Discrepancies between validated GC-FID and UHPLC-DAD methods for the analysis of Δ-9-THC and CBD in dried hemp flowers.

Herbal products for smoking containing cannabidiol (CBD) are available as "low-tetrahydrocannabinol cannabis products" in most EU countries. In Belgium, Δ9-tetrahydrocannabinol (THC) content of these products must be less than 0.2% w/w, which is also the limit for agricultural hemp. For agricultural hemp, the official and only valid method for European regulators is gas-chromatography coupled to flame ionization detector (GC-FID). There is no such method, for smoking products. Many of these herbal for smoking products are analyzed as part of their quality control and have certificate of analysis. During surveillance by official labs, discrepancies were seen between the official results and the certificate of analysis. In this study, a GC-FID method based on the European method and an ultra-high-performance liquid chromatography coupled to diode array detection (UHPLC-DAD) method were validated and applied for samples analysis in order to investigate these discrepancies. The GC-FID method shows better results for the validation parameters; notably, it has ß-expectation tolerance limits within 10% with a ß value of 95% while the validated UHPLC-DAD method has ß-expectation tolerance limits within 15% with a ß value of 90%. Furthermore, the other parameters evaluated are generally better with the GC-FID method. The statistic t test shows that the difference between both methods was significantly different for total-THC, but not significantly different for the total-CBD. The authors state that, as for agricultural hemp, the GC-FID method is to be preferred for the analysis of THC and CBD in products for smoking.

Relative response factors and multiple regression models in liquid chromatography to quantify low-dosed components using alternative standards-proof of concept: total Δ9-THC content in cannabis flowers using CBD as reference.

A classical quantitative analysis in liquid chromatography is performed using either a one-point calibration or a calibration line, prepared using a reference standard of the compound(s) of interest. However, in some cases, adequate reference standards may be very expensive, rare to obtain, or have limited shelf-life properties. Also, in herbal matrices, multiple compounds could be necessary to be quantified, needing a whole series of different (related) reference standards. In these cases, the use of relative response (sometimes called relative correction factors) factors (RRFs) towards reference standards, different of the compound to be quantified, gained attraction. This study performed a comparison of the use of RRFs and linear relative response factor models (LRRFM) for the quantification of targeted low-dosed compounds using an alternative standard, since it is known that classical RRFs often fail in lower concentration ranges. For this purpose, the determination of the total Δ9-tetrahydrocannabinol (Δ9-THC + Δ9-THC-A) content in dried cannabis flowers, using UHPLC-DAD, was used as a case study. A chromatographic method was implemented and validated, and the use of classical calibration lines, classical RRF, and the LRRFM was applied and compared, with special focus on the concentration around 0.2% (w/w) total Δ9-THC, the legal limit (in most European countries) in these products. Results showed that the newly presented and validated LRRFM approach outperformed the classical RRFs, especially in the low-concentration ranges and that concentrations obtained with the LRRFM were in accordance with the interpolation results obtained with a calibration line.