The longterm effects of ketamine on haemodynamic parameters and exogenous catecholamine requirements were studied in twenty-five critically ill patients with catecholamine-dependent heart failure. Following sedation with midazolam (0.15±0.07 mg.kg-1 .h-1) and sufentanil (0.88±0.33 µg.kg-1 .h-1), patients with impaired left ventricular function (left ventricular ejection fraction area 30±7%) were randomly assigned to receive ketamine (2.5±0.9 mg.kg-1 .h-1) and midazolam (Group A) or remained on sufentanil/midazolam (Group B). Haemodynamic measurements were performed throughout the first 24 hours after randomization. In group A cardiac index decreased by 21% (P=0.01), mean arterial pressure increased by 13% (P=0.01), mean pulmonary artery pressure by 14% (P=0.04), pulmonary capillary wedge pressure by 20% (P=0.03), and systemic vascular resistance index by 38% (P<0.001). No significant cardiovascular effects were observed in Group B. Neither group had significant changes of exogenous catecholamine requirement. In conclusion, ketamine exhibits potential negative cardiovascular effects in patients with catecholaminedependent heart failure. Therefore, ketamine should not be considered a first line drug for longterm sedation of patients with impaired left ventricular function.