Testing cannabis edibles for potency is hard enough as it is. Few standardized testing procedures exist in the emerging legal industry, and demand from consumers for high THC levels can create an incentive to falsify results. All of which makes it hard for consumers to know exactly what they’re getting or how it might make them feel.
“It’s hard to say how many products this might apply to.”
Now science has thrown yet another curveball. A particularly delicious curveball.
Chocolate, according to new findings out of California, can interfere with cannabis testing procedures and lead to inaccurate THC potency results. While the errors are small enough that they’re unlikely to pose a threat to consumers, they reveal yet-to-be understood wrinkles in how potency testing works. And in some cases, they could keep even carefully crafted edibles from reaching store shelves.
A Mystery, Not a Menace
The findings have already triggered alarm bells nationwide. The Associated Press reported Monday morning that “a chocolate labeled as 10 milligrams of THC could have far more and send someone to the emergency room with hallucinations.”
“Some people have asked, ‘Is this a public health threat?’ This is definitely not a public health threat.”
That’s a stretch, says David Dawson, the organic chemist who first uncovered the phenomenon. The discrepancies he found in THC test results were rather small, he told Leafly, and are more likely to lead to regulatory headaches than hallucinations.
“Some people have asked, ‘Is this a public health threat?’” he said. “This is definitely not a public health threat.”
But the scientific implications of his chocolate discovery could be wide-reaching, injecting further uncertainty into potency testing of other infused edibles.
“It’s hard to say how many products this might apply to,” said Dawson, who is scheduled to present what he calls the “quirky” findings of his research at this week’s meeting of the American Chemical Society in San Diego.
A Glitch in the Matrix?
When Dawson joined the staff of CW Analytical Laboratories in Oakland, California, about two years ago, he was fascinated by some of the more unusual cannabis product formats entering the market. “It seems like every week we see a new product type,” he said. “We just saw CBD protein powder for the first time. Everything’s new.”
As he worked to determine how best to prepare different kinds of edibles for testing, Dawson was drawn to chocolates, a common category of edibles and one particularly popular with new consumers.
Edibles containing chocolate were already giving chemists trouble when it came to potency testing. “We had seen some anecdotal variation, but we couldn’t put our finger on it,” Dawson recalled. “It was enough to be a red flag.”
Shop highly rated dispensaries near you
Showing you dispensaries nearSo he dove into the matrix. The chocolate matrix.
“A matrix you can think of as kind of the canvas on which your cannabinoids and your dosage exists,” he explained. “The matrix in cannabis-infused chocolate is the chocolate, which is to say it’s the fat, it’s the flavoring, it’s the chocolate sugars.”
When laboratories test edibles for potency, they’re trying to strip away that matrix and measure the amount of stuff left over.
“The name of the game for us is being to reliably and precisely extract all of the cannabinoids—or analytes, which we’re trying to analyze—and do it in the most straightforward manner,” Dawson said. “Some matrices are harder to pry it out of the cage, as it were.”
And chocolate, mysteriously, is one of them.
The Chocolate Paradox
Suppose you test just a single piece of an infused edible, then use the results from that sample to estimate the amount of THC in the edible as a whole. Statistically speaking, the bigger your sample, the more accurate your results.
Not so, Dawson says, with chocolate. As he tested larger and larger samples, his estimates actually got less accurate. “As you approach testing the entirety of the sample, you should be getting better and better representation of what’s in there,” he said, “and we seem to see the opposite.”
“It went down linearly based on how much chocolate was in it.”
When scientists test an edible, the first step is to turn it into a sort of homogenous goo. “We blend up the chocolate bar in a food processor to get a nice small, homogenous particulate matter to draw from for our samples,” Dawson explained.
From there, a sample of the chocolate mixture is analyzed. It’s put into a solvent, heated, bombarded by high-intensity sound waves, spun in a centrifuge, frozen, filtered into vials, then measured through high-pressure liquid chromatography.
The problem Dawson encountered is that the sample size—the amount of chocolate added into the solvent—seemed to influence the test result. “If you measure one gram of that ground-up chocolate bar and then do the full procedure, and then did the exact same procedure with two grams, you would get different values,” he said.
Smaller samples, Dawson found, would yield higher THC numbers. More chocolate would result in a lower test result. “At a given solvent volume,” his findings suggested, “one-gram [samples] will always give higher values than two grams.”
The discrepancies so far have been small, but they’re more than just academic. Under California cannabis rules, edibles producers can be penalized if products contain more or less than the advertised potency.
A chocolate bar labeled at 100 milligrams of THC, for example, needs to test within 10% of that amount. If it tests below that threshold—at less than 90 mg THC—it must be relabeled to show the lower amount. If it tests higher than the limit—110 mg THC—it’s deemed unsellable and unfit for consumption.
Generally speaking, the variability in chocolate testing isn’t enough to affect whether a product passes or fails. But that’s not always the case.
One of the starkest examples of the chocolate-testing issue, Dawson said, came when the lab tested a particular dark chocolate bar. “If we tested one gram [of chocolate] in 20 milliliters [of solution], it passed. If we tested two grams in 20 milliliters, it came back in the 80s.”
“The only variable that changed,” he added, “is did we scoop out one gram or did we scoop out two grams? That really is shocking that what seems like an inconsequential decision made on the bench could possibly have as wide-reaching an effect as keeping thousands of units off the shelves.”
Is the Fat Where It’s At?
The centerpiece of Dawson’s experiment involved dissolving cannabinoids in solvent to create a sort of stock solution, then adding virgin chocolate to see how it affected THC test results. “You could see that the more chocolate I added—one gram compared to two compared to three—the recovery of THC went down. And it went down linearly based on how much chocolate was in it.”
Those results suggested to Dawson that something in the chocolate was binding with the cannabinoids. “It is the chocolate essentially trapping some of your analytes,” he said. “My hunch is: It’s the fats.”
“Anything that is a fatty product might have a similar phenomenon going on.”
THC, after all, dissolves wonderfully in lipids. It’s part of why cannabis is so often mixed with butter when cooking, and it explains why exercise—which can burn body fat containing THC—can lead to a person testing positive for cannabis even weeks after consumption.
But wait. Chocolate may have a lot of fat in it, but what about other infused edibles that contain significant amounts of fat?
“That’s a fantastic question and is absolutely the immediate large-scale takeaway from this,” Dawson said. “Anything that is a fatty product might have a similar phenomenon going on.”
To Dawson, it’s an indication of the need for yet more research into how testing works and how to ensure consistent results from lab to lab. “We don’t want producers to fail compliance tests because we decided to use one gram instead of two grams. That’s insane,” he said.
At the same time, Dawson seems to accept that unforeseen hiccups in how cannabis testing works are to be expected in these early stages of the industry. “If there’s some weird quirk of testing and the matrices are interfering with our testing, we’re not aware of it until we are,” he said. “That’s the scientific method.”
