OBJECTIVE During cardiopulmonary resuscitation (CPR), elimination of CO2 was shown to be limited by low tissue perfusion, especially when very low perfusion pressures were generated. It has therefore been suggested that sodium bicarbonate (NaHCO3), by producing CO2, might aggravate the hypercarbic component of the existing acidosis and thereby worsen CPR outcome. The objectives of this study were to evaluate the effects of CO2 producing and non-CO2 producing buffers in a canine model of prolonged ventricular fibrillation followed by effective CPR. METHODS Prospective, randomized, controlled, blinded trial. METHODS Experimental animal research laboratory in a university research center. METHODS Thirty-eight adult dogs, weighing 20 to 35 kg. METHODS Animals were prepared for study with thiopental followed by halothane, diazepam, and pancuronium. Ventricular fibrillation was electrically induced, and after 10 mins, CPR was initiated, including ventilation with an FIO2 of 1.0, manual chest compressions, administration of epinephrine (0.1 mg/kg every 5 mins), and defibrillation. A dose of buffer, equivalent to 1 mmol/kg of NaHCO3, was administered every 10 mins from start of CPR. Animals were randomized to receive either NaHCO3, Carbicarb, THAM, or 0.9% sodium chloride (NaCl). CPR was continued for up to 40 mins or until return of spontaneous circulation. RESULTS Buffer-treated animals had a higher resuscitability rate compared with NaCl controls. Spontaneous circulation returned earlier and at a significantly higher rate after NaHCO3 (in seven of nine dogs), and after Carbicarb (six of ten dogs) compared with NaCl controls (two of ten dogs). Spontaneous circulation was achieved twice as fast after NaHCO3 compared with NaCl (14.6 vs. 28 mins, respectively). Hydrogen ion (H+) concentration and base excess, obtained 2 mins after the first buffer dose, were the best predictors of resuscitability. Arterial and mixed venous Pco2 did not increase after NaHCO3 or Carbicarb compared with NaCl. CONCLUSIONS Buffer therapy promotes successful resuscitation after prolonged cardiac arrest, regardless of coronary perfusion pressure. NaHCO3, and to a lesser degree, Carbicarb, are beneficial in promoting early return of spontaneous circulation. When epinephrine is used to promote tissue perfusion, there is no evidence for hypercarbic venous acidosis associated with the use of these CO2 generating buffers.