Determination of Indomethacin Crystallinity in the Presence of Excipients Using Diffuse Reflectance Near-Infrared Spectroscopy

• Published in: Pharmaceutical development and technology Vol. 6; no. 4; pp. 573 - 582
• Main Authors: Seyer, Jeffery J., Luner, Paul E.
• Format: Journal Article
• Language: English
• Published: New York, NY Informa UK Ltd 01-01-2001; Taylor & Francis; Informa Healthcare
• Subjects: Anti-Inflammatory Agents, Non-Steroidal - analysis; Biological and medical sciences; Calibration; Cellulose - analysis; Crystallization; Excipients - analysis; General pharmacology; Indomethacin; Indomethacin - analysis; Lactose - analysis; Medical sciences; Multi-component; Near-infrared spectroscopy; Percent crystallinity; Pharmaceutical technology. Pharmaceutical industry; Pharmacology. Drug treatments; Sodium Chloride - analysis; Spectroscopy, Near-Infrared - methods; Amorphous material; Drug adjuvant interaction; Pharmaceutical technology; Indometacin; Drug adjuvant; Crystallinity; Crystalline state; Near infrared spectrometry; Investigation method; Analytical method; Indolacetic acid derivatives; Quantitative analysis; Physical structure
• ISSN: 1083-7450; 1097-9867
• DOI: 10.1081/PDT-120000295

Studies were conducted to investigate the use of near-infrared spectroscopy for determining the crystallinity of indomethacin in multi-component physical mixtures. Three calibration sets of amorphous crystalline indomethacin physical mixtures were prepared over the composition range of 0-100% crystallinity. Each of the three calibration sets was diluted step-wise with increasing amounts of a single excipient (Avicel, α-lactose monohydrate, or sodium chloride). Near-infrared spectra were obtained after each round of dilutions using diffuse reflectance sampling on samples contained in glass vials. After a second derivative transformation, standard curves were constructed by plotting percent indomethacin crystallinity against the ratio of responses at two wavelengths. At dilution levels up to 75% Avicel or lactose, the calibration models demonstrated high coefficients of determination and low standard errors. Dilution with sodium chloride did not produce comparable results and it was necessary to use partial least-squares regression to achieve a similar level of error. These findings were confirmed with separate validation sets. An investigation of instrument error showed that the impact of instrument variability on quantification generally increased as a function of the dilution level.