BACKGROUND Premature infants require high intakes of Ca and P to mimic fetal accretion rates. With the current phosphate salt used, adequate amounts cannot be provided due to the precipitation of Ca and P in TPN solutions. OBJECTIVE To compare monobasic potassium phosphate (monobasic regimen) and monobasic plus dibasic potassium phosphate (dibasic regimen) on calcium phosphate solubility in 5 amino acid products, and to determine whether solubility differences observed in these products can be explained by buffering capacity. METHODS TPN solutions were prepared according to standard clinical practice. The following amino acid products were used at 3% concentrations: Primene, Vamin N, TrophAmine, Aminosyn-PF, and Travasol. Dextrose 10%, standard electrolytes, heparin, vitamins and trace elements were added. Calcium (as gluconate) and phosphate (as monobasic or dibasic regimen) were added in one-to-one molar ratios from 0-45 mmol/L. Solutions were inspected macroscopically and microscopically for precipitation under three conditions: immediately, 24 h after preparation at room temperature, and 3 h later in a 37 degrees C water bath. Buffering capacity was determined for each amino acid product by titrating with standardized 0.1 M NaOH. RESULTS Variations in Ca:P solubility and buffer capacity exist between amino acid solutions. With Primene and Vamin no macroscopic or microscopic precipitation was detected up to 45 mmol/L using monobasic regimen, compared to 25 mmol/L using dibasic regimen with Trophamine. Buffer capacity did not account for the solubility differences observed between the five amino acid products, which were related to the pH of the final solution. CONCLUSIONS These data will allow clinicians to double the current concentrations of calcium and phosphate in neonatal TPN solutions using monobasic regimen. Although this is particularly relevant to situations when fluid intake is restricted, the effect of the acid load needs to be investigated in extremely low birth weight infants.