Abstract:The quality control of the Traditional Chinese medicine (TCM) are important links in the production and application of pharmaceuticals. In this study, the quantitative analysis of rhubarb was realized by near-infrared spectroscopy (NIRs), including the rapid detection of free anthraquinone and total anthraquinone content. Combined with method validation of NIR quantitative model and release limit, a reliable real-time release testing (RTRT) method for rhubarb was constructed. The competitive adaptive weighted resampling (CARS) method was applied for characteristic variables selection, and the quantitative model was established based on partial least squares regression (PLSR) method and particle swarm optimization based least square support vector machines (PSO-LSSVM) method. The relative standard error of prediction (RSEP) values of free anthraquinone and total anthraquinone models were 10.66% and 4.95%, respectively. The accuracy profile (AP) was introduced to validate and evaluate the performance of the optimized model at different concentration levels. The relative bias, precision and linearly of the two quantitative models were all within the acceptable range. Based on the results of the method validation, the minimum release control limit for total anthraquinone content was set as 1.569% to make sure accurate release, slightly higher than the pharmacopoeia standard. The constructed RTRT system for rhubarb based on the NIR model validation can improve the efficiency and accuracy of quality control.

Quantitative Real-Time Release Testing of Rhubarb Based on Near-Infrared Spectroscopy and Method Validation