Extent of trace-element contamination from simulated compounding of total parenteral nutrient solutions.

Abstract

The extent and sources of trace-element contamination during simulated manual and automated compounding of total parenteral nutrient (TPN) solutions were studied. Four experimental groups were prepared: (1) sterile water for injection as received from the manufacturer, (2) TPN bags manually filled with sterile water for injection, (3) TPN bags automatically filled with sterile water for injection, and (4) sterile water for injection manually added directly to sample vials (i.e., bypassing the TPN bag). After the bags had been stored for 16 hours at 4 degrees C, samples were analyzed for 66 trace elements by inductively coupled plasma-mass spectrometry. Of the 66 trace elements analyzed for, 3 (zinc, boron, and aluminum) were present as contaminants in each of the experimental groups. Very low background concentrations of zinc, boron, and aluminum were found to contaminate the sterile water as received from the manufacturer. Simulated compounding (manual or automated) in TPN bags did not contribute additional boron or aluminum but did result in a mean total zinc concentration of 35.4 micrograms/L; of this, 12.4 micrograms/L came from the compounding equipment, 9.1 micrograms/L from the TPN bag, and 13.9 micrograms/L from the sterile water for injection. Although the amount of zinc contributed by compounding was significantly higher than the background level, the amount was inconsequential compared with that needed to satisfy physiological requirements. Simulated compounding of TPN solutions either manually or with an automated system contributed minimal trace-element contamination.