The effects of HCO3Na load on acid-base balance and muscle intracellular bioenergetics have been investigated using 31P-magnetic resonance spectroscopy in an experimental model of endotoxinic shock. Anesthetized, mechanically ventilated, and paralyzed rats (n = 16) were given an intravenous bolus of Escherichia coli lipopolysaccharide (15 mg/kg). When shock was established they were randomly assigned to receive either HCO3Na intravenously (2 mmol/kg in 2 min) or an equimolar saline injection. Lipopolysaccharide induced a significant decrease in the levels of mean arterial pressure (58 ± 6 vs. 120 ± 8 mmHg), arterial pH (7.20 ± .03 vs. 7.35 ± .01), intracellular pH (6.86 ± .04 vs. 7.08 ± .01), a marked hyperlactatemia (7 ± 3 vs. 1.2 ± .2 mmol/L) and a drop in the phosphocreatine-inorganic phosphate ratio. In the bicarbonate-loaded rats, mean arterial pressure further decreased whereas it remained unchanged in the saline group. Bicarbonate increased arterial pH and PaCO2 transiently. In the saline group, arterial pH decreased and PaCO2 remained stable. In both groups, intracellular pH and high energy phosphates had a similar evolution. In this model of septic shock, partial correction of arterial pH using HCO3Na did not reduce the metabolic cellular injury in skeletal muscle. Based on these results, HCO3Na may be of limited therapeutic value in severe septic metabolic acidosis.