OBJECTIVE Despite the absence of outcome evaluation, the use of sodium bicarbonate in cardiac arrest has declined based on advanced cardiac life-support guidelines. The effects of bicarbonate therapy on outcome in a canine model of ventricular fibrillation cardiac arrest of brief (5-min) and prolonged (15-min) duration were examined. METHODS Prospective, randomized, controlled trial. METHODS Experimental animal laboratory in a university medical center. METHODS Thirty-two adult dogs, weighing 10 to 17 kg. METHODS The animals were prepared with ketamine, nitrous oxide/oxygen, halothane, and pancuronium. Ventricular fibrillation was then electrically induced and maintained in arrest for 5 mins (n = 12) or 15 mins (n = 20). Canine advanced cardiac life-support protocols were instituted, including defibrillation, cardiopulmonary resuscitation (CPR), and the administration of epinephrine (0.1 mg/kg), atropine, and lidocaine. The bicarbonate group received 1 mmol/kg of sodium bicarbonate initially, and base deficit was corrected to -5 mmol/L with additional bicarbonate, whereas acidemia was untreated in the control group. Cardiopulmonary values were recorded at intervals between 5 mins and 24 hrs, and the neurologic deficit score was determined at 24 hrs after CPR. RESULTS The treatment group received an additional 2 to 3 mmol/kg of bicarbonate in the early postresuscitation phase. Compared with controls, the bicarbonate group demonstrated equivalent (with brief arrest) or improved (with prolonged arrest) return of spontaneous circulation and survival to 24 hrs, with lessened neurologic deficit. The acidosis of arrest was decreased in the prolonged arrest group without hypercarbia. Improved coronary and systemic perfusion pressures were noted in the bicarbonate group with prolonged arrest, and the epinephrine requirement for return of spontaneous circulation was decreased. CONCLUSIONS The empirical administration of bicarbonate improves the survival rate and neurologic outcome in a canine model of cardiac arrest.