In the pharmaceutical industry quality control is one of the most time-consuming activities in the production workflow. Kent T Møller, Magnus Tolleshaug and Thomas W. Jensen report on the use of near infrared spectroscopy for quality control analysis.
Near infrared spectroscopy (NIR) is suited for quality control analysis of tablets because the analysis is fast, easy to use and requires no sample preparation. In this study a NIR calibration model was developed based on an UV spectroscopic reference method and used for prediction of the content of paracetamol in intact 500-mg tablets.
NIR transmission spectra were measured using an ABB Bomem MB160 FT-NIR spectrometer equipped with a Tablet Samplir accessory. The equipment is shown in Fig. 1. Tablets were simply placed with their flat side up on the sample tray of the accessory, and their FT-NIR spectra were recorded.
The equipment was controlled by the AIRS Pharmaceutical software (ABB Bomem) in the BGRAMS/32 (Galactic Industries Corporation) environment. Calibrations were developed using PLSplus/IQ (Galactic Industries Corporation) and Unscrambler 7.6 (Camo ASA).
A Perkin Elmer Lambda 12 UV/VIS spectrometer was used for the UV-spectroscopic reference method.
The samples studied were commercially available 500 mg tablets from Nycomed Pharma. Tablets weighed 600 mg and consisted of 83 per cent m/m paracetamol. Thirty batches from normal production (98100 per cent m/m paracetamol) and spanning production variations were used for this study. Eight batches of laboratory-made tablets were added to the set in order to span a range of 80105 per cent m/m paracetamol.
Method development and validation
The calibration model included 10 production batches and eight laboratory batches and the remaining 20 production batches were used for analytical validation of the model. A Savitsky-Golay 1st derivative (11 points, 2nd polynomial order fit) followed by a normalisation were applied to the spectra prior to calibration (Fig. 2). A PLS calibration model was developed using two spectral regions and the reference values obtained by the UV reference method. The result of the segmented cross validation is presented in Fig. 3. A two-principal component model gave a correlation of 0.995 and a RMSEP of 4.7 mg/g.
The content of paracetamol determined by the reference and NIR method was compared for 20 production batches (10 tablets per batch). The full lines on Fig. 4 indicate the target and the dashed lines indicate the precision of the UV reference method. The RSEP (per centage) was calculated to 0.4 per cent.
Both methods are considered in close agreement, the offset between the UV and FT-NIR results being within the precision of the reference method.
In order to evaluate the repeatability of the method, paracetamol content was determined six times on 10 tablets from the same batch. Whether the tests were performed by one or more analysts, or on different days, the RSD stayed constant at
0.1 per cent.
The results obtained in this work reveal that the direct analysis of intact tablets by near infrared transmittance spectroscopy is an effective way of quantifying paracetamol in 500 mg tablets. The method was validated according to the recommendations of the International Conference of Harmonization, ICH guidelines. The validation shows that the NIR method is accurate and precise. The Norwegian regulatory authorities approved the NIR method described in this study in 2001.
Although resources are needed for developing and updating calibration models, the cost of NIR analysis is very small compared to UV and HPLC analysis. Furthermore, the lead-time of the quantitative analysis can be reduced from 48 hours to about 15 minutes with NIR. Because the NIR apparatus is robust and the measurement requires no laboratory facilities, the analysis can easily be integrated into the production plant for at-line determination. Thereby the pharmaceutical industry could move a step closer towards parametric release determinations.
References: Eustaquio et al. Analytica Chimica Acta, 383 (1999), 283-290; Moffat et al. Analyst,
125 (2000), 1341-1351; Trafford et Al. Analyst,
124 (1999), 163 -167; Text of validation of analytical procedures. ICH. October 1994;
Validation of Analytical Procedure: Methodology. ICH. November 1996.
Acknowledgements: The authors would like to thank Project Leader Lis Jacobsen and all the members of the NIR Quantitative project team in Nycomed Pharma. We would also like to thank
Q-Interline, Denmark for inspiration and support. u
ENQUIRY No 46
Kent T Møller is Manager NIR Department, Magnus Tolleshaug and Thomas W Jensen are research scientists with Nycomed Pharma, Roskilde, Denmark. www.nycomed.dk
