BACKGROUND: Therapeutic hypothermia was shown to facilitate resumption of spontaneous circulation when instituted during cardiac arrest. Here, we investigated whether it directly improved the chance of successful resuscitation independently of adrenaline administration in rabbits. We further evaluated the direct effect of hypothermia on vascular function in vitro. METHODS: In a first set of experiments, four groups of anesthetized rabbits were submitted to 15 min of cardiac arrest and subsequent cardiopulmonary resuscitation (CPR). The "control" group underwent CPR with only cardiac massage and defibrillation attempts. Two other groups received cold or normothermic saline infusion during CPR (20 mL/kg of NaCl 0.9% at 4°C or 38°C, respectively). In a last group, the animals received adrenaline (15 µg/kg intravenously) during CPR. In a second set of experiments, we evaluated at 32°C vs. 38°C the vascular function of aortic rings withdrawn from healthy rabbits or after cardiac arrest. RESULTS: In the first set of experiments, cardiac massage efficiency was improved by adrenaline but neither by hypothermic nor normothermic saline administration. Resumption of spontaneous circulation was observed in five of eight animals after adrenaline as compared with none of eight in other groups. Defibrillation rates were conversely similar among groups (7/8 or 8/8). In the second set of experiments, in vitro hypothermia (32°C) was not able to prevent the dramatic alteration of vascular function observed after cardiac arrest. It also did not directly modify vasocontractile or the vasodilating functions in healthy conditions. CONCLUSION: In rabbits, hypothermia did not exert a direct hemodynamic or vascular effect that might explain its beneficial effect during CPR.