| Titre : | EVALUATING DIFFERENT ASPECTS OF MATRIX EFFECT ON PHARMACEUTICAL DRUG ANALYSIS THROUGH THE ASSOCIATION OF VIBRATIONAL SPECTROSCOPY TECHNIQUES TO CHEMOMETRIC TOOLS | | Type de document : | thèse | | Auteurs : | Mohammed ALAOUI MANSOURI, Auteur | | Année de publication : | 2021 | | Langues : | Anglais (eng) | | Mots-clés : | Matrix effect Vibrational spectroscopy chemometrics pharmaceutical drug analysis Effet de matrice Chimiométrie spectroscopie vibrationnelle analyses pharmaceutiques التحليلات الصيدلانية المقاربات الكيميائية التحليل الطيفي تأثير السواغات | | Résumé : | The aim of this thesis was to investigate the ability of the association of vibrational spectroscopic techniques to chemometric tools in order to overcome three different aspects of matrix and interference effects for qualitative and quantitative pharmaceutical drug analysis. The first aspect was focused on testing the ability of applying Partial Least Squares-Discriminant Analysis (PLS-DA) to each of Fourier transform infrared spectroscopy (FT-IR) and Fourier transform near infrared spectroscopy (FT-NIR) to discriminate between different pharmaceutical products based on three main polymorphic forms of the fluconazole. PLS-DA showed its ability to discriminate between samples that have only one of three polymorphic forms of fluconazole. While in case of the matrix effect, which is represented in the existence of two polymorphic in the same sample, PLS-DA showed its limit to do the right discrimination. Thus, another approach is known by Hotelling’s T² and Q residuals joined PLS-DA to detect these samples with two polymorphic forms as outliers. The second aspect of matrix effect focused mainly on evaluating the application each of PLS regression and Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) model on FT-NIR data to quantify ciprofloxacin in different brands of pharmaceutical products. When it comes to quantify ciprofloxacin in samples with the same composition, the obtained results based on Root Mean Squares Error of Prediction (RMSEP) and Relative Errors (RE) proved that both of regression approaches of PLS regression and MCR-ALS are able to carry out the quantitation of the drug substance accurately. However, the quantitation of ciprofloxacin whether in mixtures of different matrix composition or in different brands of pharmaceutical products clearly showed the limit of PLS regression because of the matrix effect, whereas the MCR-ALS based on its low relative and prediction errors, has shown its ability to overcome the problem of the change in the matrix composition due to its second order advantage. The third and last aspect of interference effect aims to investigate a special property of Raman spectroscopy to quantify the ibuprofen in a ternary mixture through an interference polypropylene container. This property of Raman spectroscopy is known by Spatially Offset Raman Scattering (SORS), which allows the monochromatic laser to pass through the packaging and obtain the spectrum of what is inside the container. This investigation was carried out through a comparison study between backscattering (conventional) and SORS mode to quantify ibuprofen through the interference of polypropylene packaging by evaluating the PLS regression models with mean of accuracy profiles following the ICH Q2 (R1) guidelines on validation with ±15% as acceptance limits. Based on the obtained accuracy profiles, the SORS mode demonstrated its ability to quantify accurately through the interference which is not in case of backscattering mode.
