ABOVE: An employee at California-based testing company CannaSafe preps a cannabis sample for testing. © CANNASAFE

When voters in Massachusetts approved a 2012 ballot measure to legalize the sale and use of medical marijuana, it came as a relief for many patients: they now had legal access to a drug known to relieve chronic pain and muscle spasticity associat-ed with a range of conditions, including HIV and multiple sclerosis. But the new law worried Christopher Hudalla, then a chemist at Waters Corporation, a Massachusetts-based company that manufactures analytical laboratory instruments. Like other states, Massachusetts had legalized cannabis without mandating that the herb, or products derived from it, be tested for safety before being sold, which struck Hudalla as odd. 

Why would marijuana, especially as it’s being considered medical—why would that not be tested?

 —Christopher Hudalla, ProVerde Laboratories

“Anything we put on or in our body is tested, whether...

“Anything we put on or in our body is tested, whether it’s cosmetics, or lotions, or bread, or nutritional supplements,” he tells The Scientist. “So why would marijuana, especially as it’s being considered medical—why would that not be tested?” 

Some time after the law went into effect in early 2013, Hudalla visited a university library in Boston to find out about safety issues associated with cannabis. He didn’t expect acute adverse effects from the plant’s main active ingredients, tetrahydrocannabinol (THC) or cannabidiol (CBD), or the hundreds of other compounds contained cannabis. Humans have been smoking the herb for thousands of years, he adds. 

He was, however, concerned about the risks of contamination. Sure enough, he found dozens of case reports documenting people who had fallen ill—and sometimes died—after smoking cannabis products that had been contaminated with harmful substances, from pesticides, lead, and other heavy metals to molds and other microbes.

Hudalla printed out the studies and took them to the Massachusetts Department of Public Health, where officials told him they’d look into the issue. A few weeks later, in May 2013, he spotted an article in The Boston Globe reporting that Massachusetts had become the first state to mandate analytical testing of medical marijuana. Companies wanting to sell medicinal cannabis in the state would have to hire an independent lab to test their products for contaminants and for potency—the latter to ensure the accuracy of product label claims. Other states were quick to follow, typically mandating testing for both safety and potency. By 2019, 26 states had introduced mandatory testing for medical and/or recreational markets, according to Cannabis Industry Journal.

These steps have led to an explosive growth of companies specialized in cannabis testing. While some states, such as Kentucky and North Carolina, only got their first accredited cannabis testing labs last year, well-established cannabis markets such as California have dozens. In principle, these labs should act as a US-wide filter that prevents unsafe or in-accurately labeled marijuana products from reaching consumers. However, the fledgling testing industry is facing considerable challenges that compromise its important role. 

Because cannabis is still considered illegal at the federal level, the responsibility of regulating cannabis and cannabis-derived products falls to states, creating a patchwork of different testing requirements across the country. In addition, state governments issue little to no guidance about protocols for testing products for either potency or safety. Instead, labs have had to trailblaze the development of their own methods. 

Now, in a cannabis testing industry that is only a few years old, it’s evident that reports on potency can vary from lab to lab, and recalls of contaminated products happen across the country, threatening consumer trust. “It’s not really at all like any other industry I’ve worked in, in that they’re still trying to work out proficiency and certification standards,” notes Frank Conrad, an analytical chemist formerly in the biofuels industry who now runs Colorado Green Lab, a consultancy firm for the cannabis industry. 

Measuring the potency of cannabis-derived products

Knowing the concentration of active ingredients is crucial throughout the cannabis supply chain. Cultivators and retailers want to know a product’s strength because higher concentrations of CBD or THC (the more psychoactive compound of the two) usually mean higher prices. Potency information is also important for doctors who prescribe cannabis to their patients and for recreational users who wish to control the intensity of their high. Labs are tasked with testing samples at every stage of production—from the whole plant to the manufactured product, and often its final packaged form. 

In the absence of federal guidance, labs have had to develop and validate their own methods of measuring the concentrations of THC and CBD, and distinguishing between them—a particular challenge due to the molecular similarity of these two compounds to each other and to other cannabis compounds. Many labs have converged on a few analytical techniques, although the precise details of each company’s procedure are often proprietary. Hudalla, who runs the Massachusetts-based cannabis testing lab ProVerde Laboratories, says he reckons that most use high-performance liquid chromatography (HPLC) to measure concentrations of THC and CBD, plus their acidic precursors, tetrahydrocannabinolic acid and cannabidiolic acid, which degrade to THC and CBD upon being heated through smoking or preparation for oil and edible products. 

