HANDHELD RAMAN SPECTROSCOPY: AN ESSENTIAL TOOL TO TACKLE THE SUBSTANDARD MEDICINES ISSUE?

According to the World Health Organization, there is a growing concern about the quality of medicines around the world. Indeed, more and more substandard and falsified medicines are identified. That is why several spectroscopic techniques such as Raman, near- or mid-infrared spectroscopy, have gained great interest for this purpose. By means of chemometrics, interesting results have been shown in terms of elucidation of falsified medicines composition by hyperspectral imaging (L. Coic) and with benchtop spectrophotometers (O.Ye. Rodionova). However, these instruments are rather expensive, heavy and are not appropriated for low and middle-income countries. To circumvent these issues, several low-cost and middle-cost handheld spectrophotometers have been developed. In some cases of falsification, there is presence of a wrong or an absence of active pharmaceutical ingredient (API) that is often easy to prove with spectroscopy. However, the major part of the burden is constituted of lower dosed API that is trickier to evaluate in the field with conventional tools. In this study, the potential of handheld Raman spectroscopy to assay API in a solid dosage form was evaluated. For this purpose, fifteen formulations with a various proportion of mannitol and microcrystalline cellulose, seven level of concentration of ibuprofen (14 % – 26 % (m/m)) were produced thanks to a design of experiments, following the ICH Q3 guidelines. The calibration set was realized by analysing 3 tablets per formulation and each tablet was assayed using a previously validated benchtop NIR model. The PLS model was developed using PLS-Toolbox running in a Matlab® environment. The PLS model calibration has shown very nice results, with a R² of calibration / R² of cross-validation of 0.981 / 0.968 and a RMSECV of 0.83% (m/m). As explained, the validation set was projected on model and showed a RMSEP of 0.89 % (m/m). Then, the quantitative model has been validated following the total error approach with 4 series, 5 levels of concentration and 3 replicates. The acceptance limits were set at +/-15 % following the European Pharmacopeia criteria for uniformity of content. In a nutshell, handheld Raman spectrophotometer has shown very interesting results for studied formulation. Thanks to the 14 – 26 % (m/m) range, the model could be applied to get the quantitative information of the dosage of substandard medicines on the field