the portal venous blood flow without vasodilating the hepatic artery.