Root Cause Analysis of Quality Defects Using HPLC-MS Fingerprint Knowledgebase for Batch-to-batch Quality Control of Herbal Drugs.

Introduction: The batch-to-batch quality consistency of herbal drugs has always been an important issue.

Objectives: To propose a methodology for batch-to-batch quality control based on HPLC-MS fingerprints and process knowledgebase.

Methods: The extraction process of Compound E-jiao Oral Liquid was taken as a case study. After establishing the HPLC-MS fingerprint analysis method, the fingerprints of the extract solutions produced under normal and abnormal operation conditions were obtained. Multivariate statistical models were built for fault detection and a discriminant analysis model was built using the probabilistic discriminant partial-least-squares method for fault diagnosis.

Results: Based on multivariate statistical analysis, process knowledge was acquired and the cause-effect relationship between process deviations and quality defects was revealed. The quality defects were detected successfully by multivariate statistical control charts and the type of process deviations were diagnosed correctly by discriminant analysis.

Conclusion: This work has demonstrated the benefits of combining HPLC-MS fingerprints, process knowledge and multivariate analysis for the quality control of herbal drugs.