Quality control of Lonicerae Japonicae Flos using near infrared spectroscopy and chemometrics.

A set of qualitative and quantitative methods based on near infrared (NIR) spectroscopy was established for the geographical origin identification, quantitative determination of active compounds, and fingerprint analysis of Lonicerae Japonicae Flos. The spectra of 140 samples from different origins were collected with two NIR instruments from different manufacturers (Thermo Scientific and Buchi). A Soft Independent Modeling of Class Analogy model was established for the identification of Lonicerae Japonicae Flos from the genuine producing area. Using the established discriminant analysis model, 22 samples from Henan province were predicted with 100% rate of accuracy, while the 68 samples from other producing areas were predicted with 9 samples incorrectly judged. Futhermore, partial least square regression method was used for developing the quantitative calibration models with the reference values determined with a validated HPLC-UV method. The RMSEP values of external validation are 0.169%, 0.048%, 0.172%, 0.007%, 0.203%, and 0.066% for NCA, CA, CfA, 3,4-DCA, 3,5-DCA, and 4,5-DCA, respectively, which indicated that the established models possess satisfactory predictive abilities. In addition, the models established on the primary instrument can also be transferred to the secondary instrument using direct standardization algorithm, which enlarged the application scope of the established models. Since the NIR spectra can reflect the comprehensive quality information of Lonicerae Japonicae Flos, a fingerprint analysis method was finally proposed for the quality consistency check of raw materials. The proposed methods are rapid and effective, and possess good portability, which is helpful to the quality control of Lonicerae Japonicae Flos.