Profiling of illicit cannabis samples using high performance liquid chromatography and multivariate analysis

Abstract

Drug profiling is an important aspect of drug analysis. Physical and chemical properties of a seized drug samples can be accumulated, and provides intelligence information to combat drug trafficking and abuse. The purpose of this study was to profile drugs, especially illicit cannabis samples using high performance liquid chromatography (HPLC) and multivariate analysis. Herbal cannabis samples were extracted using methanol- chloroform mixture in 9:1 ratio. HPLC separation employed Onyx Monolithic C18 column. Mobile phase consisting of methanol-water (75:25) was used as eluent at a flow rate of 0.8 mL/min and analytes detected at 220 nm. Three major cannabinoids: cannabinoid (CBD), cannabinol (CBN) and tetrahydrocannbinol (THC) were separated within twenty five minutes. Peak area of the three cannabinoids obtained from HPLC analysis were then further analyzed using three types of multivariate analyses, which are Principal Component Analysis (PCA), Cluster analysis and Soft Independent Modeling Class Analogy (SIMCA) analysis. All multivariate analyses were performed using Unscrambler X 10.0. Results from these three multivariate analyses were compared to find which is more suited for profiling of illicit cannabis samples. HPLC profiling produced four groupings among fifty six cannabis extracts, these four groups were further sub-grouped by PCA into fifteen groups. Out of these fifteen groups, four main clusters were observed. Cluster analysis was also performed to confirm findings of PCA; results from Cluster analysis also revealed four clusters among the cannabis extracts. SIMCA was unsuccessful in profiling cannabis compared with PCA and Cluster analysis, since it produced different classes among the cannabis extracts. Hence, illicit cannabis samples were successfully profiled using PCA and cluster analysis to reveal the samples originating from different origins.