Accuracy of three automated compounding systems determined by end-product laboratory testing and comparison with manual preparation.

The accuracy of three parenteral nutrient (PN) automated compounding systems and the usefulness of end-product laboratory testing of PN solutions were evaluated. Ten identical PN solutions were prepared by the Nutrimix (Abbott), MicroMacro 23 (Baxa), and Automix-Micromix (Clintec) compounders and manually. Three samples were removed from each of the 40 solutions. Dextrose was measured by high-performance liquid chromatography using a modified sorbitol assay; sodium, potassium, calcium, and magnesium were measured by atomic absorption spectroscopy. Theoretical concentrations of ingredients in the PN solution were adjusted by the actual concentrations in the source containers. The solutions differed significantly in their concentrations of glucose, calcium, and magnesium but not in their concentrations of sodium and potassium. The mean glucose concentrations produced by the manual method and Nutrimix system matched the adjusted theoretical values and by the MicroMacro and Automix-Micromix compounders were within the manufactures' accuracy specifications. The sodium concentrations of all solutions were higher than the adjusted theoretical concentration but within the manufacturers' specifications. Potassium and magnesium concentrations were lower in all solutions than the adjusted theoretical concentrations. Calcium concentrations were above the adjusted theoretical value in all solutions. Differences in compounder accuracy, even when significant, were not consistent and not appreciable. Multiple samples of the same solution should be tested if laboratory analysis is part of quality control.