Study On Rapid Alternative Test Methods Of Lyophilized Intermediate Content

Freeze dried technology made the liquid samples into solid preparation, adding the appropriate medium for use, which can increase the stability of the product. This technology is favorable for storage and transport. According to the concept of the2010version of GMP quality by design, quality control of the lyophilized preparation intermediates was crucial to the product quality assurance. According to the requirements of related regulations, the lyophilized preparation needs to be tested in accordance with the Chinese pharmacopoeia or other statutory standard, while intermediate content control can make internal standards according to the specific circumstances of each unit, which provides the application of near infrared spectroscopy in the lyophilized product intermediates content feasible guarantee determination.Near infrared radiation (NIR) is a kind of electromagnetic wave between visible light (VIS) and middle infrared(MIR). NIR regions defined in the ASTM wavelength range is780-2526nm, we also separate NIR in two regions in habits, which are the short wave NIR(780-1100nm) and long wave NIR(1100-2526nm). Fast analysis of near infrared spectroscopy, without destroying the sample, which can be applied to determination of the characteristics of online. Freeze-dried products determination of intermediates can save analysis time and production cost, improving labor productivity.In this research we used the conventional HPLC method for the determination of the content of freeze dried powder, and then tested the NIR spectra parameters. We set up the relationship between the spectra parameter and the content, that is to say the standard curve. After repeated tests, we established the mature model to replace the HPLC method. The main research contents are as follows:1Near infrared spectroscopy analysis technique for rapid detection of the content of intermediate of doxofylline for injection.This study selected intermediate of doxofylline for injection as research object, collecting53samples, from which selected30samples as calibration set,15samples as validation set. In the PCA chart, the validation set samples all included in the calibration set ones, which was accord to the ideal set classification. The variance contribution of the first three principal components reached92.93%. Selecting the SG15point smooth plus first derivative as a pretreatment method, we used iPLS, correlation coefficient method and the SPA method to choose the characteristic bands. Models under the selected wavelength was set up, and then comparing the results with that of full spectrum model. In the iPLS model, selecting5928-6310cm-1and7471-7853cm-1as spectral range, the first-order derivative SG15point smooth as preprocessing method, we selected3principal fractional to set up PLS model. In the resulting model R2=0.998, RMSECV=0.40129, RMSEP=0.36604. SEP/SEC=1.01<1.2, which show that the model can be accepted, RPD=22.3>8, which illustrated the high accuracy prediction model.2Near infrared spectroscopy analysis technique for rapid detection of the content of intermediate of puerarin for injection.This study selected intermediate of puerarin for injection as research object, collecting46samples, from which selected31samples as calibration set,15samples as validation set. In the PCA chart the validation set samples was uniformly dispersed in the calibration set ones, which was accord to the ideal set classification. Selecting the SG15point smooth method as the pretreatment method, we used iPLS, variance analysis, Manual selection and the GA method to choose the characteristic bands. Models under the selected wavelength was set up, and then comparing the results with that of full spectrum model. In the iPLS model, selecting5967-6079cm-1,7124-7236cm-1and7587-7698cm-1as spectral range. In the resulting model RMSECV=0.3234, RMSEP=0.3407, which was lower than that of the full spectrum. The model was better than that of the full spectrum and other models with different variables, RPD=33.3>8, which illustrate the high accuracy prediction model.3Near infrared spectroscopy analysis technique for rapid detection of the content of intermediate of mezlocillin sodium for injection.This study selected intermediate of mezlocillin sodium for injection as research object, collecting45samples, from which selected30samples as calibration set,15samples as validation set. In the PCA chart, the validation set samples all included in the calibration set ones, which was accord to the ideal set classification. Selecting the SG5point smooth as a pretreatment method, we used iPLS, correlation coefficient method and the SPA method to choose the characteristic bands. Models under the selected wavelength is set up, and then comparing the results with that of full spectrum model. In the CARS-iPLS model,23variables were selected, RMSECV=1.81594, RMSEP=2.05057, which was lower than that of the full spectrum. The model was better than that of the full spectrum and other models with different variables, RPD=22.3>8, which illustrated the high accuracy prediction model.