A High-Throughput Method for Routine Analysis of Pesticide Residues on Cannabis

Featured Article

Cannabis sativa has long been used as a medicine, shamanic agent and euphoriant.¹ Though medical legitimacy was regained in California in 1996 with the passage of Proposition 215, much of the Cannabis grown in the state has not been subjected to any level of quality testing. With the growth of the industry and the passage of a 2016 California referendum allowing access to Cannabis for all adults, the future cultivation and sale of Cannabis will be fully regulated starting in 2018. These regulations (still being drafted by the state) along with those already existing in several other states² will likely include mandatory quality testing for both the plant material and processed products. Of particular consumer interest and regulatory concern is the question of pesticide residues. Given the quasi-legal standing of many cultivators, the impetus to apply chemical pest controls has been unimpeded. Due to its federally illegal status, no pesticide is approved for use on Cannabis, and thus cultivators have had no guidance as to application limits.

In response to this challenge, many commercial laboratories have conducted their own research as to what products are in use and balance that with their analytical capabilities. This has led to a wide disparity in target lists available from different labs, as well as regulatory lists from different states. Although the market value of Cannabis exceeds that of most crops, batch sizes are quite small, leading to slim profit margins for nonretail businesses. Intense competition between Cannabis labs has further reduced pricing in an industry where testing is largely voluntary. There will be pressure on laboratories to keep testing fees low to accommodate market demand.

Although Cannabis remains popular in its dry plant form, many other associated forms exist. These include a wide variety of food and beverage products, as well as salves, tinctures, lotions, concentrated solvent-extracted oils and hashish. Each of these products provides for unique matrix challenges that can frustrate scientific analysis.

To address these needs, testing for pesticide residues in Cannabis must be sensitive, selective, fast and affordable. This article describes a robust, reliable method developed by CW Analytical (Oakland, Calif.) and AB Sciex (Toronto, Canada) that uses HPLC with triple-quadrupole mass spectrometry.

Materials and methods

Sample preparation was conducted by homogenizing the starting material and placing 500 mg of the sample into 10 mL of LC/MS-grade acetonitrile. The sample vial was then placed in a sonicator bath (Branson, Danbury, Conn.) for 10 minutes and placed on a shaker table for 1 minute. Supernatant was syringe-filtered through a 0.2-µm polyethersulfone (PES) filter and placed into an LC autosampler vial. Separation occurred on a Sciex ExionLC AC system employing a Phenomenex (Torrance, Calif.) Kinetex 5-µm biphenyl column (150 × 3.0 mm) and 5-mm guard column held at 40 °C. The aqueous phase (A) was 0.1% formic acid and 5 mM ammonium formate; organic phase (B) was 0.1% formic acid and 5 mM ammonium formate in 98/2 acetonitrile/water. A gradient of 10% B was held for 1 minute, followed by a ramp to 80% B at 4.30 minutes and 95% B at 8.7 minutes. This was held for 1.8 minutes before returning to starting conditions and holding for 3.5 minutes. The flow rate was 0.4 mL/min, and a 10-µL injection size was used. A Sciex 3500 triple-quadrupole system with Turbo V and electrospray ionization (ESI) probe with positive polarity was used for detection. Source temperature was set to 325 °C to minimize degradation of some compounds, curtain gas was set to 20, collision gas was 9 and ion-spray voltage was 5500. The ion source gases were both set to 60.

