Studies On Consistency Evaluation Methods Of Traditional Chinese Medicinal Quality And Manufacturing Process

With the progress of modernization and internationalization of traditional Chinese medicine (TCM), both the national and international regulatory institutions have required more and more strict quality control of TCM. To improve the quality consistency of finished products, the systematic methodology for the batch-to-batch consistency evaluation of the quality and manufacturing process of TCM needs to be established. In recent years, an increasing number of modem analytical techniques have been used for the measurement of TCM raw materials, in-process materials and final products, resulting in the significant increase of data complexity and size. However, the corresponding data analysis technique is backward, and cannot satisfy the analytical demands of large amounts of high-dimensional data. As a result, it has become the bottleneck in the establishment of the consistency evaluation methods. For this reason, multivariate data analysis was combined with high-performance liquid chromatography (HPLC), chromatographic fingerprint and near infrared (NIR) spectroscopy to explore the consistency evaluation methods of the quality and manufacturing process of TCM. The main contents and achievements of this dissertation are summarized as follows:1. The method for fast quality consistency evaluation of the raw herbal materials based on NIR spectroscopy was presented. A kind of herbal material, Radix Salvia miltiorrhiza (Danshen) was adopted as the application example. The effects of the sieve size and drying time on spectral variability were studied, and these two operating parameters were optimized. The spectra of samples from different sources and of that cultivated in different years were collected, and quality consistency evaluation was achieved using multivariate statistical analysis. This study provides a fast statistical evaluation method for large amounts of herbal materials during TCM industrial production.2. To overcome the inapplicability of both single stage and univariate control method during the manufacture of TCM, a new approach, multistage multivariate statistical quality control (MMSQC), was presented. The manufacturing process of Danshen injection was adopted as the application example. Based on the data of multiple quality parameters of multistage intermediates, the multivariate control charts were established using multi-way principal component analysis (MPCA), and were used to monitor the batch-to-batch consistency of TCM manufacturing process.3. The batch-to-batch consistency evaluation method of TCM finished products based on multivariate statistical analysis was established. A peak weighting algorithm of chromatographic fingerprint was proposed, and was combined with multivariate statistical analysis. The statistics of Hotelling T2and DModX were used as the consistency criterion. The method overcame the drawbacks of the regular fingerprint similarity calculation about the weight distribution among characteristic peaks, the acquirement of the reference fingerprint, the set of similarity thresholds, and diagnosis on abnormal variations. This research provides a new statistical method for the chromatographic fingerprint-based quality evaluation of TCM.4. The real-time monitoring method based on spectroscopy for batch-to-batch consistency of the TCM manufacturing process was established. In-line NIR spectroscopy was combined with multi-way partial least squares (MPLS), and the multivariate statistical process control (MSPC) model was established to monitor the liquid-liquid extraction of Flos Lonicera japonica in real time. This approach provides the opportunities to achieve early fault diagnosis and adjustments, and is helpful for improving process consistency.5. The batch-to-batch consistency evaluation method of the TCM manufacturing process based on chromatographic fingerprint was established. The preparation of Sheng-mai-yin (SMY) oral liquid was selected as the typical example, and multivariate statistical analysis was applied on the chromatographic fingerprint data of finished products. The possible causes responsible for quality abnormality were analyzed according to the variations of characteristic peak areas. This approach can help diagnose the abnormal variations of herbal materials and operating parameters in time, and can provide guidance on adjustments of subsequent batches.