Optical emission spectrometer inductively coupled plasma: Application to chemical controls of parenteral nutrition mixtures

S.Vandewoestyne ; D.Sémély ; J-B. Desbouges ; F. Roussin ; Z.Ramjaun ; M.Tafani. Uf Pharmacotechnie, C.H.U Toulouse, Toulouse, France.

Context

The increase of activity of parenteral nutrition in our production unit has led us to acquire an optical emission spectrometer inductively coupled plasma (ICP-OES), the Avio200® (Perkin Elmer), to achieve the physicochemical control of parenteral nutrition mixtures (PNM).

Objectives

The objective is to validate the performance of an ICP-OES by analyzing the concentration of sodium (Na+) and potassium (K+) in binary and ternary PNM.

Methods

Three standard binary and ternary PNM have been produced using a qualified automated compounding system, ExactaMix® 2400 (Baxter). 9 samples were taken from each infusion bag. The procedure for analytical validation included:

Study of the matrix effect by comparing two ranges of calibration: with or without dextrose and by analyzing the concentrations of Na+ and K+ in each sample according to calibration curve. Result analysis was conducted by comparing of the coefficients of correlation according to Fischer transformation.
Determination of the linearity range, repeatability with 9 assays of each analyte on day 1 and intermediate precision with all the assays of each analyte during three days.

Results

No significant matrix effect was found (U <1.96). However, the means of relative errors on the concentrations of Na+ and K+ are lower with the dextrose calibration curve (< 5% vs < 7.5%).

The method is linear for concentrations between 20 and 500 mg/mL with the dextrose calibration curve (r=0.9999 for the Na+ and r=0.9998 for K+). The results showed coefficients of variation in repeatability and intermediate precision were less than 5% for binary and ternary PNM for both analytes.

Discussion-conclusion

This study showed the influence of the matrix on the accuracy of the results obtained dextrose. The use of dextrose calibration curve allows us to release all the preparations with a relative error below 10%. This study will be continued with more samples and other analytes (magnesium, calcium and zinc).