After the dried plant has been ground up and doused in a solvent to extract cannabinoids, the residual fluid is fed into a HPLC system which separates different compounds by using pressure to force the mixture through a granular material. The technique is based on the principle that molecules with different structures take different amounts of time to pass through this material. If done correctly, the method allows chemists to separate THC or CBD from terpenes, flavonoids, and other cannabinoids and to detect each compound’s concentration. 

However, there are many accounts of labs producing starkly different potency measurements for the same products—some results varying by as much as 40 percent, according to a 2018 report by Marijuana Business Daily. Retailer Jerred Kiloh, who owns a marijuana dispensary in Los Angeles, has himself received very different estimates from different labs testing the same products. “That’s what we deal with constantly,” he says. 

TESTING TECHNIQUES: A CannaSafe chemist uses high-performance liquid chromatography to analyze samples for potency.

Many experts blame this variability on the lack of standardized methods for determining marijuana potency. Lab protocols can differ in the specific solvents or reagents used during extraction and analysis. Then there are the instruments themselves, which may come from different manufacturers and have different calibration standards, potentially contributing additional variation. 

Complicating matters further is the variety of cannabis products on the market, some of which are easier to analyze than others. From CBD-containing tinctures, lotions, and dog treats to THC-infused chocolate bars and turkey pot pie, each product needs to undergo a custom cannabinoid extraction method before its ingredients can be tested. Data from California’s Bureau of Cannabis Control suggest that the more complex the product, the more likely it is to be inaccurately labeled. In 2018, regulators found that 10.6 percent of cannabis flower samples, 20.4 percent of inhalable oils and waxes, and 32.9 percent of edibles, tinctures, and lotions carried labels with potency estimates that were more than 10 percent different from the true value

Antonio Frazier, vice president of operations at California-based testing company CannaSafe, worries that these inaccuracies—and the recalls that often follow—make consumers wary of the industry. “People have a hard time trusting us,” he says. 

Detection of pesticides and other contaminants

The stakes are higher for contaminant testing than for potency testing, as weed contaminated during cultivation, processing, or packaging could be dangerous, especially for frail or immunocompromised patients. For growers, this sort of safety testing can be “make or break,” says Frazier. A finding of contamination could force them to discard entire batches worth hundreds of thousands of dollars or more. 

Luckily for testing companies and consumers alike, safety testing is in principle more straightforward than potency evaluation because labs can turn to methods prescribed by federal agencies for testing other botanical products. Labs often isolate pesticides using liquid chromatography and then assess their concentrations using mass spectrometry, a technique suited to detecting the often tiny traces of contaminants. Techniques such as PCR and DNA sequencing are used to look for biological contaminants such as fungi and bacteria.

A more significant obstacle in cannabis safety testing is uncertainty about what concentrations of contaminants are safe. Most research on pesticide toxicity, for example, assumes ingestion of products, but cannabis is often smoked—a mode of consumption that poses different risks, Hudalla explains. “There’s been very little study that’s focused on the thermal combustion or degradation of pesticides prior to inhalations.” 

While government and academic research is lacking, some labs have conducted their own studies on the risks of certain pesticides in cannabis cultivation. In 2015, Conrad’s firm was asked by a local Colorado consumer advocacy group if myclobutanil, a fungicide that prevents mildew on plants, posed risks to people smoking cannabis. During a routine inspection of cannabis farms a few weeks earlier, state officials had noticed workers applying the pesticide, which is considered safe at very low concentrations in agricultural crops, but is prohibited for use in tobacco cultivation. 

Conrad’s chemical analysis found that, once heated past the boiling point, myclo-butanil generates hydrogen cyanide. Though unlikely to be lethal at the levels used in cannabis cultivation, the compound should not be inhaled by people with weakened health, Conrad says. After he shared his findings with the state’s department of health, Colorado and several other states banned the use of the fungicide in cannabis cultivation. Similar cases have played out for other pesticides. “As a general rule, the labs [are] advocating mainly on the behalf of consumers that we should be doing more testing,” says Conrad.

Some data suggest that increases in this sort of testing have made cannabis products safer over time. When California first mandated pesticide testing for cannabis in 2018, more than 24 percent of products tested by CannaSafe, Frazier’s company, failed the state’s pesticide standards. By the start of 2019, that rate had dropped to 1.5 percent—to Frazier a sign that the testing system encourages growers to be more careful about the products they use. 