Keywords: Matrix effect; Polymorphism; PLS-DA; Hotelling’s T² and Q residuals; PLS regression, MCR-ALS; SORS; FT-IR; FT-NIR; Raman spectroscopy | | Numéro (Thèse ou Mémoire) : | D0262021 | | Président : | Yahia CHERRAH ; Philippe Hubert | | Directeur : | Abdelaziz Bouklouze ; Roland Marini | | Juge : | BOUATIA MUSTAPHA | | Juge : | Mohamadine ElMrabet | | Juge : | Khalid Digua ; Eric Ziemons |
EVALUATING DIFFERENT ASPECTS OF MATRIX EFFECT ON PHARMACEUTICAL DRUG ANALYSIS THROUGH THE ASSOCIATION OF VIBRATIONAL SPECTROSCOPY TECHNIQUES TO CHEMOMETRIC TOOLS [thèse] / Mohammed ALAOUI MANSOURI, Auteur . - 2021. Langues : Anglais ( eng) | Mots-clés : | Matrix effect Vibrational spectroscopy chemometrics pharmaceutical drug analysis Effet de matrice Chimiométrie spectroscopie vibrationnelle analyses pharmaceutiques التحليلات الصيدلانية المقاربات الكيميائية التحليل الطيفي تأثير السواغات | | Résumé : | The aim of this thesis was to investigate the ability of the association of vibrational spectroscopic techniques to chemometric tools in order to overcome three different aspects of matrix and interference effects for qualitative and quantitative pharmaceutical drug analysis. The first aspect was focused on testing the ability of applying Partial Least Squares-Discriminant Analysis (PLS-DA) to each of Fourier transform infrared spectroscopy (FT-IR) and Fourier transform near infrared spectroscopy (FT-NIR) to discriminate between different pharmaceutical products based on three main polymorphic forms of the fluconazole. PLS-DA showed its ability to discriminate between samples that have only one of three polymorphic forms of fluconazole. While in case of the matrix effect, which is represented in the existence of two polymorphic in the same sample, PLS-DA showed its limit to do the right discrimination. Thus, another approach is known by Hotelling’s T² and Q residuals joined PLS-DA to detect these samples with two polymorphic forms as outliers. The second aspect of matrix effect focused mainly on evaluating the application each of PLS regression and Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) model on FT-NIR data to quantify ciprofloxacin in different brands of pharmaceutical products. When it comes to quantify ciprofloxacin in samples with the same composition, the obtained results based on Root Mean Squares Error of Prediction (RMSEP) and Relative Errors (RE) proved that both of regression approaches of PLS regression and MCR-ALS are able to carry out the quantitation of the drug substance accurately. However, the quantitation of ciprofloxacin whether in mixtures of different matrix composition or in different brands of pharmaceutical products clearly showed the limit of PLS regression because of the matrix effect, whereas the MCR-ALS based on its low relative and prediction errors, has shown its ability to overcome the problem of the change in the matrix composition due to its second order advantage. The third and last aspect of interference effect aims to investigate a special property of Raman spectroscopy to quantify the ibuprofen in a ternary mixture through an interference polypropylene container. This property of Raman spectroscopy is known by Spatially Offset Raman Scattering (SORS), which allows the monochromatic laser to pass through the packaging and obtain the spectrum of what is inside the container. This investigation was carried out through a comparison study between backscattering (conventional) and SORS mode to quantify ibuprofen through the interference of polypropylene packaging by evaluating the PLS regression models with mean of accuracy profiles following the ICH Q2 (R1) guidelines on validation with ±15% as acceptance limits. Based on the obtained accuracy profiles, the SORS mode demonstrated its ability to quantify accurately through the interference which is not in case of backscattering mode.
Keywords: Matrix effect; Polymorphism; PLS-DA; Hotelling’s T² and Q residuals; PLS regression, MCR-ALS; SORS; FT-IR; FT-NIR; Raman spectroscopy | | Numéro (Thèse ou Mémoire) : | D0262021 | | Président : | Yahia CHERRAH ; Philippe Hubert | | Directeur : | Abdelaziz Bouklouze ; Roland Marini | | Juge : | BOUATIA MUSTAPHA | | Juge : | Mohamadine ElMrabet | | Juge : | Khalid Digua ; Eric Ziemons |
|
Accueil
Affiner la recherche Faire une suggestionEVALUATING DIFFERENT ASPECTS OF MATRIX EFFECT ON PHARMACEUTICAL DRUG ANALYSIS THROUGH THE ASSOCIATION OF VIBRATIONAL SPECTROSCOPY TECHNIQUES TO CHEMOMETRIC TOOLS / Mohammed ALAOUI MANSOURI
