Featured Article
The legalization of medicinal and recreational cannabis and cannabinoid products by states in the U.S. and Canada has created the need for safety, quality, and regulatory compliance testing prior to retail distribution. Each individual state that has some form of legalization has enacted regulations pertaining to safety and compliance testing. In Canada, medicinal cannabis has been regulated at the federal level for more than a decade. Recreational cannabis legalization is expected to occur in Canada soon and similarly will be regulated at the federal level.
For both states where either medicinal or recreational cannabis is legalized and Canada, chemical classes commonly measured in cannabis and related products include: psychoactive Δ9-tetrahydrocannabinol (THC) and other cannabinoids, terpenes, volatile solvents commonly used in manufacturing processes, residual pesticides, toxic metals, and mycotoxins. Other common tests include screening for microbial contamination, water activity, and moisture content.
Although there is a general commonality to the testing, each individual U.S. state and Canada has a unique set of target chemicals in each category and action levels for which these chemicals cannot exceed. Proper chemical analysis of these chemotypes requires a suite of analytical systems ranging from high-pressure liquid chromatography (HPLC) with ultraviolet (UV) detection to more sophisticated liquid and gas chromatography triple-quadrupole mass spectrometry systems (LC-MS/MS and GC-MS/MS).
Potency testing measures THC and other cannabinoids and is always required. Terpene profiling, although not required by every jurisdiction, provides information about the cultivar and the organoleptic properties of the product. Unless the measured potency level is out of specification with the product label or mandated level, terpenes and potency measurements will not remove a product from the retail sales stream. In contrast, contamination with pesticides, heavy metals, certain microbes, mycotoxins, or residual solvents can result in the failure of an entire product lot at substantial costs to growers and producers.
Of all the mandated safety and compliance tests, residual pesticide analysis is particularly challenging primarily because the amounts of pesticides retained on the plant are extremely low compared to the amounts of endogenous chemicals like cannabinoids, terpenes, flavonoids, and chlorophyll, and these “co-extractives” interfere with accurate measurement. This article will describe methodologies for residual pesticide analyses in cannabis flower with an emphasis on the larger target lists of Canada and California.
Comprehensive residual pesticide quantitation in cannabis flower
A comprehensive approach to pesticide residue analysis in cannabis flower included a single sample preparation scheme shunted to both LC-MS/MS and GC-MS/MS for the analysis of more than 210 pesticides.1 The sample preparation strategy resulted in injecting 500-fold dilutions of the weighed sample into each analytical instrument. Using highly dilute sample extracts, the researchers leveraged the sensitivity and specificity of the analytical platforms and maintained performance for extended periods, thus reducing instrument downtime and increasing productivity. The reported method included 215 target pesticides with 141 being analyzed via LC-MS/MS and the remaining 74 by GC-MS/MS. Method performance was demonstrated by injecting five replicates at the limit of quantitation (LOQ, determined as signal-to-noise >=10:1). Recoveries between 70 and 120% were determined for 72/74 compounds, and the
percent root mean square (%RSD) was less than 15% for all GC-MS/MS amenable compounds. For the LC-MS/MS compounds, recoveries of 70–120%, and %RSD <15% were determined for 138/141 of the target pesticides. An LOQ of 0.1 mg/kg was determined for all but 17 of the 215 targets. For more details see Ref. 1.
Residual pesticide testing in cannabis flower in Canada and California
With respect to the number of target pesticides and action levels, Canada has the most comprehensive list of 95 pesticides with action levels as low as 20 parts-per-billion (ppb) for dried cannabis, and 10 ppb or fresh (wet) cannabis or cannabis oils. The California list is currently the largest in the U.S., with 66 target pesticides and action levels down to 100 ppb for inhalable cannabis and other cannabis products. The Canadian list does not completely incorporate the California list with captan, chlordane, dimethomorph, and fenhexamid unique to California.
The most common analytical platform for the quantitation of residual pesticides in cannabis is LC-MS/MS, and for most U.S. states, including the 59 target pesticides in the Oregon list, this is exactly true. Except for California and Nevada (as of the date of this writing), all U.S. state pesticide lists can be analyzed using LC-MS/MS only. The Canadian list further presents at least six compounds that are not amenable to common LC-MS/MS ionization sources such as electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI). These are: endosulfan-alpha and beta, etridiazole, fenthion, kinoprene, and pentachloronitrobenzene. The reasons these compounds do not ionize, or poorly ionize in ESI or APCI, are varied and complex, but may be due to low chemical polarity, thermal instability, dissimilar proton affinities, or the absence of atoms in the structural configuration amenable to the gain or loss of a proton. However, the compounds listed above and others such as captan and chlordane are commonly, and quite successfully, analyzed via GC-MS/MS using electron ionization. Figure 1 illustrates a GC-MS/MS chromatogram collected on the Agilent 7890B/7010 GC-MS/MS system (Santa Clara, CA) for compounds contained in the Canadian and California lists that are better analyzed via gas-phase technologies.
