Near infrared spectroscopy – multivariate analysis: a proof-of-concept study for non-destructive quality control of pharmaceutical preparations

Introduction
Quality control (QC) of pharmaceutical preparations (PP) is an essential part of the pharmaceutical drug compounding activity. However, qualitative and quantitative analytical assays nearly always require either a sample of the preparation (liquid forms) and can be destructive (solid forms). Near infrared spectroscopy - multivariate analysis (NIR-MVA) is a novel technology that could possibly be used for QC of PP in their primary container (syringes or bags) or dosage forms (capsules). However, it is as of yet mostly untested in a hospital setting. The objective of this proof-of-concept study was to investigate the analytical capabilities of NIR-MVA for the QC of several PP.

Methods
A Fourier-Transformation-NIR-spectrometer (TANGO-R^®, Bruker) including patented accessories to optimize the sample presentation and the software OPUSLAB provided by Ayna Analytics^® was used. The quantitative capabilities of the system were tested using several dosage forms: spironolactone capsules, amikacin ophthalmic solution in amber glass vials, and bevacizumab solutions conditioned in polypropylene syringes or IV-bags. The discriminative capability (identification confirmation) of the system was also tested using multiple anticancer IV preparations. Two preparations of each active substance (but different dosage) where analysed to create the identification data base, then 7 preparations of undisclosed identity were submitted for analysis. Each preparation was analysed either directly in its primary packaging (IV-bags, syringes, glass vials) or individual dosage form, without extraction (capsules).

Results
The broad possibility for quantitative analysis was shown for the capsules, syringes, and bags. The system was able to quantify (within the quantification range) the content of 5 independent 3 mg capsules (average trueness bias of -0.32%). Preliminary data also showed that it was also possible to quantify the content correctly for an independent bevacizumab bag and amikacin ophthalmic solution. Concerning the identification/discrimination assay: despite the very small number of reference spectra included into the database, the system correctly identified the nivolumab, carboplatin and gemcitabine preparations.

Discussion/Conclusion
NIR-MVA seems to be a very promising analysis tool for QC in PP. In this preliminary study and despite a very low number of reference spectra included in the database, this analytical method was able to correctly identify a monoclonal antibody and a platin derivative, which are compounds know to be difficult to analyse. The quantitative capabilities of the system also looked promising. The ability to correctly quantify a compound directly in its primary packaging container or dosage form could offer new possibilities to the hospital pharmacist looking to perform fast and safe QC on PP, as it nullifies the need to withdraw a sample.

See the poster