Designing the method to overcome matrix effects

Cannabis and Cannabis-infused samples were prepared according to three methodologies: the method described above, the original unbuffered QuEChERS method, and the AOAC 2007.01 method. The extracts were each spiked at 1, 10, 40 and 80 ppb using a standard mix obtained from LGC Standards (Cumberland Foreside, Maine). To compare suppression effects, analytes were quantified using a solvent-based calibration curve rather than a matrix-matched curve. Using these three methods, distinct matrices were prepared in triplicate: baked good, hard candy, gummi, glycerin tincture, wax, beverage, cow butter, beeswax-based balm, chocolate and Cannabis flower. For the purposes of this comparison, six pesticide residues popular among Cannabis growers were quantified: avermectin B1a, bifenazate, bifenthrin, dichlorvos, imidacloprid and myclobutanil. Results of the recovered residues are shown in Figure 1. The average recovery of all residues was 54.4% for the original method, 56.4% for the AOAC method and 89.5% for the dilute-and-shoot method described above. Based on these data, the researchers decided to proceed with a simple dilute-and-shoot method for sample preparation.

Figure 1 – Recovered residues of spike recovery by three sample preparation methods. Five residues were quantified within 10 different sample types.

Further efforts to overcome ion suppression were made during method development. Each analyte began method development with a minimum of three applicable Q1→Q3 transitions. These transitions were then examined in matrix samples, and quenched transitions were omitted from the final method. Changing to a biphenyl column chemistry and extending the gradient also aided in removing suppression zones from the target retention times. The full method contains a total of 47 pesticides of concern for Cannabis patients, one synergist (piperonyl butoxide), two plant-growth regulators (paclobutrazol and daminozide) and four aflatoxins (B1, B2, G1 and G2).

Although method development plays an integral role in establishing sensitivity of the method, routine maintenance and quality control (QC) monitoring procedures are also vital to maintain this sensitivity. Fresh mobile phase was purged through the autosampler and pumps daily, and matrix-matched QC samples were injected every 25 samples to track performance. A five-point matrix-matched calibration curve was generated weekly and used for quantification. The curtain plate and orifice were cleaned weekly according to the manufacturer’s protocols, and a new guard column was installed.

Results and discussion

With over 5000 samples run, the finished method has proven robust and dependable. Daily tracking of the QC has permitted continual updates to method performance, with very little change in sensitivity, as shown in Figure 2. The instrumental response to these ions has remained within 15% of the average throughout the use of the instrument, indicating that maintenance protocols are effective. These results illustrate the reliability of an instrument and method in constant use with minimal sample preparation. By limiting costs from sample preparation and maximizing run time, the method allows for high throughput with lower operating costs for the laboratory.

Figure 2 – Performance of the QC sample as shown by stability of quantification ion transitions over time.

Several of the analytes on the target list have yet to be detected on the samples tested, although many appear to be in common use (Figure 3).

Figure 3 – Percentage of all samples received with analyte concentration exceeding 100 ppb; n = 5109; analytes with <0.1% prevalence omitted.

Powdery mildew (caused by fungi of the order Erysiphales) and various insects (mites, aphids, whiteflies, thrips and worms) are the most common targets of a chemical treatment on Cannabis. This explains the high prevalence of miticides and fungicides. Piperonyl butoxide is not generally considered a pesticide, but is frequently added to pyrethroids and carbamate formulations as a synergistic agent. Certain formulations of spinosyn A and D and pyrethrins are listed as approved for use in organic production by the Organic Materials Review Institute,⁴ suggesting that growers may be looking for an organic means of pest control. Use of plant-growth regulators also appears to be relatively common in an effort to grow flowers with “tighter” buds by decreasing internodal distances. These data present a picture of the most prevalent pesticides in use in California.

References

1. Zuardi, A. History of cannabis as a medicine: a review. Rev. Bras. Psiquiatria2006, 28(2), 157.
2. www.leafly.com/news/industry/leaflys-state-by-state-guide-to-cannabis-testing-regulations
3. AOAC International. Official Method 2007.01. In Official Methods of Analysis of AOAC International, 19th ed., 2012.
4. https://nevegetable.org/insecticides-alphabetical-listing-trade-name

David Egerton is vice president of technical operations for CW Analytical, 851 81st Ave., Ste. D, Oakland, Calif. 94621, U.S.A.; tel.: 510-306-4716; e-mail:  [email protected];www.cwanalytical.com