Figure 1 – GC-MS/MS MRM chromatogram. From left to right: etridiazole, pentachloronitrobenzene, kinoprene, fenthion, chlordane cis- and transisomers (pink, slight isobaric impurity noted at approx. 12 minutes), endosulfan-alpha, captan, and endosulfan-beta. All concentrations were 10 ppb in dry cannabis flower matrix. The y-axis is scaled to 100%. (Unpublished data courtesy of Jean-François Roy, Agilent Technologies, Santa Clara, CA.)When presented with the target lists of Canada and California and the physicochemical properties of the pesticides in those lists, experienced laboratorians immediately recognize that no single analytical platform can properly identify and quantitate the compounds in the complex cannabis matrix without compromising the end results. Simply diluting samples 50-fold in an organic solvent and injecting into a highly sensitive LC-MS/MS that “detunes” the response of most of the pesticides2 may result in ESI or APCI detection, but only at the cost of increased maintenance and decreased productivity. Another caveat of a single-platform LC-MS/MS approach is the use of a multimode source at high temperatures. Many ESI pesticides in the state lists are thermally labile, and therefore must be ionized at the lowest possible temperatures to achieve sensitivity. APCI requires higher temperatures to work properly—maybe as high as 450 °C. These two disparate properties are incompatible in a single method. Therefore, two different methods taking 30 minutes in total are required. Following this approach, a laboratory would need three LC-MS/MS systems to achieve the throughput of the comprehensive pesticide method presented above. This estimate does not incorporate the increased need for instrument maintenance that results from injecting large volumes of high-matrix samples, which will further reduce productivity and revenue generation.
Increasing productivity by including mycotoxins in the LC-MS/MS residual pesticides test
When run in parallel, analytical cycle times of 10 minutes or less will result in 5–6 samples per hour—essentially triple that of a single-platform approach. To further improve throughput, mycotoxins such as aflatoxins b1, b2, g1, g2, and ochratoxin should be added to the LC-MS/MS pesticide method, thus negating the need for a separate analysis. Figure 2 shows the analysis of the California pesticide list and five mycotoxins collected on the Agilent Infinity II Prime/Ultivo LC-MS/MS system.
Figure 2 – LC-MS/MS analysis of the California target list plus aflatoxins b1, b2, g1, 2, and ochratoxin. (Unpublished data courtesy of Peter J. Stone, Agilent Technologies, Santa Clara, CA.)Conclusion
A single-stream extraction followed by 500-fold sample dilution factors analyzed by both LC-MS/MS and GC-MS/MS leverages the benefits of each analytical platform. This approach improves quantitative accuracy and precision and decreases the need for instrument maintenance. Combined, these advantages result in fewer samples requiring reanalysis and increased throughput and revenue generation. Experts in organizations such as AOAC agree that a multiplatform approach is required to quantitate residual pesticides in the various cannabis and cannabinoid matrices, and compounds such as pentachloronitrobenzene, chlordane, captan, and at least two dozen other pesticides included in the various target lists should be analyzed by GC-MS/MS technologies. The ability to orthogonally confirm and quantitate compounds amenable to both platforms, with no reduction on performance, further demonstrates the benefit of a multiplatform approach. Moreover, laboratories equipped with both state-of-the-art LC-MS/MS and GC-MS/MS possess the analytical resources to rapidly and efficiently adapt to a perpetually changing regulatory environment.
References
- Jordan, R.; Asanuma, L. et al. A comprehensive approach to pesticide residue analysis in cannabis. Cannabis Sci. and Technol. 2018, 1(2), 26–31.
- Winkler, P. and Tran, D. Advanced strategies for comprehensive cannabis analysis: analyzing the complete California pesticide list using LC-MS/MS. The Emerald Conference. Feb. 15–16, 2018, San Diego, CA.
Anthony Macherone, Jean-François Roy, and Peter J. Stone are with the Life Science and Chemical Analysis Group, Agilent Technologies, Santa Clara, CA, U.S.A. Anthony Macherone is also with the Department of Biological Chemistry, Johns Hopkins University School of Medicine, Baltimore, MD, U.S.A., and can be contacted at Agilent Technologies, 2850 Centerville Rd., Wilmington, DE 19808, U.S.A.; tel.: 302-636-8159; fax: 302-633-8868; e-mail: [email protected]; www.agilent.com. Rick Jordan, Dan Miller, and Lilly Asanuma are with Pacific Agricultural Laboratory, Sherwood, OR, U.S.A.