(This is the first installment in an occasional series that will examine questionable cannabis laboratory testing results.)
The THC potency of marijuana flower sold in legal stores in four states is routinely and systematically inflated, sometimes by as much as 25% or more, according to an independent analysis of licensed cannabis testing-laboratory data obtained by MJBizDaily.
But perhaps even more troublingly, the analysis – conducted by Yasha Kahn of MCR Labs, a state-licensed cannabis testing laboratory in Massachusetts – also found evidence of data manipulation in lab testing for yeast and mold.
The analysis suggests tainted products that should have been failed for contaminants and destroyed or remediated instead was passed and eventually sold by legal retailers.
In addition, confirmation that THC potency inflation is running rampant – and often unchecked – across legal U.S. marijuana markets also stands to undermine consumer confidence in product labeling.
That, in turn, can harm the $34 billion legal cannabis industry’s credibility while also breaking one of marijuana legalization’s core promises of reliably tested and safe products, critics say.
Kahn, MCR Labs’ vice president of marketing and data science, obtained anonymized lab-testing results from Maryland, Massachusetts, Michigan, and Oregon.
He then applied basic data-forensic tools to identify and quantify abnormal results.
Kahn presented his findings to an online meeting of state regulators on Oct. 26, several officials who attended confirmed to MJBizDaily.
Industry observers contacted by MJBizDaily said Kahn’s analysis is the first known attempt to quantify across multiple legal markets what’s acknowledged in cannabis circles as a widespread dirty secret: The THC percentages printed on product labels in legal stores are often grossly inflated.
Principals of other labs who reviewed Kahn’s work told MJBizDaily the methods were sound – and the results are a welcome quantification of a known problem, even if the implications are troubling.
“This is the scientific approach (to quantifying lab results) we’ve been trying to get for a long time,” said Josh Swider, the co-founder and CEO of San Diego-based Infinite Chemical Analysis Labs.
“This is not ‘mistakes being made,’” Swider added. “This is fraud.”
Shopping spree
Lab shopping – in which marijuana product manufacturers and retailers seek out testing labs that will return friendly results in the form of high THC readings – has been prevalent in the cannabis industry for years.
In a study conducted in 2018-19 but published just this year in the journal PLOS One, researchers found that marijuana flower purchased in Colorado had “substantially lower” THC potency than what was advertised on the product labels.
In 23 samples obtained from 10 Colorado retailers, the “average observed THC potency was 23.1% lower than the lowest label reported values and 35.6% lower than the highest label reported values,” according to researchers at the University of Northern Colorado’s School of Biological Sciences.
Marijuana sold with unreliable THC potency “is an issue in every medical and recreational state,” Kahn told MJBizDaily via email.
Routine and runaway THC potency inflation undermines consumer trust and confidence in the regulated market.
It contributes to a situation that retailers, cultivators and sophisticated consumers acknowledge but claim to be unable to stop.
And unchecked contaminant-testing data manipulation “allows unsafe products to enter the market and puts consumer health at risk,” Kahn said.
Cause-effect situation
The findings highlight how the promises of verifiable product safety that convinced voters and lawmakers to legalize adult use and create today’s state-legal marijuana industry have led to a vicious cycle that many industry insiders accept as a normal business arrangement.
Consumers, looking for the best bargain for limited dollars, gravitate toward marijuana flower with the highest THC percentages.
They do this despite, as cannabis competition judges and other connoisseurs have insisted for years, THC percentage is not an indicator of quality or even the strength of the high.
In response, retailers purchase high-THC flower only from product suppliers who, in turn, are known to go lab shopping, where they present their flower for testing to multiple laboratories and then use the best result.
Finally, laboratories face market pressure to present friendly results – and otherwise risk losing business to labs willing to flout rules if they don’t.
Most insiders will readily admit to all of this, at least in private.
Yet, it’s a cycle no business or regulator has been able to break.
Analysis methodology
Kahn obtained data from four states via public-records requests or open data.
The data includes potency results from nearly 400,000 flower samples, tested by 70 different labs.
Some of the samples also included contaminant screening results.
The data is anonymized: No labs or product makers are identified.
