Quantitation of pharmaceutical formulations and monitoring of pharmaceutical processes using process analytical technology techniques: Near infrared and Raman spectroscopy.

摘要

This research was aimed at contributing to the Food and Drug Administration's process analytical technology (PAT) initiative that is intended to monitor pharmaceutical processes and pharmaceutical formulations ensuring the safety and efficacy of the end product. PAT tools of near infrared (NIR) and Raman spectroscopy have been adapted to aid in this study. Continuous processes of hot melt extrusion and hydrogenation reactions were monitored using these tools. Hot melt extruded films were monitored for the active ingredient both offline and online using NIR/Raman spectroscopy. The amount of dexamethasone and theophylline was determined in immediate release pellet formulations using NIR spectroscopic method. Cannabis sativa matrices were investigated for the amount of Delta9-tetrahydrocannabinol (THC) and also quantitation of THC was performed in hot-melt films using NIR spectroscopy.;The first objective of this research was to demonstrate the utility of NIR spectroscopy for quantitative analysis of a model drug in a two-component system of hot-melt extruded (HME) film formulations. The NIR method developed resulted in an assayed clotrimazole amount in the film matrix to be with in 3.5 % of the quantity determined by the reference method. These studies clearly demonstrate that NIR spectroscopy is a powerful tool for the quantitation of active drug substances contained in films produced by HME.;The other formulation category that was investigated to assess the utility of these PAT tools was immediate release pellet formulations. NIR spectroscopy was utilized for the quantitative and qualitative analysis of an active ingredient in pellet formulations. Theophylline and Dexamethasone were used as the active pharmaceutical ingredients (APIs) in the pellet formulations. In both the cases, pellets with contents varying from 2% to 20% of the respective API were prepared. The prediction error ranged from--0.29 to +0.60% for dexamethasone and -0.56 to +0.63% for theophylline. It can be inferred from the narrow window of the % error in both the cases that NIR can be a good alternative to UV method, which is a time-consuming means of laboratory analyses for these types of formulations.;It has been a challenge to investigate the utility of NIR spectroscopy for the quantitative analysis of an active like THC in dry marijuana plant leaves. Marijuana consists of about 400 other cannabinoids along with THC. This study was aimed at the applicability of NIR in a multi-component matrix. Samples from different lots of cannabis sativa containing 3.74 -- 16.94% THC were used for the calibration model. NIR spectra of the marijuana samples from 10000 -- 4000 cm-1 were obtained using a Bruker FT-IR spectrophotometer, which was also used in previous experiments. Chemometrics was applied using OPUS/QUANT2 version 2.2 software and partial least squares algorithm was used for the calibration models. A sample set of 55 was used as the training set and 50 samples were used for validation. All the samples used for both of calibration and the validation set were assayed using the standard gas chromatography (GC) reference method. The linear regression of the final calibration model yielded a R2 of 0.97 with a root mean square error of 0.75. No spectral preprocessing was performed for this model. The root mean square error of prediction was found to be 2.01. The amount of THC predicted in the cannabis sativa samples was determined to be within 5% of the value obtained using the GC reference method. This research was then extended to determine the amount of THC in film dosage forms. Solid dispersions of THC-PEO containing 0 -- 15% w/w of the drug were punched into circular polymeric matrices utilizing a hot-melt punch method. NIR spectra were obtained for all of these formulations in the entire range of 12,500 -- 4000 wavenumbers. Calibration set consisted of 0, 5, 10 and 15% THC films and the test set comprised of 2.5, 7.5 and 12.5% THC films. A very good correlation was found between the amount of THC estimated in these films using NIR model and the value obtained using HPLC analysis. (Abstract shortened by UMI.)