Use this URL to cite or link to this record in EThOS:
Title: Near-infrared spectroscopic assay transfer for active pharmaceutical ingredients in intact tablets
Author: Smith, Mark Richard
Awarding Body: University College London
Current Institution: University College London (University of London)
Date of Award: 2005
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
The transferability of quantitative near-infrared (NIR) spectroscopic methods is not well understood, though instrumental differences are considered to be the limiting factor. This gap in current knowledge questions the robustness and validity of such methods. Development at a central laboratory using a master instrument, with subsequent transfer to further testing sites or on-line applications must be demonstrated to ensure the progression of NIR spectroscopy as an analytical technique within the pharmaceutical industry. The aim of this research was to address calibration transfer and in doing so generate a set of best practice guidelines. To understand the intricacies involved a number of increasingly challenging assays were utilised, starting with a batch assay for paracetamol (84.2% m/m) between equivalent reflection instruments, progressing to a single tablet assay for piroxicam (37.7% m/m) between equivalent transmission instruments and finally a single tablet assay for atorvastatin (6.6% m/m) between optically diverse transmission instruments (off-line to on-line configurations and dispersive to Fourier transform technologies). Direct transfer (i.e. using the assay developed on one instrument directly for samples measured on a second instrument) provided inaccurate results. Different correction procedures and criteria were therefore investigated for this purpose. The mean sample residual spectrum correction (a spectral bias correction) was suitable where there was minimal difference on the wavelength scale between instruments. Further optimisation of the calibration model using a novel variable selection algorithm was necessary where such differences were present. The potential of qualitative approaches (e.g. principal components analysis in conjunction with 95% equal frequency ellipses) to gauge the outcome of calibration model transfer and for determining the samples required to be scanned on the second instrument was also demonstrated. Knowledge distilled from successive projects was exploited to generate and refine a set of transfer guidelines, algorithms and validation criteria, applicable for different instruments and pharmaceutical products.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available