The effect of sodium metabisulfite (MBS) on hydrogen peroxide (HP) production in model and commercial amino acid solutions exposed to phototherapy light was studied. Model and commercial pediatric amino acid solutions were prepared such that the amino acid concentration was 1%. MBS concentration, riboflavin concentration, and duration of exposure to phototherapy light were varied to determine the effect on HP production. Control solutions were kept in the dark. HP production was assayed in the model amino acid solutions by using potassium iodide in the presence of ammonium molybdate. In all experiments, HP production was measured at 360 nm in the presence and absence of catalase. In light-exposed solutions, HP production increased linearly for several hours and reached a plateau by eight hours. A mean maximum of 940 microM was produced (data pooled for all solutions). No detectable HP was generated in the solutions kept in the dark. After two hours of light exposure, it was necessary to add at least 10 times more MBS than is typically found in commercial total parenteral nutrient solutions to scavenge all the HP produced. An average of up to 940 microM of HP was produced in model and commercial pediatric parenteral 1% amino acid solutions in the presence of phototherapy light and clinically relevant concentrations of riboflavin and MBS. Light exposure decreased the antioxidant effect of MBS.