Kahn used basic data-forensic tools to identify discontinuity in data sets.
Many data sets will present what’s called a “normal distribution” – an array of results across a spectrum that adheres to a familiar pattern, such as the distribution of height in male U.S. residents.
In some cases, labs presented what looked like a normal distribution of THC results – but with the average result inflated by about 25% versus other labs that tested the same flower.
That suggests labs with a normal-looking data distribution could be inflating results.
But Kahn also identified discontinuities in data sets, suggesting “abnormal” distribution that’s suggestive of selective THC inflation that could be as high as 50% – meaning marijuana flower that tested at 14% THC is labeled as 21%.
Kahn also found some labs presented a spectrum of results with multiple “humps,” or peaks, suggesting selective data manipulation for particular clients or samples rather than across-the-board inflation.
Those peaks were also suspiciously tied to round numbers, such as 25% or 35%.
But he also found irregular distribution in testing for yeast and mold.
Rather than a set of results that “looks normal” when plotted on a graph, he found labs that overrepresented results right before the fail point – the legal allowable limit for a contaminant that, if exceeded, means the batch must be remediated before it’s sold or discarded altogether.
That suggests labs that succumbed to market pressures to pass products that ought to have failed, according to Kahn’s analysis.
‘Big problem here’
Kahn presented the findings via video chat on Oct. 26 to a subcommittee of the Cannabis Regulators Association (CANNRA), a nonpartisan organization of government officials.
David Standiford, the laboratory compliance coordinator at the Oregon Liquor and Cannabis Commission who also leads a CANNRA working group on lab testing, was among the attendees.
But he said he did not review the data on pesticide and contaminant data manipulation.
While regulators were aware labs were presenting questionable results, confirmation that potency inflation is a problem across the country emphasizes “there’s a big problem here,” Standiford said.
“I think a lot of state regulators recognized this as an issue,” he added.
“I think a lot of people are really excited about how we can tackle this problem.”
A spokesperson for the Massachusetts Cannabis Control Commission (CCC) said in an emailed statement that marijuana consumers s in that state “should remain confident that the regulated products they purchase in Massachusetts have passed testing for contaminants and the results are current.”
“The Commonwealth’s Independent Testing Laboratories (ITLs) are held to some of the highest standards in the nation, as outlined by the Cannabis Control Commission (Commission)’s regulations and further articulated through its testing protocols,” the statement continued.
“The agency remains vigilant in its efforts to ensure consumers and patients have access to fully tested, safe products in the legal marketplace.”
The CCC did not respond to further questions about whether potency labels on products sold in Massachusetts can be trusted.
Regulators in Maryland, one of the states whose data was analyzed by Kahn, did not respond to requests for comment.
David Harns, a spokesperson for the Michigan Cannabis Regulatory Agency, said the governing body could not comment because it is currently locked in litigation with a state-licensed lab over alleged THC potency inflation regulators flagged.
‘Race to the bottom’
What the data doesn’t show is exactly what labs are doing to produce questionable results, noted Zachary Eisenberg, vice president of California-licensed Anresco Laboratories in San Francisco.
“There are a number of games you can play in the laboratory or otherwise to inflate results,” he said.
“That puts regulatory bodies in a really hard place to regulate labs and prove that a lab is cheating.”
Eisenberg is chair of a cannabis working group at the American Council of Independent Laboratories.
The trade association representing commercial testing labs has been attempting to sound the alarm about problems with cannabis testing for years.
He and Swider, who is the group’s vice chair, have sent letters to California regulators and raised the issue at conferences – with limited results.
In the meantime, Eisenberg and Swider told MJBizDaily that their labs have lost nearly all of his cannabis-testing clients, who have fled to other labs that will fudge results in ways they will not.
Other laboratories, including Kahn’s MCR Labs, also said they’ve also lost business for refusing to inflate THC numbers or fudge data to allow tainted products to pass.
“When we have a customer test for aspergillus or pesticides, we never see that customer again,” Eisenberg said.
“Eventually, it’s just going to be a race to the bottom,” he added.
“Only the people willing to play that game will stay in business.”
Chris Roberts can be reached at chris.roberts@mjbizdaily.com.