Zhao Xuejia, Wang Ning, Zhu Minghui, Qiu Xiaodan, Sun Shengnan, Liu Yitong, Zhao Ting, Yao Jing, Shan Guangzhi
Institute of Medicinal Biotechnology, Chinese Academy of Medical Sciences & Peking Union Medical College, No. 1, Tian Tan Xi Li, Beijing 100050, China.
College of Life Science and Technology, Beijing University of Chemical Technology, North Third Ring Road 15, Beijing 100029, China.
Molecules. 2022 Mar 5;27(5):1707. doi: 10.3390/molecules27051707.
In recent years, transmission Raman spectroscopy (TRS) has emerged as a potent new tool for rapid, nondestructive quantitation in pharmaceutical manufacturing. In order to expand the applicability of TRS and enhance its use in product quality monitoring during drug production, we aimed, in the present study, to apply partial least-squares (PLS) approaches to build a model consisting of 150 handmade tablets and covering 15 levels through the use of a multifactor orthogonal design of experiment (DOE), which was used to predict concentrations of validation tablets made by hand. The difference between results according to HPLC and TRS were negligible. The model was used to predict the active pharmaceutical ingredient (API) content in four random commercial paracetamol tablets, and corrected with the spectra of the commercial tablets to obtain four corresponding models. The results show that the content relative error in the model's predictions after correction with commercially available tablets was significantly lower than that before correction. The corrected model was used to make predictions for 20 tablets from the brand Panadol. Compared with the HPLC results, the prediction relative error was basically less than 4.00%, and the relative standard deviation (RSD) of the content was 0.86%.
近年来，透射拉曼光谱（TRS）已成为制药生产中快速、无损定量分析的一种强大新工具。为了扩大TRS的适用性并加强其在药品生产过程中产品质量监测方面的应用，在本研究中，我们旨在应用偏最小二乘法（PLS）方法，通过使用多因素正交实验设计（DOE）构建一个由150片手工压制片剂组成、涵盖15个水平的模型，该模型用于预测手工制作的验证片剂的浓度。根据高效液相色谱法（HPLC）和TRS得出的结果之间差异可忽略不计。该模型用于预测四种随机市售对乙酰氨基酚片剂中的活性药物成分（API）含量，并通过市售片剂的光谱进行校正以获得四个相应模型。结果表明，用市售片剂校正后模型预测中的含量相对误差显著低于校正前。校正后的模型用于对品牌为必理通的20片片剂进行预测。与HPLC结果相比，预测相对误差基本小于4.00%，含量的相对标准偏差（RSD）为0.86%。
