We evaluated the effect of frequent, early bolus administration of low-dose sodium bicarbonate (NaHCO3) on blood gas values during ventricular fibrillation and cardiopulmonary resuscitation (CPR) compared with normal saline and standard bolus doses of NaHCO3. This was a randomized laboratory investigation involving 13 mongrel dogs and 18 experiments (5 dogs were used in a crossover manner). Each dog underwent 3 minutes of ventricular fibrillation, followed by 15 minutes of CPR. Animals were randomly assigned to one of three treatments administered early in the resuscitation effort: NaHCO3 0.5 mEq/kg at 5, 10, and 15 minutes of ventricular fibrillation (SB); NaHCO3 1 mEq/kg at 5 minutes and 0.5 mEq/kg at 15 minutes of fibrillation (B); or 0.9% NaCl 1 ml/kg at 5 minutes and 0.5 ml/kg at 15 minutes of fibrillation (P). A total of 15 experiments were included for analysis. Arterial and venous blood gases were sampled at 4, 8, 13, and 18 minutes of fibrillation. The SB group demonstrated the highest arterial partial pressures of carbon dioxide (pCO2) at each sampling point after NaHCO3, including the 18-minute sample: 42 ± 12, 29 ± 11, and 35 ± 10 torr for SB, P, and B, respectively. In addition, SB produced arterial alkalemia (pH > 7.45) after NaHCO3 administration. The arterial pH at 18 minutes of fibrillation for SB, P, and B was 7.46 ± 0.14, 7.29 ± 0.07, and 7.41 ± 0.1, respectively. Similar trends for pCO2 and pH were observed for venous samples. Early, frequent administration of low-dose NaHCO3 during CPR is associated with elevated pCO2 and pH (alkalotic) values that may be potentially detrimental in this setting. It appears that this mode of administration offers no advantage over B with regard to blood gas values during CPR in this canine model.