EVALUATION OF SPATIAL DISTRIBUTION AND PATTERNS OF TWO BENZNIDAZOLE FORMULATIONS USING NIR IMAGING

Benznidazole (BNZ) is an antiprotozoal drug used for the treatment of Chagas disease, a neglected disease endemic to Latin America. The drug has been administered by immediate release tablets, but extended release tablets can reduce the number of administrations, the toxicity and, consequently, increase the patient adherence to treatment. Two different tablets formulation for extended release were developed, only differing by the kind of hydroxypropyl methylcellulose (HPMC) used: K4M (viscosity of 4000 mPa.s) or K100M (viscosity of 100000 mPa.s). Since the distribution of the excipients is an important parameter for the modulation of the drug release, the objective of this work was to evaluate the spatial distribution of the formulation compounds and possible differences in the two formulations, using NIR hyperspectral images. The images were acquired using a SisuCHEMA (Specim), in the spectral range 1000-2500 nm; spectral resolution of 10 nm and pixel size of 30x30. Multiplicative Signal Correction (MSC) was used as preprocessing technique.The imaging analysis was made using 20 tablets (10 of each kind of HPMC used) and the isolated raw materials: BNZ, HPMC K4M or K100M, lactose, polyvinylpyrrolidone (PVP) and magnesium stearate. Multivariate Curve Resolution Alternating Least Squares (MCR-ALS) was employed to achieve the chemical distribution maps (CDM). As initial estimates, the pure spectra of the excipients were used. With the CDM, multivariate image analysis (MIA) was performed to acquire score images and loadings. PCA was also applied to the feature vectors extracted using different Image Feature Extraction (IFE) techniques (Singular value decomposition-SVD, Soft Color Texture Descriptor - SCTD and discrete wavelet transform – DWT). The CDMs did not show differences concerning both formulations, due to the similarity in their relative concentrations. When MIA was performed on the CDMs, a similar pattern for chemical internal structure was achieved for all samples. PC1 showed lactose information (major excipient, positive signal) and information of BNZ and HPMC (negative signal). PC2 provided information about the major excipients (HPMC and lactose). PC3 show a mixing pattern for HPMC and lactose. Those results showed a mixing of BNZ with the two major excipients, which can be useful to evaluate the results of the mixing step, critical for an extended release tablet. PCA was also performed with the feature vectors. The result was that the tablets of the two formulations were separated in PC1. All methods (DWT, SCTD and SVD) presented suitable results for identifying both formulations. The samples of the HPMC K100M formulation were more grouped in the PC1xPC2 plot, showing lower variability. The chemometric techniques used were suitable to evaluate spatial chemical distribution, the internal chemical correlation and to identify both formulations, improving understanding of the process and the development of the formula.