Sodium benzoate lowers serum ammonia concentrations by the activation of a non-urea cycle pathway of ammonia removal. The disposition of sodium benzoate was monitored in four hyperammonemic newborn infants, using a simple and newly developed assay for benzoate and hippurate, to assess (1) the metabolic capability of patients of this age to utilize this pathway for nitrogen removal, (2) the potential risks of benzoate toxicity at clinically achieved serum benzoate concentrations, and (3) the value of routine monitoring of serum benzoate concentrations in this patient population. In three of the four infants, more than half of the administered benzoate was converted to hippurate. Hippurate was effectively cleared by the neonatal kidney, although removal of unconjugated benzoate by peritoneal dialysis or urinary excretion was slow compared with the metabolic conversion to hippurate. There was a considerable interpatient variability in benzoate metabolism; consequently, an eight-fold range in serum benzoate concentrations (2.14 to 16.0 mM/L) was found after patients had received benzoate for longer than 24 hours. These serum benzoate concentrations were calculated to be capable of producing substantial (four to 25 times) increases in free bilirubin concentrations in jaundiced infants. Although sodium benzoate offers considerable promise for the treatment of hyperammonemia, toxicity appears likely in some infants receiving this drug in currently recommended doses. Monitoring of serum concentrations appears to be warranted. Dosage reduction in jaundiced infants and in those with demonstrated insufficiency of benzoate metabolism is recommended.