Causes determination of noncompliance in the measurement of calcium in parenteral nutrition mixtures

Context
The pharmacotechnical unit produces 5000 parenteral nutrition mixtures (PNM) per year. The biochemistry laboratory controls sodium, potassium, calcium and glucose’s concentration on samples taken from the PNM. The Ca colorimetric assay is considered as non-compliant (NC) when the difference between the target and assayed concentration is higher than 20%.
Alerts have been received for the Ca determination of ternary PNM containing SMOFLIPID^® (SMOF) or MEDIALIPID^® (MEDIA), but not for binary PNM. A difference in concentration between the first NC assay and a second control assay suggests interference during the assay and not an incident during the production. In view of the non-conformities observed, interference related to lipid emulsions (LE) could be at the origin. The objective of this work is to determine the origin of the non-conformities attributable to LE.

Methods
We performed a retrospective study of Ca determinations of ternary PNM over 70 days. The parameters studied were: possible causes of interference, type of formulation and LE, mean deviation (MD) between target and assayed Ca concentration, number of NC assays and maximum deviation (Dmax).
A student’s t test and chi-square of independence were performed to compare the ME and the proportion of NC dosages before and after interference removal.

Results
From 04/04/2022 to 15/06/2022, 26 Ca dosages were made NC (14%): 22 concerned 2 formulations with the highest concentration of SMOF ("P2T" and "P4T"), 3 formulations with MEDIA and 1 with more diluted SMOF.
According to automat’s data sheet, a lipemic index (LI) > 1000 mg/dL causes interferences during the dosage. An automatic dilution at 1/1.16 is performed by the automaton when LI>1000 mg/dL. The LI of MNP, P2T and P4T, is respectively 3313 mg/dL and 3248 mg/dL. For MNP concentrated in MEDIA: LI = 1600 mg/dL.
On this period, 52 Ca determinations were performed on P2T and P4T, the MD is 15.36%, s = 10.69%. In total, 22 assays (42%) were returned NC (Dmax = 38%).
Since 16/06, a 1/5th dilution is programmed before the P2T and P4T assays, (32 assays). The MD found is 3.20%, s = 3.54% (p=7.64×10-11).
No measured assay was made NC (p=1.85×10-5) with Dmax = 8.9%.

Discussion
MNP containing SMOF and MEDIA have too high IL for the assay. Automat’s dilution was not enough for concentrated SMOF formulations. A higher dilution results in more accurate assays and an improvement in the deviation between the target and actual measured value. The automatic dilution of the MEDIA MNP is sufficient to limit the interferences but does not allow to totally eliminate the NC assays (3 on this period). The same work should be done on these MNP to verify if a 1/5 dilution reduces NC and MD in the long term.

See the poster