However, safety lapses do happen. Last year, Colorado officials randomly sampled cannabis products sold around Denver and discovered yeast and mold on products from batches that testing labs had declared contamination-free. The findings triggered a major recall that affected 144 dispensaries around the city. It’s not clear who in the supply chain was at fault. While some experts suggest that such situations could be caused by microbes growing after products are packaged, some similar situations in the past have involved error or manipulation from labs or growers. (See below, “Bad Behavior.”) 

A standardized approach to cannabis testing?

While a certain degree of variation in product testing is inevitable, particularly for agricultural products that themselves show natural variation, efforts are afoot to tackle variability in potency and safety testing in a systematic way. In addition to conducting random testing of products that end up on shelves, a number of state regulators have raised the bar for becoming a licensed lab. Of the 26 states with mandatory testing, 18 require some form of accreditation, usually ISO 17025, a rigorous and expensive certification issued by the International Organization for Standardization for which labs have to provide extensive data to convince auditors that their methods are accurate. However, even accredited labs appear to have difficulties replicating one another’s results—something that puzzles Holly Johnson, chief science officer at the American Herbal Products Association (AHPA), which represents more than 350 companies doing business in herbal products. 

One solution could be nationwide, compulsory proficiency testing, whereby an independent third party sends lab researchers an unknown sample to see whether they can accurately analyze its ingredients. This sort of oversight is standard for US companies testing water, biofuels, agricultural goods, and many other products, Conrad notes. However, a national proficiency program for cannabis labs is infeasible because federal law prohibits the transportation of high-THC cannabis across state lines, he explains. 

Still, some states have managed to set up their own proficiency programs. In 2016, Nevada began to send cannabis samples to certified labs across the state and to evaluate the consistency of the results. And in 2014, California-based Emerald Scientific launched a proficiency test involving various products spiked with low but traceable concentrations of THC. Nearly a hundred US labs participate, Kirsten Blake, the company’s vice president, tells The Scientist in an email, either voluntarily or as a part of the ISO 17025 certification. 

In addition to improved proficiency testing, Johnson would like to see the cannabis industry adopt standardized testing procedures. Another organization she works for, a Maryland-based nonprofit called AOAC that develops standardized methods for agricultural testing, recently published two methods of analysis for evaluating THC and CBD potency in cannabis, from extraction to interpretation of results. Some labs have already adopted the standards. Doing the same for safety testing, however, will be more challenging, Johnson says, because states differ widely in how (or whether) they regulate pesticide use in cannabis cultivation.

Agreeing on what testing standards should be is also difficult, Kiloh notes. As some companies have argued, the more stringent the standards, the more expensive they’ll be to adopt, which raises the bar for newcomers and less established companies, potentially stifling industry growth. On the other hand, Kiloh says he worries that looser standards could allow companies to try to undercut competition by running tests as cheaply and as quickly as possible—a situation that could cripple the business of established, reputable labs and compromise the safety and integrity of cannabis products, not to mention consumer trust. 

“We started a billion-dollar industry not more than a couple of years ago,” Kiloh says. Finding standards that keep consumers safe and informed while also allowing the young cannabis-testing industry to thrive will take time, he says. “We still have some work to do to get to very clear standards for testing in labs.” 

Bad Behavior 

While most issues with potency and safety measurements in the cannabis industry stem from a lack of standardized methods, there have also been reports of bad behavior. 

Sometimes, it could be the cultivators who are being dishonest. In many areas, cannabis growers themselves typically choose the samples they send in for testing, which may allow them to influence testing results by cherry-picking samples known to have higher THC concentrations or to be free of contaminants. 

Other times, the testing labs are to blame. In one high-profile 2018 case, for example, California’s Bureau of Cannabis Control caught the director of Sequoia Analytical Labs in Sacramento fabricating pesticide testing results for hundreds of batches of cannabis.

Even without direct data manipulation, labs can enhance results—for example, by using testing protocols likely to inflate potency measurements. 

James MacRae, a data scientist and founder of the cannabis consultancy firm Straight Line Analytics, began investigating Washington State’s cannabis testing several years ago and found that the labs most likely to approve batches for safety also produced high scores for THC concentrations. Some results were so inflated they were virtually impossible, statistically speaking, and eventually led to regulators shutting down one of the state’s largest testing labs in 2017.

Growers can take advantage of this system to try out different labs until they find one that gives them desirable results, a practice MacRae calls “lab shopping,” which in turn benefits labs that produce inflated results. 

Katarina Zimmer is a New York–based freelance journalist. Find her on Twitter @katarinazimmer.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?